—To determine the prevalence, geographic variation, and charges to Medicare of alcohol-related hospitalizations among elderly people in the United States.
—A cross-sectional prevalence study using 1989 hospital claims data from the Health Care Financing Administration (HCFA). Rates were determined using (1) hospital claims records from the HCFA's Medicare Provider Analysis and Review Record (MEDPAR) database for all Medicare Part A beneficiaries aged 65 years and older; (2) county population estimates for 1985 from the Bureau of the Census; and (3) per capita consumption of alcohol by state in 1989 as estimated by the US Department of Health and Human Services.
—Data include all hospital inpatient Medicare Part A beneficiaries aged 65 years and older in the United States in 1989.
—The prevalence of alcohol-related hospitalizations among people aged 65 years and older nationally in 1989 was 54.7 per 10 000 population for men and 14.8 per 10000 for women. Comparison with hospital records showed that MEDPAR data had a sensitivity of 77% to detect alcohol-related hospitalizations. There was considerable geographic variation; prevalence ranged from 18.9 per 10 000 in Arkansas to 77.0 per 10 000 in Alaska. A strong correlation existed between alcohol-related hospitalizations and per capita consumption of alcohol by state (Spearman correlation coefficient,.64; P<.0001). In 1989, the hospital-associated charges to Medicare for all admissions where the primary diagnosis was alcohol related (N=33 039) totaled $233 543 500. Median charge per hospital stay was $4514.
—Alcohol-related hospitalizations among elderly people are common; rates were similar to those for myocardial infarction as detected by the same method. The charges to Medicare for this preventable problem are considerable. Ecological analysis suggests that per capita consumption in the total US population predicts alcohol-related hospitalizations in the elderly population.(JAMA. 1993;270:1222-1225)