Report No: AUS10347 Republic of Moldova Health financing Policy Note: How to Reduce Out-of-Pocket Payments (OOPs) in the Health Sector in Moldova? May, 2015 GSPDR EUROPE AND CENTRAL ASIA Standard Disclaimer: This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The find- ings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judg- ment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Copyright Statement: The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dis- semination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. 2 1. Introduction KEY ISSUES Out-of-pocket payments (OOPs) are direct (at the point of o Public financing for health fails to pro- service) financial contributions or co-payments by patients vide adequate financial protection and their families associated with consumption of medical o Every year on average about 160,000 products (such as medicines) and/or services.1 They can be people are exposed to catastrophic formal as well as informal payments.2 Together with taxa- health expenditure tion, social and private health insurance contributions, o Lack of financial protection negatively they constitute the main sources of financing for medical affects insured and uninsured, but pri- products and services in all countries of the world. This marily poor and rural residents policy note looks at prevalence and trends of OOPs in o Reducing out-of-pocket payments and Moldova during 2007-2013, evaluates their impact on the making them more transparent should population’s economic well-being (section 2), identifies become a policy priority for the new key drivers of the current OOPs (section 3), and, based on government the analysis, suggests several policy options for govern- ment’s consideration (section 4). 2. Trends and patterns in out-of-pocket expenditure in Moldova During 2007-2013, households OOPs more than doubled, from approximately 2.32 to 5.28 billion Moldo- van Leu3 or from US$ 191.4 million to 404.7 million in nominal terms. Over the same period OOPs’ growth outpaced the increase in government health spending (from 2.6 to 5 billion Leu). Consequently the share of OOPs over total health expenditure slightly increased, and at the end of the period was above 50% of total health expenditure. As a share of total household spending OOP for health services slightly increased from 5.3 to 5.7 percent from 2007 to 2013.4 In International comparisons, Moldova stands out as a country with high percentage spent on OOP (ap- proximately 50 percent), and at the same time a relatively high percentage of GDP been absorbed by gov- ernment health expenditure (above 5 percent). Figure 1: Health financing in the East Europe and Central Asia region, 2012 1 They can be paid either before the service is provided, and in fact be a precondition for receiving the service, or after the service is provided, in which case they may be seen in some cases as a “gratuity” or gift by the patient to the health worker for the med- ical services she/he received. 2 Which OOP payments are formal and which are informal is determined in each country by the set of existing social entitlements. Any direct payment for services which is above the level determined by law or existing regulations is to be considered informal. For example, if for a specific blood test the co-payment is set at 10 Leu, and the patient pays 15 Leu, 10 Leu is considered the formal OOP payment, and 5 Leu the informal payment. It is extremely difficult to empirically distinguish between formal and in- formal OOP payments. For example, if a patient seeks care in a public hospital for a service which is provided for free at the point of service, but the doctor asks him/her to go instead to a private clinic the doctor owns, the price paid for the same service in the private clinic would be categorized as a formal OOP payment (if the patient is not entitled to that service for free in a pri- vate clinic), but in reality it should be considered as an informal payment. 3 Please see in Annex 1 a short note explaining the sources and basis for our estimates. 1 Moldovan Leu equals 0.055 US Dollar as st of March 1 . 4 National Bureau of Statistics http://statbank.statistica.md/pxweb/Database/EN/databasetree.asp (Accessed on November 2nd, 2014). 3 Source: Getting Better, Regional ECA Health Report, World Bank, 2013 In addition, over the period 2007-2013 the number of individuals who incurred catastrophic health-related payments - defined as OOPs greater than 40 percent of their total non-food consumption- did not improve and approximately 160,000 individuals, or 4.5 percent of the population, were affected annually. Therefore, one can conclude that over the period 2007-2013 the level of financial protection in Moldova did not improve, in spite of the significant increase in government health expenditure, and the repeated attempts implemented by the government since the inception of the social insurance system in 2004 to expand insurance coverage to a broader set of socio-economic groups. Figure 2: Total HH OOP Health Expenditure by Quintile Expenditure Groups 2007-2013 6,000,000 HH Health Spending 1,000 Lei 5,000,000 Average annual increase 21.5% 4,000,000 3,000,000 2,000,000 1,000,000 - 2007 2008 2009 2010 2011 2012 2,013 Lowest quintile 2 3 4 Highest quintile 4 Source: Household Budget Survey, 2007 -2013 2.1 Distribution of OOPs across socio-economic groups in Moldova By looking at the distribution of OOP across different socio-economic groups in Moldova (Figure 2 above), one can see that the richest forty percent of the society accounted for approximately sixty five percent of the total spent on OOPs in 20135, while the poorest 40 percent accounted for 18 percent. The doubling of OOPs between 2007 and 2013 was largely determined by an increased in OOPs made by the richest 40% of the population. Over the same period, the per-capita OOP in the poorest quintile grew “only” at an annual rate of 11.4%, while the average annual growth rate of OOPs was 21.5%. As a share of total household spending OOP for health services increased over the period from 2007 to 2013 for all quintiles but the poorest. The largest increase was registered for the second poorest quintile, from 5 to 5.8 percent of their total consumption expenditure. The share of those facing catastrophic payments was also lowest among the poorest 20% of the population – 2.7%, and highest among the richest 20% - 5.5%. 