Seeking consensus on universal health coverage indicators in the sustainable development goals

Creating indicators for SDG Target 3.8

The Inter-Agency Expert Group on SDG Indicators (IAEG SDG) was established by the United Nations Statistical Commission in 2015 to create and review the global indicator framework to be adopted by the Economic and Social Council (ECOSOC) and the General Assembly. ³ The IAEG SDG includes considerably more interlocutors than the previous Inter-Agency and Expert Group on the Millennium Development Goals (IAEG MDG) which had similar responsibilities for selecting the MDG indicators. The MDGs were criticized for being formulated top-down led by international agencies and a small number of influential countries. ⁷

The IAEG SDG, therefore, includes not only more but also a wider cross-section of participants, thereby increasing the need for compromises to reach agreement on indicators. In 2015, an open consultation which allowed input from countries and organizations that were not members of the IAEG SDG gathered contributions from more than 600 contributors. ⁸

Chronology of changes to the indicators

The indicators for universal health coverage were not easy to finalize because of conceptual and measurement-related issues. The two indicators were initially among the ‘grey’ ones which needed considerable discussion before submission for ECOSOC approval, ⁹ and were included in the category of indictors for which data are not easily available, even if there was an established method for obtaining the data. ¹⁰

Universal health coverage is fundamentally concerned with ensuring that all ‘people should have access to the health services they need without risk of financial ruin or impoverishment’. ¹¹ Universal health coverage ‘is defined as all people and all communities receiving the services they need, while ensuring that they are protected from financial hardship’. ¹² Therefore, it was originally proposed to the IAEG SDG that the universal health coverage indicators should cover two main areas – access to high quality health services and financial risk protection − which would then be reflected in the two indicators initially advanced for discussion. ¹³

When submitted to ECOSOC in June 2016, the first indicator (3.8.1) was proposed as ‘coverage of essential health services (defined as the average coverage of essential services based on tracer interventions that include reproductive, maternal, newborn and child health, infectious diseases, non-communicable diseases and service capacity and access, among the general and the most disadvantaged population)’ (Table 1). This first indicator had undergone a change from the previous iteration to include a definition and list of essential health services. Tracer indicators were pre-selected based on the criteria of relevance to the epidemiological situation, the presence of cost-effective interventions, measurability, and the presence of comparable data across countries. ¹⁴

OPEN IN VIEWER

Table 1.

Changes to Indicators for Target 3.8 before submission to the United Nations Statistical Commission, March 2017.

Target 3.8 Achieve universal health coverage, including financial risk protection, access to quality essential health-care services and access to safe, effective, quality and affordable essential medicines and vaccines for all.
	January 2016	March 2016	November 2016
Indicator 3.8.1	Coverage of tracer interventions (e.g. child full immunization, ARV therapy, TB treatment, hypertension treatment, skilled attendant at birth, etc.)	Coverage of essential health services (defined as the average coverage of essential services based on tracer interventions that include reproductive, maternal, newborn and child health, infectious diseases, non-communicable diseases and service capacity and access, among the general and the most disadvantaged population)	(No further change proposed)
Indicator 3.8.2	Fraction of the population protected against catastrophic/impoverishing out-of-pocket health expenditure	Number of people covered by health insurance or a public health system per 1,000 population	Proportion of population with large household expenditures on health as a share of total household expenditure or income

The second indicator (3.8.2.) generated considerably more debate. Civil society, in particular, expressed concerns that the proposed indicator did not adequately measure financial protection.^15,16 By March 2016, the indicator had been changed from the ‘fraction of the population protected against catastrophic/impoverishing out-of-pocket health expenditure’ ¹⁷ to the ‘number of people covered by health insurance or a public health system per 1000 population’ (Table 1). ¹⁸ Predictably, this indicator was further revised as both the World Bank and World Health Organization expressed reservations. ¹⁹ The subsequent proposed indicator in November 2016 for the meeting of the IAEG SDG was the ‘proportion of population with large household expenditures on health as a share of total household expenditure or income’. ²⁰ Citing the need for robust measurement, researchers supported this change. ²¹

The two universal health coverage indicators were categorized as Tier III, i.e. those for which no established method exists or for which standards and methods are being developed or tested. ²² One of the factors that influenced the change from the original indicator on impoverishing and catastrophic expenditure for health care was data availability. According to the World Bank, data on who personally pays for health insurance is available for 142 countries, while only 89 countries have data on protection against catastrophic or impoverishing out-of-pocket health expenditure through household surveys. ¹⁰ Limitations in data availability for global monitoring have been well documented^23,24 and are likely to be an issue for other indicators in the global indicator framework. Therefore, there has been commitment to strengthening statistical capacity for the 2030 Agenda for Sustainable Development, particularly in low income countries. ²⁰

