The Measurement and Evolution of Health Inequality

06/01/2005
Featured in print Digest

During 1987-2001, low-income households experienced an increase of 78 percent ($2624) in per capita expenditures on healthcare, double the increase for the highest income group of 34 percent ($1214)...But survival for the lowest income group during the 1990s grew by 0.2 years, compared to 0.8 years for the highest income group.

Has U.S. health care become more equitable for the elderly during the past several decades? If equality is measured by Medicare expenditures, the answer is yes. During 1987-2001, low-income households experienced an increase of 78 percent ($2624) in per capita expenditures on healthcare, double the increase for the highest income group of 34 percent ($1214). However, if equality is measured by life expectancy, the answer is no. Survival for the lowest income group during the 1990s grew by 0.2 years, compared to 0.8 years for the highest income group. The fact that the two measures deliver such discordant messages may reflect their intrinsic shortcomings: expenditures depend on preferences, health status, and prices, while outcomes are strongly affected by health behavior and past illness.

In The Measurement and Evolution of Health Inequality: Evidence from the U.S. Medicare Population (NBER Working Paper No. 10842), authors Jonathan Skinner and Weiping Zhou -- using U.S. data from the elderly Medicare population over age 65 -- consider an alternative approach to measuring inequality. The authors suggest comparing quality of care across income groups -- as measured by the use of effective treatments with proven effects on health outcomes -to capture the degree of inequality in the health care system. This approach avoids complications encountered in other measures of inequality such as differences in underlying health status or preferences. The efficacy of these treatments is so well proven that nearly everyone in the relevant population, regardless of health or preferences, should be receiving these treatments.

Using Medicare claims data matched to zip code and income, the authors find greater use of mammography screening, diabetic eye exams, and alpha-blockers and reperfusion following heart attacks among the higher income households. These differences in care appear to be stable or growing slowly over time. The rapid relative growth in health care expenditures among low-income elderly people has not translated into relative improvement in either survival rates or rates of effective care.

The authors caution, though, that the magnitudes of the differences in effective care observed in the data would not be expected to have a large impact on overall mortality rates. Ensuring that low income households are as likely to receive beta-blockers as are high income households would increase survival among heart attack patients by only about 0.2 percentage points. (The impact on population health would be even smaller, since heart attack patients make up a small fraction of the total elderly population.) The fact that these measures of effective care account for a small fraction of overall expenditures, and a small fraction of the overall variation in health outcomes, motivates the authors' interest in whether other measures of quality are correlated with mammography rates or beta-blocker use.

The authors also note some important limitations of their study. For example, using outcome data, they focus only on survival and not on quality-adjusted or "healthy life years." To capture a fuller measure of health, it would be necessary to include income-based differentials in treatments with proven effectiveness in improving functioning rather than survival per se. Some examples include hip or knee replacements for the treatment of osteoarthritis or the use of angioplasty for patients with ischemic heart disease. Another limitation is that this study is confined to the over-65 population. Focusing just on income-based differences in mammography rates within the Medicare program ignores the fact that Medicare itself contributes to a substantial increase in mammography rates at age 65 among those previously uncovered by insurance or in lower educational groups.

The authors point out that a singular advantage of focusing on equality in effective care (or quality of care) is that there are reasonable approaches to fixing the problem. Monitoring claims data in real time with the objective of raising rates to ideal levels of near 100 percent among appropriate candidates is one sure way to at least reduce income-based inequality in health outcomes. Inequality in outcomes may continue for many years, but at least such differences would not be exacerbated by inequality in health care. Indeed, one could imagine "non-discrimination" rules like those developed for 401(k) pension plans in which hospitals or health care systems would experience a partial loss in Medicare funding if effective care measures for their low income patients fell too far below those for their high income patients, or too far below those for all patients at high quality hospitals.

Quote: "During 1987-2001, low-income households experienced an increase of 78 percent ($2624) in per capita expenditures on healthcare, double the increase for the highest income group of 34 percent."

-- Les Picker