2.2 Should OOPs be of concern in Moldova? Following the above definition of “excessive” health-related OOP payments (40 percent of total non-food household expenditures spent on health services), Moldova does not do too well, because currently four and a half percent of the population are exposed to catastrophic health expenditures. However, if one on- ly focused on the poorer segments of society, at first it does appear that the poorest 20% households are better protected, because they are less likely to incur OOPs and face catastrophic health expenditure. Such conclusion would be mistaken, due to two main reasons. First, several sources of evidence, including the household budget surveys with extended health module6 conducted by the National Bureau of Statis- tics in 2008, 2010 and 2013, reveal that inpatient and outpatient service utilization is lowest among these 20% poorest households, in fact two times lower than among the richest 20% (Figure 3). While inequalities in outpatient service utilization noted in 2008 were somewhat reduced over the following five years, utili- zation rates for both outpatient and inpatient services continued to be strongly positively correlated with households’ consumption, income, and wealth. The same conclusion (low utilization of services by the poor) can be drawn by looking at the multiple indicator cluster survey (MICS)7 results. According to this survey, in 2012 the poorest 20% women were four times more likely forego antenatal care during preg- nancy, and, if they sought care, they were 5.5 times more likely to receive inadequate antenatal supervi- sion (i.e. less than 4 antenatal visits). Obviously, such behavior negatively affects maternal health out- comes and as a consequence, the poorest 20% of women were 2.7 times more likely to have longer post- partum stay in the hospital due to delivery complications. In addition, they continued to utilize health ser- vices less than the rich post-partum; they had 0.6 times lower probability of visiting a health provider after birth when compared with the richest 20%. All of these findings point to the fact that the poorest seg- ments of the population use the least amount of health services, and frequently forego treatment when in need, among other reasons due to lack of adequate insurance coverage and due to financial reasons,8 which obviously has negative impact on their health outcomes. 5 In per capita terms, in 2013 average monthly health OOP spending varied from Lei 437 for the poorest quintile to Lei 3,117 for the richest quintile. 6 People's access to health services: Household survey results. Summary report, 2013. National Bureau of Statistics of the Republic of Moldova. 7 National Centre of Public Health of the Ministry of Health of the Republic of Moldova, United Nations Children’s Fu nd (UNICEF), 2014. 2012 Republic of Moldova Multiple Indicator Cluster Survey, Final Report. Chișinău, Republic of Moldova. 8 People's access to health services: Household survey results. Summary report. 2013. National Bureau of Statistics of the Republic of Moldova. 5 Figure 3: Service utilization ratios between richest and poorest fifth of population during 2008-20139 Ratio: richest 20 per- cent/ poorest 20 per- cent Second, since poorer households are closer to the poverty line and have minimal living standards, even relatively low health-related OOP payments can push them into poverty, as Figure 4 below indicates. Figure 4: Effects of Health Payments on Households Consumption in 2013 8 pre-OOP consumption post-OOP consumption 6 4 2 0 0 .2 .4 .6 .8 1 Cumulative proportion of population, ranked from poorest to richest Figure 4 (called “Pen Parade”) shows that the poorer 40 percent households in Moldova are more likely to be pushed into poverty by OOPs, although the latter are more prevalent among richer segments of the population. In addition, the health care financing system in Moldova is failing to protect from catastrophic health payment -and consequently from impoverishment- mainly rural residents and residents of the northern and southern regions. In fact, during the period from 2007 to 2013, especially regional inequities 9 Ibid. 6 (although inequalities also exist between urban-rural locations) widened and geographic equity gaps be- came more pronounced, as Figures 5 and 6 indicate. Along with health financing issues, these inequalities could be due to lack of information and awareness, or inadequacy of a supply network and other supply side factors, which we were not able to fully explore in this analysis. Figure 5: Urban – Rural Distribution of HH OOP Health Spending Urban-Rural Equity gap has widened 7.0% Share of health expenditure in HH Total 6.5% 6.0% Consumption 5.5% 5.0% 4.5% 4.0% 2007 2008 2009 2010 2011 2012 2013 2014 Urban Rural Figure 6: Regional Distribution of HH OOP Health Spending Regional Equity gap has widened 7.0% Share of health expenditure in HH Total 6.5% 6.0% Consumption 5.5% 5.0% 4.5% 4.0% 2007 2008 2009 2010 2011 2012 2013 2014 Nord Centru Sud Chisinau 2.3 What is being purchased with OOPs? Numerous household surveys conducted in Moldova over Formal and informal payments are frequent during hospital stays. In addi- the past several years indicate that OOPs are a widespread tion, patients are being asked by hos- phenomenon in public health facilities, especially in hospi- pital staff to bring drugs and medical consumables tals, and negatively affect utilization, as they impose con- siderable financial barriers on uninsured and, to a lesser 7 extent, insured individuals.10,11 In addition to paying doctors and nurses informally, people are required to pay for consumables and drugs not available and/or provided by the hospital where they seek treatment. These hospital payments can be either formal or informal. Formal payment usually occurs if a patient wants better hotel conditions, which are provided by hospitals for a fee, or if they are paying for their treatment (hospitals have the right to charge fees from patients without referral, or without insurance). An informal payment happens when doctors or nurses are asking patients to pay “under the table” before treating them, or ask them bring certain consumables or drugs (arguments could differ – “We are out of stock today, but you need it now”; or: “We have a lower quality of medications and in the pharmacy across the street there is a better drug”, etc.), or when administrators, doctors or nurses are given a “bribe”, or a gratuity, as a direct payment for their services after providing them. For example, in 2012 out of all hospitalized, 14.4% had to purchase and bring medical consumables, and 40.5% drugs; furthermore, 16% brought their food, and 35.6% bed linen. Food and