What’s the “universal health coverage” push really about?


This page in:

L. Garrett, A. Chowdhury, A. Pablos-Méndez. “All for universal health coverage”, The Lancet, 2009, 374(9697), pp. 1294-1299

Maps, like pictures, are often worth a thousand words. The map above comes from an article in the medical journal The Lancet. The article is part of a campaign to make 2010 the year of a big push toward universal health coverage. Over the course of the next three days, over a thousand health systems researchers are gathering in Switzerland for the First Global Symposium on Health Systems Research; the theme of the symposium is “science to accelerate universal health coverage”. Next week, health ministers from around the world will gather at an international ministerial conference on “Health Systems Financing – Key to Universal Coverage” hosted by the German government. At the conference, the World Health Organization will launch its 2010 World Health Report entitled “Health Systems Financing: The Path to Universal Coverage”.

The Lancet’s map—ostensibly showing the fraction of the population currently with health insurance coverage—is a neat tool for the participants in these exercises to see where they’re starting from. Or it would be if it were accurate. Unfortunately, it’s not. According to the map, the US already has universal health coverage! In which case, what was President Obama doing risking his political capital by trying to expand insurance coverage? And according to the map, Brazil has incomplete coverage despite the fact that in 1988 the government amended the constitution to guarantee all citizens access to health care and introduced a tax-financed universal health system. In fact, all Latin Americans would probably be rather confused by the map, since they all live in a country whose ministry of health (MOH) operates facilities that are open to everyone. The same is true of African countries too. And India. And Indonesia. And the Philippines. In fact, come to think of it, it’s true of most countries.

Unfortunately, then, the map is completely misleading. The problem with the map—and indeed with much if not most of the debate on universal coverage—is that it portrays the universal health coverage challenge as an either-or problem. People either have coverage or they don’t. In actual fact, everyone everywhere has some coverage. The stark reality, though, is that in many—if not most—countries there are large inequalities in coverage, typically mirroring pretty closely the income distribution.

The challenge, it seems to me, is not to cover everyone (already achieved). Or even to give everyone the same cover (desirable but equality of effective coverage is best seen as a long-term goal). Rather, the coverage challenge to my mind is really about narrowing inequalities in coverage.

Many countries have segmented health systems. Government facilities often charge for services, so while people have access to them, they have to pay for them. And in some countries, when people arrive at government facilities for treatment, they find no staff, no drugs, and no equipment. So, people have access to services that don’t actually exist!

By contrast, more privileged sections of the population—civil servants and formal-sector employees—are often covered by one or more social health insurance (SHI) schemes. These either reduce out-of-pocket payments in government facilities or give enrollees access to an altogether separate network of (public or private) providers. These providers may or may not charge a lower price, but at least when patients arrive, they find doctors, nurses, drugs and equipment.

It is this segmentation that countries are trying to reduce. They’re doing so in different ways.

Some—like the Philippines and Vietnam—have tried to bring informal-sector workers and their families into the SHI scheme, often offering partial or complete subsidies for the poor. Other countries have set up insurance schemes that operate in parallel to the SHI schemes. Sometimes, as in China and Mexico, these are located in the health ministry. In others, like India, the labor ministry operates the parallel scheme.

Irrespective of who runs it, voluntary schemes have proved vulnerable to adverse selection, and to low enrollment rates and/or low revenues from contributions (on paper some governments apply means-testing but in practice haven’t done so rigorously). And enforcing mandatory enrollment has proved hard.

An alternative model is the Thai model. Thailand set up what is in effect a parallel insurance scheme within its health ministry for those outside of the SHI programs; however, everyone is covered, and the costs are covered through additional taxes.

All these parallel schemes—however financed and wherever located—leave open the possibility of narrowing gaps in coverage but not eliminating them, a useful strategy for a government with limited revenues to play with.

But they all raise the question of what enrollees gain from the ‘insurance’ process, especially when the health ministry operates the scheme. What’s different from the original MOH model that was seen as the cause of poor quality care and high out-of-pocket spending in the first place?

One answer seems to be that the schemes give their enrollees a card that explicitly entitles them to a specified set of services—a type of patients’ charter. Providers have to deliver.

Another advantage seems to be that it provides governments with an opportunity to change the way providers are paid. They can shift from salaries and budgets (that do nothing to incentivize providers to turn up to work and make sure drugs and equipment are in stock), to higher-powered payment methods like fee-for-service or payments per case (which encourage doctors to show up for work and to make sure they have drugs and equipment).

