100949 v2 Trends in Maternal Mortality: 1990 to 2015 Estimates by WHO, UNICEF, UNFPA, World Bank Group and the United Nations Population Division Trends in maternal mortality: 1990 to 2015 Estimates by WHO, UNICEF, UNFPA, World Bank Group and the United Nations Population Division WHO Library Cataloguing-in-Publication Data Trends in maternal mortality: 1990 to 2015: estimates by WHO, UNICEF, UNFPA, World Bank Group and the United Nations Population Division. 1.Maternal Mortality - trends. 2.Maternal Welfare. 3.Data Collection - methods. 4.Models, Statistical. I.World Health Organization. II.World Bank. III.UNICEF. IV.United Nations Population Fund. ISBN 978 92 4 156514 1 (NLM classification: WQ 16) PRE-PUBLICATION VERSION © World Health Organization 2015 All rights reserved. Publications of the World Health Organization are available on the WHO website (www.who.int) or can be purchased from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: bookorders@who.int). Requests for permission to reproduce or translate WHO publications –whether for sale or for non-commercial distribution– should be addressed to WHO Press through the WHO website (www.who.int/about/licensing/ copyright_form/en/index.html). The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines for which there may not yet be full agreement. The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. All reasonable precautions have been taken by the World Health Organization to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall the World Health Organization be liable for damages arising from its use. Contents Acknowledgments ........................................................................................................................ iv Acronyms and abbreviations .................................................................................................... vi Executive summary .................................................................................................................... viii 1 Introduction .............................................................................................................................. 1 2 Methodology for the 1990–2015 estimates of maternal mortality ........................ 3 2.1 Methodological refinements ........................................................................................................ 3 2.2 Model input variables .................................................................................................................. 4 2.3 Country data on maternal mortality used for the 1990–2015 estimates .................................... 5 2.4 Statistical modelling to estimate 1990–2015 maternal mortality ............................................. 11 2.5 Maternal mortality indicators estimated by the model ............................................................. 13 2.6 Uncertainty assessment ............................................................................................................. 13 2.7 Model validation ........................................................................................................................ 14 3 Analysis and interpretation of the 2015 estimates ................................................... 16 3.1 Maternal mortality estimates for 2015 ...................................................................................... 16 3.3 Comparison with previous maternal mortality estimates ......................................................... 26 4 Assessing progress and setting a trajectory towards ending preventable maternal mortality ...................................................................................................................... 27 4.1 Millennium Development Goal (MDG) 5 outcomes ................................................................... 27 4.2 Looking towards the future ........................................................................................................ 28 4.3 A call to action ............................................................................................................................ 33 References ...................................................................................................................................... 34 Annexes ........................................................................................................................................... 38 List of tables Table 1. Availability of maternal mortality data records by source type and country for use in generating maternal mortality ratio estimates (MMR, maternal deaths per 100 000 live births) for 2015 Table 2. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk, by United Nations Millennium Development Goal (MDG) region, 2015 Table 3. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths and AIDS-related indirect maternal deaths, by United Nations Millennium Development Goal (MDG) region, 2015 Table 4. Comparison of maternal mortality ratio (MMR, maternal deaths per 100 000 live births) and number of maternal deaths, by United Nations Millennium Development Goal (MDG) region, 1990 and 2015 ii List of annexes Annex 1. Summary of the country consultations 2015 Annex 2. Measuring maternal mortality Annex 3. Methods used to derive a complete series of annual estimates for each covariate, 1985– 2015 Annex 4. Adjustment factor to account for misclassification of maternal deaths in civil registration, literature review of reports and articles Annex 5. Usability assessment of civil registration data for selected years (1990, 1995, 2000, 2005, 2010 and latest available year) Annex 6. Estimation of AIDS-related indirect maternal deaths Annex 7. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, lifetime risk and percentage of AIDS-related indirect maternal deaths, 2015 Annex 8. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk by WHO region, 2015 Annex 9. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by WHO region, 1990–2015 Annex 10. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk by UNICEF region, 2015 Annex 11. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by UNICEF region, 1990–2015 Annex 12. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk by UNFPA region, 2015 Annex 13. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by UNFPA region, 1990–2015 Annex 14. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk by World Bank Group region and income group, 2015 Annex 15. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by World Bank Group region and income group, 1990–2015 Annex 16. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk by UNPD region, 2015 Annex 17. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by UNPD region, 1990–2015 Annex 18. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by United Nations Millennium Development Goal region (indicated in bold) and other grouping, 1990–2015 Annex 19. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by country, 1990–2015 iii Acknowledgments The Maternal Mortality Estimation Inter-Agency Group (MMEIG), together with Leontine Alkema of the National University of Singapore, Singapore, and the University of Massachusetts Amherst, United States of America (USA), Sanqian Zhang of the National University of Singapore, Singapore, and Alison Gemmill of the University of California at Berkeley, USA, collaborated in developing these maternal mortality estimates. The MMEIG consists of the following individuals, listed in alphabetical order: Mohamed Ali of the World Health Organization (WHO); Agbessi Amouzou of the United Nations Children’s Fund (UNICEF); Ties Boerma of WHO; Liliana Caravajal of UNICEF; Doris Chou of WHO; Patrick Gerland of the United Nations Population Division (UNPD); Daniel Hogan of WHO; Victor Gaigbe-Togbe of the UNPD; Edilberto Loaiza of the United Nations Population Fund (UNFPA); Matthews Mathai of WHO; Colin Mathers of WHO; Samuel Mills of the World Bank Group; Holly Newby of UNICEF; Lale Say of WHO; Emi Suzuki of the World Bank Group; and Marleen Temmerman of WHO. Leontine Alkema is the lead adviser to the MMEIG. Flavia Bustreo of WHO oversaw the overall work and process of developing the estimates. An external technical advisory group (TAG) provided independent technical advice. The members of the TAG are: Saifuddin Ahmed of Johns Hopkins University, USA; David Braunholz, independent consultant, United Kingdom of Great Britain and Northern Ireland; Peter Byass of Umeå University, Sweden; Namuunda Mutombo of the African Population and Health Research Centre, Kenya; and Thomas Pullum of ICF Macro, USA. We are also grateful to Jeffrey Eaton of Imperial College London, United Kingdom, Bilal Barakat of the Vienna Institute of Demography/International Institute for Applied Systems Analysis (IIASA), Austria, and Emily Peterson of the University of Massachusetts Amherst, USA, for discussion of the analyses. The Department of Governing Bodies and External Relations of WHO, WHO country offices, UNFPA country offices and UNICEF country offices are all gratefully acknowledged for facilitating the country consultations. Thanks are also due to the following WHO regional office staff: Regional Office for Africa: Phanuel Habimana, Derege Kebede, Tigest Ketsela Mengestu, Peter Mbondji, Gisele Carole Wabo Nitcheu, Triphonie Nkurunziza, Leopold Ouedraogo Regional Office for the Americas: Gerardo de Cosio, Patricia Lorena Ruiz Luna, Cuauhtemoc Ruiz Matus, Bremen De Mucio, Antonio Sanhueza, Suzanne Serruya Regional Office for South East Asia: Mark Landry, Neena Raina, Sunil Senanayake, Arun Thapa Regional Office for Europe: Gauden Galea, Gunta Lazdane, Ivo Rakovac, Claudia Elisabeth Stein Regional Office for the Eastern Mediterranean: Mohamed Mahmoud Ali, Haifa Madi, Ramez Khairi Mahaini Regional Office for the Western Pacific: Jun Gao, Susan P. Mercado, Mari Nagai, Teret Reginaldo, Howard Sobel. iv In addition, Maria Barreix, Dmitri Botcheliouk, Lauri Jalanti and Karin Stein of WHO provided translation during the country consultations. Thanks to all focal points of governments who reviewed the preliminary estimates of maternal mortality ratios and provided valuable feedback. Financial support was provided by WHO, through the Department of Reproductive Health and Research and HRP (the UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction), the United States Agency for International Development (USAID) and the National University of Singapore. This report was prepared by Leontine Alkema, Elena Broaddus, Doris Chou, Daniel Hogan, Colin Mathers, Ann-Beth Moller, Lale Say and Sanqian Zhang. Many thanks to Maria Barreix, Sara Cottler and Karin Stein for extensive work during the final preparation of the report. Contact persons: Doris Chou (e-mail: choud@who.int) and Lale Say (e-mail: sayl@who.int), Department of Reproductive Health and Research, WHO. Editing: Green Ink (www.greenink.co.uk) v Acronyms and abbreviations AIHW Australian Institute of Health and Welfare ARR annual rate of reduction (of MMR) BMat Bayesian maternal mortality estimation model CEMD Confidential Enquiry into Maternal Deaths CMACE Centre for Maternal and Child Enquiries COIA Commission on Information and Accountability CRVS civil registration and vital statistics DHS Demographic and Health Survey EPMM ending preventable maternal mortality GDP gross domestic product per capita based on PPP conversion1 GFR general fertility rate ICD-10 International statistical classification of diseases and related health problems, 10th edition ICD-MM Application of ICD-10 to deaths during pregnancy, childbirth and the puerperium: ICD maternal mortality MDG Millennium Development Goal MDG 5 Improve maternal health MDG 5A Reduce by three quarters, between 1990 and 2015, the maternal mortality ratio MICS Multiple Indicator Cluster Survey MMEIG Maternal Mortality Estimation Inter-Agency Group MMR maternal mortality ratio (maternal deaths per 100 000 live births) MMRate maternal mortality rate (the number of maternal deaths divided by person-years lived by women of reproductive age) PM proportion of deaths among women of reproductive age that are due to maternal causes 1 as used in this report. vi PMMRC Perinatal and Maternal Mortality Review Committee (New Zealand) PPP purchasing power parity RAMOS reproductive-age mortality study SAB skilled attendant(s) at birth SDG Sustainable Development Goal SDG 3.1 By 2030, reduce the global maternal mortality ratio to less than 70 per 100 000 live births TAG technical advisory group UI uncertainty interval UN United Nations UNAIDS Joint United Nations Programme on HIV/AIDS UNFPA United Nations Population Fund UNICEF United Nations Children’s Fund UNPD United Nations Population Division (in the Department of Economic and Social Affairs) USA United States of America VR vital registration (VR data come from CRVS systems) WHO World Health Organization vii Executive summary In 2000, the United Nations (UN) Member States pledged to work towards a series of Millennium Development Goals (MDGs), including the target of a three-quarters reduction in the 1990 maternal mortality ratio (MMR; maternal deaths per 100 000 live births), to be achieved by 2015. This target (MDG 5A) and that of achieving universal access to reproductive health (MDG 5B) together formed the two targets for MDG 5: Improve maternal health. In the five years counting down to the conclusion of the MDGs, a number of initiatives were established to galvanize efforts towards reducing maternal mortality. These included the UN Secretary-General’s Global Strategy for Women’s and Children’s Health, which mobilized efforts towards achieving MDG 4 (Improve child health) as well as MDG 5, and the high-level Commission on Information and Accountability (COIA), which promoted “global reporting, oversight, and accountability on women’s and children’s health”. Now, building on the momentum generated by MDG 5, the Sustainable Development Goals (SDGs) establish a transformative new agenda for maternal health towards ending preventable maternal mortality; target 3.1 of SDG 3 is to reduce the global MMR to less than 70 per 100 000 live births by 2030. Planning and accountability for improving maternal health, and assessment of MDG 5 and SDG targets, require accurate and internationally comparable measures of maternal mortality. Countries have made notable progress in collecting data through civil registration systems, surveys, censuses and specialized studies over the past decade. Yet, many still lack comprehensive systems for capturing vital events data, and underreporting continues to pose a major challenge to data accuracy. Given the challenges of obtaining accurate and standardized direct measures of maternal mortality, the Maternal Mortality Estimation Inter-Agency Group (MMEIG) – comprising the World Health Organization (WHO), the United Nations Children’s Fund (UNICEF), the United Nations Population Fund (UNFPA), World Bank Group and the United Nations Population Division (UNPD) – partnered with a team at the University of Massachusetts Amherst, United States of America (USA), the National University of Singapore, Singapore, and the University of California at Berkeley, USA, to generate internationally comparable MMR estimates with independent advice from a technical advisory group that includes scientists and academics with experience in measuring maternal mortality. The estimates for 1990 to 2015 presented in this report are the eighth in a series of analyses by the MMEIG to examine global, regional and country progress in reducing maternal mortality. To provide increasingly accurate maternal mortality estimates, the previous estimation methods have been refined to optimize use of country-level data and estimation of uncertainty around the estimates. The methodology used in this round by the MMEIG builds directly upon previous methods, but now provides estimates that are more informed by national data. It also supports more realistic assessments of uncertainty about those estimates, based on the quality of data used to produce them. The statistical code and full database used to produce the current estimates are publicly available online.2 2 Available at: http://www.who.int/reproductivehealth/publications/monitoring/maternal-mortality-2015/en/ viii Globally, the MMR fell by nearly 44% over the past 25 years, to an estimated 216 (80% uncertainty interval [UI]3 207 to 249) maternal deaths per 100 000 live births in 2015, from an MMR of 385 (UI 359 to 427) in 1990. The annual number of maternal deaths decreased by 43% from approximately 532 000 (UI 496 000 to 590 000) in 1990 to an estimated 303 000 (UI 291 000 to 349 000) in 2015. The approximate global lifetime risk of a maternal death fell considerably from 1 in 73 to 1 in 180. Developing regions account for approximately 99% (302 000) of the global maternal deaths in 2015, with sub-Saharan Africa alone accounting for roughly 66% (201 000), followed by Southern Asia (66 000). Estimated MMR declined across all MDG regions4 between 1990 and 2015, although the magnitude of the reduction differed substantially between regions. The greatest decline over that period was observed in Eastern Asia (72%). As of 2015, the two regions with highest MMR are sub-Saharan Africa (546; UI 511 to 652) and Oceania (187; UI 95 to 381). At the country level, Nigeria and India are estimated to account for over one third of all maternal deaths worldwide in 2015, with an approximate 58 000 maternal deaths (19%) and 45 000 maternal deaths (15%), respectively. Sierra Leone is estimated to have the highest MMR at 1360 (UI 999 to 1980). Eighteen other countries, all in sub-Saharan Africa, are estimated to have very high MMR in 2015, with estimates ranging from 999 down to 500 deaths per 100 000 live births: Central African Republic (881; UI 508 to 1500), Chad (856; UI 560 to 1350), Nigeria (814; UI 596 to 1180), South Sudan (789; UI 523 to 1150), Somalia (732; UI 361 to 1390), Liberia (725; UI 527 to 1030), Burundi (712; UI 471 to 1050), Gambia (706; UI 484 to 1030), Democratic Republic of the Congo (693; UI 509 to 1010), Guinea (679; UI 504 to 927), Côte d’Ivoire (645; UI 458 to 909), Malawi (634; UI 422 to 1080), Mauritania (602; UI 399 to 984), Cameroon (596; UI 440 to 881), Mali (587; UI 448 to 823), Niger (553; UI 411 to 752), Guinea-Bissau (549; UI 273 to 1090) and Kenya (510; UI 344 to 754). The two countries with the highest estimated lifetime risk of maternal mortality are Sierra Leone with an approximate risk of 1 in 17, and Chad with an approximate risk of 1 in 18. The estimated lifetime risk of maternal mortality in high-income countries is 1 in 3300 in comparison with 1 in 41 in low-income countries. Emergent humanitarian settings and situations of conflict, post-conflict and disaster significantly hinder the progress of maternal mortality reduction. In such situations, the breakdown of health systems can cause a dramatic rise in deaths due to complications that would be easily treatable under stable conditions. In countries designated as fragile states, the estimated lifetime risk of maternal mortality is 1 in 54. Globally, approximately 1.6% (4700) of all maternal deaths are estimated to be AIDS-related indirect maternal deaths. In sub-Saharan Africa, 2.0% of all maternal deaths are estimated to be AIDS-related indirect maternal deaths, yielding an AIDS-related indirect MMR of 11 maternal deaths per 100 000 live births. In 2015 there are five countries where 10% or more of maternal deaths are estimated to be AIDS-related indirect maternal deaths: South Africa (32%), Swaziland (19%), Botswana (18%), Lesotho (13%) and Mozambique (11%). 3 The uncertainty intervals (UI) computed for all the estimates refer to the 80% uncertainty intervals (10th and 90th percentiles of the posterior distributions). This was chosen as opposed to the more standard 95% intervals because of the substantial uncertainty inherent in maternal mortality outcomes. 4 An explanation of the MDG regions is available at: http://mdgs.un.org/unsd/mdg/Host.aspx?Content=Data/RegionalGroupings.htm (a list of the MDG regions is also provided in the full report). ix Nine countries that had MMR of more than 100 in 1990 are now categorized as having “achieved MDG 5A” based on MMR reduction point-estimates indicating a reduction of at least 75% between 1990 and 2015: Bhutan, Cambodia, Cabo Verde, the Islamic Republic of Iran, the Lao People’s Democratic Republic, Maldives, Mongolia, Rwanda and Timor-Leste. Based on MMR reduction point-estimates and uncertainty intervals for the same period, an additional 39 countries are categorized as “making progress”, 21 are categorized as having made “insufficient progress”, and 26 are categorized as having made “no progress”. Achieving the SDG target of a global MMR below 70 will require reducing global MMR by an average of 7.5% each year between 2016 and 2030. This will require more than three times the 2.3% annual rate of reduction observed globally between 1990 and 2015. Accurate measurement of maternal mortality levels remains an immense challenge, but the overall message is clear: hundreds of thousands of women are still dying due to complications of pregnancy and/or childbirth each year. Many of these deaths go uncounted. Working towards SDG 3.1 and ultimately towards ending preventable maternal mortality requires amplifying the efforts and progress catalysed by MDG 5. Among countries where maternal deaths remain high, efforts to save lives must be accelerated and must also be paired with country-driven efforts to accurately register births and deaths, including cause of death certification. Strengthening civil registration and vital statistics will support measurement efforts and help track progress towards reaching SDG 3.1. Among those countries with low overall maternal mortality, the next challenge is measuring and amending inequities among subpopulations. The new Global Strategy for Women’s, Children’s and Adolescents’ Health will spearhead an enhanced global collaborative response aimed at ending all preventable maternal deaths. x 1 Introduction When the global commitment was first made in 2000 to achieve the Millennium Development Goals (MDGs), United Nations (UN) Member States pledged to work towards a three-quarters reduction in the 1990 maternal mortality ratio (MMR; maternal deaths per 100 000 live births) by 2015. This objective (MDG 5A), along with achieving universal access to reproductive health (MDG 5B), formed the two targets for MDG 5: Improve maternal health. In the years counting down to the conclusion of the MDGs, a number of initiatives were established to galvanize efforts towards reducing maternal mortality. These included the UN Secretary-General’s Global Strategy for Women’s and Children’s Health, which mobilized efforts towards achieving MDG 4 (Improve child health) as well as MDG 5, and the high-level Commission on Information and Accountability (COIA), which promoted “global reporting, oversight, and accountability on women’s and children’s health” (1, 2). To build upon the momentum generated by MDG 5, a transformative new agenda for maternal health has been laid out as part of the Sustainable Development Goals (SDGs): to reduce the global MMR to less than 70 per 100 000 live births by 2030 (SDG 3.1) (3). The recent World Health Organization (WHO) publication, Strategies toward ending preventable maternal mortality (EPMM), establishes a supplementary national target that no country should have an MMR greater than 140 per 100 000 live births, and outlines a strategic framework for achieving these ambitious targets by 2030 (4). Planning and accountability for improving maternal health, and assessment of MDG 5 and SDG targets, require accurate and internationally comparable measures of maternal mortality. Many countries have made notable progress in collecting data through civil registration systems, surveys, censuses and specialized studies over the past decade. This laudable increase in efforts to document maternal deaths provides valuable new data, but the diversity of methods used to assess maternal mortality in the absence of civil registration systems prevents direct comparisons among indicators generated. The 2011 COIA recommendations emphasized the need for countries to establish civil registration systems for recording births, deaths and causes of death (2). Insufficient progress has been made, as many countries still lack civil registration systems and where such systems do exist, underreporting continues to pose a major challenge to data accuracy (5). Looking ahead, one cross-cutting action towards EPMM is to “improve metrics, measurement systems and data quality to ensure that all maternal and newborn deaths are counted” (4). Given the challenges of obtaining accurate and standardized direct measures of maternal mortality, the Maternal Mortality Estimation Inter-Agency Group (MMEIG) – comprising WHO, the United Nations Children’s Fund (UNICEF), the United Nations Population Fund (UNFPA), the World Bank Group, and the UN Population Division (UNPD) of the Department of Economic and Social Affairs – has been working together with a team at the University of Massachusetts Amherst, United States of America (USA), the National University of Singapore, Singapore, and the University of California at Berkeley, USA, to