bed linen are usually provided by hospitals, but the quality is sub-optimal, and therefore patients prefer to bring their own12. Patients treated in rayon and municipal hospitals were more likely to face such informal payments and in- kind contributions compared to those treated in republican facilities13. Both poor and rich face OOPs, though prevalence of informal payments among poorest 20% is lower compared to richest 20% (for exam- ple, respectively, 24.3% of the poorest vs. 54.4% of the richest 20% paid for drugs while in hospital). Figure 7: Structure of Household Spending on Health Services 100% 90% 80% 70% 61% 71% 72% 73% 71% 60% 77% 50% 40% 30% 22% 20% 18% 18% 17% 18% 17% 10% 15% 10% 10% 8% 10% 5% 0% 2008 2009 2010 2011 2012 5-year Average Other services Inpatient curative services Outpatient treatment Medicines and medical consumables 10 Turcanu G, Domente S, Buga M, Richardson E. Republic of Moldova: Health System Review. Health Systems in Transition, 2012, 14(7):1–151. 11 PAS (2011). Monitorul Sanatatii: accesul si calitatea serviciilor medicale spitalicesti in perceptia populatiei din Republica Moldova [Health Monitor: access and quality of hospital services in the perception of the population from the Republic of Moldova, Re- sults of the national poll, 2011]. Chişinău, PAS. 12 People's access to health services: Household survey results. Summary report. 2013. National Bureau of Statistics of the Republic of Moldova. 13 Ibid. 8 Furthermore, all patients have to pay for medical goods and services not included in the benefit package covered by Mandatory Health Insurance (MHI), or for those subject to co-payment. By far the principal item among these goods and services not included in the benefit package are medicines purchased on an outpatient basis, most of them over the counter without doctors’ prescription. Based on the data reported by the National Bureau of Statistics14 about the composition of households’ payment for health, the largest share of OOPs is being spent on drugs – 49%, followed by medical consum- ables – 22% (altogether, 71% in 2013). Provider payments for outpatient care services (including dentistry services) account for 18% of OOPs, and the payments at hospital level account for approximately 10% of OOPs (Figure 7). The latter payments can be very significant or even catastrophic for those subject to them, by they account for a small share overall because of the lower utilization of inpatient services in compari- son to outpatient services. 3. Drivers of OOP in Moldova Recent studies implemented by the World Bank, World Health Organization (WHO)15, and by others help understand the nature as well as the drivers of the OOPs payment for health, and inform possible policy choices for the government to consider for increasing the level of financial protection. The following seem to be the principal immediate factors at play, which have so far impeded the achieve- ment of better levels of financial protection in Moldova: 1. Inadequate breadth of coverage with the Mandatory Health Insurance (MHI), especially for the poor, self-employed, and agricultural workers.16 While the government in 2010 introduced dis- count schemes to facilitate purchase of insurance coverage (50% off for self-employed and 75% off for agricultural workers), based on the household surveys12, these measures did not produce the expected results in the time frame expected. Coverage did increase over time, but there are still significant gaps for the following reasons: o Widespread informal OOP payments limit the attractiveness of the MHI (see below). o Demand for insurance by the poor is very price elastic, and it is often difficult to cover the poor through the Treasury contribution to MHI due to their employment status – i.e. they are self-employed, or employed in the shadow economy, and therefore cannot benefit from unemployment benefit or from the Ajutor Social program (the targeted scheme that protects the unemployed poor whose beneficiaries are also automatically given health in- surance for free), and are therefore not covered by MHI. 2. Inadequate depth of coverage, with high levels of formal (co)payments by patients . Specifically, the MHI benefits package does not cover sufficiently cost of pharmaceuticals as the “Reimbursed Drugs List” is very limited, only drugs prescribed in an inpatient setting are supposedly fully cov- ered (albeit patients are still requested to bring drugs to the hospitals; see above), while nearly all drugs prescribed in outpatient care are purchased directly by patients at full cost17. In addition, most of the compensated drugs, which are included in the Basic Benefit Package (BBP), are not generic and therefore even partial reimbursement is costly for the patients. Patients af- fected by chronic conditions are especially at disadvantage, and rural residents are burdened dis- 14 National Bureau of Statistics http://statbank.statistica.md/pxweb/Database/EN/databasetree.asp (Accessed on November 2nd, 2014) 15 Vian T, Feeley FG, Domente S. 2014. Framework for addressing out-of-pocket and informal payments for health services in the Republic of Moldova. Health Policy Paper Series No. 16. WHO Moldova. 16 Richardson E., Roberts B., Sava V., Menon R., McKee M. Health insurance coverage and health care access in Moldova. Health Policy and Planning 2011:1–9. 17 Kutzin J, Jakab M, Shishkin S. From scheme to system: social health insurance funds and the transformation of health financing in Kyrgyzstan and Moldova. Advances in Health Economics and Health Services Research 2009; 21:291-312. 9 proportionally as not all medicines are available, and pharmaceuticals in private pharmacies are more expensive in rural areas than in the cities18. 