In the developing world, a hospitalization is one of the things that families – especially poor ones – fear most. This came through in country after country in the World Bank’s Voices of the Poor exercise. Here are just some examples:
A man from Ghana is quoted as saying: “Take the death of this small boy this morning, for example. The boy died of measles. We all know he could have been cured at the hospital. But the parents had no money and so the boy died a slow and painful death, not of measles, but out of poverty.”
The researchers write that in Lahore, Pakistan, “a father explained that it had taken him eight years to repay debts acquired after he, his wife, and two of their children had been hospitalized.”
In Orissa, India, the authors found that “the hospital is rated by the poor as one of the most important institutions – and also as one of the most prohibitively expensive.”
In Macedonia, the authors were told: “Doctors openly told one man from Debar whose wife needed an ulcer operation, “If you have a thick envelope it is all right, if not, scram.”
In Uganda, the team was told about a girl who died because the family didn’t have the money to seek treatment: “By the beginning of May, Grace had become very weak. Her parents then sold some chickens for Shs. 2,500 and, with the help of neighbors, took her to Ngora Hospital where she was immediately admitted. She was seriously anemic and required urgent blood transfusion. However, the family was asked to pay Shs. 5,000 that they did not have. They went back home to try and look for money. It was too late. She died on 8 May and was buried the following day.”
These stories tug at the heartstrings of course. But they also raise questions. Is a hospitalization more affordable in some countries than others? What are the causes of unaffordable inpatient care? And how will universal health coverage (UHC) reforms help make it more affordable?
An index of inpatient care (un)affordability
An obvious way to measure the (un)affordability of a hospitalization is to take the typical out-of-pocket cost of a hospitalization in a public hospital, express it as a share of a family's “discretionary” expenditure, and multiply by 100. This index focuses on the cost a family would incur in the event of a hospitalization, not the family’s actual expenditure, which will be zero for families who find the price prohibitive, and zero for those who don’t require a hospitalization. This distinguishes my hospitalization (un)affordability index from other measures of financial protection in health.
Both of the two key variables in the index – the typical cost of a hospitalization and discretionary spending – require some thought. For "typical" do we use the mean or the median? Do we strip out outliers? What type of hospital do we mean? I use the mean out-of-pocket cost of a hospitalization, I drop outliers only in the case of one country where it was clearly an issue, and I look at the out-of-pocket cost in public hospitals run by the ministry of health; these are open to everyone in contrast to hospitals run by other ministries such as the ministry of social security and the ministry of defense which can be accessed only by people covered by the insurance scheme run by that ministry. The denominator is less straightforward. What’s “discretionary” spending? And what period do we measure discretionary spending over? I’ve chosen the family’s nonfood spending – excluding health spending – on the grounds that a family needs to eat to survive, and health spending is largely involuntary. I've picked a year as the time period.
My index has a simple interpretation. The higher the index value, the less affordable inpatient care is. A value of 100 means that a hospitalization would absorb an entire year's worth of the household's nonfood spending, leaving it with nothing left for shelter, power, transport, entertainment, and so on. That would clearly be devastating. But even a value of 25 would also be very serious – a hospitalization would eat up a quarter of the family's annual discretionary spending.
I've used data from the World Health Survey to compute this affordability index for all 49 countries that have the required data. The WHS was fielded in the early 2000's so the data are rather old, and they predate a lot of the UHC reforms of the 2000's. Still, I think they're interesting.
The picture for the population as a whole (Fig 1) is pretty shocking. It’s true that at the population level no country comes out above 50. However, five countries come out with an affordability index between 25 and 50. China comes at the top of the global hospitalization unaffordability league table with a score of 49. Georgia (43) is next, then Mali (30), India (27), and Chad (26). Asia scores badly on inpatient care affordability. Nearly half the countries with an unaffordability index of 10 or more are in E Asia or S Asia. They include (in descending order of unaffordability, and in addition to China and India) Vietnam, Myanmar, Nepal, Pakistan, the Philippines, Lao PDR, and Bangladesh. Other countries in the 10-25 group are Ukraine, Mexico, Ecuador, and Burkina Faso.