generate internationally comparable MMR estimates. An independent technical advisory group (TAG), composed of demographers, epidemiologists and statisticians, provides technical advice. The estimates for 1990–2015 presented in this report are the eighth in a series of analyses by the MMEIG to examine the global, regional and country progress of maternal mortality (6–11). To provide increasingly accurate estimates of MMR, the previous estimation methods have been refined to optimize use of country-level data. Consultations with countries were carried out following the development of preliminary MMR estimates for the 1990–2015 period. Consultations aimed to: give countries the opportunity to 1 review the country estimates, data sources and methods; obtain additional primary data sources that may not have been previously reported or used in the analyses; and build mutual understanding of the strengths and weaknesses of available data and ensure broad ownership of the results. These consultations generated substantial additional data for inclusion in the estimation model, demonstrating widespread expansion of in-country efforts to monitor maternal mortality. Annex 1 presents a summary of the process and results of the 2015 country consultations. This report presents global, regional and country-level estimates of trends in maternal mortality between 1990 and 2015. Chapter 2 describes in detail the methodology employed to generate the estimates. Chapter 3 provides the estimates and interpretation of the findings. Chapter 4 assesses performance in terms of MDG 5, discusses implications of the estimates for future efforts towards achieving SDG 3.1, and closes by underlining the importance of improved data quality for estimating maternal mortality. The annexes to this report present an overview of the definitions and common approaches for measuring maternal mortality, the sources of data for the country estimates, and MMR estimates for the different regional groupings for WHO, UNICEF, UNFPA, the World Bank Group and the UNPD. 2 2 Methodology for the 1990–2015 estimates of maternal mortality The methodology employed by the Maternal Mortality Estimation Inter-Agency Group (MMEIG) in this round followed an improved approach that built directly on methods used to produce the 1990– 2008, 1990–2010 and 1990–2013 maternal mortality estimates (9–13). Estimates for this round were generated using a Bayesian approach, referred to as the Bayesian maternal mortality estimation model, or BMat model (14, 15). This enhanced methodology uses the same core estimation method as in those previous rounds, but adds refinements to optimize the use of country-specific data sources. It provides estimates that are directly informed by country-specific data points, and uncertainty assessments that account for the varying levels of uncertainty associated with the different data points. There were two key methodological changes, described in section 2.1. Combined with the updated global maternal mortality database, the BMat model provides the most up-to-date maternal mortality estimates yet for the entire 1990–2015 timespan. These results supersede all previously published estimates for years within that time period, and differences with previously published estimates should not be interpreted as representing time trends. The full database, country profiles and all model specification code used are available online.5 2.1 Methodological refinements First, the improved model estimates data-driven trends for all countries with national data, better utilizing the substantial amount of data now available from recently established or strengthened civil registration systems, population-based surveys, specialized studies, surveillance studies and censuses. Given the historical scarcity of data, the previous iteration of the MMEIG model generated estimates for countries without well established civil registration and vital statistics (CRVS) systems from country-level covariate information (i.e. gross domestic product per capita based on purchasing power parity conversion [GDP], general fertility rate [GFR], and coverage of skilled attendants at birth [SAB]). The new model still incorporates these covariates, but the regression model has been modified to prioritize country-level data on maternal mortality, whenever available, to determine national trends in maternal mortality. Second, the improved methodology gives data from higher quality sources more weight in the model so that they influence the final estimates more than data that are less precise or accurate. Final estimates convey information about the overall quality of all of the data contributing to them through the size of their uncertainty interval – those informed by higher quality data are more certain, and those informed by lower quality data are less certain. Many of the key components of the estimation process, including data adjustments, covariates for informing estimates in settings with sparse data, and how AIDS-related indirect maternal deaths are estimated, remain very similar in the BMat model. In the future, as data quality and modelling 5 Available at: http://www.who.int/reproductivehealth/publications/monitoring/maternal-mortality-2015/en/ 3 methods improve, refinement of the methodology will continue. The following sections give an overview of all variables, data sources and statistical models involved in the estimations, and highlight the updated components. 2.2 Model input variables Maternal mortality measures Maternal mortality measures were obtained from country-specific data sources. Several data inputs on maternal mortality were included in the analysis: the absolute number of maternal deaths; the number of maternal deaths per 100 000 live births (i.e. the maternal mortality ratio or MMR); and the proportion of maternal deaths out of all deaths among women aged 15–49 years (PM).6 The PM was the primary input of analysis, because it is less affected by underreporting of all-cause deaths. In cases where only the MMR was reported, this was converted to a PM using the UN Population Division’s estimates of live births for 2015 (16) and all-cause mortality among women of reproductive age from WHO life tables (17). In some cases a reported PM also includes pregnancy-related deaths (i.e. accidental or incidental deaths not caused by the pregnancy) in the ratio, which requires adjustment. The absolute number of maternal deaths reported was used as the model input for a small subset of specialized studies that assessed the completeness of deaths recorded (including confidential enquiries and those studies which reported maternal deaths only). Details on the types of country-level maternal mortality data sources, the type of variable extracted from each, and the limitations of each type and consequent adjustments made are described in Box 1 and section 2.3. Types of data sources, variables extracted, and adjustments were similar to those made during the previous estimation round. Covariates To inform projection of trends across periods where data were sparse, or for countries with little or no data, the model included factors known to be associated with maternal mortality as predictor covariates. These predictor covariates were originally chosen by the MMEIG in 2010 from a broader list of potential predictor variables that fell into three groups: indicators of social and economic development (such as GDP, human development index and life expectancy), process variables (SAB, antenatal care, proportion of institutional births, etc.) and risk exposure (fertility level). The specific variables selected were: GDP, GFR and the proportion of births with SAB. Data for each of these variables were obtained respectively from: the World Bank Group (18), the UNPD (16) and UNICEF (19). Methods used to derive a complete series of annual estimates (1990–2015) for each covariate are described in detail in Annex 3. The most recent data from each source were used to update covariates; otherwise little changed from the previous estimation round. Other model inputs Estimating the MMR and other maternal mortality indicators required that country-year estimates for live births, and both all-cause deaths and deaths due to HIV/AIDS among women aged 15–49 years be included in the model. Sources for these data were the same as in the last round, but live 6 More information on these measures and precise definitions for terms used are provided in Annex 2. 4 births were updated following the release of UNPD’s World population prospects: 2015 revision in July 2015 (16). WHO life tables provided all-cause mortality estimates (17), and UNAIDS provided AIDS-related mortality estimates (20). Details on the methodology used to produce these estimates are provided in the references cited after each (see Annex 4). Box 1 Data source types, measures extracted from each, and sources of error Data source type Information used to Sources of systematic error Sources of random error accounted construct maternal accounted for in analysis for in analysis mortality estimates • Underreporting of maternal • Stochastic errors due to the CRVS PM deaths general rarity of maternal deaths Number of maternal • Stochastic errors due to the • None Specialized studies general rarity of maternal deaths deaths, PM or MMR • Underreporting of Other data sources pregnancy-related deaths reporting on • Over-reporting of maternal • Sampling error pregnancy-related PM or MMR deaths due to the inclusion • Error during data collection and of deaths which are data processing mortality (including accidental or incidental to surveys) pregnancy • Error during data collection and • Underreporting of maternal data processing Other data sources Pregnancy-related PM deaths • Stochastic errors due to the reporting on or pregnancy-related general rarity of maternal deaths maternal mortality MMR • Additional random error CRVS: civil registration and vital statistics; MMR: maternal mortality ratio, i.e. maternal deaths per 100 000 live births; PM: the proportion of maternal deaths out of all deaths among women aged 15–49 years. 2.3 Country data on maternal mortality used for the 1990–2015 estimates This section addresses the different sources of maternal mortality data obtained from countries, describing for each source: the types of measures extracted, the adjustments made to each and the sources of error. Detailed descriptions of each type of data source are provided in Annex 2. Box 1 summarizes the measures extracted from each data source and the sources of error, and Table 1 provides an overview of data availability by type and by country. Availability varies greatly; among the 183 countries covered in this analysis (i.e. all countries with a population higher than 100 000), 12 countries had no data available. Overall, 2608 records providing 36347 country-years of data on maternal mortality were included in this analysis. 7 The sum of country-years of data has been rounded. 5 Table 1. Availability of maternal mortality data records by source type and country for use in generating maternal mortality ratio estimates (MMR, maternal deaths per 100 000 live births) for 2015 Source type # records # country-years A. CRVS 2025 years of reporting 2025 B. Specialized studies 224 studies 364 C. Other sources – reporting on maternal mortality 178 reports/studies 206 D. Other sources – reporting on pregnancy-related 181 reports/studies 1038 mortality All 2608 records 3634a CRVS: civil registration and vital statistics. a The sum of country-years of data has been rounded. CRVS systems are the primary (and generally preferred) source of data on maternal mortality. However, many countries lack such a system, or have one that is not nationally representative. In such situations, other data sources can provide valuable information. These alternate data sources include specialized studies on maternal mortality, population-based surveys and miscellaneous studies. All data on maternal mortality include a degree of uncertainty associated with the error in a particular data source. Some data are always (systematically) lower or higher than the true value of the variable being measured. For example, the numbers of deaths reported in CRVS records will be systematically lower than the true number, because there will always be deaths that go unreported. This is referred to as systematic error. Estimates of the amount of systematic error for a given data source were calculated based on past analyses that assessed the extent to which data from that source differed from the truth (as determined by rigorous specialized studies which looked to determine underreporting of maternal mortality, see Annex 4). Based on these assessments, adjustments were then applied to the data to account for systematic error and bring it closer to the “truth” using methods similar to the previous estimation round. These adjustments contribute uncertainty to the final estimates of maternal mortality, since no adjustment is based on perfect information. Data may also differ from the truth in a direction that is unpredictable. For example, human error when recording information and entering it into a database may cause data to deviate from the truth in either direction (higher or lower). This is referred to as random error, and it cannot be adjusted for but also adds uncertainty to the final maternal mortality estimates. Uncertainty due to random error and uncertainty due to adjustments is communicated in the data’s error variance. Generally speaking, inputs (usually PMs) from data sources with less random error and less uncertainty in systematic error (and corresponding adjustments) had smaller variances than inputs from data sources with more error. In turn, inputs with smaller variances carried greater weight in determining the final maternal mortality estimates. In this way, all data sources could be included, with higher quality data (containing less uncertainty) having a greater influence on estimated country-specific trends as compared to lower quality data. Box 2 discusses the implications for the trend estimates of countries that have been improving the 6 quality of their data over time. For more details on the data models and variance estimation, please see the paper by Alkema et al. (15). The subsections below include discussion of sources of both systematic and random error for each type of data source, and how the model accounted for them. Box 2 Estimating trends for countries with improving data quality The MMR trend lines for Cuba, a country with consistently high-quality civil registration and vital statistics (CRVS) data, and Peru, a country with improving data, illustrate how data quality influences the estimates generated by the updated model: Cuba has had a complete CRVS system established since before 1985 that consistently provides high-quality data for estimation of maternal mortality. As shown in the figure above, the estimated MMR trend line closely tracks the CRVS data points throughout the 1990–2015 time period. The shaded region around the trend line, which represents the 80% uncertainty interval (UI), remains roughly the same width throughout. In contrast, Peru had little data of adequate quality available prior to 2000, but since 2000 has established a more 7 Box 2 Estimating trends for countries with improving data quality robust CRVS system, and has conducted numerous additional studies. The estimated trend line is therefore influenced by covariate information prior to 2000, but tracks the data points from the high-quality data sources closely after 2000. Four DHS studies were conducted in Peru during the 1990–2015 period, and data points from these studies also influence the trend line. However, given the lower reliability of the data from these studies, they exert less influence (the line does not track them as closely) compared to the CRVS data points. Finally, the shaded region around the trend line narrows dramatically as time progresses. This represents the narrowing of the 80% UI as data quality improves and allows estimates to become more precise. Like Peru, many countries have recently established CRVS systems, or have substantially improved the quality of data collected by their CRVS systems. The new model takes advantage of these new data, allowing these countries’ trend lines to be more influenced by the data during the period after the system was established, and increasingly so as the quality improves. Civil registration and vital statistics data National CRVS systems are meant to record all births, deaths and causes of death within a country. The data retrieved from CRVS systems are referred to as vital registration (VR) data. For VR data, the observed proportion of maternal deaths among all deaths to women aged 15–49 was included as the data input. For VR country-years based on ICD-9, deaths coded to 630-676 were used and for those based upon ICD-10, deaths coded to codes O00-O95, O98-O99 and A34 were used (which include only those maternal deaths for which the timing corresponds to the definition of a maternal death)8. Under ideal circumstances, CRVS systems provide perfect data on the number of maternal deaths within a country. In reality, however, deaths often go unrecorded (resulting in incompleteness) or the causes of death are incorrectly recorded (resulting in misclassification) both of which contribute to underreporting of maternal deaths. The extent of underreporting determines a civil registration record’s usability in the analysis. Usability is defined as the percentage of all deaths among women of reproductive age in the country-year for which a cause of death has been recorded. It is calculated by multiplying the system’s completeness (proportion of all-cause deaths that were registered in the system) by the proportion of deaths registered in the system that were assigned a specific ICD cause (see Annex 5 for details on calculating usability). Additionally, the number of data-years available from a CRVS system in a given time period was used as a proxy for the data’s reliability, with regular data reporting across years indicating a high-functioning system. Given these factors, each country-year of VR data was placed into one of three categories (type I, II or III) depending on its usability and the number of available years with data. Box 3 summarizes the criteria for each category. The category determined whether or not the record for that country-year of data was included in analysis, and if included, how it was adjusted to account for misclassification. 8 A maternal death is defined as the death of a woman while pregnant or within 42 days of termination of pregnancy, irrespective of the duration and site of the pregnancy, from any cause related to or aggravated by the pregnancy or its management (from direct or indirect obstetric death), but not from accidental or incidental causes. 8 This method of categorizing each year of a country’s VR data, rather than placing all of a country’s data into the same category (as in the previous estimation round), takes into account changes in data quality over time. For example, if a country strengthens its CRVS system, data from years after the system improvement can be categorized as type I, even if data from earlier years were classified as type II. Annex 5 includes a table listing the calculated data usability for selected years of VR data, by country. Box 3 Categorization of VR data retrieved from CRVS systems (country-year records) based on usability and availability Category Criteria Type I • Usability > 80% AND • Part of a continuous string of three or more country-year records with > 60% usability and no more than one year gap in between records Type II • Usability > 60% AND • Part of a continuous string of three or more country-year records with > 60% usability and no more than one year gap in between records Type III • Other data from registration and mortality reporting systems. For these data points, data quality cannot be assessed as the countries have not submitted data to the relevant WHO office. Excluded • Usability < 60% OR • Not part of a continuous string of three or more country-year records with > 60% usability and no more than one year gap in between records Initial adjustment factors for all VR data (types I, II and III) were determined using procedures similar to those used in previous estimation rounds. For countries with type I data that have not conducted specialized studies (to assess the extent of systematic error in VR data; see next subsection for further information), the number of maternal deaths was multiplied by an adjustment factor of 1.5, as determined by a review of findings from 49 specialized studies, which was conducted in 2014 (the findings are summarized in Annex 4). However, for countries with type I data that have conducted at least one specialized study, the findings from the specialized study informed the adjustment factor applied to that country’s VR data. Calculation of adjustment factors was based on the approach used in the last estimation round, and the methods are described in the paper by Alkema et al. (15). Any civil registration records covering the same periods for which specialized study data were available were excluded to avoid double counting of the same information. For countries with type II data, a similar procedure was used as described for countries with type I data to obtain initial estimates of adjustments factors for civil registration records (either 1.5 or values indicated by specialized studies). However, for type II and III data, the model set-up included the possibility of higher adjustment factors depending on data quality, with the possibility of estimating a larger adjustment factor decreasing as usability increases (15). In addition to the systematic errors described above, and the uncertainty associated with those adjustments, the observed PMs obtained from civil registration records are subject to stochastic error, attributed to 9 maternal mortality being a generally rare event. Specialized studies on maternal mortality A number of countries reporting maternal deaths via CRVS systems also conducted specialized studies to determine if maternal deaths were underreported. While the methodology for these studies varies, any nationally representative study that documented corrections to data previously submitted to the WHO mortality database was considered a specialized study. These studies were used to inform maternal mortality estimates as well as VR data misclassification adjustment factors. Examples include those conducted in Guatemala and the United Kingdom, which reviewed a representative sample of the population using methods such as verbal autopsy to identify and correctly categorize causes of death; or studies such as those conducted in Australia, Mexico and the United Kingdom, which used the Confidential Enquiry system to review the classification and completeness of death reporting for deaths among women of reproductive age in a vital events database. Information from specialized studies was summarized into an observed PM. The PM or MMR reported in the study was generally used, except for Confidential Enquiries or other specialized studies reporting on maternal deaths only, which addressed both potential underreporting of maternal deaths as well as the total deaths among women of reproductive age during the study time period; for those studies, the absolute number of maternal deaths observed was used directly as a model input. All data inputs from specialized studies were used to inform the modelled maternal mortality estimates, without further adjustments. The only studies excluded from analysis were those that did not report the total number of all-cause deaths among women of reproductive age or associated births within the study period, and for which that information was not available from the CRVS system. Model inputs from specialized studies were assumed to have no systematic error. Sources of random error are the same as those for VR data. Population-based surveys and other data sources Examples of population-based surveys include the Demographic and Health Surveys (DHS), Multiple Indicator Cluster Surveys – Round 4 (MICS4), and Reproductive Health Surveys. Other data sources include censuses and surveillance systems. Many surveys include questions inquiring whether deceased women of reproductive age died during pregnancy or shortly after. For example, DHS and MICS both use the direct “sisterhood” method in which they ask respondents about the survival of all of their siblings. Such surveys therefore collect data on pregnancy-related deaths, which are used to compute the pregnancy-related PM. Other studies obtain and report the PM, and some may report a pregnancy-related MMR rather than PM if information on births is collected and information on all causes of deaths among women of reproductive age is not collected. Specialized studies indicate that there is some underreporting of maternal or pregnancy-related deaths in PMs derived from sources such as population-based surveys, censuses and surveillance studies, particularly since respondents may be unaware of the pregnancy status of their sisters or other women in the household. If no specific adjustments were reported, estimates for these data sources were revised to increase the number of maternal or pregnancy-related deaths by 10% to correct for underreporting. When pregnancy-related deaths were reported, the number was adjusted downward by 10% for sub-Saharan African countries and 15% in other low- and middle- 10 income countries to correct for inclusion of incidental and accidental deaths (21). As in previous estimation rounds, for studies that excluded deaths due to accidents when calculating pregnancy-related PMs, the calculated PMs were taken and used as model inputs without any further adjustment. In addition to the sources of systematic error discussed above, sources of random error for model inputs derived from surveys, censuses and other types of studies include sampling error and errors occurring during the data collection and data administration processes. 