3. High drug prices. Prices in both public and private pharmacies are about five times compared to in- ternational reference prices19, due to weak regulations and/or weak enforcement of pharmaceuti- cal market regulations in the country, as well as due to small size of the market, which does not makes it attractive for big pharmacy players to enter, and reduces market competition. Secondly, doctors, “motivated” by pharmaceutical companies, have an incentive to prescribe expensive drugs, as well as potentially over-prescribe.20 Supplier-induced demand is a significant driver of drug costs. In addition, there are some regulatory gaps. For example, generic substitution for pre- scription medicines is not allowed either for reimbursed or non-reimbursed medicines. For pre- scription medicines, if the doctor has prescribed a branded medicine, the only way to obtain a ge- neric equivalent is to get another prescription from another doctor.21 4. Quite widespread phenomenon of informal payments (IPs) affecting those who are insured, and for services included in the BBP. These IPs are generated by a combination of factors, including: o Low pay of medical staff. While public financing of wages increased by 54% during 2008- 201222, starting levels were so low that this growth was not adequate to meet health care workers’ expectations, according to the providers’ surveys conducted by WHO. o Tokenistic sanction system. To reduce informal payments the government introduced fines for health facilities where informal payments are paid, but they only constitute 10% of the case-based payment if the facility is found charging informally. All the above immediate “drivers” of OOPs have to be looked in the context of broader organization of health system issues. We have termed these second set of issues “contextual factors”, and we believe that although they are not immediate causes of OOPs, they have played and will continue to play a critical role in maintaining the Moldova health system trapped in what we have called a “vicious cycle” characterized by high levels of OOPs, depicted in Figure 8. These factors include the following: o Excessive (in size and number) capacity of health care facilities relative to the available re- sources. Moldova has 45 rayon/municipal and 16 republican hospitals, complemented with 230 public outpatient facilities23. This creates a situation where the limited public resources are spread thinly across a large number of facilities, and used to simply maintain all these facilities at a mini- mal level of functionality. The National Health Insurance Company (Compania Nationala de Asig- urari in Medicina or CNAM) has never been empowered to purchase services more strategically and from a fewer number of health care facilities. Therefore, CNAM has been unable to increase reimbursement prices and/or total allocation for any of the selected facilities sufficiently, thus so- lidifying over time a situation where the available public budget is intended to pay for capital maintenance and few essential recurrent costs, such as energy, while labor and other recurrent costs are covered mainly through direct patients’ OOP contributions. The excess capacity is espe- cially notable in the capital city were 40.3% of the national hospital bed capacity is located in the 18 Medicine prices, availability, affordability and price components in the Republic of Moldova 2011. Health Policy Paper Series No. 11. WHO Regional Office for Europe 2013. 19 Sautenkova N., Ferrario A., Bolokhovets A., Kanavos P. Availability and affordability of medicines and assessment of quality sys- tems for prescription of medicines in the republic of Moldova. WHO Regional Office for Europe, 2012, Health Policy Paper Series No. 6. 20 MacLehose L, in: McKee M, ed. Health care systems in transition: Republic of Moldova. Copenhagen, WHO Regional Office for Europe, 2002, 4(5). 21 Sautenkova N., Ferrario A., Bolokhovets A., Kanavos P. Availability and affordability of medicines and assessment of quality sys- tems for prescription of medicines in the republic of Moldova. WHO Regional Office for Europe, 2012, Health Policy Paper Series No. 6 22 BOOST Analysis 2008-2012, presented to MoH during May 2014 workshop. 23 And approximately 530 private establishments, including 511 outpatient ones (Statistical Yearbook 2013. Resursele Ocrotirii Sănătăţii, Ministerul Sănătății al Republicii Moldova Centrul Național de Management în Sănătate). 10 republican institutions, though some rayons are affected as well24. In addition, in the capital repub- lican, municipal and some limited private providers converge and, due to their political influence, draw close to 20% of CNAM resources, therefore reducing funds available for rayon facilities. o Increased unfunded mandates for CNAM. Over the last few years, CNAM has been asked to ex- pand the set of benefits for insured and non-insured population, without being given correspond- ing additional funds. Unfunded mandates in the benefit package have grown, which have further limited the resources available to CNAM (in relative terms) to adequately finance the preexisting BBP. Also, senior doctors and administrators in highly specialized facilities are able to lobby for high-specialty services to be included in the BBP, even if they may not be a priority for the nation’s health, given the tight fiscal context. o Lack of transparency and accountability in management of health care establishments, regulato- ry practices, and of information about patients’ entitlements. Lack of transparency in public ten- ders, for example, contribute to high variability and, on several instances, high prices of medicines and medical equipment procured by government, informal charges, corrupt practices, etc.25 In ad- dition, the regulatory and pricing environment for medicines and medical equipment is not trans- parent, leading to huge variability in the prices charged for similar services and medicines across the country.26 Furthermore, the Mandatory Health Insurance (MHI) BBP is not well publicized and/or well-known and/or understood by the majority of the population. Patients are often una- ware about their entitlements, and therefore accept to pay directly for services that are covered by insurance according to the Law. Figure 8: Vicious Cycle in which the Moldova Health System is trapped 6. Catastrophic expenditures and lack of access to health services by the poor 5. Increased Context: 1. Excess supply financial access of providers Overcapacity, espe- barriers cially at the hospital level Gap between de- clared promises and Government’s ability to fully fund existing 4. Increased demand for 2. Downward BBP informal payments pressure on and payment for providers' public Unclear definition of drugs and reimbursement BBP and unfunded consumables mandates Weak accountability 3. Low official revenue of of providers to com- health providers munities 24 Statistical Yearbook 2013. Ministerul Sănătății al Republicii Moldova Centrul Național de Management în Sănătate. 25 rd In 2014 Moldova was on 103 place out of 175 countries on the Corruption Perception Index published by Transparency Interna- tional. 26 Medicine prices, availability, affordability and price components in the Republic of Moldova 2011. Health Policy Paper Series No. 11. WHO Regional Office for Europe 2013. 