Published on Let's Talk Development

Health and the SDGs: Out of the doldrums, heading for the rapids

This page in:
Until quite recently, things were looking good for health in the SDG process. It wasn’t always so. Two and a half years ago, at the time of the high-level panel report on the SDGs, the health SDG discussion was actually stuck in the doldrums. Health was the only area to get less column inches than in the MDGs. The proposed goals and targets were pretty much business as usual. The only real hint of any new thinking was the addition of a target to reduce non-communicable diseases, but it was subsumed within an old target and looked very much like an afterthought.

Then much to everyone’s surprise, the health SDGs got an infusion of new energy, and in crept the issue that everyone in the health field had been talking about for some time: universal health coverage (UHC).

UHC is, in fact, the obvious umbrella for all the health SDGs, combining a concern that everyone—poor and rich alike—gets the health services they need, with a concern that getting needed health services doesn’t cause families undue financial hardship.

Jumping on the tailwinds of WHO/World Bank Ministerial-level Meeting on Universal Health Coverage in February 2013 in Geneva, the World Health Organization and the World Bank got busy consulting with civil society and technical experts, and ended up proposing two indicators, one getting at the first side of the UHC coin (service coverage), the other at the second (financial protection). The latter was to be captured by looking at household health expenditures, i.e. out-of-pocket spending on health care. The Bank and WHO proposed counting the fraction of households whose annual out-of-pocket spending might be said to be ‘catastrophic’ (exceeding, say, 10% or 20% of household income) and the fraction whose spending was ‘impoverishing’ (sufficiently big to push the household below the poverty line). These measures were introduced some time ago, and had already been widely used in studies of specific countries and regions, as well as in global studies.

UHC and the SDGs are natural bedfellows. The SDGs have as their objectives “to end extreme poverty” and “to ensure that every person achieves a basic standard of wellbeing”. The financial protection angle to UHC speaks directly to this theme. As Kaushik Basu, chief economist at the World Bank, put it: “…high levels of impoverishment, which happen when poor people have to pay out of pocket for their own emergency health care, pose a major threat to the goal of eliminating extreme poverty,”

Out of the doldrums. Sailing nicely?

Well no, actually. The SDG winds recently changed, this time seemingly pushing the health agenda not into the doldrums but headlong toward the rapids.

The process by which the SDG indicators are finalized is, like everything to do with the SDG process, planned as a broad-based bottom-up process. Technical experts in the relevant field and civil society offer advice, but the ultimate decisions are taken by representatives of the national statistical offices of 28 countries, who pore over and take decisions on literally hundreds of indicators from 12 broad areas as diverse as poverty and “stability and peace”. A daunting task, especially given the tendency for sectoral experts to use terminology that even clever people like statisticians may find quite impenetrable. Not surprising perhaps, then, that an odd decision might emerge.

And emerge it did. It seems the World Bank-WHO proposed indicator for financial protection didn’t resonate with the statisticians. The proposal is to replace it by “the number of people covered by health insurance or a public health system per 1,000 population.”

On the face of it, the replacement seems to make sense. Covering people by health insurance or by a public health system means ensuring they can get health services without financial hardship, surely? And it’s easy to count who’s covered and who isn’t covered, surely?

4 reasons why assessing financial protection isn’t that simple

First, everyone everywhere has recourse to their health ministry’s network of health facilities, i.e. their country’s “public health system”, where prices are zero or at least heavily subsidized. So every country in the world would score 100% on the proposed indicator. Makes monitoring easy, but hardly very informative!

Second, there’s a difference—often a big one—between theory and practice. According to their national policies, both Azerbaijan and the UK operate a national health service (NHS) health system, financed largely through general revenues, with patients paying nothing out-of-pocket at the point of service. That’s the theory. In practice, the share of total health spending in these two countries financed by out-of-pocket payments differs dramatically: in Azerbaijan it’s more than 70%; in the UK, it’s under 10%. We could go badly wrong, then, if we tried to assess financial protection by looking at what’s promised rather than at what people pay in practice.

Third, in many countries, there are formal health insurance schemes that sit on top of the “public health system”. In countries like Brazil and S Africa, they’re voluntary private health insurance schemes. In many countries, they’re social health insurance schemes where formal-sector workers (and their employers) make mandatory contributions, and other groups often have the chance to enroll voluntarily. In some cases, other groups, such as the poor and near-poor, are enrolled at the taxpayer’s expense. This additional layer of coverage creates a gap between those without the additional layer and those with it, with people in the latter group, who are often the better off, getting access to different facilities and/or paying different amounts out-of-pocket. A tiered segmented system would do well in terms of the proposed new indicator—a lot of people would have coverage by both the public health system and a health insurance scheme. But it’s likely to perform badly when assessed in terms of UHC, i.e. everyone getting the services they need, and people not being exposed to financial hardship when they seek care.

Fourth, affiliation to an insurance scheme is not necessarily a good indicator of financial protection.  Korea reached 100% population coverage in 1989, but in that year about 65% of total health spending came from out-of-pocket payments.  By 2007, this share fell to 38%, suggesting an improvement had occurred with no change in population insurance coverage (still 100%). And as China saw in the early days of its rural health insurance scheme, expanding formal insurance coverage can actually increase out-of-pocket payments. So greater scheme affiliation doesn’t necessarily mean better financial protection; and protection can get better (or worse) without any change in insurance coverage.

Time to change tack

A good monitoring indicator would measure the objective of interest and avoid risking encouraging countries to implement misguided reforms because it would make them look good on SDG performance. The proposed indicator does neither.

One option would be to stick with what was originally proposed: count the fraction of the households whose annual out-of-pocket spending might be said to be ‘catastrophic’ and/or ‘impoverishing’. On the plus side: the indicators measure what we want to measure; and they’re easy enough to compute with a household expenditure survey that contains information on out-of-pocket spending and total household income (or expenditure or consumption), and a poverty line. On the minus side, it’s probably true they’re a little harder to get your head round than, say, “the fraction of the population with access to safe water”, and the Bank and WHO for sure could have done a better job at explaining them.

Another option that has been suggested as a compromise is to go for something that’s very accessible but still gets at the idea we need to know how much people spend in practice: “household health expenditure as a share of total household income (or total expenditure or consumption)”. This can also be computed easily from a typical household expenditure survey. It’s simple—simpler than catastrophic and impoverishing spending—and it correlates quite highly with the incidence of catastrophic spending.

It’s true this alternative indicator correlates less strongly with impoverishing spending, because it doesn’t tell us whether it’s the near poor or the well-off who spend out-of-pocket, and therefore whether it’s a large or small fraction of the population that’s plunged into poverty because of the money they spend getting health care. But that’s not an insurmountable problem. Just as in the MDGs, there will inevitably be supplemental indicators to the main indicators in the SDGs. So we could make ‘impoverishing expenditures’ a supplemental monitoring indicator for the financial protection angle to UHC.

The important thing is we get away from the rapids—from thinking we can capture UHC by counting “the number of people covered by health insurance or a public health system” to focusing on households’ spending on health services.


Adam Wagstaff

Research Manager, Development Research Group, World Bank

Joseph Kutzin

Coordinator, Health Financing Policy, World Health Organization

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000