Comment Introducing the World Bank’s 2018 Health Equity and Financial Protection Indicators database Among the many shifts of emphasis that have been expenses (those exceeding a specified proportion Published Online October 22, 2018 evident in global health over the past 25 years or so, of a household’s total consumption or income) or http://dx.doi.org/10.1016/ two stand out: a concern over the poor lagging behind impoverishing expenses (expenses without which the S2214-109X(18)30437-6 the better off in progress towards global goals; and a household would have been above the poverty line, but concern to look beyond whether people get the services because of the expenditures is below the poverty line). they need to the affordability of the out-of-pocket The health indicators include both MDG-era indicators expenditures associated with these services. These and SDG-era (eg, NCD) indicators, and the financial concerns over health equity and financial protection protection indicators include those that reflect the SDG were absent from the Millennium Development catastrophic expenditure threshold (10%) and the SDG Goals (MDGs), but are integral to the Sustainable international poverty line (US$1.90 per day). Development Goals (SDGs). The data are calculated from household surveys, The World Bank’s 2018 Health Equity and Financial identified mostly through searches of data catalogues Protection Indicators (HEFPI) database1 is a new global and websites of multicountry survey initiatives. None resource for tracking progress on both fronts. It is, in come from official reports by national governments, effect, the fourth in the series of such databases. The in part because such data do not lend themselves to first two2,3 (2000 and 2007) focused on maternal and disaggregation by household living standards, and in child health and communicable diseases, and drew on part because of concerns about accuracy, especially data from Demographic and Health Surveys. The third4 where governments do not face incentives to report (2012) added data from Multiple Indicator Cluster accurate numbers.5–7 Where we have been able to Surveys and World Health Surveys, non-communicable access the raw microdata, we have done so, mostly disease (NCD) and financial protection indicators, and because indicator definitions can vary from one survey high-income countries. The 2018 database continues family to another, and sometimes even within a survey this broadening-out, including more health indicators, family. The estimates we report are simply direct (re) more countries, and more years of data. Disaggregated calculations of the quantities reported in the survey health data by wealth quintile are reported in all four reports, harmonised as much as possible across surveys datasets. subject to the constraints imposed by the wordings of The 2018 database includes 18 indicators of service use the original questions. In line with the growing concerns (12 preventative, six curative) and 28 health outcome about the use of modelling in global health datasets,8,9 indicators. The financial protection indicators capture we do not produce forecasts for country-years where the proportions of the population incurring catastrophic there is no survey. Nor do we replace estimates directly Number of Mean number of Mean number of Mean number of Mean number of indicators datapoints countries datapoints countries (population)* (population)* (quintiles)† (quintiles)† Service coverage (MDGs: MCH, etc) 10 279 92 271 88 Service coverage (SDGs: NCDs) 17 182 81 122 55 Health outcomes (MDGs: MCH, etc) 5 217 75 217 75 Health outcomes (SDGs: NCDs) 14 157 96 84 48 Catastrophic expenditure 2 575 142 101 35 Impoverishing expenditure 3 548 141 96 34 Total 51 234 93 148 59 MDGs=Millennium Development Goals. MCH=maternal and child health. SDGs=Sustainable Development Goals. NCDs=non-communicable diseases. *Estimates reported for the population as a whole. †Disaggregated data reported at the level of the (wealth or consumption) quintile. Table: Data availability in 2018 Health Equity and Financial Protection Indicators dataset www.thelancet.com/lancetgh Vol 7 January 2019 e22 Comment calculated from the survey microdata by modelled paper1 and highlight the differences between our estimates. The downside is that our dataset is full of definitions and others’; we also provide the essential gaps. The upside is that, insofar as the surveys we use are computer code used to produce the estimates. reliable, differences over time or across countries ought to reflect reality