11 All the above immediate and contextual factors act simultaneously and drive the system deeper in the vi- cious cycle (see Figure 8 above), where pressure on the resources managed by CNAM relative to the com- mitments and expectations forces CNAM to reduce reimbursement rates below cost of service provision, which obviously reduces provider incomes and fuels demand for informal charges and/or requests for medical supplies and drugs to be purchased and brought in by patients. These requests impose additional costs onto the households and increase financial access barriers to care, particularly for the poor and popu- lation lacking insurance coverage. All of this reduces services’ utilization, again particularly by the poor, and maintains excess capacity relative to the demand. Excessive infrastructure thinly spreads the limited public resources and further increases downward pressure on the case-based reimbursement payments, especial- ly for the rayon-level facilities, where almost half of medical services are being delivered to the popula- tion.27, 28 It is essential to break the vicious cycle described above, using a multi-prong policy approach, briefly de- scribed in the following. If not addressed, the above structural issues will lead to progressively more severe negative consequences, because underlying dynamic tendencies are leading to an increasing gap be- tween available public resources and population’s health needs and expectations. These tendencies in- clude: (i) new expensive technology penetration in the health care market, which make health services in Moldova, as in the rest of the world, ever more costly; (ii) population aging and out-migration, which in Moldova are rapid phenomena and contribute to raising dependency ratios and burden of non- communicable diseases. 4. Policy options to address OOPs Based on the information presented, it would be important to achieve two critical objectives: a. Reduce burden for medicines and consumables’ expenditure, which would require numerous actions aimed at drug price reduction (for government purchased and market supplied pharmaceuticals), optimiza- tion of drug consumption, generic substitution and expansion of targeted drug benefits covered by CNAM. These policies will allow at least partial tackling 49% of OOPs currently spent on drugs by the population. b. Reduce and minimize OOPs paid for inpatient and outpatient services, arising from formal and informal payments for medical services and for drugs and consumables bought and brought in by the patients. These OOPs account for 28% of the total amount of OOPs. 4.1 Reduce burden imposed by drug expenditure Reduction of the burden imposed by drug expenditure, while may seem too ambitious in the current fiscal context, could in fact be at least partially achieved through incremental steps that can be taken by the gov- ernment immediately as well as in a medium to long term perspective. All the policy options could be grouped into three strategic directions: a) introduce transparency requirements to assure effective public procurement and monitoring; b) improve and enforce pharmaceutical market regulations; and c) improve content of the benefit package and assure better targeting of the poor. These policies need to be imple- mented in conjunction and not individually as they are mutually reinforcing, although implementation se- 27 People's access to health services: Household survey results. Summary report. 2013. National Bureau of Statistics of the Republic of Moldova. 28 PAS (2011). Monitorul Sanatatii: accesul si calitatea serviciilor medicale spitalicesti in perceptia populatiei din Republica Moldova [Health Monitor: access and quality of hospital services in the perception of the population from the Republic of Moldova, Re- sults of the national poll, 2011]. Chişinău, PAS. 12 quencing and close monitoring will be necessary. The details of the policy recommendations include the following29: o Introduce pharmaceutical audits by CNAM: if CNAM expands the drug benefit package, this should go hand in hand with good implementation of pharmaceutical audits to review prescribing pat- terns and to reduce corrupt practices such as side-payments from pharmacies to providers which encourage over-prescribing. When the drug benefit package is minimal, as it has been until recent- ly, the lack of good pharmaceutical audits by CNAM is more or less inconsequential, but once the package is expanded significantly, CNAM needs to take on greater pharmaceutical audits responsi- bilities. o Strengthen regulation of pharmacies: The regulation of private sector pharmacies should be strengthened, using certification and licensing mechanisms to promote quality control and mini- mum standards; and when they are found to provide side-payments to doctors to over-prescribe or to prescribe specific costly drugs their licenses should be revoked. o Introduce International Nonproprietary Names (INN) prescribing mandate for outpatient drugs. Use of International Nonproprietary Names is expected to limit providers’ ability to promote cer- tain drugs, in which they may have vested financial interests and may help reduce cost to the pa- tients. o Promote the use of low-priced quality-assured generics that could be achieved through appropri- ate state regulations and in-conjunction with INN prescription mandates. Promoting non-branded generics may help CNAM generate savings when a less expensive non-branded generic is dis- pensed instead of a more expensive branded ones. Considering that the market share of branded generics is 30–35% versus 60–65% for non-branded generics, this is likely to have some impact in terms of potential savings. o Institutionalize e-prescription to better monitor prescription practices, control prices, eliminate low-standard drugs in the market, as well as to reinforce promotion of generic drugs on the mar- ket. o Introduce transparency requirements and public monitoring of the government tenders for drugs. This could be achieved through public disclosure of the government’s purchasing decisions with the help of the Internet. o Expand outpatient drug benefit under the national health insurance, especially for the poor and for the prevalent health conditions. Additional budget revenues and/or realized savings by CNAM, if any, should be reinvested in drug package expansion for prevalent health problems, instead of funding high tech costly interventions benefiting a limited number of patients. However, unless the pharma regulation and drug pricing issues are addressed first, expanding the medicines benefit would create very little incentive for (and even risk disincentivizing) regulatory reform and getting greater pricing efficiency. 