So, what should the Lancet’s map have shown?

It wouldn’t be straightforward to calculate but one possibility would be the amount of spending needed to bring everyone up to the de facto coverage enjoyed by the group with the most generous benefit package. Does anyone have any other ideas?

And am I right that the coverage challenge is not actually about achieving universal coverage, but rather about reducing inequalities in coverage? If so, where do we go wrong in policy discussions—if at all—by misleadingly talking of reaching “universal coverage” when everyone already has some coverage?


Adam Wagstaff

Research Manager, Development Research Group, World Bank

Joe Kutzin
November 17, 2010

You're absolutely right here -- this way of looking at coverage has a bias towards "formal coverage" and results in a very biased view. Sri Lanka appears to have 0-10% coverage even though the EQUITAP data show low levels of catastrophic, low levels of inequity in utilization, and levels of utilization close to the OECD average. But no contributory-based insurance, so no coverage according to this map.

In the background paper (I think it's a "long" HNP working paper, from 2000) to my 2001 article, the last section of the conclusions was called "measuring coverage". As I am long-winded, repeating it here:

"Defining insurance as a function rather than as membership in a scheme raises questions of measurement: how can a country determine the proportion of its population that is effectively covered, and how can changes in this coverage be assessed over time? If insurance is defined as participation in a scheme, measurement simply involves calculating the percentage of the population in schemes. This neglects the possibilities that (1) persons who are in a scheme may not be effectively covered, and (2) persons not in a scheme may be effectively covered. What is needed is a way to measure both the breadth and depth of coverage, and it may not be possible to capture these two elements in a single index measure (i.e. percent “covered”).

Since the insurance function is concerned with access to effective services and financial protection, methods are needed to measure each of these. This poses many difficulties, one of which is that conceptually, there are many degrees of access and protection; they are not discrete variables. Measures of access will need to include assessments of physical and financial access to care. In terms of physical access, it may be possible for countries to examine access to key ‘tracer’ services, such as basic primary care, emergency services (e.g. emergency obstetric services), and referral hospitals. Financial access can probably best be measured with the help of data on care seeking behavior and out-of-pocket health care expenditures derived from household surveys, although indirect information gleaned from health facilities (e.g. changes in the number of people exempted from fees) may be of some use. It will also be necessary to have a consistent definition of what constitutes “good” or “adequate” health care. Financial risk protection may be examined at the policy level (e.g. is there an out-of-pocket maximum?), but the analysis of actual financial risk protection also needs to involve analysis of household survey data showing, for example, changes over time in the percentage of total household expenditure devoted to health care, in the distribution of financial risks for health care expenditures (see Pradhan and Prescott 1999), and in the level (in absolute terms or relative to income) at which the financial risk for health care expenditures is ‘truncated’, if at all.

The challenge in measuring coverage will be to develop reasonably low cost and accurate methods. The potential payoff from such work is great, because it would shift the analytic focus from measuring the implementation of reform instruments to measuring the effects of these instruments on health system objectives."

I believe that in the forthcoming World Health Report, we have stayed away from this bias, and among other things, give the message that contributory-based entitlement has limits and general budget revenues have to play a key role. But you can read for yourself next week and see if we have done this well enough or are still trapped in this institutional bias.

In any case, we should write something about this (again).

Prashanth NS
December 01, 2010

Universal health coverage for me is a rebranding of the primary health care agenda. It seems sometimes that in public health, we like to re-package existing cliches as a revolutionary new innovation. Yes, primary health care as it was formulated, was indeed idealistic. How does the migration to a new terminology help?

Also, I am not sure I agree when you say that "some coverage" is universal. (Or am I reading you wrong?). There are still several regions in India (even!) where primary health care is a far cry. Of course, I speak of a large proportion of the 8 per cent tribal population. But on the other hand, we have at least over a 100 million people below poverty line, many living amongst the sprawling nine per cent economic growth who cannot afford even care for outpatient conditions! So, it is indeed a long way India still needs to go!

What I do like though is the fervor surrounding the call for universal health coverage. Perhaps, it may invigorate the policy makers into action. I am also impressed with the coverage on the ill effects of user fees and the stress on financing of health systems which seem to be neglected functions of the primary health care movement.

Anyways, my disjoint comments to your nicely written blog! Thanks!

Doctor Michel ODIKA
January 21, 2011

Now more than ever, better information systems remain essential in expanding the "universal health coverage". Why?