2.4 Statistical modelling to estimate 1990–2015 maternal mortality Summary of methods Limited data availability for many countries, and the limitations of the data that are available, mean that statistical models are needed for generating comparable estimates of maternal mortality across countries. The BMat model is flexible enough to account for differences in data availability and quality. Therefore, the same statistical model can now be used to generate estimates for all countries. As in previous MMEIG estimation rounds, the MMR for each country-year is modelled as the sum of the AIDS-related indirect MMR and the non-AIDS-related MMR: MMR = non-AIDS-related MMR + AIDS-related indirect MMR, where non-AIDS-related maternal deaths refer to maternal deaths due to direct obstetric causes or to indirect causes other than HIV, while AIDS-related indirect maternal deaths are those AIDS-related deaths for which pregnancy was a substantial aggravating factor. The estimation of the AIDS-related indirect MMR follows the same procedure as used in previous publications (9–11) and is described in detail in Annex 6. The expected non-AIDS-related MMR for the year 1990, and expected changes in the non-AIDS-related MMR from 1990 to 2015, are obtained through the multilevel regression model that was used in previous estimation rounds (explained in more detail below in this subsection). However, this existing model was extended to enable it to capture country-specific data-driven trends. To do this, it now includes information from the data via a country-year-specific multiplier. The result of this approach is that in country-year periods where high-quality data exist, the data dominate (i.e. the estimates produced are closer to the data), and in cases where there are no data, the regression determines the level and trend of estimates. In between, both sources of information inform the estimate of a country’s level and trend. For countries with high-quality VR data, the model tracks the data very closely, while providing some smoothing of the curve over time to remove stochastic fluctuations in the data. In the new model, the non-AIDS-related MMR is estimated for all countries as follows: Non-AIDS-related MMR(t) = expected non-AIDS-related MMR(t) x data-driven multiplier(t), where “expected non-AIDS-related MMR(t)” is estimated from the multilevel regression model, and the “data-driven multiplier(t)” allows for differences in the rate of change in MMR implied by the “expected non-AIDS-related MMR” and country-year-specific data points. For example, if data suggested that the non-AIDS-related MMR decreased much faster in year t than expected based on covariates, the data-driven multiplier for that year is estimated to be greater than 1, allowing the 11 model to produce estimates that closely track country data. This data-driven multiplier is modelled with a flexible time series model, which fluctuates around 1, such that the covariates determine the estimated change when data are absent (for further details on the multiplier please see the technical paper [15]). The extension of the non-AIDS-related MMR to allow for country-specific data trends was the main revision in the MMEIG model, as compared to the previous estimation approach. The second significant change to the model was the use of integrated data models to allow for uncertainty around data inputs to be incorporated into the estimates. For example, the PM from a DHS with a small sample size is assumed to be less precise than a PM from a DHS with a large sample size. As explained in section 2.3, this uncertainty is taken into account by the model when generating PM and thus MMR estimates; observations with smaller error variances are more informative of the true PM and thus will carry a greater weight in determining the estimates as compared to observations with larger error variances. All analyses were conducted using JAGS 3·3·0 and R; both are open-source statistical software packages (22, 23). Statistical code can be accessed online.9 Multilevel regression model A multilevel regression model was used to obtain the expected number of non-AIDS-related maternal deaths for each country-year. The model predicts maternal mortality using three predictor variables described in section 2.2. The model can be described as follows: log(PMina) = αi – β1 log(GDPi) + β2 log(GFRi) – β3 SABi with random country intercepts modelled hierarchically within regions: αi ~ N(αregion, σ2country), αr ~ N(αworld, σ2region) 2 meaning country intercepts (αi) are distributed normally with a country-specific variance (σ country) around random region intercepts (αregion), and random region intercepts (αregion) are distributed 2 normally with a region-specific variance (σ region) around a world intercept (αworld); and: GDPi = gross domestic product per capita (in 2011 PPP dollars) GFRi = general fertility rate (live births per woman aged 15–49 years) SABi = skilled attendant at birth (as a proportion of total births). For countries with data available on maternal mortality, the expected proportion of non-AIDS-related maternal deaths was based on country and regional random effects, whereas for countries with no data available, predictions were derived using regional random effects only. 9 Available at: http://www.who.int/reproductivehealth/publications/monitoring/maternal-mortality-2015/en 12 2.5 Maternal mortality indicators estimated by the model The immediate outputs of the BMat model were estimates in the form of PMs. These values were then converted to estimates of the MMR as follows: MMR = PM(D/B), where D is the number of deaths in women aged 15–49 years and B is the number of live births for the country-year corresponding to the estimate. Based on MMR estimates, the annual rate of MMR reduction (ARR) and the maternal mortality rate (MMRate; the number of maternal deaths divided by person-years lived by women of reproductive age [13]) were calculated. The ARR was calculated as follows: ARR = log(MMRt2/MMRt1)/(t1–t2), where t1 and t2 refer to different years with t1 < t2. The MMRate was calculated by using the number of maternal deaths divided by the number of women aged 15–49 in the population, as estimated by UNPD in World population prospects: 2015 revision (16). The MMRate was used to calculate the adult lifetime risk of maternal mortality (i.e. the probability that a 15-year-old woman will die eventually from a maternal cause). In countries where there is a high risk of maternal death, there is also an elevated likelihood of girls dying before reaching reproductive age. For this reason, it makes sense to consider the lifetime risk of maternal mortality conditional on a girl’s survival to adulthood. The formula used yields an estimate of the lifetime risk that takes into account competing causes of death: Lifetime risk of maternal mortality = (T15-T50)/ ℓ15 x MMRate, where ℓ15 equals the probability of survival from birth until age 15 years, and (T15 – T50)/ℓ15 equals the average number of years lived between ages 15 and 50 years (up to a maximum of 35 years) among survivors to age 15 years. The values for ℓ15, T15 and T50 are life-table quantities for the female population during the period in question. Regional maternal mortality estimates (according to the MDG, UNFPA, UNICEF, UNPD, WHO and the World Bank Group regional groupings) were also computed. The MMR in a given region was computed as the estimated total number of maternal deaths divided by the number of live births for that region. Additionally, the lifetime risk of maternal mortality was based on the weighted average of (T15 – T50)/ℓ15 for a given region, multiplied by the MMRate of that region. 2.6 Uncertainty assessment Accurately estimating maternal mortality proves challenging due to many countries’ limited data availability, and due to quality issues affecting the data that are available. The improved model provides a more realistic assessment of uncertainty around the estimates based on the amount and quality of input data. It allows for greater precision when more and better data are available and indicates the extent of estimate uncertainty in cases where there the amount of data is insufficient or the data are from sources more susceptible to error. It should be noted, however, that the uncertainty assessment does not include the uncertainty in covariates or other model input variables other than maternal mortality data. Model input data quality decreases with increasing systematic error and random error (discussed for 13 each data type in section 2.3), introducing uncertainty. This uncertainty is then carried through to the final estimates. Bayesian models allow for accurate assessment of the extent of uncertainty for a given estimated indicator by generating a posterior distribution of that indicator’s potential values. A Markov Chain Monte Carlo (MCMC) algorithm was used to generate samples of the posterior distributions of all model parameters (24). The sampling algorithm produced a set of trajectories of the MMR for each country, from which other indicators and aggregate outcomes were derived. This distribution is then used to compute a point-estimate and uncertainty interval (UI) for the indicator. In this case 80% UIs were calculated (rather than the standard 95%) because of the substantial uncertainty inherent in maternal mortality outcomes. The extent of uncertainty about a particular estimate, indicated by the size of the 80% UI, is determined by the amount and quality of data used to produce that estimate. For a country with very accurate sources of maternal mortality data, the MMR can be estimated with greater precision, and the 80% UI will be smaller than for a country with little data, or with data from less reliable sources. 2.7 Model validation The BMat model’s predictive validity was assessed by cross-validation. This procedure involves removing a subset of records from the data set, re-fitting the model to that smaller data set, and then seeing how well the model’s new estimates match the records that were removed (taking into account systematic errors). If the model’s new estimates are similar to the dropped data, it provides evidence that the model can accurately predict the values of missing data, which is important because data on maternal mortality is very limited for many countries. Another variation was also run in which data from the most recent time period were dropped and then estimates were produced using the remaining data. Results from this validation process indicate that the model is robust and adequately calibrated to generate the estimates for global maternal mortality indicators. Box 4 Accurately interpreting point-estimates and uncertainty intervals All maternal mortality indicators derived from the 2015 estimation round include a point-estimate and an 80% uncertainty interval (UI). For those indicators where only point-estimates are reported in the 10 text or tables, UIs can be obtained from supplementary material online. Both point-estimates and 80% UIs should be taken into account when assessing estimates. For example: The estimated 2015 global MMR is 216 (UI 207 to 249) This means: • The point-estimate is 216 and the 80% uncertainty interval ranges 207 to 249. • There is a 50% chance that the true 2015 global MMR lies above 216, and a 50% chance that 10 Available at: http://www.who.int/reproductivehealth/publications/monitoring/maternal-mortality-2015/en 14 Box 4 Accurately interpreting point-estimates and uncertainty intervals the true value lies below 216. • There is an 80% chance that the true 2015 global MMR lies between 207 and 249. • There is still a 10% chance that the true 2015 global MMR lies above 249, and a 10% chance that the true value lies below 207. Other accurate interpretations include: • We are 90% certain that the true 2015 global MMR is at least 207. • We are 90% certain that the true 2015 global MMR is 249 or less. The amount of data available for estimating an indicator and the quality of that data determine the width of an indicator’s UI. As data availability and quality improve, the certainty increases that an indicator’s true value lies close to the point-estimate. 15 3 Analysis and interpretation of the 2015 estimates Globally, the maternal mortality ratio (MMR; number of maternal deaths per 100 000 live births) fell by approximately 44% over the past 25 years; this falls short of the Millennium Development Goal (MDG) target MDG 5A which called for a reduction of at least 75% in MMR. All MDG regions11 of the world have experienced considerable reductions in maternal mortality. This section describes estimated MMRs, global maternal deaths, and adult lifetime risk of maternal mortality (i.e. the probability that a 15-year-old woman will die eventually from a maternal cause). It then examines trends in these indicators since 1990. The numbers provided are the most accurate point-estimates possible given the available data. However, these calculations still contain a level of uncertainty that varies depending on the amount and quality of available data used to produce them. The range that an estimated indicator’s true value most likely falls within is captured by its 80% uncertainty interval (see Box 4, Chapter 2). Uncertainty intervals (UI) are therefore given after all MMR point-estimates and MMR reduction point-estimates below. 3.1 Maternal mortality estimates for 2015 An estimated 303 000 maternal deaths will occur globally in 2015, yielding an overall MMR of 216 (UI 207 to 249) maternal deaths per 100 000 live births for the 183 countries and territories covered in this analysis (i.e. all those with a population higher than 100 000) (see Table 2). The global lifetime risk of maternal mortality is approximately 1 in 180 for 2015. Table 2 provides point-estimates of global and regional maternal mortality indicators, and the range of uncertainty for each MMR point-estimate. For the purpose of categorization, MMR is considered to be high if it is 300–499, very high if it is 500–999 and extremely high if it is ≥ 1000 maternal deaths per 100 000 live births. 11 An explanation of the MDG regions is available at: http://mdgs.un.org/unsd/mdg/Host.aspx?Content=Data/REgionalGroupings.htm (a list of the MDG regions is also provided in the full report). 16 Table 2. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk, by United Nations Millennium Development Goal (MDG) region, 2015 MDG region MMRa Range of MMR Number of Lifetime risk of uncertainty (80% UI) maternal maternal death, deathsb 1 in:c Lower Upper estimate estimate World 216 207 249 303 000 180 Developed regionsd 12 11 14 1 700 4 900 Developing regions 239 229 275 302 000 150 Northern Africae 70 56 92 3 100 450 Sub-Saharan Africaf 546 511 652 201 000 36 Eastern Asiag 27 23 33 4 800 2 300 Eastern Asia excluding China 43 24 86 378 1 500 Southern Asiah 176 153 216 66 000 210 Southern Asia excluding India 180 147 249 21 000 190 South-eastern Asiai 110 95 142 13 000 380 Western Asiaj 91 73 125 4 700 360 Caucasus and Central Asiak 33 27 45 610 1 100 Latin America and the Caribbean 67 64 77 7 300 670 Latin Americal 60 57 66 6 600 760 Caribbeanm 175 130 265 1 300 250 Oceanian 187 95 381 500 150 UI: uncertainty interval. a. MMR estimates have been rounded according to the following scheme: < 100 rounded to nearest 1; 100–999 rounded to nearest 1; and ≥ 1000 rounded to nearest 10. b. Numbers of maternal deaths have been rounded according to the following scheme: < 100 rounded to nearest 1; 100– 999 rounded to nearest 10; 1000–9999 rounded to nearest 100; and ≥ 10 000 rounded to nearest 1000. 17 c. Lifetime risk numbers have been rounded according to the following scheme: < 100 rounded to nearest 1; 100–999 rounded to nearest 10; and ≥ 1000 rounded to nearest 100. d. Albania, Australia, Austria, Belarus, Belgium, Bosnia and Herzegovina, Bulgaria, Canada, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan, Latvia, Lithuania, Luxembourg, Malta, Montenegro, Netherlands, New Zealand, Norway, Poland, Portugal, Republic of Moldova, Romania, Russian Federation, Serbia, Slovakia, Slovenia, Spain, Sweden, Switzerland, The former Yugoslav Republic of Macedonia, Ukraine, United Kingdom, United States of America. e. Algeria, Egypt, Libya, Morocco, Tunisia. f. Angola, Benin, Botswana, Burkina Faso, Burundi, Cameroon, Cabo Verde, Central African Republic, Chad, Comoros, Congo, Côte d’Ivoire, Democratic Republic of the Congo, Djibouti, Equatorial Guinea, Eritrea, Ethiopia, Gabon, Gambia, Ghana, Guinea, Guinea-Bissau, Kenya, Lesotho, Liberia, Madagascar, Malawi, Mali, Mauritania, Mauritius, Mozambique, Namibia, Niger, Nigeria, Rwanda, Sao Tome and Principe, Senegal, Sierra Leone, Somalia, South Africa, South Sudan, Sudan, Swaziland, Togo, Uganda, United Republic of Tanzania, Zambia, Zimbabwe. g. China, Democratic People’s Republic of Korea, Mongolia, Republic of Korea. h. Afghanistan, Bangladesh, Bhutan, India, Iran (Islamic Republic of), Maldives, Nepal, Pakistan, Sri Lanka. i. Brunei Darussalam, Cambodia, Indonesia, Lao People’s Democratic Republic, Malaysia, Myanmar, Philippines, Singapore, Thailand, Timor-Leste, Viet Nam. j. Bahrain, Iraq, Jordan, Kuwait, Lebanon, Occupied Palestinian Territory, Oman, Qatar, Saudi Arabia, Syrian Arab Republic, Turkey, United Arab Emirates, Yemen. k. Armenia, Azerbaijan, Georgia, Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, Uzbekistan. l. Argentina, Belize, Bolivia (Plurinational State of), Brazil, Chile, Colombia, Costa Rica, Ecuador, El Salvador, Guatemala, Guyana, Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru, Suriname, Uruguay, Venezuela (Bolivarian Republic of). m. Bahamas, Barbados, Cuba, Dominican Republic, Grenada, Haiti, Jamaica, Puerto Rico, Saint Lucia, Saint Vincent and the Grenadines, Trinidad and Tobago. n. Fiji, Kiribati, Micronesia (Federated States of), Papua New Guinea, Samoa, Solomon Islands, Tonga, Vanuatu. Regional estimates The overall MMR in developing regions is 239 (UI 229 to 275), which is roughly 20 times higher than that of developed regions, where it is just 12 (UI 11 to 14) (see Table 2). Sub-Saharan Africa has a very high MMR12 with a point-estimate of 546 (UI 511 to 652). Three regions – Oceania (187; UI 95 to 381), Southern Asia (176; UI 153 to 216) and South-eastern Asia (110; UI 95 to 142) – have moderate MMR. The remaining five regions have low MMR. Developing regions account for approximately 99% (302 000) of the estimated global maternal deaths in 2015, with sub-Saharan Africa alone accounting for roughly 66% (201 000), followed by Southern Asia (66 000). Among the developing regions, the fewest maternal deaths (an estimated 500) occurred in Oceania. The lifetime risk of maternal mortality is estimated at 1 in 36 in sub-Saharan Africa, contrasting sharply with approximately 1 in 4900 in developed countries. Developing regions with the lowest lifetime risk are Eastern Asia (1 in 2300) and Caucasus and Central Asia (1 in 1100). Table 3 shows the number of maternal deaths, MMR and percentage of AIDS-related indirect maternal deaths by MDG region in 2015. Annex 7 provides the percentage of AIDS-related indirect maternal deaths by country, for countries with an HIV prevalence of 5% or more among adults aged 12 Extremely high MMR (maternal deaths per 100 000 live births) is considered to be ≥ 1000, very high MMR is 500–999, high MMR is 300–499, moderate MMR is 100–299, and low MMR is < 100. 18 15–49 years between 1990 and 2015. Sub-Saharan Africa accounts for the largest proportion (85%) of the nearly 4700 AIDS-related indirect maternal deaths globally in 2015. The proportion of AIDS-related indirect maternal deaths in sub-Saharan Africa is 2.0%, yielding an AIDS-related indirect MMR for sub-Saharan Africa of 11 maternal deaths per 100 000 live births. Without HIV, the MMR for sub-Saharan Africa in 2015 would be 535 maternal deaths per 100 000 live births. Two other regions are estimated to have had more than 100 maternal deaths attributed to HIV in 2015: Southern Asia (310) and South-eastern Asia (150). Table 3. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths and AIDS-related indirect maternal deaths, by United Nations Millennium Development Goal (MDG) region, 2015 a MDG region MMR Number of AIDS-related Number of Percentage of c maternal indirect MMR AIDS-related AIDS-related b deaths indirect indirect maternal maternal deaths deaths World 216 303 000 3 4 700 1.6 d Developed regions 12 1 700 1 87 5.1 Developing regions 239 302 000 4 4 600 1.5 e Northern Africa 70 3 100 0 10 0.3 f Sub-Saharan Africa 546 201 000 11 4 000 2.0 g Eastern Asia 27 4 800 0 43 0.9 Eastern Asia 378 0 0 excluding China 43 0.0 h Southern Asia 176 66 000 1 310 0.5 Southern Asia 21 000 0 25 excluding India 180 0.1 i South-eastern Asia 110 13 000 1 150 1.2 j Western Asia 91 4 700 0 5 0.1 Caucasus and Central k 610 0 8 Asia 33 1.3 19 a MDG region MMR Number of AIDS-related Number of Percentage of c maternal indirect MMR AIDS-related AIDS-related b deaths indirect indirect maternal maternal deaths deaths Latin America and the 7 300 1 71 Caribbean 67 0.9 l Latin Americai 60 6 000 1 51 0.9 m Caribbean 175 1 300 3 20 1.5 n Oceania 187 500 1 3 0.6 a. MMR estimates have been rounded according to the following scheme: < 100 rounded to nearest 1; 100–999 rounded to nearest 1; and ≥ 1000 rounded to nearest 10. b. Numbers of maternal deaths have been rounded according to the following scheme: < 100 rounded to nearest 1; 100– 999 rounded to nearest 10; 1000–9999 rounded to nearest 100; and ≥ 10 000 rounded to nearest 1000. c. According to the Joint United Nations Programme on HIV/AIDS (UNAIDS), AIDS-related deaths (including AIDS-related indirect maternal deaths) include the estimated number of deaths related to HIV infection, including deaths that occur before reaching the clinical stage classified as AIDS. d–n see footnotes for Table 2. Annexes 8, 9, 10, 11, 12, 13, 14, 15, 16 and 17 present the MMR point-estimates, range of uncertainty, numbers of maternal deaths and lifetime risk for WHO, UNICEF, UNFPA, World Bank Group and UNPD regions, respectively. Country-level estimates Annex 7 provides each country’s 2015 maternal mortality indicator point-estimates, and MMR uncertainty intervals. Figure 1 displays a map with all countries shaded according to MMR levels. 20 Figure 1. Maternal mortality ratio (MMR, maternal deaths per 100 000 live births), 2015 Sierra Leone is estimated to have the highest MMR at 1360 (UI 999 to 1980) deaths per 100 000 live births in 2015. Eighteen other countries, all in sub-Saharan Africa, are estimated to have very high MMR in 2015, with estimates ranging from 999 down to 500: Central African Republic (882; UI 508 to 1500), Chad (856; UI 560 to 1350), Nigeria (814; UI 596 to 1180), South Sudan (789; UI 523 to 1150), Somalia (732; UI 361 to 1390), Liberia (725; UI 527 to 1030), Burundi (712; UI 471 to 1050), Gambia (706; UI 484 to 1030), Democratic Republic of the Congo (693; UI 509 to 1010), Guinea (679; UI 504 to 927), Côte d’Ivoire (645; UI 458 to 909), Malawi (634; UI 422 to 1080), Mauritania (602; UI 399 to 984), Cameroon (596; UI 440 to 881), Mali (587; UI 448 to 823), Niger (553; UI 411 to 752), Guinea-Bissau (549; UI 273 to 1090) and Kenya (510; UI 344 to 754). Only two countries in sub-Saharan Africa – Mauritius (53; UI 38 to 77) and Cabo Verde (42; UI 20 to 95) – have low MMR. Three countries outside the sub-Saharan African region have high MMR: Afghanistan (396; UI 253 to 620), Yemen (385; UI 274 to 582) and Haiti (359; UI 236 to 601). Nigeria and India account for over one third of all global maternal deaths in 2015, with an approximate 58 000 (UI 42 000 to 84 000) maternal deaths (19%) and 45 000 (UI 36 000 to 56 000) maternal deaths (15%), respectively. Ten countries account for nearly 59% of global maternal deaths. In addition to Nigeria and India, they are (in descending order of numbers of maternal deaths): Democratic Republic of the Congo (22 000; UI 16 000 to 33 000), Ethiopia (11 000; UI 7900 to 18 000), Pakistan (9700; UI 6100 to 15 000), United Republic of Tanzania (8200; UI 5800 to 12 000), Kenya (8000; UI 5400 to 12 000), Indonesia (6400; UI 4700 to 9000), Uganda (5700; UI 4100 to 8200) and Bangladesh (5500; UI 3900 to 8800). Regarding lifetime risk of maternal mortality, the two countries with the highest estimates are 21 Sierra Leone with an approximate lifetime risk of 1 in 17, and Chad with an approximate lifetime risk of 1 in 18. The estimated risk in high-income countries is 1 in 3300 in comparison with 1 in 41 in low-income countries. Annex 7 presents the percentage of AIDS-related indirect maternal deaths by country for countries with an HIV prevalence of at least 5.0% among adults aged 15–49 years, between 1990 and 2015. Although at a regional level the overall proportions of AIDS-related indirect maternal deaths are relatively small, for countries with high HIV prevalence they are substantial. In 2015, there are five countries where 10% or more of maternal deaths are estimated to be AIDS-related indirect maternal deaths: South Africa (32%), Swaziland (19%), Botswana (18%), Lesotho (13%) and Mozambique (11%). 