4.2 Minimize informal OOPs paid to health services providers Reduction of the OOPs at provider level could be achieved through supply and demand side interventions, and again through increased transparency and accountability in the governance of health establishment and the health system. The following concrete interventions are proposed: o Rationalize the health care delivery system (as a precondition to increase reimbursements rates), which could be done through more strategic purchasing of services, and from eventual reduction of excessive provider capacity in the capital and rayons, where necessary. Without getting into the 29 See also: Sautenkova N., Ferrario A., Bolokhovets A., Kanavos P. Availability and affordability of medicines and assessment of quality systems for prescription of medicines in the republic of Moldova. WHO Regional Office for Europe, 2012, Health Policy Paper Series No. 6. 13 details of hospital reforms here, the possible measures have been identified long ago: they include, among others, the need for putting Chisinau hospitals under common management, networking of hospital, reducing hospital beds, consolidate the hospital network in the 35 rayons, and consoli- date the laboratory system. o Use additional financial resources and adjust reimbursement levels for essential treatments for priority/prevalent health conditions. Adequate reimbursement of providers could provide neces- sary funds for essential inputs (drugs, consumables and salaries), which has the potential to reduce demands for informal payments and/or demand on patients to purchase and bring necessary in- puts to the facility they are utilizing. However, such increase public financing to selected providers and services should be closely linked to their performance. For example, pay-for-performance in hospital and primary care should be used as an instrument to influence the quality of care and as a counteraction to informal OOPs paid to doctors in the form of bribe. o Improve MHI targeting for the poor by trying to progressively limit MHI free coverage, using as a basis the means testing approach utilized for the “Ajutor Social” program (where the poor are identified using a proxy means testing approach). The strict income thresholds used for Ajutor So- cial could be increased and specific incomes disregard introduced, as it was done when the pro- gram was used to determine eligibility for the gas tariff subsidies in 2012. Reaching the poor and affording higher benefits to them (drugs and, where applicable, medical services) will help improve financial protection, as well as should facilitate increased utilization and improved health out- comes for disadvantaged groups, at the same time raising revenue for CNAM from health insur- ance premium payments. o Improve BBP understanding by the covered population with the aim of empowering consumers and increasing demand at provider level for free or subsidized services. This could be achieved through better description of the covered benefits in a layperson’s language, through branding MHI, and distributing insurance policies/cards clearly listing the benefits afforded to covered indi- viduals. o Empower patients through simplified complaint filing procedures via hot line and/or web-based and through effective use of this information to drive public accountability of providers: e.g. using patients’ feedback to rate medical establishments and placing facilities’ ranking on measures of performance and transparency in public domain. These interventions have the potential to facili- tate increased provider accountability towards the communities they serve. Transparency proved to be powerful tool in many health systems, including the National Health Service in the UK, USA, etc.30 Publicly reported data: (i) could lead to increased involvement of each medical facility’s lead- ership in performance improvement; (ii) could create a sense of accountability to both internal and external customers; (iii) could contribute to a heightened awareness of performance measures throughout the staff of the same medical facilities; (iv) could influence or re-focus organizational priorities31; finally (v) patient complaints could also be used as one of a measure/indicator for pay for performance arrangements. For example, “informal payment free” providers according to the patients’ feedback could be given a bonus, or a recognition. o Continue regularly collecting, analyzing, and publishing information on patients’ out of pocket expenditure, and use the available tools (ADEPT, BOOST, NHA and others) to analyze private and public health expenditure from an equity/distributional and efficiency point of view. o Introduce and enforce stricter fines on the providers that charge informally. This would require making sanctions financially burdensome enough to bring about a real discouraging effect. 30 Henke N., Kelsey T., Whately H. Transparency – the most powerful driver of health care improvement? Health International. McKinsey 2011. 31 Hafner JM, Williams SC, Koss RG, Tschurtz BA, Schmaltz SP, Loeb JM. The Perceived Impact of Public Reporting Hospital Perfor- mance Data. Int J Qual Health Care. 2011; 23(6):697-704. 14 5. Conclusions In synthesis, the high levels of OOP in Moldova signal a “value for money” challenge in the health system which needs to be addressed. By looking at the evidence across countries, one can broadly see two sce- narios where OOPs is high: a. Government simply does not spend enough on health; b. Government does spend a significant amount on health, but the money is not well-spent. As a result, public expenditure does not translate into good financial protection outcomes for the people. Moldova seems to fall mainly under the second scenario, since public expenditure on health exceeds 5 percent of GDP and is significantly higher than the regional average for CIS countries – 3.5% and compares well with richer countries like Croatia, Slovenia, and Ireland32. Therefore, increasing allocative and technical efficiency of existing public expenditure on health and creating fiscal space for better use of the existing public resources for health have to be key ingredients of the overall strategy to reduce OOPs. The other two key ingredients are politicians’ leadership and fiscal responsibility (having the courage to tell the popu- lation honestly what the government can afford to pay and what it cannot pay), and transparency, for ex- ample about the services which are actually