First of all, information and communication technologies (ICT) remain essential in improving resource use and service supply. However, the relationship between ICT and governance is a powerful but nuanced one. For instance, governance reforms (1) need to be informed, not just by performing ICT, but also by basic data and strategic information obtained through a stark departure from traditional views on the architecture and the scope of conventional information systems. How to do that depends on context and background. Why?

The paradigm shift required to make ICT instrumental to governance reforms is to focus on what is effective and efficient in building a critical mass of capacity for positive change. Unfortunately, regardless of whether or not they are controlled by the public sector or by the business world, many, if not most, information systems in low- and high-income countries can be characterized as closed administrative structures through which there is limited flow of data on resource allocation for service delivery. They are often only used to a limited extent by a limited number of top officials at national and global level when formulating policy reforms, while little use is made of critical and strategic information that could be extracted from other tools and sources – e.g. opinion surveys, NGOs, professional associations, academic institutions, research centres, etc. -, many of which are located outside the public sector and (sometimes) far away from the business world…
Now more than ever, from a policy point of view, the crucial information is that which allows identification of the operational and systemic constraints. In this context, the multiplication of information needs and users implies that the way information is generated, shared and processed also has to evolve. This critically depends on transparency (availability and accessibility), for example, by making pertinent information readily accessible via the Internet (2).

Today, governance reforms call for open and collaborative models, such as “Malaria Observatories” (3,4), to ensure that all the best sources of data are tapped and information flows quickly to those who can translate it into appropriate action. Once established, these state/non-state multi-stakeholder networks can play a key role in complementing and improving routine information systems, by directly linking the production and dissemination of intelligence on specific issues to the sharing of best practices. Generally speaking, these innovative structures reflect the increasing value given to cross-agency work: they thus institutionalize the linkages between local governance and country-wide policy-making…

There is need for making information and communication technologies (ICT) instrumental to governance reforms. Unfortunately, the institutional capacities to meet this ambitious requirement are typically weak in countries classified as low-income. However, even in countries with well-resourced information systems and sophisticated communication networks, there is still need for far-reaching improvements and groundbreaking innovations in terms of architecture, scope, multisectoral response and multidimensional approach…

Doctor Michel ODIKA (Congo-Brazzaville)

1. Governance reforms: balanced approaches to be found (http://blogs.worldbank.org/governance/comment/reply/841/1845)
2. Site internet pour le Ministère de la Santé (http://www.slideboom.com/presentations/169290/Site-internet-pour-le-Min…)
3. Advocacy for a Malaria Observatory in Congo-Brazzaville (http://www.booksie.com/health_and_fitness/article/michel_odika/advocacy…)
4. Observatoire du Paludisme: capital au service d’un idéal (http://www.slideshare.net/Odika/observatoire-du-paludisme-capital-au-se…)

Doctor Michel ODIKA
January 22, 2011

Recently, many countries have gone through exercises to define the package of benefits they feel should be available to their citizens. In some cases, this conceptual leap has been one of the key strategies in improving the quality of health systems in terms safety, efficiency, responsiveness and cost-effectiveness.
In most cases, however, attempts to rationalize service delivery by defining packages have not been successful. Due to inadequate or unsatisfactory information systems, for example, the scope in many, if not most, sub-Saharan African countries has been limited to maternal and child health care, and to health issues (infectious diseases) regarded as health priorities.
Another high-profile example is, not only the lack of anticipation, but equally the lack of attention to emerging, or growing, challenges in developing countries. Otherwise said, urbanization and globalization combine to make chronic and non-communicable diseases – e.g. diabetes mellitus, cardiovascular diseases and cancers – important causes of mortality and morbidity in countries classified as low-income, most of them in sub-Saharan Africa. As a result, there is now a striking shift in distribution of death and disease from younger to older ages and from infectious, prenatal and maternal causes to non-communicable diseases. For example, tobacco-related deaths are expected to overtake HIV/AIDS-induced deaths in many countries, including in… Africa.

Today, more sophisticated approaches are required to make the notion of “Universal Health Coverage” more relevant and context-sensitive. This suggests, not only principles of good practice, but also facilitators of evidence-informed health policymaking such as… CREDIBLE INFORMATION SYSTEMS.

Doctor Michel ODIKA (Congo-Brazzaville)

Xenia Scheil-Adlung
July 13, 2011

Dear Adam,

Thanks for the blog - nice discussion! I fully agree with your comments and others. You might be interested to see ILO’s approach towards measuring universal health coverage at http://onlinelibrary.wiley.com/doi/10.1111/j.1468-246X.2011.01400.x/pdf