3.2 Trends in MMR from 1990 to 2015 An estimated global total of 13.6 million women have died in the 25 years between 1990 and 2015 due to maternal causes. Over the course of that time, however, the world has made steady progress in reducing maternal mortality. The global MMR has fallen by 44% (UI 33.1% to 47.5%), from the 1990 level of 385 (UI 359 to 427) to the 2015 level of 216 (UI 207 to 249). This translates to a decrease of over 43% in the estimated annual number of maternal deaths, from 532 000 (UI 496 000 to 590 000) in 1990 to 303 000 (UI 291 000 to 349000) in 2015, and a more than halving of the approximate global lifetime risk of a maternal death from 1 in 73 to 1 in 180. Worldwide, MMR declined by an average of 3.0% (UI 2.1% to 3.4%) per year between 2005 and 2015, more than doubling the estimated average annual decline of 1.2% (UI 0.5% to 2.0%) between 1990 and 2000. Table 4 compares estimates of MMR and numbers of maternal deaths at the global and regional levels for 1990 and 2015. 22 Table 4. Comparison of maternal mortality ratio (MMR, maternal deaths per 100 000 live births) and number of maternal deaths, by United Nations Millennium Development Goal (MDG) region, 1990 and 2015 MDG region 1990 2015 % change Average Average Average in MMR annual % annual % annual % between change in change in change in MMRa Maternal MMR Maternal deathsb deaths 1990 and MMR MMR MMR c 2015 between between between 1990 and 1990 and 2000 and 2015 2000 2015 World 385 532 000 216 303 000 44 2.3 1.2 3.0 Developed d regions 23 3 500 12 1 700 48 2.6 3.3 2.2 Developing regions 430 529 000 239 302 000 44 2.4 1.3 3.1 e Northern Africa 171 6 400 70 3 100 59 3.6 4.1 3.2 Sub-Saharan f Africa 987 223 000 546 201 000 45 2.4 1.5 2.9 g Eastern Asia 95 26 000 27 4 800 72 5.0 4.8 5.0 Eastern Asia excluding China 51 590 43 380 16 0.7 –3.0 3.1 h Southern Asia 538 210 000 176 66 000 67 4.5 3.6 5.1 Southern Asia excluding India 495 57 800 180 21 000 64 4.1 2.5 5.1 South-eastern i Asia 320 39 000 110 13 000 66 4.3 4.7 4.0 j Western Asia 160 6 700 91 4 700 43 2.2 2.7 1.9 Caucasus and k Central Asia 69 1 300 33 610 52 3.0 3.1 2.9 Latin America and the 135 16 000 67 7 300 50 2.8 3.1 2.6 23 MDG region 1990 2015 % change Average Average Average in MMR annual % annual % annual % between change in change in change in MMRa Maternal MMR Maternal deathsb deaths 1990 and MMR MMR MMR 2015c between between between 1990 and 1990 and 2000 and 2015 2000 2015 Caribbean l Latin America 124 14 000 60 6 000 52 2.9 3.1 2.8 m Caribbean 276 2 300 175 1 300 37 1.8 2.5 1.4 n Oceania 391 780 187 500 52 3.0 2.9 3.0 a. MMR estimates have been rounded according to the following scheme: < 100 rounded to nearest 1; 100–999 rounded to nearest 1; and ≥ 1000 rounded to nearest 10. b. Numbers of maternal deaths have been rounded according to the following scheme: < 100 rounded to nearest 1; 100– 999 rounded to nearest 10; 1000–9999 rounded to nearest 100; and ≥ 10 000 rounded to nearest 1000. c. Overall change. d–n see footnote in Table 2. Regional estimates Estimated MMR declined across all MDG regions between 1990 and 2015, although the magnitude of the reduction differed substantially between regions (Annex 18). When interpreting change in MMR, one should take into consideration that it is easier to reduce MMR when levels are high than when they are low. The highest decline between 1990 and 2015 was observed in Eastern Asia (72%), followed by Southern Asia (67%), South-eastern Asia (66%), Northern Africa (59%), Caucasus and Central Asia (52%), Oceania (52%), Latin America and the Caribbean (50%), sub-Saharan Africa (45%) and Western Asia (43%). The decline in developed regions was 48%. In the developing regions, the annual rate of MMR reduction was 1.3% (UI 0.6% to 2.0%) between 1990 and 2000, and progress accelerated to an annual rate of 3.1% (UI 2.2% to 3.5%) between 2000 and 2015. Overall, this translates to an estimated 2.4% (UI 1.7% to 2.7%) average yearly reduction over the past 25 years. Eastern Asia experienced the highest estimated annual rate of decline with an average yearly MMR decrease of 5.0% (UI 4.0% to 6.0%) between 1990 and 2015. The lowest estimated annual rate of decline occurred in Western Asia, where MMR decreased by 2.2% (UI 0.8% to 3.4%) per year during the same period. In 1990 there were approximately 1500 AIDS-related indirect maternal deaths in sub-Saharan Africa. Following the trend of the epidemic, these AIDS-related indirect maternal deaths increased in number until 2005 when there were an estimated 12 370 AIDS-related indirect maternal deaths, before decline to an estimated 4700 in 2015. Annexes 8, 10, 12, 14 and 16 present the MMR trends, reduction in MMR between 1990 and 2015, range of uncertainty for reduction estimates, and average annual change in MMR between 1990 and 2015 for WHO, UNICEF, UNFPA, World Bank Group and UNPD regions, respectively. 24 25 Country estimates Annex 19 provides information on MMR trends from 1990 to 2015 for each country. Assessments of national-level progress towards achieving MDG 5A13 (see categories explained in Box 5) were conducted for those 95 countries that started the evaluation period in 1990 with the highest MMR (100 or greater). This cut-off was chosen in order to focus the assessment of progress on those countries with the greatest maternal mortality burden, and due to the difficulty of reducing MMR further in countries where levels were already relatively low in 1990. Of these 95 countries, results strongly14 indicate that 58 experienced a decline in MMR between 1990 and 2015. For the remaining 26 countries, it cannot be confidently concluded whether MMR increased or decreased, however point-estimates suggest that 22 of them likely experienced a decrease and 4 likely experienced an increase. Point-estimates indicate that nine countries achieved at least a 75% reduction in MMR over the 25-year period, meaning that they achieved MDG 5A. These countries are: Maldives (90% reduction in MMR), Bhutan (84%), Cambodia (84%), Cabo Verde (84%), the Islamic Republic of Iran (80%), Timor-Leste (80%), the Lao People’s Democratic Republic (78%), Rwanda (78%) and Mongolia (76%). 3.3 Comparison with previous maternal mortality estimates The results described in this report are the most accurate maternal mortality estimates yet for all years in the 1990–2015 period. Therefore, these 2015 estimates should be used for the interpretation of trends in MMR from 1990 to 2015, rather than extrapolating estimates from previously published estimates. As explained in Chapter 2, these estimates were generated using an improved approach that built directly upon the methods used to produce previously published estimates. In addition to the refined model, updated data and a larger overall global database informed the 2015 estimates, as compared to those previously published. Notably, the publication of new population-based studies from the Democratic Republic of the Congo, Nigeria and, to a lesser extent, Sierra Leone all indicated much higher MMR than was previously estimated for those counties. Given the large populations in the Democratic Republic of the Congo and Nigeria, this has impacted the global-level estimates. The updated methodology adds refinements that allow country-level data to drive estimates as much as possible (rather than the covariates GDP, fertility rate and skilled attendants at birth coverage), and ensure that higher quality data influences estimates more than lower quality data. 13 Reduce by three quarters, between 1990 and 2015, the maternal mortality ratio. 14 With a confidence level of ≥ 90%. 26 4 Assessing progress and setting a trajectory towards ending preventable maternal mortality 4.1 Millennium Development Goal (MDG) 5 outcomes With the aim of improving maternal health, MDG 5 established a target of reducing the 1990 global maternal mortality ratio (MMR) by 75% by 2015 (MDG 5A). Assessing country-level progress towards this target requires examining estimated reductions, while also taking into consideration the range of uncertainty around those estimates. For example, Nigeria’s estimated MMR reduction between 1990 and 2015 is 39.6%, but the 80% uncertainty interval (UI) for that point-estimate spans zero (–5% to 56.3%), which implies that there is a greater than 10% chance that no reduction in Nigeria’s MMR has occurred. There is, therefore, not enough reliable information on maternal mortality in Nigeria to conclude with confidence that the country has made any progress towards the MDG 5A target. Due to this need to consider estimation uncertainty when evaluating progress, the 95 countries with an MMR above 10015 in 1990 have been categorized based on both MMR reduction point-estimates and 80% UI. Box 5 lists the categories and describes the criteria used to assign countries to categories. Countries were placed into the highest category for which they met the criteria. Box 5 Categorization of countries based on evidence for progress in reducing the MMR between 1990 and 2015 Category Criteria Achieved MDG 5A • MMR reduction point-estimate of ≥ 75% • MMR reduction point-estimate of ≥ 50% Making progress AND • ≥ 90% probability of an MMR reduction of ≥ 25% • MMR reduction point-estimate of ≥ 25% Insufficient progress AND • ≥ 90% probability of an MMR reduction of ≥ 0% • MMR reduction point-estimate of < 25% OR No progress • a 90% probability that there has been no reduction in MMR, or there has been an increase in MMR Among those 95 countries, the 9 countries with an estimated MMR reduction between 1990 and 2015 of 75% or more have achieved MDG 5A – they have been placed in the first category. The second category, those countries that are making progress, includes 39 countries with an estimated MMR reduction of 50% or more, and at least a 90% chance that the true reduction was above 25%. The third category, countries making insufficient progress, comprises 21 countries with an 15 The MMR cut-off of 100 maternal deaths per 100 000 live births was chosen in order to focus the assessment of progress on countries that started with a relatively high level of maternal mortality in 1990, and due to the difficulty of reducing MMR further in countries where levels were already relatively low (< 100) in 1990. 27 estimated MMR reduction of 25% or more, and at least a 90% chance that the true reduction was above zero. The fourth and final category includes 26 countries that have made no progress; they have an estimated MMR reduction of less than 25%, or a greater than 10% chance that no reduction has occurred at all. Given the variability of maternal mortality reporting methods and data quality, these categories provide the best possible assessment of likely performance on the MDG 5A target. Annex 18 displays category labels for all 95 countries. The nine countries which are considered to have achieved MDG 5A based on point-estimates are: Bhutan, Cambodia, Cabo Verde, the Islamic Republic of Iran, the Lao People’s Democratic Republic, Maldives, Mongolia, Rwanda and Timor-Leste. Yet, among these countries there is substantial variation in the level of certainty of this achievement. As indicated by uncertainty intervals (only Cambodia and Maldives have a greater than 90% likelihood of having a true MMR reduction of 75% or more. For the other seven, a 10% or greater chance of not having achieved the target persists. The consideration of uncertainty regarding rates of reduction is intended to demonstrate the need for more rigorous data collection. Differences in the sizes of UIs are due to differences in the quality of data used to inform estimates. For example, the Islamic Republic of Iran and Maldives had substantial maternal mortality data from civil registration and vital statistics (CRVS) systems and surveillance studies available for inclusion in the estimation model, while others, such as Cabo Verde, Lao People’s Democratic Republic and Timor-Leste, had little to no country-level data. While no MDG region achieved the target of reducing maternal mortality by 75% (see Table 4), all demonstrated substantial progress, particularly after announcement of the MDGs in 2000 – the estimated global 2000–2015 annual reduction rate of 3% was significantly increased in comparison to the 1990–2000 rate of 1.2%. This acceleration of progress reflects a widespread escalation of efforts to reduce maternal mortality, stimulated by MDG 5. Maternal mortality has proved to be a valuable indicator both for tracking development progress and for spurring action to improve maternal health. 4.2 Looking towards the future The Sustainable Development Goals (SDGs) now call for an acceleration of current progress in order to achieve a global MMR of 70 maternal deaths per 100 000 live births, or less, by 2030, working towards a vision of ending all preventable maternal mortality. Achieving this global goal will require countries to reduce their MMR by at least 7.5% each year between 2016 and 2030. Based on their point-estimates for average annual reduction, three countries with an MMR greater than 100 nearly reached or exceeded this reduction rate between 2000 and 2015: Cambodia (7.4%; UI 5.4% to 9.5%), Rwanda (8.4%; UI 6.5% to 10.6%) and Timor-Leste (7.8%; UI 5.7% to 10.2 %). The recent success of these countries in rapidly reducing maternal mortality demonstrates that this goal is achievable. Global targets for ending preventable maternal mortality (EPMM): By 2030, every country should reduce its maternal mortality ratio (MMR) by at least two thirds from the 2010 baseline, and no country should have an MMR higher than 140 deaths per 100 000 live births (twice the global target) (4). 28 While differing contexts make issuing prescribed reduction strategies impossible, examining the strategies employed by successful countries can illuminate routes that other countries may find useful. However, the 30 countries with the highest MMRs in 2015 will have to achieve substantially higher annual rates of reduction to attain MMRs below 140 in 2030. Projections indicate that accomplishing this target will result in over 60% fewer deaths in 2030 than the estimated number in 2015, and will save a cumulative 2.5 million women’s lives between 2016 and 2030, as compared to a situation in which current reduction trajectories remain unchanged (14). Strategies for success and challenges to address Drivers of success in reducing maternal mortality range from making improvements at the provider and health system level to implementing interventions aimed at reducing social and structural barriers. Box 6 describes several key strategies used by countries that have demonstrated success in improving maternal survival. These strategies are situated within a recently developed strategic framework for policy and programme planning that is informed by the guiding principles of: (1) empowering women, girls and communities, (2) protecting and supporting the mother–baby dyad, (3) ensuring country ownership, leadership and supportive legal, technical and financial frameworks, and (4) applying a human rights framework to ensure that high-quality reproductive, maternal and newborn health care is available, accessible and acceptable to all who need it (4). 29 Box 6 Strategies driving success in reducing maternal mortality WHO’s recently published Strategies towards ending preventable maternal mortality (EPMM) establishes a strategic framework that specifies five objectives (4). Below, for each of these objectives, examples are presented of strategies implemented by countries that have made significant reductions in maternal mortality: 1. Addressing inequities in access to and quality of sexual, reproductive, maternal and newborn health care • Ethiopia trained women’s association members in strategies for addressing social and structural barriers to sexual, reproductive, maternal and newborn health, and also trained health managers on gender mainstreaming in their areas of work (25). • Viet Nam developed sexual and reproductive health services specifically for adolescents and youths (25). 2. Ensuring universal health coverage for comprehensive sexual, reproductive, maternal and newborn health care • Rwanda used a community-based health insurance scheme to ensure vulnerable populations’ access to maternal and child health services (26). • Bangladesh expanded access to maternity services in new, private-sector health-care facilities (27). 3. Addressing all causes of maternal mortality, reproductive and maternal morbidities, and related disabilities • Nepal expanded access to modern family planning methods, and increased school attendance and literacy rates among women and girls (28). • The Maldives strengthened emergency obstetric care, including basic care and comprehensive emergency obstetric care throughout the country’s health system (29). 4. Strengthening heath systems to respond to the needs and priorities of women and girls • Indonesia invested in the training of midwives and the creation of dedicated, village-level delivery points for maternal health services (30). • Cambodia invested in transport infrastructure and construction of health-care facilities staffed with an expanded cadre of trained midwives throughout the country, including maternity waiting houses and extended delivery rooms (31). 5. Ensuring accountability to improve quality of care and equity • Mongolia introduced procedures at the facility, provincial and ministerial levels to ensure maternal deaths were reported within a 24-hour period and transmitted to the Ministry of Health for review (32). • India developed guidelines for maternal death audits and near-miss analyses (25). Examining countries that experienced little to no reduction in maternal mortality since 1990 reveals several prevalent factors that impede progress. Among the 27 countries categorized as likely having made “no progress”, 23 are particularly impacted by the HIV epidemic. Despite the recent positive influence of antiretroviral medications on AIDS-related indirect maternal mortality, overall the epidemic poses immense challenges to maternal mortality reduction due to the strain it places on 30 health systems and infrastructure, in addition to its direct health impacts. Emergent humanitarian settings and situations of conflict, post-conflict and disaster also significantly hinder progress. Indeed, 76% of high maternal mortality countries (those with MMR ≥ 300) are defined as fragile states (33). In such situations, the breakdown of health systems can cause a dramatic rise in deaths due to complications that would be easily treatable under stable conditions. At the peak of the 2014–2015 Ebola virus disease outbreak in Liberia, for example, maternal health service utilization dropped precipitously and common obstetric complications went untreated out of fear of disease transmission (34). Compounding the tragedy of lives lost in crisis settings, many of these deaths go unrecorded. Settings where the needs are greatest are also those with the least evidence and analysis. In countries designated as fragile states, the estimated lifetime risk of maternal mortality is 1 in 54. Many of the most vulnerable populations are not represented in the current global data. Moreover, even within countries with good overall progress indicators, the optimistic numbers often mask extreme disparities. Australia, for example, determined through a specialized study that the MMR among Aboriginal and Torres Strait Islander women was over twice that of non-indigenous women. Marginalized subpopulations often lack representation in the data, and disparities may not be evident without disaggregating data. This lack of accurate information makes it nearly impossible to determine how to best address the maternal health needs among the most vulnerable. An emerging challenge is increasing late maternal mortality, a phenomenon referred to as part of the “obstetric transition” (35). Late maternal mortality refers to maternal deaths that occur more than 42 days but less than one year after termination of pregnancy. As health systems improve and are better able to manage immediate childbirth complications, deaths within the first 48 hours of delivery may be averted, but the proportion of morbidity and mortality caused by late maternal sequelae or late maternal complications can also increase. This trend has been observed in several countries, such as Mexico where late maternal deaths account for up to 15% of overall maternal mortality (36). Further analyses of this subset of deaths is warranted. Monitoring all maternal deaths thus proves increasingly important for ensuring accurate documentation to detect shifting dynamics in maternal health. Need for improved measurement and data Impressive efforts to establish and improve CRVS systems or implement alternative methods of rigorously recording maternal deaths have been made in recent years. Box 7 provides examples of several methods countries are using to dramatically improve data collection. The high-quality data generated even prompted the use for this report of a refined estimation methodology, one that fully utilizes country-level data to produce a more accurate and realistic picture of global maternal mortality trends than ever before. 31 Box 7 Tools for improving data collection Confidential Enquiry into Maternal Deaths (CEMD) Within established civil registration and vital statistics (CRVS) systems, CEMD facilitates investigation of and correction for underreporting of maternal deaths due to misclassification. Developed in England and Wales and conducted continuously there since 1952 (37), CEMD involves having multiple experts review all potential maternal mortality cases in detail, assessing the accuracy of classifications applied as well as examining the circumstances of the death. It thus also helps to identify areas for action to prevent future deaths. Kazakhstan and South Africa both recently conducted CEMD studies, identifying 29% and 40% more maternal deaths, respectively, than were initially recorded in the CRVS system. Maternal Death Surveillance and Response (MDSR) At the health-care facility level, MDSR systems promote a continuous action cycle for monitoring of maternal deaths, identifying trends in and causes of maternal mortality, and acting to prevent future deaths (38). Information generated by MDSR can be communicated upwards from facilities, to be aggregated at the regional and national levels. Where national CRVS systems have not yet been established, MDSR serves as a building block for a comprehensive, national-level data collection system. Countries that have recently established, strengthened or expanded MDSR systems include Cameroon, the Democratic Republic of the Congo, India, Morocco, Nigeria and Togo (25). Digital innovations Given the high percentage of births and maternal deaths that occur outside of health-care facilities, there is a critical need to obtain and communicate vital events data from the community level. Digital solutions delivered via mobile devices (mHealth tools) that connect frontline health workers to national health systems can simultaneously improve health-care service delivery, strengthen accountability, and generate real-time data (39). A growing proportion of these digital tools focus on registration of pregnancies and notification of births and deaths, linking information directly to facility-, district- and national-level health management and vital events statistical systems (40). One example is the Open Smart Register Platform, or OpenSRP (41). Pilot tests of OpenSRP and similar digital tools are under way in Bangladesh, India, Indonesia, Pakistan and South Africa. Yet, while the estimates presented in this report provide valuable policy and programme planning guidance, they cannot change the fact that many women who die from maternal causes still go uncounted. Taking effective action to prevent future maternal deaths requires knowing who has died and why they died. Respect for human rights and human life necessitates improved record-keeping so that all births, deaths and causes of death are officially accounted for. For these reasons, improving metrics, measurement systems and data quality is a crucial cross-cutting action for all strategies aimed at ensuring maternal survival (4). The broad uncertainty intervals associated with the estimates presented throughout this report directly reflect the critical need for better data on maternal mortality. Governments are called upon to establish well functioning CRVS systems with accurate attribution of cause of death. Improvements in measurement must be driven by action at the country level, with governments creating systems to capture data specific to their information needs; systems that must also meet the standards required for international comparability. Globally, standardized methods for 32 preventing underreporting should be established to enhance international comparability. Finally, data that can be disaggregated to examine trends and measure the mortality burden within the most vulnerable and most frequently overlooked populations are critical for implementing strategies to address inequities and accelerate progress towards maternal mortality reduction. Populations requiring particular attention include refugees and groups that face discrimination or stigma. Better data on the maternal mortality burden among adolescent girls is also needed; maternal causes rank second among causes of death for girls aged 15–19 (42). Several countries, particularly those in Latin America and the Caribbean, and in South-East Asia, have already begun reporting data for women and girls outside the standard 15–49 year age interval, documenting the disturbing fact that maternal deaths are occurring among girls even younger than 15. 