The most affordable group (where a hospitalization costs less than 1 percent of a family’s annual nonfood expenditure) include most of the high-income countries (Latvia is the exception). Slovenia and the Czech Republic are the countries with the most affordable inpatient care with index values of just 0.02 and 0.003. But this top group of countries also includes plenty of less prosperous countries. It includes some upper middle-income countries such as Brazil and Mauritius, a couple of lower middle-income countries (Tunisia and Swaziland), and even some low-income countries (Malawi and Mauritania). Several countries come in at between 1 and 2 percent, notably Bosnia and Herzegovina, Malaysia, Namibia, Senegal, South Africa, Sri Lanka, Zambia and Zimbabwe.
A hospitalization is, of course, much less affordable for a poor family. Fig 2 shows the (un)affordability index for the poorest fifth of the population (as measured by a synthetic wealth index). Here there are countries with an index of over 50. China again comes out worst with an index of 129. For a family in the poorest fifth of China’s population, a hospitalization eats up as much as 129 percent of a year’s worth of discretionary spending. The regional differences are striking. In eighty percent of the countries of E Asia and S Asia the poor can expect to spend more than 25 percent of their annual nonfood consumption on a typical inpatient admission.
What causes hospitalization unaffordability?
At the end of the day inpatient care can be unaffordable at the national level for either or both of two reasons: the population is paying a large share of the cost of a hospitalization out of their own pocket at the point of use; and the cost of an inpatient admission is high relative to per capita income.
Neither is a fixture. It is no accident that China scored so badly at the time the WHS was fielded. As I and my coauthors explained in our book on China , at that time only formal-sector workers and civil servants had any insurance coverage, and government subsidies to health facilities delivering inpatient care were fairly small. Patients themselves paid most of the cost of an inpatient admission out of their own pocket. Moreover, health providers were paid via fee-for-service through a price schedule that allowed them to make profits on expensive care (drugs, tests, and the like) and forced them to make a loss on basic care. Unsurprisingly providers focused on delivering expensive care – new technology was adopted rapidly even in small facilities, drugs were overprescribed, and health care costs increased rapidly from one year to the next. While China’s economy was already large at this time, its per capita income was still relatively low. Yet its style of care was fast resembling that of a much richer country.
The story is likely to vary from one country to the next, but I suspect countries that do well on the hospitalization affordability index do a mix of two things. First, they’re likely to have in place a financial protection scheme that covers the costs of hospitalization so that inpatient care is financed only minimally through out-of-pocket payments at the point of use. This scheme will likely be coupled with an initiative to root out under-the-table payments demanded by corrupt providers. Second, they’re likely to have policies in place that keep hospital costs under control and in line with what the country can afford. Examples include a hospital payment system that encourages cost consciousness, an emphasis on the use of clinical guidelines and protocols, and an institution that ensures that technology and drugs are introduced in a systematic way with benefits being rigorously assessed and compared with what the country can afford. There are likely other policies that will also help keep hospitalization costs in check.
It’s probably no coincidence that much of the push toward UHC in the 2000’s came from the region with the most red and yellow in Fig 1 – Asia. And it is probably no coincidence that inpatient care costs have featured in prominently in these initiatives. In fact, in several countries – notably China  and India , but also the Philippines  – cover has focused exclusively or almost exclusively on inpatient care.
One concern is whether the coverage is deep enough to make enough of a dent in the typical out-of-pocket cost of a hospital admission. In the early days of China’s reforms coverage was markedly shallow, and in India today families covered by Aarogyasri (a subnational scheme focusing on tertiary care that started out in Andhra Pradesh) or RSBY (a national scheme focusing on secondary inpatient care) face incomplete coverage even for inpatient care.
If I’m right in my hunches about what makes countries more likely to be red or yellow than green in the two maps, UHC reforms also need to devote a lot of attention to putting downward pressure on the overall cost of hospital care. In the early days of China’s reforms there were concerns that insurance might steer people to more costly hospitals – ones that may or may not deliver more appropriate care. In the Philippines too concerns have been expressed that, given the hospital payment system, insurance may simply encourage hospitals to deliver more costly care. China has now started to introduce reforms that attempt to discourage hospitals from delivering unnecessary and unnecessarily costly care, but the pace of reform has been slow.
As I reported in an earlier post , several Asian countries have sought to put downward pressure on hospital costs through a purchaser-provider split, and through contracting with the private sector. Whether this will help remains to be seen. Even if it does, I suspect there are lots of other parts of the hospital cost puzzle – such as moderating the diffusion of new drugs and technology – that UHC reforms will need to get to grips with if they’re going to be successful in reducing the fear that poor families have of requiring a hospitalization.