rather than modelling assumptions; *Adam Wagstaff, Patrick Eozenou, Sven Neelsen, conversely, when real changes occur on the ground, they Marc-Francois Smitz ought to get reflected in our numbers, rather than being Development Research Group (AW) and Health, Nutrition and Population Global Practice (PE, SN, M-FS), World Bank, smoothed away by the modelling process. Washington DC 20433, USA The health data were checked against the reports awagstaff@worldbank.org and websites of the original surveys where possible; We declare no competing interests. differences are typically small and due to our © 2018 The World Bank. Published by Elsevier. This is an Open Access article harmonisation of definitions across surveys. The health published under the CC BY 3.0 IGO license which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is data were also checked to make to make sure they lie in properly cited. In any use of this article, there should be no suggestion that The the required range. The financial protection estimates World Bank endorses any specific organisation, products or services. The use of The World Bank logo is not permitted. This notice should be preserved along were subject to several internal and external checks, with the article’s original URL. which led to us to drop several household surveys, 1 Wagstaff A, Eozenou P, Neelsen S, Smitz M. The 2018 health equity and including several entire survey families. financial protection indicators database: overview and insights. Washington, DC: World Bank, 2018. http://documents.worldbank.org/ The HEFPI database now covers 193 countries, up curated/en/582871539784481127/The-2018-Health-Equity-and- Financial-Protection-Indicators-Database-Overview-and-Insights. from 109 previously, and draws on over 1600 surveys, 2 Gwatkin D, Rutstein S, Johnson K, Pande R, Wagstaff A. Socioeconomic up from just 285. The table shows the variation differences in health, nutrition, and population within developing countries: an overview. Washington, DC: World Bank, 2000. across groups of indicators in terms of the number of 3 Gwatkin D, Rutstein S, Johnson K, Suliman E, Wagstaff A, Amouzou A. datapoints and countries with data, for the population Socio-economic differences in health, nutrition, and population within developing countries: an overview. Washington, DC: World Bank, 2007. and quintile data. On average, across the 51 indicators, 4 Bredenkamp C, Wagstaff A, Buisman L, Prencipe L, Rohr D. Health equity we have population data for just over 90 countries, and financial protection datasheet : Sub-Saharan Africa. Washington, DC: World Bank, 2012. with an average of 2·5 years of data per country. For 5 Sandefur J, Glassman A. The political economy of bad data: evidence from African Survey and Administrative Statistics. J Dev Stud 2015; 51: 116–32. the financial protection indicators, we have data for 6 Lim SS, Stein DB, Charrow A, Murray CJL. Tracking progress towards over 140 countries; for the SDG health service coverage universal childhood immunisation and the impact of global initiatives: a systematic analysis of three-dose diphtheria, tetanus, and pertussis indicators, we have fewer countries. We also have less immunisation coverage. Lancet 2008; 372: 2031–46. disaggregated data, since we report disaggregated 7 Murray CJL, Shengelia B, Gupta N, Moussavi S, Tandon A, Thieren M. Validity of reported vaccination coverage in 45 countries. Lancet 2003; financial protection data only for high-income countries, 362: 1022–27. and some NCD surveys do not collect the data necessary 8 Boerma T, Victora C, Abouzahr C. Monitoring country progress and achievements by making global predictions: is the tail wagging the dog? to disaggregate by household living standards. Lancet 2018; 392: 607–09. 9 AbouZahr C, Boerma T, Hogan D. Global estimates of country health For the dataset see The 2018 HEFPI dataset is freely downloadable, and indicators: useful, unnecessary, inevitable? Glob Health Action 2017; https://datacatalog.worldbank. org/dataset/hefpi a data visualisation tool is also available. To ensure the 10 (suppl 1): 1290370. 10 Stevens GA, Alkema L, Black RE, et al. Guidelines for Accurate and For the data visualisation tool data are reproducible, and in line with the Guidelines for Transparent Health Estimates Reporting: the GATHER statement. Lancet see http://datatopics.worldbank. Accurate and Transparent Health Estimates Reporting,10 2016; 388: e19–23. org/hefpi we document our methods thoroughly in a working e23 www.thelancet.com/lancetgh Vol 7 January 2019