4.3 A call to action The announcement of MDG 5 in 2000 attracted intense scrutiny of the shamefully high numbers of women dying from maternal causes. It initiated an unprecedented and ongoing global conversation about how maternal mortality should be measured, what strategies could be employed to save lives, and how the progress of these reduction efforts would be assessed. Accurate measurement of maternal mortality levels remains an immense challenge, but the overall message is clear: hundreds of thousands of women are still dying during childbirth or from pregnancy-related causes each year. The goal of ending preventable maternal mortality is a call to action across all regions of the globe, developed and developing, including areas where substantial progress has already been made. Among countries where maternal death counts remain high, the challenge is clear. Efforts to save lives must be accelerated and must also be paired with country-driven efforts to accurately count lives and record deaths. Among those countries with low overall maternal mortality indicators, the next challenge is measuring and amending inequities among subpopulations. Across varying settings, strategies must be both context-specific and thoroughly grounded in a human rights approach. With rapid acceleration of the efforts and progress catalysed by MDG 5, ending preventable maternal mortality on a global level can be achieved by 2030. 33 References 1. Ki-Moon B. Global strategy for women’s and children’s health. New York (NY): United Nations; 2010. 2. Keeping promises, measuring results: commission on information and accountability for women’s and children’s health. Geneva: World Health Organization; 2011. 3. Transforming our world: the 2030 Agenda for Sustainable Development 2015. Resolution adopted by the General Assembly on 25 September 2015. United Nations General Assembly, Seventieth session; 2015 (A/RES/70/1; http://www.un.org/ga/search/view_doc.asp?symbol=A/RES/70/1, accessed 5 November 2015). 4. Strategies towards ending preventable maternal mortality (EPMM). Geneva: World Health Organization; 2015 (http://www.everywomaneverychild.org/images/EPMM_final_report_2015.pdf, accessed 5 November 2015). 5. Every woman, every child: from commitments to action: the first report of the independent Expert Review Group (iERG) on Information and Accountability for Women's and Children's Health. Geneva: World Health Organization; 2012 (www.who.int/woman_child_accountability/ierg/reports/2012/IERG_report_low_resolution.pdf, accessed 5 November 2015). 6. World Health Organization (WHO), United Nations Children’s Fund (UNICEF). Revised 1990 estimates of maternal mortality: a new approach by WHO and UNICEF. Geneva: WHO; 1996 (http://apps.who.int/iris/bitstream/10665/63597/1/WHO_FRH_MSM_96.11.pdf, accessed 5 November 2015). 7. AbouZahr C, Wardlaw T, Hill K. Maternal mortality in 1995: estimates developed by WHO UNICEF UNFPA. Geneva: World Health Organization; 2001. 8. AbouZahr C, Wardlaw TM, Hill K, Choi Y. Maternal mortality in 2000: estimates developed by WHO, UNICEF and UNFPA. Geneva: World Health Organization; 2004 (http://apps.who.int/iris/bitstream/10665/68382/1/a81531.pdf, accessed 5 November 2015). 9. World Health Organization (WHO), United Nations Children’s Fund (UNICEF), United Nations Population Fund (UNFPA) and The World Bank. Trends in maternal mortality: 1990 to 2008. Estimates developed by WHO, UNICEF, UNFPA and The World Bank. Geneva: WHO; 2010 (http://apps.who.int/iris/bitstream/10665/44423/1/9789241500265_eng.pdf, accessed 5 November 2015). 10. World Health Organization (WHO), United Nations Children’s Fund (UNICEF), United Nations Population Fund (UNFPA), The World Bank. Trends in maternal mortality: 1990 to 2010: WHO, UNICEF, UNFPA and The World Bank estimates. Geneva: WHO; 2012 (http://apps.who.int/iris/bitstream/10665/44874/1/9789241503631_eng.pdf, accessed 5 November 2015). 11. World Health Organization (WHO), United Nations Children’s Fund (UNICEF), United Nations Population Fund (UNFPA), The World Bank, United Nations Population Division. Trends in maternal mortality: 1990 to 2013. Estimates by WHO, UNICEF, UNFPA, The World Bank and the United Nations Population Division. Geneva: WHO; 2014 (http://apps.who.int/iris/bitstream/10665/112682/2/9789241507226_eng.pdf, accessed 5 November 34 2015). 12. Wilmoth J, Mizoguchi N, Oestergaard M, Say L, Mathers C, Zureick-Brown S, et al. Levels and trends of maternal mortality in the world: the development of new estimates by the United Nations. Technical report (submitted to the WHO, UNICEF, UNFPA, and The World Bank). 2010 (http://www.who.int/reproductivehealth/publications/monitoring/MMR_technical_report.pdf, accessed 12 November 2015). 13. Wilmoth J, Mizoguchi N, Oestergaard M, Say L, Mathers C, Zureick-Brown S, et al. A new method for deriving global estimates of maternal mortality: supplemental report. Stat Politics Policy. 2012;3(2):1-38. 14. Alkema L, Chou D, Hogan D, Zhang S, Moller A, Gemmill A, et al. National, regional, and global levels and trends in maternal mortality between 1990 and 2015 with scenario-based projections to 2030: a systematic analysis by the United Nations Maternal Mortality Estimation Inter-Agency Group. Lancet. 2015 (in press). 15. Alkema L, Zhang S, Chou D, Gemmill A, Moller A, Ma Fat D, et al. A Beyesian approach to the global estimation of maternal mortality. 2015 (submitted for peer review; http://arxiv.org/abs/1511.03330). 16. World population prospects: the 2015 revision. New York (NY): United Nations, Department of Economic and Social Affairs, Population Division; 2015 (http://esa.un.org/unpd/wpp/, accessed 9 November 2015). 17. Life tables for WHO Member States 1990–2012. Geneva: World Health Organization; 2014. 18. Data Catalog. Washington (DC): The World Bank; 2013. 19. UNICEF Data: Monitoring the Situation of Children and Women [website]. New York (NY): United Nations Children’s Fund; 2015 (http://data.unicef.org/, accessed 5 November 2015). 20. Global report: UNAIDS report on the global AIDS epidemic 2013. Geneva: Joint United Nations Programme on HIV/AIDS; 2013. 21. Mortality and burden of disease estimates for WHO Member States in 2008. Geneva: World Health Organization; 2011. 22. Plummer M, editor. JAGS: a program for analysis of Bayesian graphical models using Gibbs sampling. In: Proceedings of the 3rd international workshop on distributed statistical computing. Vienna: Technische Universität Wien; 2003. 23. R Core Team. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2013 (http://www.R-project.org, accessed 15 September 2015). 24. Gelman A, Hill J. Data analysis using regression and multilevel/hierarchical models. Cambridge University Press; 2006. 25. H4+ Partnership. The H4+ partnership: joint country support to improve women’s and children’s health: progress report. Geneva: World Health Organization; 2015. 26. Worley H. Rwanda's success in improving maternal health 2015. In: Population Reference Bureau (PRB) Publications [website]. Washington (DC): PRB; 2015 (http://www.prb.org/Publications/Articles/2015/rwanda-maternal-health.aspx, accessed 5 35 November 2015). 27. El Arifeen S, Hill K, Ahsan KZ, Jamil K, Nahar Q, Streatfield PK. Maternal mortality in Bangladesh: a Countdown to 2015 country case study. Lancet. 2014;384(9951):1366-74. 28. Nepal Ministry of Health and Population Nepal, Partnership for Maternal, Newborn & Child Health, World Health Organization (WHO), World Bank and Alliance for Health Policy and Systems Research. Success factors for women's and children's health: Nepal. Geneva: WHO; 2015 (http://www.who.int/pmnch/knowledge/publications/nepal_country_report.pdf, accessed 5 November 2015). 29. Maternal and Perinatal Morbidity and Mortality Review Committee. Maternal deaths in the Maldives: 2009-2011. The Maldives Government; 2011. 30. Van Lerberghe W, Matthews Z, Achadi E, Ancona C, Campbell J, Channon A, et al. Country experience with strengthening of health systems and deployment of midwives in countries with high maternal mortality. Lancet. 2014;384(9949):1215-25. 31. Cambodia reduces maternal mortality. In: WHO in the Western Pacific [website]. Manila: World Health Organization Western Pacific Regional Office; 2015 (http://www.wpro.who.int/about/administration_structure/dhs/story_cambodia_reduces_maternal _mortality/en/, accessed 5 November 2015). 32. Yadamsuren B, Merialdi M, Davaadorj I, Requejo JH, Betrán AP, Ahmad A, et al. Tracking maternal mortality declines in Mongolia between 1992 and 2007: the importance of collaboration. Bull World Health Organ. 2010;88(3):192-8. 33. Organisation for Economic Co-operation and Development (OECD). States of fragility 2015: Meeting post-2015 ambitions. Paris: OECD Publishing; 2015. 34. Iyengar P, Kerber K, Howe CJ, Dahn B. Services for mothers and newborns during the Ebola outbreak in Liberia: the need for improvement in emergencies. PLoS currents. 2014;7. 35. Souza J, Tunçalp Ö, Vogel J, Bohren M, Widmer M, Oladapo O, et al. Obstetric transition: the pathway towards ending preventable maternal deaths. BJOG. 2014;121(s1):1-4. 36. Búsqueda Intencionada de Muertes Maternas en México. Informe 2011. Mexico; Secretaría de Salud México; 2013. 37. Knight M, Kenyon S, Brocklehurst P, Neilson J, Shakespeare J, Kurinczuk J, editors, on behalf of MBRRACE-UK. Saving lives, improving mothers’ care: lessons learned to inform future maternity care from the UK and Ireland Confidential Enquiries into Maternal Deaths and Morbidity 2009-2012. Oxford: National Perinatal Epidemiology Unit, University of Oxford; 2014 (https://www.npeu.ox.ac.uk/downloads/files/mbrrace-uk/reports/Saving%20Lives%20Improving%2 0Mothers%20Care%20report%202014%20Full.pdf, accessed 5 November 2015). 38. MDSR Working Group (Canadian Network for Maternal, Newborn & Child Health, International Federation of Gynecology and Obstetrics, International Stillbirth Alliance, Department for International Development UK, United Nations Population Fund, United States Centers for Disease Control and Prevention and the World Health Organization. Maternal death surveillance and response: technical guidance information for action to prevent maternal death. Geneva: World Health Organization; 2013 (https://www.unfpa.org/sites/default/files/pub-pdf/Maternal_Death_Surveillance_and_Response_0. 36 pdf, accessed 5 November 2015). 39. Mehl G, Labrique A. Prioritizing integrated mHealth strategies for universal health coverage. Science. 2014;345(6202):1284-7. 40. Labrique AB, Pereira S, Christian P, Murthy N, Bartlett L, Mehl G. Pregnancy registration systems can enhance health systems, increase accountability and reduce mortality. Reprod Health Matters. 2012;20(39):113-7. 41. Open Smart Register Platform (OpenSRP) [website]. 2015 (www.smartregister.org, accessed 5 November 2015). 42. Health for the world's adolescents: a second chance in the second decade. Geneva: World Health Organization; 2014. 37 Annexes 38 Annex 1. Summary of the country consultations 2015 The  generation  of  global,  regional  and  country-­‐level  estimates  and  trends  in  morbidity  and  mortality   is  one  of  the  core  functions  of  WHO,  which  is  the  agency  within  the  UN  system  that  leads  the   production  of  updated  maternal  mortality  estimates.  In  2001,  the  WHO  Executive  Board  endorsed  a   resolution  (EB.107.R8)  seeking  to  “establish  a  technical  consultation  process  bringing  together   personnel  and  perspectives  from  Member  States  in  different  WHO  regions”.  A  key  objective  of  this   consultation  process  is  “to  ensure  that  each  Member  State  is  consulted  on  the  best  data  to  be  used”.   Since  the  process  is  an  integral  step  in  the  overall  estimation  strategy,  it  is  described  here  in  brief.   The  country  consultation  process  entails  an  exchange  between  WHO  and  technical  focal  person(s)  in   each  country.  It  is  carried  out  prior  to  the  publication  of  estimates.  During  the  consultation  period,   WHO  invites  focal  person(s)  to  review  input  data  sources,  methods  for  estimation  and  the   preliminary  estimates.  Focal  person(s)  are  encouraged  to  submit  additional  data  that  may  not  have   been  taken  into  account  in  the  preliminary  estimates.   The  country  consultation  process  for  the  2015  round  of  maternal  mortality  estimates  was  initiated   with  an  official  communication  from  WHO  to  all  Member  States  on  25  August  2014.  This  letter   informed  Member  States  of  the  forthcoming  exercise  to  estimate  maternal  mortality  and  requested   the  designation  of  an  official  contact  (typically  within  the  national  health  ministry  and/or  the  central   statistics  office)  to  participate  in  the  consultation.  The  designated  officials  received  the  following   items  by  email:  (1)  a  copy  of  the  official  communication;  (2)  draft  estimates  and  data  sources;  and   (3)  a  summary  of  the  methodology  used.  WHO  regional  offices  actively  collaborated  in  identifying   focal  persons  through  their  networks.   The  formal  consultation  process  was  officially  completed  by  24  July  2015.  Of  the  183  Member  States   included  in  the  analysis,  WHO  received  nominations  of  designated  officials  from  125  –  Regional   Office  for  Africa  (17),  Regional  Office  for  the  Americas  (24),  Regional  Office  for  South-­‐East  Asia  (6),   Regional  Office  for  Europe  (39),  Regional  Office  for  the  Eastern  Mediterranean  (19),  Regional  Office   for  the  Western  Pacific  (20)  –  and  received  feedback,  comments  or  data  from  60  Member  States.   During  the  consultation  period,  new  data  submitted  by  countries  were  reviewed  to  determine   whether  they  met  the  study’s  inclusion  criteria.  Data  were  considered  acceptable  to  use  as  new   input  if  they  were  representative  of  the  national  population  and  referred  to  a  specific  time  interval   within  the  period  from  1985  to  2015.   As  a  result  of  the  country  consultation  and  updated  vital  registration  data,  234  new  or  modified   entries  were  included.  Thus,  the  current  estimates  are  based  on  2608  observations  corresponding  to   3634  country-­‐years  of  information  in  171  countries.   As  in  the  previous  country  consultation,  the  new  observations  were  from  civil  registration  systems   and  surveys;  however,  an  increase  in  number  of  other  new  observations  shows  that  countries   lacking  functioning  civil  registration  systems  are  increasingly  investing  in  monitoring  maternal   mortality  with  empirical  data  from  alternative  sources.     Annex 2. Measuring maternal mortality Concepts and definitions In the International statistical classification of diseases and related health problems, 10th revision (ICD-10),1 WHO defines maternal death as: The death of a woman while pregnant, or within 42 days of termination of pregnancy, irrespective of the duration and the site of the pregnancy, from any cause related to or aggravated by the pregnancy or its management (from direct or indirect obstetric death), but not from accidental or incidental causes. This definition allows identification of maternal deaths, based on their causes, as either direct or indirect. Direct maternal deaths are those resulting from obstetric complications of the pregnant state (i.e. pregnancy, delivery and postpartum), interventions, omissions, incorrect treatment, or a chain of events resulting from any of the above. Deaths due to, for example, obstetric haemorrhage or hypertensive disorders in pregnancy, or those due to complications of anaesthesia or caesarean section are classified as direct maternal deaths. Indirect maternal deaths are those resulting from previously existing diseases, or from diseases that developed during pregnancy and that were not due to direct obstetric causes but aggravated by physiological effects of pregnancy. For example, deaths due to aggravation of an existing cardiac or renal disease are considered indirect maternal deaths. The concept of death during pregnancy, childbirth and the puerperium is included in the ICD-10 and is defined as any death temporal to pregnancy, childbirth or the postpartum period, even if it is due to accidental or incidental causes (this was formerly referred to as “pregnancy-related death”, see Box 1). This alternative definition allows measurement of deaths that are related to pregnancy, even though they do not strictly conform to the standard “maternal death” concept, in settings where accurate information about causes of death based on medical certificates is unavailable. For instance, in population-based surveys, respondents provide information on the pregnancy status of a reproductive-aged sibling at the time of death, but no further information is elicited on the cause of death. These surveys – for example, the Demographic and Health Surveys and Multiple Indicator Cluster Surveys – therefore, usually provide measures of pregnancy-related deaths rather than maternal deaths. Further, complications of pregnancy or childbirth can lead to death beyond the six weeks postpartum period, and the increased availability of modern life-sustaining procedures and technologies enables more women to survive adverse outcomes of pregnancy and delivery, and to delay death beyond 42 days postpartum. Despite being caused by pregnancy-related events, these deaths do not count as maternal deaths in routine civil registration systems. Specific codes for “late maternal deaths” are included in the ICD-10 (O96 and O97) to capture delayed maternal deaths occurring between six weeks and one year postpartum (see Box A2.1). Some countries, particularly those with more developed civil registration systems, use this definition.                                                                                                                         1 International statistical classification of diseases and related health problems, tenth revision. Vol. 2: Instruction manual. Geneva: World Health Organization; 2010. Box A2.1 Definitions related to maternal death in ICD-10 Maternal death The death of a woman while pregnant or within 42 days of termination of pregnancy, irrespective of the duration and site of the pregnancy, from any cause related to or aggravated by the pregnancy or its management (from direct or indirect obstetric death), but not from accidental or incidental causes. Pregnancy-related death The death of a woman while pregnant or within 42 days of termination of pregnancy, irrespective of the cause of death. Late maternal death The death of a woman from direct or indirect obstetric causes, more than 42 days, but less than one year after termination of pregnancy. Coding of maternal deaths Despite the standard definitions noted above, accurate identification of the causes of maternal deaths is not always possible. It can be a challenge for medical certifiers to correctly attribute cause of death to direct or indirect maternal causes, or to accidental or incidental events, particularly in settings where most deliveries occur at home. While several countries apply the ICD- 10 in civil registration systems, the identification and classification of causes of death during pregnancy, childbirth and the puerperium remain inconsistent across countries. With the publication of the ICD-10, WHO recommended adding a checkbox on the death certificate for recording a woman’s pregnancy status at the time of death.2 This was to help identify indirect maternal deaths, but it has not been implemented in many countries. For countries using ICD-10 coding for registered deaths, all deaths coded to the maternal chapter (O codes) and maternal tetanus (A34) are counted as maternal deaths. In 2012, WHO published Application of ICD-10 to deaths during pregnancy, childbirth and the puerperium: ICD maternal mortality (ICD-MM) to guide countries to reduce errors in coding maternal deaths and to improve the attribution of cause of maternal death.3 The ICD-MM is to be used together with the three ICD-10 volumes. For example, the ICD-MM clarifies that the coding of maternal deaths among HIV-positive women may be due to one of the following. • Obstetric causes: Such as haemorrhage or hypertensive disorders in pregnancy – these should be identified as direct maternal deaths. • The interaction between human immunodeficiency virus (HIV) and pregnancy: In these cases, there is an aggravating effect of pregnancy on HIV and the interaction between pregnancy                                                                                                                         2 International statistical classification of diseases and related health problems, tenth revision. Vol. 2: Instruction manual. Geneva: World Health Organization; 2010. 3 Application of ICD-10 to deaths during pregnancy, childbirth and the puerperium: ICD maternal mortality (ICD-MM). Geneva: World Health Organization; 2012. and HIV is the underlying cause of death. These deaths are considered as indirect maternal deaths. In this report, they are referred to as “AIDS-related indirect maternal deaths”, and in the ICD those deaths are coded to O98.7 and categorized in Group 7 (non-obstetric complications) in the ICD-MM. • Acquired immunodeficiency syndrome (AIDS): In these cases, the woman’s pregnancy status is incidental to the course of her HIV infection and her death is a result of an HIV complication, as described by ICD-10 codes B20–24. These are not considered maternal deaths. Thus, proper reporting of the mutual influence of HIV or AIDS and pregnancy in Part 1 of the death certificate will facilitate the coding and identification of these deaths. Measures of maternal mortality The extent of maternal mortality in a population is essentially the combination of two factors: (i) The risk of death in a single pregnancy or a single live birth. (ii) The fertility level (i.e. the number of pregnancies or births that are experienced by women of reproductive age). The MMR is defined as the number of maternal deaths during a given time period per 100 000 live births during the same time period. It depicts the risk of maternal death relative to the number of live births and essentially captures (i) above. By contrast, the maternal mortality rate (MMRate) is defined as the number of maternal deaths in a population divided by the number of women aged 15–49 years (or woman-years lived at ages 15– 49 years). The MMRate captures both the risk of maternal death per pregnancy or per total birth (live birth or stillbirth), and the level of fertility in the population. In addition to the MMR and the MMRate, it is possible to calculate the adult lifetime risk of maternal mortality for women in the population (see Box A2). An alternative measure of maternal mortality, the proportion of maternal deaths among deaths of women of reproductive age (PM), is calculated as the number of maternal deaths divided by the total deaths among women aged 15–49 years. Box A2.2 Statistical measures of maternal mortality Maternal mortality ratio (MMR) Number of maternal deaths during a given time period per 100 000 live births during the same time period. Maternal mortality rate (MMRate) 4 Number of maternal deaths divided by person-years lived by women of reproductive age.                                                                                                                         4 Wilmoth J, Mizoguchi N, Oestergaard M, Say L, Mathers C, Zureick-Brown S, et al. A new method for deriving global estimates of maternal mortality: supplemental report. Stat Politics Policy. 