of the conditions chosen for this analysis. Finally, in terms of differences in disease prevalence, the research group looked at asthma and diabetes—two conditions for which data on prevalence are available through the NHIS. Depending on the age group, differences in prevalence between high– and low-income populations ranged from 1.35 to 2.36 times higher for low-income age cohorts with asthma and from 1.15 to 2.96 times higher for low-income cohorts with diabetes. Prevalence thus explains only a portion of the four– to fivefold differences between income groupings for these two conditions.
Hospital Discharge Data Systems. States that do not have centralized hospital discharge data bases should develop them. In addition to their value for the types of analyses suggested by this committee, the data bases will be useful for future research on costs and quality of care.
Expanding Data Elements in the Discharge Abstract. As recommended previously, states should consider the feasibility of adding additional elements to the discharge abstract, especially information to measure the severity of illness and income.
Further Research. The committee believes that more detailed studies of patients and admitting physicians are needed to sort out the relative contribution of the various factors, including access to primary care, that lead to hospitalization for chronic disease-related conditions. Items of particular interest include the timeliness and quality of outpatient care, patient characteristics, and physician admitting practices. Studies focusing on better measurement of continuity of care and the effect of site of care (walk-in clinics, physician's offices, hospital clinics, community health centers, emergency departments) also should be considered.
The access-related excess mortality rate is the number of deaths per 100,000 population that are thought to be the result of access problems. The estimate is based on a comparison of two groups in the population—one that is believed to have relatively good access and one that is considered likely to experience barriers to access. Because data are available (Stoto, 1992), the population groups of particular focus for this measure are blacks and whites.
It has been well documented that, compared with whites, blacks in the United States have a disproportionately high mortality rate from chronic disease. Some of the difference may be the result of increased levels of