Conceptual challenges and changes to indicators for universal health coverage

The lead agencies for indicators for Target 3.8, the World Bank and World Health Organization, have repeatedly supported two indicators − one measuring coverage and the other measuring financial protection (with disaggregation when possible).^13,19 The indicators combined should be able to measure the three dimensions of the so called universal coverage ‘cube’ – population coverage, service coverage for those who need it and financial protection. ²⁵ Deciding on a coverage indicator (that includes tracer interventions) which is satisfactory to all countries and their people is notably difficult to produce. In order for the indicator to be meaningful, it would have to address what matters most to each country in terms of epidemiological and demographic realties, and how the need for health care is perceived by patients and providers.^26,27

Although the specific aspect of ‘effective coverage’ (defined as whether health services are obtained in a timely manner and at the quality necessary to obtain the desired effect and potential health gains) ²⁸ is a policy ideal, ²⁹ it is difficult to measure.^14,29 The coverage indicator will not include effectiveness, and according to metadata for 3.8.1 updated in May 2016, coverage will be calculated as the number of people receiving the intervention as a percentage of people who need the intervention. This is limited to the number of people who come into contact with the health system, rather than the whole population. This shortcoming could be addressed by more descriptive country-level evaluations.

There were also some difficulties with the first iteration of 3.8.2. Health expenditure is considered to be impoverishing when it pushes households below the poverty line, or further into poverty if they are already there, and catastrophic if they spend more than a specific portion of their income on health care. ³⁰ Metadata for the revised 3.8.2 indicator will not be released before its approval by the Statistical Commission in 2017 and until the definition of large health care expenditure is resolved. Although the compromise on this indicator leans towards incorporating catastrophic expenditure, the revised indicator omits explicit mention of impoverishing or catastrophic expenditure. In the original metadata proposal, ¹⁷ there was some arbitrariness in measurements of catastrophic health expenditure which was proposed as 25% of total household expenditure or, using an international poverty line of $1.25 or $2.00 per day that arguably had no relevance for middle and higher income countries. ³¹ In the revised indicator, the figure of 25% of household expenditure was again floated but without unanimity, as some considered 10% to be an alternate benchmark of health care expenditure. ³²

The previous version had seemed to raise anxieties about lop-sided monitoring of universal health coverage that emphasized reducing out-of-pocket payments for health care, and the over-simplified assumption that health financing systems guarantee access to health services. ³³ In response to this critique, there have been efforts at dispelling myths that suggest that universal health coverage is only about health financing. ³⁴

While it might appear reasonable that insurance be used to measure the financial risk protection component of universal health coverage, the way insurance interacts with economic and policy realities makes it an unreliable metric for universal health coverage. Firstly, there is no uniformity in what is covered by health insurance across countries or even between insurance carriers within countries.^35,36Secondly, health insurance is also known to exacerbate inequalities − since the groups that typically benefit from targeted insurance programs especially voluntary, private health insurance − tend to be either employed or economically better-off.^37,38 Social health insurance programs could similarly fail to meet population-level universal health coverage goals, particularly when mandatory enrolment places greater emphasis on the participation of employed persons because of their ability to contribute payroll deductions. ²⁹ Thirdly, health insurance itself is influenced by local market forces, is less likely to be extended to more risky clients (who most need insurance coverage), and is susceptible to differences in services covered, out-of-pocket deductibles, benefits and compensation. ³⁶ Public health system coverage (which was previously proposed as a metric for 3.8.2) presents similar concerns because of unpredictable differences in the range and quality of health services offered in different countries. ³⁹ In some cases, administrative data showing high utilization of public health glosses over the fact that there could be a parallel requirement for mandatory out-of-pocket spending which runs contrary to the aims of universal health coverage. ⁴⁰

Indicators that do not reflect the concept of access to health care will miss the mark since universal health coverage is premised on the reality of access for all. ⁴¹ Health care access relies on important utilization, satisfaction and consumer awareness concepts that hinge on services being available in adequate supply, personally acceptable and affordable. ⁴²

Universal coverage indicators in the future

Negotiation and consensus-building in selecting indicators are routine requirements in formulating global health policy. As indicators are developed, political interests, development aspirations and scientific method have to be effectively balanced. The number of interested constituencies (such as human rights proponents, economists and patient advocacy organizations) in the discussion on indicators of universal health coverage is evidence of the importance of health and sustainable development across a cross-section of interests. While the diversity of perspectives is a strength, this complexity makes consensus difficult to achieve. In contrast, national evaluation frameworks are formulated mindful of national priorities and therefore tend to be more cohesive and quicker to establish.

There is a need for more discussion and research on international measures of poverty and their applicability to health care which would be useful not only for Target 3.8, but also the other SDG targets that measure poverty and social risk protection. Among the SDG indicators, data disaggregation ⁴³ is particularly important for Goal 3, where coverage and health expenditure indicators might apply to different groups in the population who might have significantly different realities.

SDG indicators have to be chosen carefully, not only because they provide benchmarks for development and progress in achieving access to health care but also because they, like the Millennium Development Goals, have priority-setting influence. ⁴⁴ Amidst the various voices and iterations of the indicators, the pre-eminent goal is to ensure that universal health coverage is attained for all people. ²⁵