2012;3(2):1–38. Box A2.2 Statistical measures of maternal mortality Adult lifetime risk of maternal death The probability that a 15-year-old woman will die eventually from a maternal cause. The proportion of maternal deaths among deaths of women of reproductive age (PM) The number of maternal deaths in a given time period divided by the total deaths among women aged 15–49 years. Approaches for measuring maternal mortality Ideally, civil registration systems with good attribution of cause of death provide accurate data on the level of maternal mortality and the causes of maternal deaths. In countries with incomplete civil registration systems, it is difficult to accurately measure levels of maternal mortality. First, it is challenging to identify maternal deaths precisely, as the deaths of women of reproductive age might not be recorded at all. Second, even if such deaths were recorded, the pregnancy status or cause of death may not have been known and the deaths would therefore not have been reported as maternal deaths. Third, in most developing-country settings where medical certification of cause of death does not exist, accurate attribution of a female death as a maternal death is difficult. Even in developed countries where routine registration of deaths is in place, maternal deaths may be underreported due to misclassification of ICD-10 coding, and identification of the true numbers of maternal deaths may require additional special investigations into the causes of death. A specific example of such an investigation is the Confidential Enquiry into Maternal Deaths (CEMD), a system established in England and Wales in 1928.5,6,7 The most recent report of the CEMD (for 2009–2011) identified 79% more maternal deaths than were reported in the routine civil registration system.8 Other studies on the accuracy of the number of maternal deaths reported in civil registration systems have shown that the true number of maternal deaths could be twice as high as indicated by routine reports, or even more.9,10 Annex 6 summarizes the results of a                                                                                                                         5 Lewis G, editor. Why mothers die 2000–2002: the confidential enquiries into maternal deaths in the United Kingdom. London: RCOG Press; 2004. 6 Lewis G, editor. Saving mothers’ lives: reviewing maternal deaths to make motherhood safer 2003–2005. The seventh report on confidential enquiries into maternal deaths in the United Kingdom. London: Confidential Enquiry into Maternal and Child Health (CEMAH); 2007. 7 Centre for Maternal and Child Enquiries (CMACE). Saving mothers’ lives: reviewing maternal deaths to make motherhood safer: 2006–2008. The eighth report on confidential enquiries into maternal deaths in the United Kingdom. BJOG. 2011;118(Suppl.1):1–203. doi:10.1111/j.1471-0528.2010.02847.x. 8 Knight M, Kenyon S, Brocklehurst P, Neilson J, Shakespeare J, Kurinczuk JJ, editors (on behalf of MBRRACE-UK). Saving lives, improving mothers’ care – lessons learned to inform future maternity care from the UK and Ireland Confidential Enquiries into Maternal Deaths and Morbidity 2009–12. Oxford: National Perinatal Epidemiology Unit, University of Oxford; 2014. 9 Deneux-Tharaux C et al. Underreporting of pregnancy-related mortality in the United States and Europe. Obstet Gynecol. 2005;106:684–92. 10 Atrash HK, Alexander S, Berg CJ. Maternal mortality in developed countries: not just a concern of the past. Obstet literature review (updated January 2014) for such studies where misclassification on coding in civil registration could be identified. These studies are diverse in terms of the definition of maternal mortality used, the sources considered (death certificates, other vital event certificates, medical records, questionnaires or autopsy reports) and the way maternal deaths are identified (record linkage or assessment from experts). In addition, the system of reporting causes of death to a civil registry differs from one country to another, depending on the death certificate forms, the type of certifiers and the coding practice. These studies have estimated underreporting of maternal mortality due to misclassification in death registration data, ranging from 0.85 to 5.0, with a median value of 1.5 (i.e. a misclassification rate of 50%). Underreporting of maternal deaths was more common among: • early pregnancy deaths, including those not linked to a reportable birth outcome; • deaths in the later postpartum period (these were less likely to be reported than early postpartum deaths); • deaths at extremes of maternal age (youngest and oldest); • miscoding by the ICD-9 or ICD-10, most often seen in cases of deaths caused by: o cerebrovascular diseases; o cardiovascular diseases. Potential reasons cited for underreporting and/or misclassification include: • inadequate understanding of the ICD rules (either ICD-9 or ICD-10); • death certificates completed without mention of pregnancy status; • desire to avoid litigation; • desire to suppress information (especially as related to abortion deaths). The definitions of misclassification, incompleteness and underreporting of maternal deaths are shown in Box A2.3. Box A2.3 Definitions of misclassification, incompleteness and underreporting Misclassification Refers to incorrect coding in civil registration, due either to error in the medical certification of cause of death or error in applying the correct code. Incompleteness Refers to incomplete death registration. Includes both the identification of individual deaths in each country and the national coverage of the register.                                                                                                                                                                                                                                                                                                                                                                                   Gynecol. 1995;86(4 pt 2):700–5. Box A2.3 Definitions of misclassification, incompleteness and underreporting Underreporting Is a combination of misclassification and incompleteness. In the absence of complete and accurate civil registration systems, MMR estimates are based on data from a variety of sources – including censuses, household surveys, reproductive-age mortality studies (RAMOS) and verbal autopsies. Each of these methods has limitations in estimating the true levels of maternal mortality. Brief descriptions of these methods together with their limitations are shown in Box A2.4. Box A2.4 Approaches to measuring maternal mortality 8,9,11 Civil registration system This approach involves routine registration of births and deaths. Ideally, maternal mortality statistics should be obtained through civil registration data. However, even where coverage is complete and the causes of all deaths are identified based on standard medical certificates, in the absence of active case finding, maternal deaths may be missed or misclassified; and therefore confidential enquiries are used to identify the extent of misclassification and underreporting. 12,13 Household surveys Demographic and Health Surveys (DHS) and Multiple Indicator Cluster Surveys – Round 4 (MICS) use the direct “sisterhood” method using household survey data. This method obtains information by interviewing a representative sample of respondents about the survival of all their siblings (to determine the age of all siblings, how many are alive, how many are dead, age at death and year of death of those dead, and among sisters who reached reproductive age, how many died during pregnancy, delivery or within two months of pregnancy). This approach has the following limitations. • It identifies pregnancy-related deaths, rather than maternal deaths. • It produces estimates with wide confidence intervals, thereby diminishing opportunities for trend analysis. • It provides a retrospective rather than a current maternal mortality estimate (referring to a                                                                                                                         11 Knight M, Kenyon S, Brocklehurst P, Neilson J, Shakespeare J, Kurinczuk JJ, editors (on behalf of MBRRACE-UK). Saving lives, improving mothers’ care – lessons learned to inform future maternity care from the UK and Ireland Confidential Enquiries into Maternal Deaths and Morbidity 2009–12. Oxford: National Perinatal Epidemiology Unit, University of Oxford; 2014. 12 Hill K et al. How should we measure maternal mortality in the developing world? A comparison of household deaths and sibling history approaches. Bull World Health Organ. 2006;84:173–80. 13 Stanton C, Abderrahim N, Hill K. DHS maternal mortality indicators: an assessment of data quality and implications for data use (DHS Analytical Report No. 4). Calverton (MD): Macro International; 1997. Box A2.4 Approaches to measuring maternal mortality period approximately five years prior to the survey); the analysis is more complicated. 14,15 Census A national census, with the addition of a limited number of questions, could produce estimates of maternal mortality. This approach eliminates sampling errors (because all women are covered) and hence allows a more detailed breakdown of the results, including trend analysis, geographic subdivisions and social strata. • This approach allows identification of deaths in the household in a relatively short reference period (1–2 years), thereby providing recent maternal mortality estimates, but is conducted at 10-year intervals and therefore limits monitoring of maternal mortality. • It identifies pregnancy-related deaths (not maternal deaths); however, if combined with verbal autopsy, maternal deaths could be identified. • Training of enumerators is crucial, since census activities collect information on a range of other topics unrelated to maternal deaths. • Results must be adjusted for characteristics such as completeness of death and birth statistics and population structures, in order to arrive at reliable estimates. 11,12 Reproductive-age mortality studies (RAMOS) This approach involves identifying and investigating the causes of all deaths of women of reproductive age in a defined area or population, by using multiple sources of data (e.g. interviews of family members, civil registrations, health-care facility records, burial records, traditional birth attendants), and has the following characteristics. • Multiple and diverse sources of information must be used to identify deaths of women of reproductive age; no single source identifies all the deaths. • Interviews with household members and health-care providers and reviews of facility records are used to classify the deaths as maternal or otherwise. • If properly conducted, this approach provides a fairly complete estimation of maternal mortality (in the absence of reliable routine registration systems) and could provide subnational MMRs. However, inadequate identification of all deaths of reproductive-aged women results in underestimation of maternal mortality levels. • This approach can be complicated, time-consuming and expensive to undertake – particularly on a large scale. • The number of live births used in the computation may not be accurate, especially in settings where most women deliver at home.                                                                                                                         14 Stanton C et al. Every death counts: measurement of maternal mortality via a census. Bull World Health Organ. 2001;79:657–64. 15 WHO guidance for measuring maternal mortality from a census. Geneva: World Health Organization; 2013. Box A2.4 Approaches to measuring maternal mortality 16,17,18 Verbal autopsy This approach is used to assign cause of death through interviews with family or community members, where medical certification of cause of death is not available. Verbal autopsies may be conducted as part of a demographic surveillance system maintained by research institutions that collect records of births and deaths periodically among small populations (typically in a district). This approach may also be combined with household surveys or censuses. In special versions, and in combination with software that helps to identify the diagnosis, verbal autopsy is suitable for routine use as an inexpensive method in populations where no other method of assessing the cause of death is in place. The following limitations characterize this approach. • Misclassification of causes of deaths in women of reproductive age is not uncommon with this technique. • It may fail to identify correctly a group of maternal deaths, particularly those occurring early in pregnancy (e.g. ectopic, abortion-related) and indirect causes of maternal death (e.g. malaria). • The accuracy of the estimates depends on the extent of family members’ knowledge of the events leading to the death, the skill of the interviewers, and the competence of physicians who do the diagnosis and coding. The latter two factors are largely overcome by the use of software. • Detailed verbal autopsy for research purposes that aims to identify the cause of death of an individual requires physician assessment and long interviews. Such systems are expensive to maintain, and the findings cannot be extrapolated to obtain national MMRs. This limitation does not exist where simplified verbal autopsy is aiming to identify causes at a population level and where software helps to formulate the diagnoses.                                                                                                                         16 Chandramohan D et al. The validity of verbal autopsies for assessing the causes of institutional maternal death. Stud Fam Plann. 1998;29:414–22. 17 Chandramohan D, Stetal P, Quigley M. Misclassification error in verbal autopsy: can it be adjusted? Int J Epidemiol. 2001;30:509–14. 18 Leitao J et al. Revising the WHO verbal autopsy instrument to facilitate routine cause-of-death monitoring. Global Health Action. 2013;6:21518. Annex 3. Methods used to derive a complete series of annual estimates for each covariate, 1985–2015 A  complete  series  of  annual  estimates  for  each  of  the  three  covariates  was  obtained  or  constructed   between  1985  and  2015.   GDP  per  capita  measured  in  purchasing  power  parity  (PPP)  equivalent  dollars  using  2011  as  the   baseline  year  were  taken  from  World  Bank  Group19  with  estimates  from  other  sources  (e.g.  IMF,   OECD,  WHO  National  Health  Accounts  and  the  Institute  for  Health  Metrics  and  Evaluation)  used  to   inform  trends  in  instances  with  missing  country-­‐years  in  the  World  Bank  Group  data  set.  A  five-­‐year   moving  average  was  applied  to  this  GDP  series  to  smooth  year-­‐to-­‐year  GDP  fluctuations.   General  fertility  rate  (GFR)  estimates  were  calculated  using  annual  series  of  live  births  and  the   populations  of  women  aged  15–49  years,  which  were  constructed  using  estimates  from  UNPD.20   Skilled  attendant  at  birth  (SAB)  coverage  estimates  consist  of  time  series  derived  using  data  from   household  surveys  and  other  sources,  obtained  from  a  database  maintained  by  UNICEF.21  Although   other  sources  of  SAB  data  were  consulted,  only  the  UNICEF  data  were  used  because  they  adhere   strictly  to  the  indicator’s  definition.22  For  countries  with  any  value  of  SAB  less  than  95%  and  with   four  or  more  observations,  annual  series  were  estimated  by  fitting  a  regression  model  with  time  as   the  sole  predictor  for  the  logit  (log-­‐odds)  of  SAB;  such  a  model  was  estimated  separately  for  each   country.  For  all  other  countries,  including  those  with  no  available  SAB  data,  the  SAB  annual  series   were  estimated  using  a  multilevel  model.  In  the  multilevel  model,  logit  (or  log-­‐odds)  of  observed  SAB   proportions  for  all  countries  were  regressed  against  time.  The  model  included  region-­‐  and  country-­‐ specific  intercepts  and  slopes.                                                                                                                               19 GDP per capita measured in purchasing power parity (PPP) equivalent dollars, reported as constant 2011 international dollars, based on estimates published by World Bank Group. International Comparison Program database. Washington (DC): World Bank Group; 2014. 20 World population prospects: the 2015 revision. New York: United Nations, Department of Economic and Social Affairs, Population Division; 2015. 21 UNICEF Data: Monitoring the Situation of Children and Women [website]. New York: United Nations Children’s Fund; 2015 (http://data.unicef.org/). 22 Making pregnancy safer: the critical role of the skilled attendant: a joint statement by WHO, ICM and FIGO. Geneva: World Health Organization; 2014. Annex 4. Adjustment factor to account for misclassification of maternal deaths in civil registration, literature review of reports and articles   Country Period/year Adjustment factor Australiaa 1994–1996 1.23 Australiab 1997–1999 1.80 Australiac 2000–2002 1.97 Australiad 2003–2005 2.03 Austriae 1980–1998 1.61 Brazilf 2002 1.40 Canadag 1988–1992 1.69 Canadah 1997–2000 1.52 Denmarki 1985–1994 1.94 Denmarkj 2002–2006 1.04 Finlandk 1987–1994 0.94 Francel Dec 1988 to 2.38 March 1989 Francem 1999 1.29 Francen 2001–2006 1.21 Franceo 2007–2009 1.21 Guatemalap 1989 1.84 Guatemalap 1996–1998 1.84 Guatemalaq 2000 1.88 Guatemalar 2007 1.73 Irelands 2009–2011 3.40 Japant 2005 1.35 Mexicou 2008 0.99 Netherlandsv 1983–1992 1.34 Netherlandsx 1993–2005 1.48 New Zealandy 2006 1.11 New Zealandz 2007 0.85 New Zealandaa 2008 1.00 Country Period/year Adjustment factor New Zealandbb 2009 0.92 New Zealandcc 2010 1.00 Portugaldd 2001–2007 2.04 Serbiaee 2007–2010 1.86 Singaporeff 1990–1999 1.79 Sloveniagg 2003–2005 5.00 South Africahh 1999–2001 0.98 South Africaii 2002–2004 1.16 South Africaii 2005–2007 0.90 Swedenjj 1997–2005 1.33 Swedenkk 1988–2007 1.68 United Kingdomll 1988–1990 1.39 United Kingdomll 1991–1993 1.52 United Kingdomll 1994–1996 1.64 United Kingdomll 1997–1999 1.77 United Kingdomll 2000–2002 1.80 United Kingdomll 2003–2005 1.86 United Kingdomll 2006–2008 1.60 United Statesmm 1991–1997 1.48 United Statesnn 1995–1997 1.54 United Statesoo 1999–2002 1.59 United Statesoo 2003–2005 1.41 Median 1.5 a AIHW, NHMRC. Report on maternal deaths in Australia 1994–96. Cat. no. PER 17. Canberra: AIHW; 2001 (). b Slaytor EK, Sullivan EA, King JF. Maternal deaths in Australia 1997–1999. Cat. No. PER 24. Sydney: AIHW National Perinatal Statistics Unit; 2004 (Maternal Deaths Series, No. 1). c Sullivan EA, King JF, editors. Maternal deaths in Australia 2000–2002. Cat. no. PER 32. Sydney: AIHW National Perinatal Statistics Unit; 2006 (Maternal Deaths Series, No. 2). d Sullivan EA, Hall B, King JF. Maternal deaths in Australia 2003–2005. Cat. no. PER 42. Sydney: AIHW National Perinatal Statistics Unit; 2007 (Maternal Deaths Series, No. 3). e Johnson S, Bonello MR, Li Z, Hilder L, Sullivan EA. Maternal deaths in Australia 2006–2010. Cat. no. PER 61. Canberra: AIHW; 2014 (Maternal Deaths Series, No. 4). f Brasil Ministério da Saúde, Secretaria de Atenção à Saúde, Departamento de Ações Programáticas Estratégicas. Estudo da mortalidade de mulheres de 10 a 49 anos, com ênfase na mortalidade materna: relatório final. Brasilia: Ministério da Saúde, Secretaria de Atenção à Saúde, Departamento de Ações Programáticas Estratégicas, Editora do Ministério da Saúde; 2006. g Turner LA et al. Underreporting of maternal mortality in Canada: a question of definition. Chronic Dis Can. 2002;23:22– 30. h Health Canada. Special report on maternal mortality and severe morbidity in Canada – enhanced surveillance: the path to prevention. Ottawa: Minister of Public Works and Government Services Canada; 2004. i Andersen BR et al. Maternal mortality in Denmark 1985–1994. Eur J Obstet Gynecol Reprod Biol. 2009;42:124–8. j Bødker B et al. Maternal deaths in Denmark 2002–2006. Acta Obstet Gynecol Scand. 2009;88:556–62. k Gissler M et al. Pregnancy-associated deaths in Finland 1987–1994 definition problems and benefits of record linkage. Acta Obstet Gynecol Scand. 1997;76(7):651–7. l Bouvier-Colle MH et al. Reasons for the underreporting of maternal mortality in France, as indicated by a survey of all deaths among women of childbearing age. Int J Epidemiol. 1991;20:717–21. m Bouvier-Colle MH et al. Estimation de la mortalité maternelle en France : une nouvelle méthode. J Gynecol Obstet Biol Reprod. 2004;33(5):421–9. n Rapport du Comité national d’experts sur la mortalité maternelle (CNEMM) 2001–2006. Saint-Maurice: Institut de veille sanitaire; 2010. o Rapport du comité national d’experts sur la mortalité maternelle (CNEMM). Enquête nationale confidentielle sur les morts maternelles France, 2007–2009 Inserm, France: Institut national de la santé et de la recherche médicale; 2013. p Schieber B, Stanton C. Estimación de la mortalidad materna en Guatemala período 1996–1998. Guatemala; 2000. q Línea basal de mortalidad materna para el año 2000. Informe final. Guatemala: Ministerio de Salud Pública y Asistencia Social; 2003. r Estudio nacional de mortalidad materna. Informe final. Guatemala: Secretaría de Planificación y Programación de la Presidencia Ministerio de Salud Pública y Asistencia Social; 2011. s Confidential Maternal Death Enquiry in Ireland, report for triennium 2009–2011. Cork: Maternal Death Enquiry; 2012. t Health Sciences Research Grant. Analysis and recommendations of the causes of maternal mortality and infant mortality. Tomoaki I, principal investigator. Research Report 2006–2008. Osaka: Department of Perinatology, National Cardiovascular Center; 2009 [in Japanese]. Hidaka A et al. [Causes and ratio of maternal mortality, and its reliability]. Sanfujinkachiryou [Treatment in obstetrics and gynaecology]. 2009;99(1):85–95 [in Japanese]. u Búsqueda intencionada de muertes maternas en México. Informe 2008. Mexico, DF: Dirección General de Información en Salud, Secretaría de Salud; 2010. v Schuitemaker N et al. Confidential enquiry into maternal deaths in the Netherlands 1983–1992. Eur J Obstet Gynecol Reprod Biol. 1998;79(1):57–62. x Schutte J et al. Rise in maternal mortality in the Netherlands. BJOG. 2010;117(4):399–406. y PMMRC. Perinatal and maternal mortality in New Zealand 2006: second report to the Minister of Health. Wellington: Ministry of Health; 2009. z PMMRC. Perinatal and maternal mortality in New Zealand 2007: third report to the Minister of Health July 2008 to June 2009. Wellington: Ministry of Health; 2009. aa PMMRC. Perinatal and maternal mortality in New Zealand 2008: fourth report to the Minister of Health July 2009 to June 2010. Wellington: Ministry of Health; 2010. bb PMMRC. Fifth annual report of the Perinatal and Maternal Mortality Review Committee: reporting mortality 2009. Wellington: Health Quality and Safety Commission; 2011. cc PMMRC. Sixth annual report of the Perinatal and Maternal Mortality Review Committee: reporting mortality 2010. Wellington: Health Quality and Safety Commission; 2012. dd Gomes MC, Ventura MT, Nunes RS. How many maternal deaths are there in Portugal? J Matern Fetal Neonatal Med. 2012;25(10):1975–9. ee Krstic M et al. Maternal deaths – methodology for cases registration and reporting. Belgrade; 2008 [unpublished paper]. ff Lau G. Are maternal deaths on the ascent in Singapore? A review of maternal mortality as reflected by coronial casework from 1990 to 1999. Ann Acad Med Singapore. 2002;31(3):261–75. gg Kralj E, Mihevc-Ponikvar B, Premru-Sršenc T, Balažica J. Maternal mortality in Slovenia: case report and the method of identifying pregnancy-associated deaths. Forensic Sci Int Suppl Ser. 2009;1(1):52–7. hh Moodley J. Saving mothers: 1999–2001. S Afr Med J. 2003;93(5):364–6. ii Saving mothers 2008–2010: fifth report on the confidential enquiries into maternal deaths in South Africa. Comprehensive report. South Africa: Department of Health, National Committee on Confidential Enquires into Maternal Deaths; 2012. jj Grunewald C et al. Modradodligheten underskattad i Sverige. Lakartidningen. 2008;34(105):2250–3. kk Esscher A et al., Maternal mortality in Sweden 1988–2007: more deaths than officially reported. Acta Obstet Gynecol Scand. 2012;92:40–6. ll Centre for Maternal and Child Enquiries (CMACE). Saving mothers’ lives: reviewing maternal deaths to make motherhood safer: 2006–2008. The eighth report on confidential enquiries into maternal deaths in the United Kingdom. BJOG. 2011;118(Suppl.1):1–203. mm Berg CJ et al. Pregnancy-related mortality in the United States, 1991–1997. Obstet Gynecol. 2003;101(2):289–96. nn MacKay AP et al. An assessment of pregnancy-related mortality in the United States. Paediatr Perinat Epidemiol. 2005;19(3):206–14. oo MacKay AP et al. Changes in pregnancy mortality ascertainment United States, 1999–2005. Obstet Gynecol. 2011;118:104–10. Annex 5. Usability assessment of civil registration data for selected years (1990, 1995, 2000, 2005, 2010 and latest available year) Assessment of civil registration data (VR data) quality – usability   National  civil  registration  and  vital  statistics  (CRVS)  systems  are  meant  to  record  all  births,  deaths   and  causes  of  death  within  a  country.  The  data  retrieved  from  CRVS  systems  are  referred  to  as  vital   registration  (VR)  data.       For  the  VR  data,  the  usability,  referred  to  as  (!,! )  for  country  c  in  year  t,  was  defined  as  the   proportion  of  all  deaths  to  women  of  reproductive  ages  in  the  country-­‐year  for  which  causes  have   been  assessed  in  the  VR  data  set.  