covered by the mandatory insurance package and should be provided for free, in a context of a new, resolute fight against corruption. Finally, the lack of progress on financial protection over the past seven years proves the need for better and more frequent monitoring of the impact of government-enacted policies on the population’s accessi- bility to quality services, particularly poor and vulnerable groups. Such monitoring could be assured through several means. An important one, among others, could be the household budget surveys with ex- panded health module, implemented by the National Bureau of Statistics every two years under the Health Services and Social Assistance World Bank project (2008-2014). In addition, the BOOST and ADEPT tools could be used to rapidly show the evolution of government as well as OOP expenditures and their distribu- tion across different socio-economic groups in the sector over time. The World Bank has supported the introduction of these tools and provided training on their use on several occasions. Policy makers need to continue supporting these surveys and to pay much more attention to the evidence, which emerges from these surveys. Such evidence clearly shows that Moldova should do more and better to improve accessibil- ity to quality services, particularly for the poor, and increase financial protection of its population. The promises and commitments of 2004, when the country started social insurance, have yet to be fully real- ized. 32 WHO EURO Health For All database 2011 data http://data.euro.who.int/hfadb 15 Annex 1: Note on Estimating Household Expenditure on Health The main source of information for this note was the Household Budget Survey, conducted every year by the National Bureau of Statistics (NBS) and published on the following website. http://statbank.statistica.md/pxweb/database/EN/databasetree.asp . The total annual household expendi- ture on health was calculated using this data. Our estimates used the data from the following sections of the NBS databank: a) “Household expenditures” and b) “National accounts”. Based on this data, the table below was produced, using two different approaches: i) using annual final household consumption (row A) and reported share spent on health (row E), which produced estimates for 2008-2013 ranging between 2.32 – 5.28 billion Lei and ii) using per capita monthly spending on health (row C) multiplied by 12 month and by population (row F), which produced estimates between 2.60-4.31 billion respectively for the period 2008-2013. Data Element 2007 2008 2009 2010 2011 2012 2013 42,832 52,053 53,117 58,242 78,104 83,664 92,910 A. Final annual household consumption Lei 1'000 ,602 ,324 ,367 ,820 ,000 ,000 ,404 B. Annual growth index for final HH Con- 122 102 110 134 107 111 sumption (previous year =100) Lei, average month- C. Expenditure on medical care and health 60.7 68.5 75.8 87.3 82.5 89.9 100.9 ly per capita D. Household consumption expenditures - Lei, average month- 1119.1 1227.5 1217.4 1371.7 1534.1 1598.6 1775.8 total ly per capita E. Health Expenditure as a % for household Percent of HH 5.42% 5.58% 6.23% 6.36% 5.38% 5.62% 5.68% monthly consumption Monthly spending F. Population Individuals 3,581, 3,572, 3,567, 3,563, 3,560, 3,559, 3,559, 110 703 512 695 430 541 497 Total HH Health Expenditure (calculated as % of 2,323, 2,904, 3,307, 3,706, 4,200, 4,704, 5,279, Lei 1'000 Final HH Consumption) 241 809 291 786 235 988 119 Total HH Health Expenditure (calculated based on 2,608, 2,936, 3,245, 3,733, 3,524, 3,840, 4,309, Lei 1'000 Per Capita Monthly Consumption) 481 762 009 327 826 033 839 There seems to be several limitations with these estimates that have to be noted: 1. These two methods for the period 2008 – 2010 produced comparable estimates for household ex- penditure on health and differences for this period ranged between -1.1% to 1.9% at most. How- ever, since 2011 significantly bigger differences, in the magnitude of 19% - 23%, emerge, which may be an indicator of lack of consistency in the macro data reported by NBS (see below). 2. Another limitation is related to estimates for household contribution to CNAM. The household budget survey tool, used by NBS, does not include questions about household’s mandatory contri- butions to CNAM, and all questions ask information only about out-of-pocket payments for health services and products consumed. Therefore, estimated volumes presented in the table above should be treated as pure out-of-pocket expenditure on health, unless NBS adjusts final household and health expenditure data for these contributions, using different data sources. 3. Finally, when row – D is multiplied by 12 months and by population from row F, the amounts of household annual consumption calculated through these means is comparable with those reported 16 by NBS for 2008-2010, but significant differences emerge since 2011 in magnitude of 16-18% de- pending on the year. 17 Annex 2: Why excessive OOP in the health sector concern society? There are three main reasons excessive OOPs are a concern: o OOPs impose significant risk on households, as OOP can result in “catastrophic” or “impover- ishing” episodes for households; o They cause inequality in access to care between rich and poor; o They may reflect “rent-seeking” by providers at the expense of the population. These three reasons are explained in the following, drawing upon existing economic theory and empirical evidence. We know from economics that for any rival, excludible33 product or service (including all individual medical products and services) the optimal level of consumption is to be found where marginal cost (which, at each level of consumption, is a measure of the value of the additional resources society has to commit to pro- vide an extra-unit of that product or service), and aggregate demand (the sum of individual demands, which is a measure of social marginal willingness to pay for that product or service) are equal. In general, such optimal level of consumption is achieved in correspondence to a price above zero (unless the margin- al cost of production is zero, as it happens for non-rival public goods). The Figure below illustrates the optimal level of consumption for “pharmaceutical product” Q, and corre- sponding “optimal price” at respectively, points Q* and P*. Pareto Optimal Equilibrium, quantity consumed at zero price, and quantity consumed with co- Price of Demand drug Q * Supply = Marginal Cost P M Q* Qm Q0 Quantity of drug Q payment Suppose Q represented the level of consumption of the specific drug/pharmaceutical product in the graph above, and that marginal cost (cost of providing each additional unit) were constant, so that supply would be infinitely elastic. In such situation, efficiency requires that equilibrium consumption be at the level Q*; however, if the drug is available free at the point of service, its consumption is increased up to quantity Q0. 