Article by Eric French and Elaine Kelly, Institute for Fiscal Studies (MethodsNews Spring 2015)
There are significant differences in how health services are financed and provided across the developed world. Yet for almost all countries, the past few decades have been marked by a rise in the share of health care spending as a percentage of GDP. This trend only looks set to continue, as medical technology improves and the population ages.
Existing cross-country work has tended to focus on levels of spending, often identifying the US as the outlier. Yet relatively little is known about how the money is spent or how patterns of spending vary across countries. Documenting and understanding these differences has the potential to help explain variation in patient outcomes and aggregate spending levels, and inform policy makers about possible methods of improving the efficiency and cost effectiveness of their own health care systems.
The ‘Medical spending across the world’ project measures patterns of individual level spending across countries, from the patient upwards, using medical records made available by governments and insurers. In particular, we examine who receives health care, and when they receive it. This is not the first project to attempt to address these questions, but is the first in several dimensions.
Previous cross-country projects used only cross sectional data. Our project exploits datasets that track the same people over many years, which will allow us to better understand medical spending from a lifetime perspective. Furthermore, we now have much better measures of medical spending, as much previous work was based mostly on survey data, whereas our project includes high quality administrative data on total spending from all sources. Work from the project will be published as an issue of Fiscal Studies in early 2016 and will bring together research from the UK, USA, the Netherlands, Denmark, Germany, France, Canada, Taiwan, and Japan.
In March 2015 researchers from the different countries gathered for a conference was held at the Institute for Fiscal Studies in London. Key conclusions from the conference were:
- The concentration of medical spending varies substantially across countries. For example, in the United States, 50% of all medical spending goes to the top 10% of all spenders. In Denmark, the Netherlands, and Taiwan the corresponding number is close to 80%, and in France and Japan it is 60%. In England the share going to the top spenders is lower, at closer to 40%, although the measure of medical spending for England is limited to inpatient (hospital) medical care. Interestingly, many commentators cite the high level of concentration of medical spending in the US as a problem with the delivery of medical care in the US. However, as it turns out, medical spending is less concentrated in the US than in many other countries.
- Medical spending is highly persistent in all countries. High medical spending in one year strongly predicts medical spending in the following year. For most countries, the probability of being in the top spending quintile in a given year conditional on being in the top spending quintile in the previous year is approximately 50%.
- There is no clear relationship between how health care is provided and the correlation between medical spending and income. In most countries, including the US and England, total medical spending is negatively correlated with income. Taiwan is the only country in our project where medical spending is positively correlated with income. This is perhaps surprising as health care in the United States is mostly privately provided, whereas Taiwan has nationalized health care.
The next step for the project is to standardize measures of individual medical spending, so as to provide the most accurate comparisons across countries. In the Fiscal Studies edition we will use these standardized measures to establish a series of common facts about medical spending across the project countries, exploring areas in which countries in differ, and identifying features that most countries share. Once this is done, we will attempt to exploit cross country comparisons to better understand who benefits from medical care, how delivery of medical care varies across countries, and what patterns of medical care delivery are most effective.