Essentially,  (!,! )  is  the  product  of  the  completeness  of  the  VR  data   and  the  percentage  of  deaths  with  a  well-­‐defined  cause:     (!"#$%&'& ) (!"" )   !,!  =  !,!  ×  (1 − !,! )     (!"#$%&'& ) (!"" ) where  ! ,! refers  to  the  completeness  of  the  VR,  and  ! ,!  refers  to  the  proportion  of  VR   deaths  with  ill-­‐defined  causes  (as  reported).     The  completeness  is  assessed  by  comparing  the  total  number  of  deaths  among  women  of   reproductive  age  recorded  in  the  VR  database  (WHO  Mortality  Database)23  to  the  WHO  estimate  of   the  total  number  of  deaths  among  women  of  reproductive  age,24  i.e.:       (!"#$%&'& )   !,!  =  VR  total  deaths  /  WHO  total  deaths     (!"#$%&'& ) with  !,! = 1  if  the  VR  total  deaths  exceeds  the  WHO  estimate  of  total  deaths.     Based  on  the  assessment  of  data  quality  and  data  source,  VR  data  are  grouped  into  three  categories.   These  categories  affect  how  much  uncertainty  is  assumed  to  surround  each  data  point  obtained   from  the  VR  system.  The  categories  are  as  follows.     • Type  I:  good  quality  VR  data  with  usability  >  80%.   • Type  II:  VR  data  from  a  string  of  decent  VR  data  with  usability  between  60%  and  80%.   • Type  III:  other  data  from  registration  and  mortality  reporting  systems.  For  these  data  points,   data  quality  cannot  be  assessed  as  the  countries  have  not  submitted  data  to  the  relevant  WHO   office.     Please  refer  to  Table  A5.1  for  the  usability  assessment  by  country  for  selected  years.                                                                                                                               23 WHO Mortality Database (http://www.who.int/healthinfo/mortality_data/en/). 24 Life tables for WHO Member States 1990–2012. Geneva: World Health Organization; 2014. Table  A5.1.  Usability  assessment  of  civil  registration  data  for  selected  years  (1990,  1995,  2000,   2005,  2010  and  latest  available  year)   Latest   available   Country   1990   1995   2000   2005   2010   year     42   Albania   NA      56   49   55   (2009)     95   Argentina   96      97   94   94   94   (2013)   82   Armenia   67      86   91   84   (2012)     98   Australia   99      96   98   98   (2011)     97   Austria   99   100   100   100   98   (2014)   87   Azerbaijan   60      64   80       (2007)   99   Bahamas      99   84   92   82   (2012)     94   Bahrain       98   94   93   (2013)   100   Barbados   83   100   98   98   100   (2012)   98   Belarus   99      98   98   (2011)       94   Belgium   94      96   98   97   95   (2012)   100   Belize   83      85   98   100   99   (2013)   21   Bolivia  (Plurinational  State  of)       15       (2003)     91   Bosnia  and  Herzegovina   88   (2011)           92   Brazil   69      72   75   81   83   (2013)   97   Brunei  Darussalam   88   97   83   (2012)       93   Bulgaria   97      98   96   96   94   (2014)   97   Cabo  Verde   (2012)             93   Canada   92      97   97   97   93   (2011)   98   Chile   97      98   98   98   98   (2013)   82   Colombia   85      82   84   83   81   (2012)   87   Costa  Rica   89      90   91   91   90   (2013)   Latest   available   Country   1990   1995   2000   2005   2010   year     99   Croatia   95      88   99   98   99   (2014)   98   Cuba   99      99   99   99   98   (2013)   71   Cyprus   37   65   (2012)         91   Czech  Republic   100      99   99   98   97   (2013)   87   Denmark   96      94   98   97   93   (2012)   65   Dominican  Republic   44      44   45   48   48   (2012)   72   Ecuador   72      74   75   76   78   (2013)   91   Egypt   80   82   91   (2013)       55   El  Salvador   61      63   65   65   64   (2012)   99   Estonia   99      99   99   98   98   (2012)   100   Fiji   99   (2012)           98   Finland   98      99   99   97   96   (2013)   90   France   93      94   92   91   90   (2011)   73   Georgia   96      89   78   87   49   (2014)   93   Germany   93      95   93   94   93   (2013)   96   Greece   96      94   95   98   96   (2012)   96   Grenada   91      87   92   100   98   (2013)   76   Guatemala   76      78   85   93   81   (2013)   65   Guyana   81      84   85   73   (2011)     15   Honduras   54   14   (2013)         97   Hungary   100   100   99   99   100   (2014)   93   Iceland   93      92   97   93   95   (2012)   Ireland   98      99   99   99   99     100   Israel   98      96   98   93   (2009)     Latest   available   Country   1990   1995   2000   2005   2010   year     97   Italy   98      98   98   98   (2012)     77   Jamaica   48   53   63   73   (2011)     88   Japan   99      99   98   97   97   (2013)   49   Jordan   50   (2011)           86   Kazakhstan   86   887   82   83   85   (2012)   53   Kiribati      61   56   (2001)         99   Kuwait      85   99   98   99   (2013)     82   Kyrgyzstan   82      80   85   89   89   (2013)   99   Latvia   98   100   98   99   99   (2012)   92   Lithuania   100      99   99   97   96   (2012)   98   Luxembourg   96      90   94   89   94   (2013)   85   Malaysia   85   81   (2008)         77   Maldives   51   63   71   (2011)       97   Malta   74      87   89   91   79   (2012)   99   Mauritius   96      96   97   99   100   (2013)   88   Mexico   96      92   90   93   94   (2013)   86   Montenegro         84   92   (2009)       15   Morocco   16   (2012)           94   Netherlands   93   94   93   95   95   (2013)   99   New  Zealand   99   100   100   99   100   (2011)   66   Nicaragua   62   63   63   64   (2013)     89   Norway   98   97   97   97   95   (2013)   Oman   56             81   Panama   79   83   75   (2013)       Latest   available   Country   1990   1995   2000   2005   2010   year     76   Paraguay   75   74   77   79   (2013)     63   Peru   43   48   55   62   61   (2013)   88   Philippines   83   85   (2008)         88   Poland   95   94   94   93   93   (2013)   80   Portugal   89   89   85   88   (2013)     89   Puerto  Rico   99   100   99   99   95   (2013)   65   Qatar   98   86   71   (2012)       96   Republic  of  Korea   85   95   96   97   96   (2012)   99   Republic  of  Moldova   100   100   89   97   100   (2013)   83   Romania   100   97   99   95   93   (2012)   96   Russian  Federation   98   97   95   95   96   (2011)   100   Saint  Lucia   98   94   85   87   95   (2012)   98   Saint  Vincent  and  the  Grenadines   97   100   83   79   93   (2013)   42   Saudi  Arabia   (2012)             66   Serbia   67   69   71   (2013)       74   Singapore   86   85   82   78   74   (2014)   91   Slovakia   NA   99   98   96   95   (2014)   Slovenia   97   95   96   95   97     67   South  Africa   NA   70   88   87   83   (2013)   91   Spain   99   99   97   97   95   (2013)   79   Sri  Lanka   72   (2006)           94   Suriname   92   69   72   75   96   (2012)   97   Sweden   99   99   99   98   94   (2013)   91   Switzerland   97   95   95   96   95   (2012)   Latest   available   Country   1990   1995   2000   2005   2010   year     Syria   90   Tajikistan     64     64     65     67         80   Thailand   67   82   77   75   (2006)   The  former  Yugoslav  Republic  of     Macedonia   NA   88   91   89   90     99   Trinidad  and  Tobago   99   98   98   99   (2009)     24   Tunisia   (2013)             45   Turkey   40   (2013)           53   Turkmenistan   74   79   (2013)         96   Ukraine   99   98   97   98   98   (2012)   United  Arab  Emirates   79   53           97   United  Kingdom   100   99   99   98   (2013)     98   United  States  of  America   95   96   96   97   96   (2013)   94   Uruguay   95   95   93   91   89   (2013)   Uzbekistan   88   89   88   90   Venezuela  (Bolivarian  Republic       96   of)   89   89   89   94   (2012)   Zimbabwe   36                 Annex 6. Estimation of AIDS-related indirect maternal deaths In this estimation process, the full model has two parts, the first part to separately estimate maternal deaths not related to AIDS (discussed in section 2.4 of the main report) and the second part to estimate AIDS-related indirect maternal deaths. AIDS-related indirect maternal deaths refer to HIV-positive women who have died because of the aggravating effect of pregnancy on HIV; where the interaction between pregnancy and HIV becomes the underlying cause of death, these are counted as indirect maternal deaths. It is important to note that direct maternal deaths among HIV-positive women are not estimated separately but are rather included within the first part of the model. Thus, the final PM estimates are the result of adding the results of this two-part model: the estimated number of non-AIDS-related maternal deaths and the estimated number of AIDS- related indirect maternal deaths: PM = (1 – a)PMna + aPMa (A6.1) where PMna is the proportion of non-AIDS-related maternal deaths among all non-AIDS-related deaths (women aged 15–49 years); PMa is the proportion of AIDS-related indirect maternal deaths among all AIDS-related deaths (women aged 15–49 years); and a is the proportion of AIDS-related deaths among all deaths (women aged 15–49 years). This appendix describes the second part of the two-part model, that is, the estimation of AIDS- related indirect maternal deaths, PMa. The sources of data for estimating the fraction of AIDS- related indirect maternal deaths are the UNAIDS 2013 estimates of AIDS-related deaths25 and the total number of deaths estimated by WHO from its life tables. The approach used to estimate the proportion of AIDS-related deaths that qualify as indirect maternal deaths, PMa, is the product of two quantities: PMa = υu (A6.2) where υ is the proportion of AIDS deaths in women aged 15–49 years that occur during pregnancy or the childbirth period, computed as follows: ckGFR υ= (A6.3) 1 + c(k − 1)GFR u is the fraction of AIDS-related deaths among pregnant women that qualify as maternal because of some causal relationship with the pregnancy, delivery or postpartum period; GFR is the general fertility rate; c is the average woman-years lived in the maternal risk period per live birth (set equal to 1 year, including the 9 month gestation, plus 42 days postpartum, and an additional 1.5 months to account for pregnancies not ending in a live birth); k  is the relative risk of dying from AIDS for a pregnant versus non-pregnant woman. In the 2013 estimates, updated values for k and u were used, in light of new data from the network for Analyzing Longitudinal Population-based HIV/AIDS data on Africa (ALPHA).26 Based on the                                                                                                                         25 According to the Joint United Nations Programme on HIV/AIDS (UNAIDS), AIDS-related deaths (including AIDS-related indirect maternal deaths) include the estimated number of deaths related to HIV infection, including deaths that occur before reaching the clinical stage classified as AIDS. 26 Zaba B et al. Effect of HIV infection on pregnancy-related mortality in sub-Saharan Africa: secondary analyses of pooled community-based data from the network for Analyzing Longitudinal Population-based HIV/AIDS data on Africa (ALPHA). Lancet. 2013;381(9879):1763–71. doi:10.1016/S0140-6736(13)60803-X. findings in the paper and further exploration of the data, both k and u were set equal to 0.3. The uncertainty distributions for both parameters were updated as well, the standard deviation for k was set to 0.1 and for u, a uniform distribution with outcomes between 0.1 and 0.5 was used.   Annex 7. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, lifetime risk and percentage of AIDS-related indirect maternal deaths, 2015a Range  of  MMR   %  of   uncertainty     Lifetime   AIDS-­‐ Range  of  PM   (UI  80%)   Number   risk  of   related   uncertainty   of   maternal   indirect   Lower   Upper   maternal   death:     maternal   Lower   Upper   b c d e Country MMR   estimate   estimate   deaths   1  in   deaths   PM   estimate   estimate   Afghanistan   396 253 620 4  300 52 – 17.7 11.3 27.7 Albania   29 16 46 11 1  900 – 1.3 0.7 2.1 Algeria   140 82 244 1  300 240 – 8.0 4.7 14.0 Angola   477 221 988 5  400 32 – 18.3 8.5 37.8 Argentina   52 44 63 390 790 – 3.8 3.2 4.6 Armenia   25 21 31 10 2  300   – 1.4 1.1 1.7   Australia   6 5 7 19 8  700 – 0.6 0.5 0.8 Austria   4 3 5 3 18  200 – 0.3 0.2 0.4 Azerbaijan   25 17 35 48 1  600 – 2.0 1.3 2.7 Bahamas   80 53 124 5 660 – 3.8 2.5 5.9 Bahrain   15 12 19 3 3  000 – 1.9 1.5 2.4 Bangladesh   176 125 280 5  500 240 – 8.6 6.1 13.6 Barbados   27   19 37 1 2  100 – 1.5 1.0 2.0 Belarus   4 3 6 5 13  800 – 0.2 0.1 0.2 Belgium   7 5 10 9 8  000 – 0.5 0.4 0.7 Belize   28 20 36 2 1  300 – 2.7 2.0 3.5 Benin   405 279 633 1  600 51 – 14.8 10.2 23.1 Bhutan   148 101 241 20 310 – 3.6 2.4 5.8 Bolivia   (Pluri-­‐ national   State  of)   206 140 351 520 160 – 7.9 5.4 13.4 Bosnia  and   Herzegovina   11 7   17 4 6  800 – 0.7 0.4 1.0 Botswana   129 102 172 72 270 18 3.1 2.5 4.2 Brazil   44 36 54 1  300 1  200 – 2.0 1.6 2.5 Brunei   Darussalam   23 15 30 2 2  300 – 1.7 1.2 2.3 Bulgaria   11 8 14 7 6  200 – 0.4 0.3 0.6 Burkina   Faso   371 257 509 2  700 48 – 14.2 9.8 19.5 Burundi   712 471 1  050 1  350 23 – 27.0 17.9 39.8 Cabo  Verde   42 20 95 5 900 – 5.0 2.3 11.2 Cambodia   161 117 213 590 210 – 6.4 4.7 8.5 Cameroon   596 440 881 5  100 35 – 15.2 11.2 22.5 f Canada   7 5 9 27 8  800 – 0.5 0.4 0.7 Central   African   Republic   882 508 1  500 1  400 27 – 15.0 8.7 25.6 Chad   856 560 1  350 5  400 18 – 28.3 18.5 44.5 Chile   22 18 26 52 2  600 – 1.8 1.5 2.2 China   27 22 32 4  400 2  400 – 1.3 1.1 1.6 Range  of  MMR   %  of   uncertainty     Lifetime   AIDS-­‐ Range  of  PM   (UI  80%)   Number   risk  of   related   uncertainty   of   maternal   indirect   Lower   Upper   maternal   death:     maternal   Lower   Upper   b c d e Country MMR   estimate   estimate   deaths   1  in   deaths   PM   estimate   estimate   Colombia   64 56 81 480 800 –   3.8 3.3 4.7 Comoros   335 207 536 88 66 – 13.4 8.3 21.3 Congo   442   300 638 740 45 – 12.8 8.7 18.4 Costa  Rica   25 20 29 18 2  100 – 1.8 1.4 2.1 Côte   d’Ivoire   645 458 909 5  400 32 – 13.4 9.5 18.9 Croatia   8 6 11 3 7  900 – 0.6 0.4 0.7 Cuba   39 33 47 45 1  800 – 1.8 1.5 2.1 Cyprus   7 4 12 1 9  400 – 0.8 0.4 1.4 Czech   Republic   4 3 6 5 14  800 –   0.3 0.2 0.4 Democratic   People’s   Republic  of   Korea   82 37 190 300 660 – 2.8 1.3 6.5 Democratic   Republic  of   the  Congo   693 509 1  010 22  000 24 – 22.3 16.4 32.5 Denmark   6 5 9 4 9  500 –   0.5 0.4 0.7 Djibouti   229 111 482 50 140 – 5.4 2.6 11.3 Dominican   Republic   92 77 111 200 400 – 3.7 3.1 4.5 Ecuador   64 57 71 210 580 – 4.4 3.9 4.9 Egypt   33 26 39 820 810 – 3.5 2.8 4.1 El  Salvador   54 40 69 57 890 – 1.9 1.4 2.4 Equatorial   Guinea   342 207 542 100 61 5.6 8.8   5.3 13.9 Eritrea   501 332 750 880 43 – 20.5 13.6 30.6 Estonia   9 6 14 1 6  300 – 0.5 0.3 0.7 Ethiopia   353 247 567 11  000 64 – 16.7 11.7 26.8 Fiji   30 23 41 5 1  200 – 1.5 1.1 2.0 Finland   3 2 3 2 21  700 – 0.2 0.2 0.3 France   8 7 10 66 6  100   –   0.7 0.6 0.9 Gabon   291 197 442 150   85 – 8.6 5.8 13.1 Gambia   706 484 1  030 590 24   – 31.1 21.4 45.5 Georgia   36 28 47 19 1  500 – 2.3 1.8 3.0 Germany   6 5 8 42 11  700 – 0.4 0.3 0.5 Ghana   319 216 458 2  800 74 – 11.3 7.6 16.2 Greece   3 2 4 3 23  700 – 0.2 0.2 0.3 Grenada   27 19 42 1 1  500 – 1.7 1.2 2.7 Guatemala   88 77 100 380 330 – 5.3 4.7 6.0   Guinea   679 504 927 3  100 29 – 23.3 17.3 31.8 Guinea-­‐ Bissau   549 273 1  090 370 38 – 13.3 6.6 26.3 Guyana   229 184 301 34 170 – 4.7 3.8 6.2 Haiti   359 236 601 950 90 – 10.1 6.6 16.9 Honduras   129 99 166 220 300 – 5.7 4.4 7.3 Hungary   17 12 22   15 4  400 –   0.7 0.5 0.9 Iceland   3 2 6 0 14  600 – 0.4 0.2 0.7 Range  of  MMR   %  of   uncertainty     Lifetime   AIDS-­‐ Range  of  PM   (UI  80%)   Number   risk  of   related   uncertainty   of   maternal   indirect   Lower   Upper   maternal   death:     maternal   Lower   Upper   b c d e Country MMR   estimate   estimate   deaths   1  in   deaths   PM   estimate   estimate   India   174 139 217   45  000 220   – 6.2   5.0 7.7 Indonesia   126 93 179 6  400 320 – 6.3 4.6   8.9 Iran  (Islamic   Republic  of)   25 21 31 340 2  000 – 1.5 1.2 1.8 Iraq   50 35 69 620 420 – 6.2 4.3 8.5 Ireland   8 6 11 5 6  100 – 0.8 0.6 1.2 Israel   5 4 6 9 6  200 – 1.2 0.9 1.4 Italy   4 3 5 18 21  970   –   0.3 0.2 0.4 Jamaica   89 70 115 43 520 – 3.8 3.0 4.9 Japan   5 4 7   56 13  400 – 0.4   0.3 0.5 Jordan   58 44 75 110 490 – 5.2 4.0 6.8 Kazakhstan   12 10 15 45 3  000 – 0.6 0.4 0.7 Kenya   510 344 754 8  000 42 2.3 17.4 11.7 25.7 Kiribati   90 51 152 3 300 – 6.6 3.8 11.2 Kuwait   4 3 6 3 10  300 – 0.9 0.7 1.2 Kyrgyzstan   76   59 96 120 390 – 5.2 4.1 6.5 Lao  People’s   Democratic   Republic   197 136   307 350 150 – 10.3 7.1 16.1 Latvia   18 13 26 4 3  500 – 0.7 0.5 1.0 Lebanon   15 10 22 13 3  700 – 1.8 1.3 2.8 Lesotho   487 310 871 300 61 12.8 5.9 3.8 10.6 Liberia   725 527 1  030 1  100 28 – 31.5 22.9 44.9 Libya   9 6 15 12 4  200 – 0.7 0.5 1.2 Lithuania   10 7 14 3 6  300 – 0.4 0.3 0.5 Luxembourg   10 7 16   1 6  500 – 0.8   0.6 1.4 Madagascar   353 256 484 2  900 60 – 16.4 11.9 22.5 Malawi   634 422 1  080 4  200 29 2.9 22.3 14.9 38.1 Malaysia   40 32 53 200 1  200 – 2.8 2.3 3.7 Maldives   68 45 108 5 600 – 11.4 7.6 18.2 Mali   587 448 823 4  400 27 – 25.2 19.2 35.3 Malta   9 6 15 0 8  300 – 0.8 0.5 1.4 Mauritania   602 399 984 810 36 – 27.4 18.2   44.8 Mauritius   53 38 77 7 1  300 – 2.2 1.5 3.1 Mexico   38 34 42 890 1  100 – 2.5 2.2 2.8 Micronesia   100 46 211 2 310 – 5.4 2.5 11.5 Mongolia   44 35 55 30 800 – 2.3 1.8 2.9 Montenegro   7 4 12 1 8  300 – 0.4 0.2 0.7 Morocco   121 93 142 850   320 – 6.3 4.8 7.4 Mozambiqu e   489 360 686 5  300 40 10.7 9.5 7.0 13.4 Myanmar   178 121 284 1  700 260 –   3.9 2.6 6.2 Namibia   265 172 423 190 100 4.3 11.1 7.2 17.8 Nepal   258 176 425 1  500 150 – 9.8 6.7 16.2 Netherlands   7 5 9 12 8  700 – 0.6 0.4 0.7 New   Zealand   11 9 14 7 4  500 –   0.9 0.7 1.1 Nicaragua   150 115 196 180 270 – 8.5 6.5 11.1 Niger   553 411 752   5  400 23 – 34.3 25.5 46.6 Range  of  MMR   %  of   uncertainty     Lifetime   AIDS-­‐ Range  of  PM   (UI  80%)   Number   risk  of   related   uncertainty   of   maternal   indirect   Lower   Upper   maternal   death:     maternal   Lower   Upper   b c d e Country MMR   estimate   estimate   deaths   1  in   deaths   PM   estimate   estimate   Nigeria   814 596 1  180 58  000 22 – 25.6 18.7 37.0 Norway   5 4 6 3 11  500 –   0.5 0.4 0.6 Occupied   Palestinian   g Territory   45 21 99 69 490 – 6.1 2.8 13.2 Oman   17 13 24 14 1  900 – 2.8 2.0 3.9 Pakistan   178 111 283 9  700 140 – 10.9 6.8 17.3 Panama   94 77   121 71 420 – 6.3 5.1 8.0 Papua  New   Guinea   215 98 457 460 120 –   7.4 3.4   15.8 Paraguay   132 107 163 190 270 – 9.3 7.6 11.6 Peru   68 54 80 420 570 – 4.7 3.7 5.5 Philippines   114 87 175 2  700 280 – 6.3 4.8 9.7 Poland   3 2 4 12 22  100 – 0.2 0.1 0.3 Portugal   10 9 13 8 8  200 –   0.5 0.4 0.6 Puerto  Rico   14 10 18 6 4  300 – 0.8 0.6 1.0 Qatar   13 9 19   3 3  500 – 2.6   1.8 3.9 Republic  of   Korea   11 9 13 50 7  200 – 0.7 0.6   0.9 Republic  of   Moldova   23 19 28 10 3  200 – 1.0 0.8 1.3 Romania   31 22 44 56 2  300 – 1.1 0.8 1.5 Russian   Federation   25 18 33 450 2  300 – 0.7 0.5 1.0 Rwanda   290 208 389 1  100 85 – 11.4 8.2 15.3 Saint  Lucia   48 32 72 1 1  100 – 2.7 1.8 4.0 Saint   Vincent  and   the   Grenadines   45 34 63 1 1  100 – 2.0 1.5 2.8 Samoa   51 24 115 2 500 – 6.2 2.9 13.8 Sao  Tome   and  Principe   156 83 268 10 140 – 8.0 4.2 13.7 Saudi  Arabia   12 7 20 72 3  100 –   1.6 0.9 2.7   Senegal   315 214 468 1  800 61 – 16.3 11.1 24.2 Serbia   17 12 24 15 3  900 – 0.8 0.6 1.1 Sierra  Leone   1  360 999 1  980 3  100 17 – 21.0   15.4 30.6 Singapore   10 6 17 5 8  200 – 0.8 0.5 1.2 Slovakia   6 4 7 3 12  100 – 0.3 0.3 0.4 Slovenia   9 6 14 2 7  000 – 0.8 0.5 1.2 Solomon   Islands   114 75 175 19 220 – 6.6 4.4 10.1 Somalia   732 361 13  900 3  400 22 – 27.6 13.6 52.5 South  Africa   138 124 154 1  500 300 32.1 1.7 1.5 1.8 South  Sudan   789 523 1  150 3  500 26 –   22.7 15.1 33.1 Spain   5 4 6 21 14  700 – 0.4 0.3 0.5 Sri  Lanka   30 26 38 98 1  580 – 1.9 1.7 2.4 Sudan   311 214 433 4  100 72 – 12.5 8.6 17.4 Suriname   155 110 220   15 270 –   7.4 5.2 10.4 Range  of  MMR   %  of   uncertainty     Lifetime   AIDS-­‐ Range  of  PM   (UI  80%)   Number   risk  of   related   uncertainty   of   maternal   indirect   Lower   Upper   maternal   death:     maternal   Lower   Upper   b c d e Country MMR   estimate   estimate   deaths   1  in   deaths   PM   estimate   estimate   Swaziland   389 251 627 150 76 18.6 4.2 2.7 6.7 Sweden   4 3 5 5 12  900 – 0.5 0.4 0.6 Switzerland   5 4 7 4 12  400 – 0.5 0.4 0.7 Syrian  Arab   Republic   68 48 97 300 400 – 6.7 4.7 9.6 Tajikistan   32 19 51 82 790 – 2.9 1.7 4.6 Thailand   20 14   32 140 3  600 – 0.6 0.4 0.9 The  former   Yugoslav   Republic  of   Macedonia   8 5 10 2 8  500 – 0.5 0.3 0.6 Timor-­‐Leste   215 150 300 94 82 – 21.8 15.3 30.4 Togo   368 255 518 940 58 – 10.7 7.4 15.1 Tonga   124 57 270 3   230 – 5.2 2.4 11.3 Trinidad  and   Tobago   63 49 80 12 860 –   2.1 1.6 2.7 Tunisia   62 42 92 130 710 – 5.0 3.4 7.4 Turkey   16 12 21 210 3  000 – 0.9 0.7 1.2 Turkmenista n   42 20 73 47 940 – 1.3 0.6 2.3 Uganda   343 247 493 5  700   47 3.1 13.4 9.7 19.3 Ukraine   24 19 32 120 2  600 –   0.7 0.5 0.9 United  Arab   Emirates   6 3 11 6 7  900 – 0.7 0.4 1.4 United   Kingdom   9 8 11   74 5  800 – 0.8 0.6 0.9 United   Republic  of   Tanzania   398 281 570 8  200 45 2.4 18.4 13.0 26.3 United   States  of   America   14 12 16 550 3  800 – 0.8   0.7 0.9 Uruguay   15 11 19 7 3  300 – 0.9 0.7 1.2 Uzbekistan   36 20 65 240 1  000 – 2.2 1.2 4.0 Vanuatu   78 36 169 5 360 – 6.8 3.1 14.7 Venezuela   (Bolivarian   Republic  of)   95 77 124 570 420 – 6.3 5.1 8.2 Viet  Nam   54 41 74 860 870 – 4.0 3.0 5.5 Yemen   385 274 582 3  300 60 – 17.4 12.3 26.2 Zambia   224 162 306 1  400 79 9.4 8.3 6.0 11.3 Zimbabwe   443 363 563 2  400 52 4.7 13.2 10.8 16.7 PM:  proportion  of  deaths  among  women  of  reproductive  age  that  are  due  to  maternal  causes;  UI:  uncertainty   interval.   a  Estimates  have  been  computed  to  ensure  comparability  across  countries,  thus  they  are  not  necessarily  the   same  as  official  statistics  of  the  countries,  which  may  use  alternative  rigorous  methods.   b   MMR  estimates  have  been  rounded  according  to  the  following  scheme:  <  100  rounded  to  nearest  1;  100–999   rounded  to  nearest  1;  and  ≥  1000  rounded  to  nearest  10.   c  Numbers  of  maternal  deaths  have  been  rounded  according  to  the  following  scheme:  <  100  rounded  to   nearest  1;  100–999  rounded  to  nearest  10;  1000–9999  rounded  to  nearest  100;  and  ≥  10  000  rounded  to   nearest  1000.   d  Life  time  risk  has  been  rounded  according  to  the  following  scheme:  <  100  rounded  to  nearest  1;  100–999   rounded  to  nearest  10;  and  ≥  1000  rounded  to  nearest  100.   e  Percentage  of  AIDS-­‐related  indirect  maternal  deaths  are  presented  only  for  countries  with  an  HIV  prevalence   ≥5.0%  in  2014  (How  AIDS  changed  everything.  MDG  6:  15  years,  15  lessons  of  hope  from  the  AIDS  response.   UNAIDS;  2015).   f  Vital  registration  data  were  available  for  analysis  only  up  to  2011.  Recent  hospital  surveillance  data  for   Canada  excluding  Quebec  indicate  a  decline  of  maternal  deaths  per  100  000  deliveries  from  8.8  in  2007/2008– 2008/2009  to  5.1  in  2011/2012.  Some  98%  of  deliveries  in  Canada  occur  in  hospitals.   g  Refers  to  a  territory.       Annex 8. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk by WHO region, 2015 Range  of  MMR   Lifetime   uncertainty risk  of   Number  of   maternal   Lower Upper   maternal   death: WHO  region MMR estimate estimate deaths 1  in Africa   542 506 650 195  000 37 Americas   52 49 59 7  900 920 South-­‐East  Asia 164 141 199 61  000   240 Europe   16 15 19   1  800 3  400 Eastern   Mediterranean   166 142 216 28  000 170 Western  Pacific   41 37   50   9  800 1  400 World 216 207 249 303  000 180 Annex 9. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by WHO region, 1990–2015 MMR   %  change   Average   in  MMR   annual  %   between   change  in  MMR     1990  and   between  1990   WHO  region   1990   1995   2000   2005   2010   2015   2015   and  2015   Africa   965   914   840   712   620   542   44   2.3   Americas   102      89      76      67      62      52   49   2.7   South-­‐East  Asia   525   438   352   268   206   164   69   4.7   Europe      44      42      33      26      19      16   64   4.0   Eastern   Mediterranean   362   340   304   250   199   166   54   3.1   Western  Pacific   114      89      75      63      50      41   64   4.1   World   385   369   341   288   246   216   44   2.3   Annex 10. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk by UNICEF region, 2015 Range  of  MMR  uncertainty   Number  of   Lifetime  risk  of   Lower   Upper     maternal   maternal  death:   Region   MMR   estimate   estimate   deaths   1  in   Sub-­‐Saharan  Africa 546 511 652 201  000 36 Eastern  and  Southern  Africa 417 387 512 70  000 51 West  and  Central  Africa 679 599 849 127  000 27 Middle  East  and  North  Africa 110 95 137 12  000 280 South  Asia 182 157 223 66  000 200 East  Asia  and  the  Pacific 62 56 76 18  000 880 Latin  America  and  Caribbean 68 64 77 7  300 670 Central  and  Eastern  Europe  and  the   Commonwealth  of  Independent   States 25 22 30 1  500 2  000 Least  developed  countries 436   207   514   135  000   52   World 216 207 249 303  000 180     Annex 11. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by UNICEF region, 1990–2015 %   Average   change   annual  %   in  MMR   change  in   MMR   between   MMR   1990   between   and   1990  and   UNICEF  region   1990   1995   2000   2005   2010   2015   2015   2015   Sub-­‐Saharan  Africa 987    928   846   717   624   546   45   2.4   Eastern  and   Southern  Africa 926    858   755   636   509   417   55   3.2   West  and  Central   Africa 1070   1020   956   814   749   679   37   1.8   Middle  East  and   North  Africa 221    198   170   145   122   110   50   2.8   South  Asia 558    476   388   296   228   182   67   4.5   East  Asia  and  the   Pacific 165    134   118    98    78    62   62   3.9   Latin  America  and   Caribbean   135    117    99    88    81    68   49   2.8   Central  and  Eastern   Europe  and  the   Commonwealth  of   Independent   States    69    71    56    43    29    25   64   4.2   Least  developed   countries   903   832   732   614   519   436   52   2.9   World   385   369   341   288   246   216   44   2.3   Annex 12. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk by UNFPA region, 2015 Range  of  MMR   uncertainty   Number  of   Lifetime  risk  of   Lower   Upper   maternal   maternal  death:   UNFPA  region   MMR   estimate   estimate   deaths   1  in:   Arab  States   162   138   212    15  000    170   Asia  and  the  Pacific   127   114   151    84  000    350   Eastern  and  Southern  Africa   407   377   501    66  000    52   Eastern  Europe  and  Central  Asia    25    22    30    1  490   2  000   Latin  America  and  the  Caribbean    68    64    77    7  290    670   West  and  Central  Africa   679   599   849   127  000    27                       World   216 207 249   303  000  180   Annex 13. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by UNFPA region, 1990–2015 Average   MMR   annual  %   %  change   change  in   in  MMR   MMR   between   between   1990  and   1990  and   UNFPA  region   1990   1995   2000   2005   2010   2015   2015   2015    Arab  States   306   285   250   216   181   162   47 2.5    Asia  and  the  Pacific   353   316   271   209   160   127   64 4.1   Eastern  and  Southern  Africa   918   848   746   627   500   407   56 3.3   Eastern  Europe  and  Central  Asia   70   71   56   44   29   25   64 4.2   Latin  America  and  the   Caribbean   135   117   99   88   81   68   49 2.8   West  and  Central  Africa   1070   1020   956   814   749   679   37 1.8                 World   385   369   341   288   246   216   44   2.3   Annex 14. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk by World Bank Group region and income group, 2015 Range  of  MMR  uncertainty Lifetime  risk   Number  of   of  maternal   World  Bank  Group  region  and   Lower Upper   maternal   death: income  group MMR estimate estimate deaths 1  in: Low  income 495 468 586 113  000        41   Middle  income 185 170 221 188  000    220   Lower  middle  income 253 229 305 169  000    130   Upper  middle  income    55      47      73      19  000    970   Low  and  middle  income 242 232 279 300  000    150   East  Asia  and  Pacific    63      57      77      18000   860   Europe  and  Central  Asia    25      22      30          1000   1900 Latin  America  and  the  Caribbean    69      65      79   6200    670   Middle  East  and  North  Africa    90      78   116 7800    350   South  Asia 182 157 223 66000   200   Sub-­‐Saharan  Africa 547 512 653    201000          36   High  income    17      16      19   2800   3300 World 216 207 249 303  000    180       Annex 15. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by World Bank Group region and income group, 1990–2015 MMR   Average   annual  %   %  change   change  in   in  MMR   MMR   between   between   World  Bank  Group  region  and   1990  and   1990  and   income  group   1990   1995   2000   2005   2010   2015   2015   2015   Low  income   1020   944   839   705   593   495   51   2.9   Middle  income      356   330   299   248   210   185   48   2.6   Lower  middle  income      532   470   411   337   287   253   52   3.0   Upper  middle  income      117   101      88      75      64      55   53   3.0   Low  and  middle  income      435   416   383   324   276   242   44   2.3   East  Asia  and  Pacific      168   137   120   100      79      63   63   3.9   Europe  and  Central  Asia          71      67      55      43      29      25   65   4.3   Latin  America  and  the   Caribbean      138   120   101      90      83      69   50   2.8   Middle  East  and  North  Africa      181   152   125   110      99      90   50   2.8   South  Asia      558   476   388   296   228   182   67   4.5   Sub-­‐Saharan  Africa      987   928   846   717   625   547   45   2.4   High  income          27      26      22      20      19   17   37   1.9   World      385   369   341   288   246   216   44   2.3           Annex 16. Estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), number of maternal deaths, and lifetime risk by UNPD region, 2015 Range  of  MMR  uncertainty   Lifetime  risk   Number  of   of  maternal   Lower Upper   maternal   death: UNPD  region MMR   estimate estimate deaths 1  in: Africa 495 464 590 204  000          42   Sub-­‐Saharan  Africa     555 518 664 197  000          35   Asia 119 108 141    90  000        370   Europe    13      11      15          1  000   4  800 Latin  America  and  the   Caribbean    67      64      77          7  300        670   Northern  America    13      11      15                580   4  100 Oceania    82      44   163                530        510   More  Developed  Regions      12      11      14          1  700   4  900 Less  Developed  Regions   238 157 210 302  000      150   Least  developed  countries     436 418 514 135  000          52   Less  developed  regions,   excluding  least  developed   countries   174 157 210 167  000        230   World 216 207 249 303  000      180       Annex 17. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by UNPD region, 1990–2015 MMR %  change  in   Average  annual   MMR   %  change  in   between  1990   MMR  between   Region 1990 1995 2000 2005 2010 2015 and  2015 1990  and  2015 Africa 870 834 770 654 565 495 43 2.3 Sub-­‐Saharan   Africa 996 939 858 728 635 555 44 2.3 Asia 329 293 251 195 149 119 64 4.1 Europe    31      30      21      17      14      13   58 3.6 Latin  America   and  the   Caribbean 135 117    99      88      81      67   50 2.8 Northern   America    11      11      12      13      14      13   –18       –0.6     Oceania 159 138 134 108    91      82   48 2.7 More   Developed   Regions    23      22      17      15      13      12   48 2.6 Less   Developed   Regions 430 409 377   319 272 238 45 2.4   Least   developed   countries 903 832 732 614 519 436  52 2.9   Less   developed   regions,   excluding   least   developed   countries 328 303 276 230 196 174 47 2.5   World 385   369   341   288   246   216   44   2.3         Annex 18. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by United Nations Millennium Development Goal region (indicated in bold) and other grouping, 1990–2015 MMR Average   annual  %   %  change   change  in   in  MMR   MMR   between   between   1990  and   1990  and   MDG  region 1990 1995 2000 2005 2010 2015 2015 2015 World 385   369   341   288   246   216   44 2.3 Developed  regionsa 23 22 17 15 13 12 48   2.6 Developing  regions 430 409 377 319 273 239 44 2.4 Africa 870 834 770 654 565 495 43 2.3   Northern  Africab 171   141   113   95   82   70   59 3.6 Sub-­‐Saharan  Africa   987   928   846   717   624   546   45 2.4 Eastern  Africac 995 906 790 659 521 424 57 3.4 Middle  Africad 958 978 911 799 748 650 32 1.6 e Southern  Africa   161 115 144 171 189 167 –4 –0.2 Western  Africaf 1120 1050 974 812 734 675 40 2.0 Asia 341 303 259 201 154 123 64 4.1 Eastern  Asiag 95   71   59   48   36   27   72 5.0 Eastern  Asia   excluding  China 51 51 68 57 52 43 16   0.7 Southern  Asiah 538   461   377   288   221   176   67 4.5 Southern  Asia   excluding  India 495 438 384 306 235 180 64 4.1 i South-­‐eastern  Asia 320   241   201   166   136   110   66 4.3 Western  Asiaj 160   141   122   110   96   91   43 2.2 Caucasus  and  Central   Asiak 69   68   50   46   37   33   52 3.0 Latin  America  and  the   Caribbean 135 117 99 88 81 67 50 2.8 l Latin  America 124   107   91   80   74   60   52 2.9 MMR Average   annual  %   %  change   change  in   in  MMR   MMR   between   between   1990  and   1990  and   MDG  region 1990 1995 2000 2005 2010 2015 2015 2015 Caribbeanm 276   257   214   198   180   175   37 1.8 n Oceania 391   320   292   239   206   187   52 3.0 a  Albania,  Australia,  Austria,  Belarus,  Belgium,  Bosnia  and  Herzegovina,  Bulgaria,  Canada,  Croatia,  Cyprus,   Czech  Republic,  Denmark,  Estonia,  Finland,  France,  Germany,  Greece,  Hungary,  Iceland,  Ireland,  Israel,  Italy,   Japan,  Latvia,  Lithuania,  Luxembourg,  Malta,  Montenegro,  Netherlands,  New  Zealand,  Norway,  Poland,   Portugal,  Republic  of  Moldova,  Romania,  Russian  Federation,  Serbia,  Slovakia,  Slovenia,  Spain,  Sweden,   Switzerland,  the  former  Yugoslav  Republic  of  Macedonia,  Ukraine,  United  Kingdom,  United  States  of  America.   b  Algeria,  Egypt,  Libya,  Morocco,  Tunisia.   c  Burundi,  Comoros,  Djibouti,  Eritrea,  Ethiopia,  Kenya,  Madagascar,  Malawi,  Mauritius,  Mozambique,  Rwanda,   Somalia,  South  Sudan,  Sudan,  Uganda,  United  Republic  of  Tanzania,  Zambia,  Zimbabwe.   d  Angola,  Cameroon,  Central  African  Republic,  Chad,  Congo,  Democratic  Republic  of  the  Congo,  Equatorial   Guinea,  Gabon,  Sao  Tome  and  Principe.   e  Botswana,  Lesotho,  Namibia,  South  Africa,  Swaziland.   f  Benin,  Burkina  Faso,  Cabo  Verde,  Côte  d’Ivoire,  Gambia,  Ghana,  Guinea,  Guinea-­‐Bissau,  Liberia,  Mali,   Mauritania,  Niger,  Nigeria,  Senegal,  Sierra  Leone,  Togo.   g  China,  Democratic  People’s  Republic  of  Korea,  Mongolia,  Republic  of  Korea.   h  Afghanistan,  Bangladesh,  Bhutan,  India,  Iran  (Islamic  Republic  of),  Maldives,  Nepal,  Pakistan,  Sri  Lanka.   i  Brunei  Darussalam,  Cambodia,  Indonesia,  Lao  People’s  Democratic  Republic,  Malaysia,  Myanmar,  Philippines,   Singapore,  Thailand,  Timor-­‐Leste,  Viet  Nam.   j  Bahrain,  Iraq,  Jordan,  Kuwait,  Lebanon,  Occupied  Palestinian  Territory,  Oman,  Qatar,  Saudi  Arabia,  Syrian   Arab  Republic,  Turkey,  United  Arab  Emirates,  Yemen.   k  Armenia,  Azerbaijan,  Georgia,  Kazakhstan,  Kyrgyzstan,  Tajikistan,  Turkmenistan,  Uzbekistan.   l  Argentina,  Belize,  Bolivia  (Plurinational  State  of),  Brazil,  Chile,  Colombia,  Costa  Rica,  Ecuador,  El  Salvador,   Guatemala,  Guyana,  Honduras,  Mexico,  Nicaragua,  Panama,  Paraguay,  Peru,  Suriname,  Uruguay,  Venezuela   (Bolivarian  Republic  of).   m  Bahamas,  Barbados,  Cuba,  Dominican  Republic,  Grenada,  Haiti,  Jamaica,  Puerto  Rico,  Saint  Lucia,  Saint   Vincent  and  the  Grenadines,  Trinidad  and  Tobago.   n  Fiji,  Kiribati,  Micronesia  (Federated  States  of),  Papua  New  Guinea,  Samoa,  Solomon  Islands,  Tonga,  Vanuatu.         Annex 19. Trends in estimates of maternal mortality ratio (MMR, maternal deaths per 100 000 live births), by country, 1990–2015 Average Range of annual % uncertainty on % change change in annual % change MMRb in MMR MMR in MMR (80% UI) between between Progress 1990 and 1990 and Lower Upper towards Countrya 1990 1995 2000 2005 2010 2015 2015c 2015 estimate estimate MDG 5Ad Making Afghanistan 1340 1270 1100 821 584 396 70.4 4.9 3.0 6.4 progress Albania 71 53 43 30 30 29 59.2 3.7 1.6 6.2 NA Algeria 216 192 170 148 147 140 35.2 1.8 –0.8 3.5 No progress Making Angola 1160 1150 924 705 561 477 58.9 3.5 1.5 5.5 progress Argentina 72 63 60 58 58 52 27.8 1.3 0.3 2.0 NA Armenia 58 50 40 40 33 25 56.9 3.3 2.4 4.2 NA Australia 8 8 9 7 6 6 25.0 1.3 0.1 2.0 NA Austria 8 6 5 5 4 4 50.0 2.9 2.0 4.2 NA Azerbaijan 64 86 48 34 27 25 60.9 3.8 2.3 5.4 NA Bahamas 46 49 61 74 85 80 –73.9 –2.2 –4.4 –0.1 NA Bahrain 26 22 21 20 16 15 42.3 2.1 0.7 3.2 NA Making Bangladesh 569 479 399 319 242 176 69.1 4.7 2.5 6.1 progress Barbados 58 49 48 40 33 27 53.4 3.0 1.8 4.8 NA Belarus 33 33 26 13 5 4 87.9 8.1 6.4 9.6 NA Belgium 9 10 9 8 8 7 22.2 0.8 –0.8 1.9 NA Belize 54 55 53 52 37 28 48.1 2.7 1.6 4.0 NA Benin 576 550 572 502 446 405 29.7 1.4 –0.6 2.8 No progress Bhutan 945 636 423 308 204 148 84.3 7.4 5.0 9.1 Achieved Bolivia (Plurinational Insufficient State of) 425 390 334 305 253 206 51.5 2.9 0.5 4.5 progress Bosnia and Herzegovina 28 22 17 14 13 11 60.7 3.6 2.1 5.4 NA Insufficient Botswana 243 238 311 276 169 129 46.9 2.5 0.1 4.2 progress Making Brazil 104 84 66 67 65 44 57.7 3.5 2.5 4.5 progress Brunei Darussalam 35 33 31 30 27 23 34.3 1.8 0.3 3.7 NA Bulgaria 25 24 21 15 11 11 56.0 3.3 2.0 4.6 NA Insufficient Burkina Faso 727 636 547 468 417 371 49.0 2.7 1.3 4.4 progress Insufficient Burundi 1220 1210 954 863 808 712 41.6 2.2 0.6 3.7 progress Cabo Verde 256 150 83 54 51 42 83.6 7.2 5.2 9.2 Achieved Cambodia 1020 730 484 315 202 161 84.2 7.4 5.6 8.9 Achieved Cameroon 728 749 750 729 676 596 18.1 0.8 –1.0 2.0 No progress Canadae 7 9 9 9 8 7 0.0 0.3 –0.9 1.6 NA Average Range of annual % uncertainty on % change change in annual % change MMRb in MMR MMR in MMR (80% UI) between between Progress 1990 and 1990 and Lower Upper towards Countrya 1990 1995 2000 2005 2010 2015 2015c 2015 estimate estimate MDG 5Ad Central African Republic 1290 1300 1200 1060 909 882 31.6 1.5 –0.4 3.4 No progress Insufficient Chad 1450 1430 1370 1170 1040 856 41.0 2.1 0.2 3.7 progress Chile 57 41 31 27 26 22 61.4 3.8 3.0 4.7 NA China 97 72 58 48 35 27 72.2 5.2 4.2 6.3 NA Insufficient Colombia 118 105 97 80 72 64 45.8 2.4 1.0 3.3 progress Insufficient Comoros 635 563 499 436 388 335 47.2 2.6 1.0 4.2 progress Congo 603 634 653 596 509 442 26.7 1.2 –0.3 2.7 No progress Costa Rica 43 44 38 31 29 25 41.9 2.2 1.5 3.1 NA Côte d’Ivoire 745 711 671 742 717 645 13.4 0.6 –0.7 1.9 No progress Croatia 10 12 11 11 10 8 20.0 0.6 –0.8 1.9 NA Cuba 58 55 43 41 44 39 32.8 1.6 0.7 2.5 NA Cyprus 16 17 15 12 8 7 56.3 3.3 1.7 5.4 NA Czech Republic 14 10 7 6 5 4 71.4 4.8 3.3 6.4 NA Democratic People’s Republic of Korea 75 81 128 105 97 82 –9.3 –0.4 –2.3 1.6 NA Democratic Republic of the Congo 879 914 874 787 794 693 21.2 1.0 –1.1 2.4 No progress Denmark 11 11 9 8 7 6 38.8 2.0 0.6 2.9 NA Making Djibouti 517 452 401 341 275 229 55.7 3.3 1.4 5.1 progress Dominican Making Republic 198 198 79 64 75 92 53.5 3.1 1.3 4.7 progress Making Ecuador 185 131 103 74 75 64 65.4 4.3 3.6 5.0 progress Making Egypt 106 83 63 52 40 33 68.9 4.7 3.8 5.9 progress Making El Salvador 157 118 84 68 59 54 65.5 4.3 3.0 5.7 progress Making Equatorial Guinea 1310 1050 702 483 379 342 73.9 5.4 3.6 7.0 progress Making Eritrea 1590 1100 733 619 579 501 68.5 4.6 3.0 6.0 progress Estonia 42 43 26 15 8 9 78.6 6.1 4.3 7.9 NA Making Ethiopia 1250 1080 897 743 523 353 71.8 5.0 2.7 6.5 progress Fiji 63 51 42 39 34 30 52.2 3.0 1.6 5.0 NA Finland 6 5 5 4 3 3 50.0 3.3 2.1 5.1 NA France 15 15 12 10 9 8 46.7 2.2 1.2 3.4 NA Gabon 422 405 405 370 322 291 31.0 1.5 –0.5 2.9 No progress Gambia 1030 977 887 807 753 706 31.5 1.5 –0.4 2.9 No progress Georgia 34 35 37 37 40 36 –5.9 –0.2 –1.4 1.0 NA Average Range of annual % uncertainty on % change change in annual % change MMRb in MMR MMR in MMR (80% UI) between between Progress 1990 and 1990 and Lower Upper towards Countrya 1990 1995 2000 2005 2010 2015 2015c 2015 estimate estimate MDG 5Ad Germany 11 9 8 7 7 6 45.5 2.3 1.5 3.2 NA Insufficient Ghana 634 532 467 376 325 319 49.7 2.7 1.3 4.4 progress Greece 5 4 4 3 3 3 40.0 1.8 0.6 3.3 NA Grenada 41 37 29 25 27 27 34.1 1.7 –0.4 3.0 NA Making Guatemala 205 173 178 120 109 88 57.1 3.4 2.8 4.0 progress Insufficient Guinea 1040 964 976 831 720 679 34.7 1.7 0.2 2.9 progress Insufficient Guinea-Bissau 907 780 800 714 570 549 39.5 2.0 0.2 3.8 progress Guyana 171 205 210 232 241 229 –33.9 –1.2 –2.6 –0.3 No progress Haiti 625 544 505 459 389 359 42.6 2.2 –0.2 3.8 No progress Making Honduras 272 166 133 150 155 129 52.6 3.0 2.0 4.1 progress Hungary 24 20 15 14 15 17 29.2 1.5 0.2 2.7 NA Iceland 7 6 5 4 4 3 57.1 2.6 1.1 4.8 NA Making India 556 471 374 280 215 174 68.7 4.6 3.5 5.7 progress Making Indonesia 446 326 265 212 165 126 71.7 5.0 3.4 6.3 progress Iran (Islamic Republic of) 123 80 51 34 27 25 79.7 6.4 5.3 7.8 Achieved Making Iraq 107 87 63 54 51 50 53.3 3.1 1.5 5.2 progress Ireland 11 10 9 8 7 8 27.3 1.5 –0.1 2.4 NA Israel 11 10 8 7 6 5 54.5 3.0 2.1 3.9 NA Italy 8 7 5 4 4 4 50.0 3.0 1.8 4.4 NA Jamaica 79 81 89 92 93 89 –12.7 –0.4 –1.9 0.8 NA Japan 14 11 10 7 6 5 64.3 3.6 2.6 4.8 NA Insufficient Jordan 110 93 77 62 59 58 47.3 2.6 1.2 4.1 progress Kazakhstan 78 92 65 44 20 12 84.6 7.5 6.5 8.5 NA Kenya 687 698 759 728 605 510 25.8 1.2 –0.5 2.8 No progress Making Kiribati 234 207 166 135 109 90 61.5 3.8 2.0 6.0 progress Kuwait 7 9 7 6 5 4 42.9 2.0 0.4 3.0 NA Kyrgyzstan 80 92 74 85 84 76 5.0 0.2 –0.9 1.3 NA Lao People’s Democratic Republic 905 695 546 418 294 197 78.2 6.1 3.9 7.7 Achieved Latvia 48 54 30 22 19 18 62.5 3.9 2.3 5.4 NA Lebanon 74 54 42 27 19 15 79.7 6.4 4.6 7.8 NA Lesotho 629 525 649 746 587 487 22.5 1.0 –1.9 2.9 No progress Insufficient Liberia 1500 1800 1270 1020 811 725 51.7 2.9 0.8 4.2 progress Libya 39 25 17 11 9 9 76.9 5.7 2.8 8.8 NA Average Range of annual % uncertainty on % change change in annual % change MMRb in MMR MMR in MMR (80% UI) between between Progress 1990 and 1990 and Lower Upper towards Countrya 1990 1995 2000 2005 2010 2015 2015c 2015 estimate estimate MDG 5Ad Lithuania 29 28 16 12 9 10 65.5 4.3 2.8 5.8 NA Luxembourg 12 13 13 13 11 10 16.7 0.8 –1.6 2.6 NA Making Madagascar 778 644 536 508 436 353 54.6 3.2 1.8 4.5 progress Malawi 957 953 890 648 629 634 33.8 1.6 –0.7 3.3 No progress Malaysia 79 68 58 52 48 40 49.4 2.7 0.8 3.9 NA Maldives 677 340 163 101 87 68 90.0 9.2 6.2 11.6 Achieved Mali 1010 911 834 714 630 587 41.9 2.2 0.6 3.2 Insufficient progress Malta 13 14 15 13 11 9 30.8 1.6 –0.9 3.3 NA Mauritania 859 824 813 750 723 602 29.9 1.4 –1.2 3.2 No progress Mauritius 81 60 40 39 59 53 34.6 1.6 0.1 3.1 NA Mexico 90 85 77 54 45 38 57.8 3.4 3.0 3.9 NA Micronesia 183 166 153 134 115 100 45.4 2.4 0.4 4.4 Insufficient (Federated States of) progress Mongolia 186 205 161 95 63 44 76.3 5.8 4.4 7.1 Achieved Montenegro 10 12 11 9 8 7 30.0 1.3 –0.5 3.9 NA Making Morocco 317 257 221 190 153 121 61.8 3.8 2.7 5.1 progress Making Mozambique 1390 1150 915 762 619 489 64.8 4.2 2.5 5.5 progress Making Myanmar 453 376 308 248 205 178 60.7 3.7 1.6 5.3 progress Namibia 338 320 352 390 319 265 21.6 1.0 –1.3 3.1 No progress Making Nepal 901 660 548 444 349 258 71.4 5.0 2.6 6.8 progress Netherlands 12 13 14 11 8 7 41.7 2.0 1.1 3.3 NA New Zealand 18 15 12 14 13 11 38.9 1.9 0.8 2.9 NA Nicaragua 173 212 202 190 166 150 13.3 0.6 –0.7 1.9 No progress Insufficient Niger 873 828 794 723 657 553 36.7 1.8 0.4 3.0 progress Nigeria 1350 1250 1170 946 867 814 39.7 2.0 –0.2 3.3 No progress Norway 7 7 7 7 6 5 28.6 1.5 0.3 2.5 NA Occupied Palestinian Making Territoryf 118 96 72 62 54 45 61.9 3.8 1.8 5.8 progress Oman 30 20 20 20 18 17 43.2 2.3 0.6 3.8 NA Making Pakistan 431 363 306 249 211 178 58.7 3.5 1.8 5.1 progress Panama 102 94 82 87 101 94 7.8 0.3 –1.0 1.4 No progress Papua New Insufficient Guinea 470 377 342 277 238 215 54.3 3.1 1.1 5.3 progress Paraguay 150 147 158 159 139 132 12.0 0.5 –0.7 1.6 No progress Making Peru 251 206 140 114 92 68 72.9 5.2 4.2 6.7 progress Average Range of annual % uncertainty on % change change in annual % change MMRb in MMR MMR in MMR (80% UI) between between Progress 1990 and 1990 and Lower Upper towards Countrya 1990 1995 2000 2005 2010 2015 2015c 2015 estimate estimate MDG 5Ad Philippines 152 122 124 127 129 114 25.0 1.1 –0.8 2.4 No progress Poland 17 13 8 6 4 3 82.4 6.8 5.4 8.2 NA Portugal 17 15 13 12 11 10 41.2 2.1 1.1 2.9 NA Puerto Rico 26 25 22 19 16 14 46.2 2.4 1.5 3.9 NA Qatar 29 28 24 21 16 13 55.2 3.3 0.8 4.9 NA Republic of Korea 21 19 16 14 15 11 47.6 2.6 1.8 3.5 NA Republic of Moldova 51 66 49 39 34 23 54.9 3.2 2.3 4.2 NA Making Romania 124 77 51 33 30 31 75.0 5.5 4.0 6.9 progress Russian Federation 63 82 57 42 29 25 60.3 3.8 2.5 5.1 NA Rwanda 1300 1260 1020 567 381 290 77.7 6.0 4.5 7.5 Achieved Saint Lucia 45 43 54 67 54 48 –6.7 –0.2 –2.1 1.6 NA Saint Vincent and the Grenadines 58 81 74 50 50 45 22.4 1.1 –0.5 2.4 NA Making Samoa 156 119 93 77 64 51 67.3 4.4 2.4 6.3 progress Sao Tome and Making Principe 330 263 222 181 162 156 52.7 3.0 1.2 5.4 progress Saudi Arabia 46 33 23 18 14 12 73.9 5.5 3.7 7.5 NA Insufficient Senegal 540 509 488 427 375 315 41.7 2.2 0.7 3.6 progress Serbia 14 15 17 15 16 17 –21.4 –0.8 –2.8 0.9 NA Insufficient Sierra Leone 2630 2900 2650 1990 1630 1360 48.3 2.6 0.5 4.0 progress Singapore 12 13 18 16 11 10 16.7 0.8 –1.4 2.9 NA Slovakia 11 9 8 7 6 6 45.5 2.8 1.8 4.0 NA Slovenia 12 12 12 11 9 9 25.0 1.2 –1.0 2.6 NA Making Solomon Islands 364 273 214 164 136 114 68.7 4.6 3.1 6.4 progress Somalia 1210 1190 1080 939 820 732 39.5 2.0 0.3 3.9 Insufficient progress South Africa 108 62 85 112 154 138 –27.8 –1.0 –2.5 0.6 No progress South Sudan 1730 1530 1310 1090 876 789 54.4 3.1 1.4 4.7 Making progress Spain 6 6 5 5 5 5 16.7 1.0 –0.1 1.8 NA Sri Lanka 75 70 57 43 35 30 60.0 3.6 2.6 4.5 NA Sudan 744 648 544 440 349 311 58.2 3.5 2.0 5.4 Making progress Suriname 127 177 259 223 169 155 –22.0 –0.8 –2.4 0.8 No progress Swaziland 635 537 586 595 436 389 38.7 2.0 –0.1 3.4 No progress Sweden 8 6 5 5 4 4 50.0 2.5 1.2 3.3 NA Switzerland 8 8 7 7 6 5 37.5 1.8 0.3 2.8 NA Syrian Arab 123 89 73 58 49 68 44.7 2.4 0.3 3.9 Insufficient Republic progress Average Range of annual % uncertainty on % change change in annual % change MMRb in MMR MMR in MMR (80% UI) between between Progress 1990 and 1990 and Lower Upper towards Countrya 1990 1995 2000 2005 2010 2015 2015c 2015 estimate estimate MDG 5Ad Tajikistan 107 129 68 46 35 32 70.1 4.8 2.9 7.0 Making progress Thailand 40 23 25 26 23 20 50.0 2.7 0.8 4.3 NA The former 14 13 12 10 8 8 42.9 2.4 1.2 4.1 NA Yugoslav Republic of Macedonia Timor-Leste 1080 897 694 506 317 215 80.1 6.5 4.8 8.0 Achieved Togo 568 563 491 427 393 368 35.2 1.7 0.5 3.2 Insufficient progress Tonga 75 100 97 114 130 124 –65.3 –2.0 –4.0 0.0 NA Trinidad and 90 77 62 62 65 63 30.0 1.5 0.5 2.5 NA Tobago Tunisia 131 112 84 74 67 62 52.7 3.0 1.4 4.3 Making progress Turkey 97 86 79 57 23 16 83.5 7.2 5.2 9.1 NA Turkmenistan 82 74 59 53 46 42 48.8 2.7 0.4 5.8 NA Uganda 687 684 620 504 420 343 50.1 2.8 1.3 4.1 Making progress Ukraine 46 52 34 30 26 24 47.8 2.6 1.4 3.7 NA United Arab Emirates 17 12 8 6 6 6 64.7 4.1 2.2 6.8 NA United Kingdom 10 11 12 12 10 9 10.0 0.4 –0.3 1.2 NA United Republic of Making Tanzania 997 961 842 687 514 398 60.1 3.7 2.2 5.0 progress United States of America 12 12 12 13 14 14 –16.7 –0.6 –1.4 0.1 NA Uruguay 37 36 31 26 19 15 59.5 3.7 2.4 5.1 NA Uzbekistan 54 32 34 42 39 36 33.3 1.6 –0.8 4.0 NA Making Vanuatu 225 184 144 116 94 78 65.3 4.2 2.3 6.2 progress Venezuela 94 90 90 93 99 95 –1.1 –0.1 –1.3 0.9 NA Making Viet Nam 139 107 81 61 58 54 61.2 3.8 1.6 5.2 progress Yemen 547 498 440 428 416 385 29.6 1.4 –0.8 3.0 No progress Making Zambia 577 596 541 372 262 224 61.2 3.8 2.6 5.2 progress Zimbabwe 440 449 590 629 446 443 –0.7 0.0 –1.4 0.9 No progress MDG:  Millennium    Development  Goal;  NA:  data  not  available;  UI:  uncertainty    interval.   a  Estimates  have  been  computed  to  ensure  comparability  across  countries,  thus  they  are  not  necessarily  the   same  as  official  statistics  of  the  countries,  which  may  use  alternative  rigorous  methods.   b  MMR  estimates  have  been  rounded  according  to  the  following  scheme:  <  100  rounded  to  nearest  1;  100–999   rounded  to  nearest  1;  and  ≥  1000  rounded  to  nearest  10.   c  Percentage  change  in  MMR  is  based  on  rounded  numbers.   d  Progress  towards  MDG  5A  (i.e.  to  reduce  MMR  by  75%  between  1990  and  2015)  was  assessed  for  the  95   countries  with  an  MMR  higher  than  100  in  1990.  See  section  4.1  and  Box  5  for  additional  details  in  the  full   report:  World  Health  Organization  (WHO),  United  Nations  Children’s  Fund  (UNICEF),  United  Nations   Population  Fund  (UNFPA),  World  Bank  Group,  United  Nations    Population    Division  (UNPD).  Trends  in  maternal   mortality:  1990  to  2015.  Geneva:  WHO;  2015  (available  from:   http://www.who.int/reproductivehealth/publications/monitoring/maternal-­‐mortality-­‐2015/en/).   e  Vital  registration    data  were  available  for  analysis  only  up  to  2011.  Recent  hospital  surveillance  data  for   Canada  (excluding  Quebec)  indicate  a  decline  of  maternal  deaths  per  100  000  deliveries  from  8.8  in   2007/2008–2008/2009    to  5.1  in  2011/2012;  some  98%  of  deliveries  in  Canada  occur  in  hospitals.   f  Refers  to  a  territory.     http://www.who.int/reproductivehealth