33 A good or service is rival when the consumption by one individual excludes the possibility of consumption of the same good by another individual, and it is excludible when those who do not contribute to its financing can be excluded from its consumption. Products and services (“goods”) which are non-rival and non-excludible are called public goods. Examples of public goods in the health sector are sanitation and disease surveillance. 18 In other words, people consume “too much” of the fully subsidized drug.34 In the above situation, a co- payment (in this case, any type of co-payment, including formal/informal payment) equal to M is able to reduce the level of “excess” consumption to a level closer to the optimal level. A second key finding of economic analysis applied to the health sector is that outside of routine care, med- ical expenses are relatively large and occur randomly, precisely the sort of problems for which there is a large demand for insurance (typically across all countries around 10 percent of the population accounts for 60-65 percent of the total health expenditure in any given year). Therefore, medical products and services (in short, health services) are a set of services for which redistributing the financial risk away from risk- averse and liquidity-constrained patients towards risk neutral agents typically enhances efficiency. That is the reason why insurance (which does not necessarily imply establishing social or private insurance, since also general taxation financing or direct public provision can provide such “insurance”), plays such an essential role in the health sector in most countries. That is the theory. And reality? First, quite extensive empirical evidence (following the seminal results of the Rand Experiment Group study; see: Manning et al. 198735) indicates that, while people may “abuse” consumption of certain health services such as pharmaceutical products, outpatient specialist visits, or hospital elective care, when they are offered totally for free at the point of service, for most other types of health services demand is not so much influenced by price (on average, the price elasticity of demand for health services is low –roughly equal to - 0.2). When elasticity is low, the demand containment effect of co- payments is negligible, as the figure below shows. Pareto Optimal Equilibrium, quantity consumed at zero price, and quantity consumed with co- Price of Demand emergency services Marginal Cost P* M E* E0 Emergency services payment for an inelastic service More importantly, the key empirical question is whether co-payments are discouraging unnecessary re- course to medical treatment, or are they excluding those who are unable to pay? The extent of demand for health services and its elasticity are influenced by information, wealth, and cultural attitudes, so that it var- ies significantly across different socio-economic groups. Thus, individuals who share the same health con- 34 In fact some people may still have to sustain transportation costs and other costs not considered here. 35 W. Manning et al. (1987), Health insurance and the Demand for Medical Care: Evidence from a Randomized Experiment, Ameri- can Economic Review 77: 251-277. 19 dition (have the same health needs) and face the same prices, in reality use health services very differently. For instance, poor people or people in rural areas may not use health services even if a small co-payment is imposed, because they are not aware of the benefits they could receive, or simply because being in “poor health” is accepted as a normal condition, rather than an accident to be cured. Several empirical studies show that higher co-payments for health services “discouraged” the poor much more than they did the rich. Therefore, the “demand discouragement” effect associated with co-payments, far from being a posi- tive, moral-hazard reducing effect, was in fact signaling a perverse, negative equity effect, mainly affecting the poor. In synthesis, while in general economic theory concludes that having positive co-payments for rival goods, including all individual health services, may be economically efficient, such conclusion may not hold for medical services, particularly when such co-payments are “excessive”. For two reasons: - The first is precisely an efficiency reason. Health systems predominantly financed on an OOP basis are inefficient, because there are potential welfare-enhancing redistributions of health-related fi- nancial risks through insurance, which are not exploited. - The second and more important one is an equity reason. The majority of individuals in our socie- ties perceive that health services should be consumed mainly according to need and not ability to pay. Therefore, they are concerned that high levels of OOPs may constitute a barrier to access health services, particularly for the poor, and are concerned about the potential impoverishing ef- fects of high levels of out of pocket expenditure. We could summarize these equity arguments as follows: “no one should forgo necessary health care because s/he cannot afford it”, and: “no one should become poor because of illness”. As to the question of which level of OOP health-related payments could be considered acceptable (and for some services even perform a positive “demand screening” function), and which is to be considered “e x- cessive”, there is no rigorous scientific answer. A threshold generally used in the literature for individual households is 25 percent of their total household expenditure, or 40 percent of their total non-food ex- penditure. According to this threshold, it is claimed that any given health system provides an optimal level of financial protection if no households incur a level of OOP exceeding 25 percent of their total expendi- ture, or 40 percent of their total non-food expenditure. In addition, as the above analysis indicates, it is essential to study the distribution of so-called catastrophic health expenditures disaggregated by socio- economic groups, because it is mainly the poorer segments of society that are of concern. The idea is that no household should be brought below the poverty line by OOPs for health services. 20