Use of emergency departments by elder patients: A five-year follow-up study

Objective: To compare the use of emergency medical care by elders in the Un ited States in 1995 with that previously described for 1990. Methods: A com puterized billing database of 88 EDs in 21 states was retrospectively revie wed for 1995, comparing elder and nonelder patients, estimating national us e of emergency medical services by elders, and comparing the 1995 data with previously published results for 1990. Results: From 1990 to 1995, the num ber of ED visits in the United States increased from 92 million to 100 mill ion. The number of visits made by patients aged 65 years or older increased from 13,639,400 (15%) to 15,666,300 (15.7%), but this increase did not rea ch statistical significance (p = 0.17). The admission rate for elder ED pat ients increased from 32% to 46% over the five-year interval (p < 0.01). Thi s represents more than 7 million hospital admissions for elder patients in 1995. The rate of intensive care unit (ICU) admission for elders decreased from 7% to 6% over the five-year interval (p = 0.56), compared with 1.3% fo r nonelder patients for both years. Thirty percent of elder ED patients arr ived by ambulance in 1990, compared with 33% in 1995 (p = 0.02). Based on 1 995 data, elders comprised 39% of patients arriving by ambulance [odds rati o (OR) 4.75, 95% confidence interval (CI) = 4.71 to 4.79], 43% of all admis sions (OR 6.59, 95% CI = 6.54 to 6.64), and 47% of ICU admissions (OR 5.00, 95% CI = 4.91 to 5.09). The comparable ORs in 1990 were 4.4, 5.6, and 5.5, respectively. Conclusions: From 1990 to 1995, the overall number of ED vis its increased. The rate of increase was somewhat greater for elder patients . The use of ambulance services also disproportionately grew among elder pa tients, as did the rate of hospital admission. The overall rate of ICU admi ssion was stable, but actually fell modestly for elder patients. Of these c hanges, only the increase in the rate of hospital admission for elders reac hed statistical significance.