Epidemic tuberculosis in a nursing home: a retrospective cohort study.

In a retrospective study, the results of tuberculin skin tests done in a nursing home were examined, where most residents admitted during 1972-1981 were tested using purified protein derivative of tuberculin administered intradermally. Of 514 residents who were tested at least once during the study period, results of the admission skin test were available for 254 and follow-up skin test results were available for 226. On admission, 13 per cent (35/254) were skin-test positive (greater than or equal to 10 mm induration at 48 hours). Skin test positivity for males was 16 per cent, females, 11 per cent, nonwhites, 19 per cent, and whites, 12 per cent. Highest skin test positivity was for persons received as transfers from other nursing homes (24 per cent) and lowest was for those entering from individual homes (8 per cent, P = 0.016, Fisher's Exact Test). On follow-up, 38/226 (17 per cent) residents who had been tuberculin-negative on at least two previous occasions were found to be positive; 24 (63 per cent) of these conversions occurred in a single year (1975) following detection of an infectious patient. The infection rate for persons residing in the nursing home during that year was 28 per cent (28/99) compared with 7.9 per cent (10/127) for persons either discharged before or admitted after 1975 (relative risk = 3.6, P less than 0.001). Ten to 15 per cent of new residents are tuberculin-positive (harbour a dormant tuberculous infection), leaving 85-90 per cent of newly admitted persons who are tuberculin-negative and thus susceptible to infection if exposed. The study shows that tuberculosis must be considered as a potential nosocomial infection in nursing homes.(ABSTRACT TRUNCATED AT 250 WORDS)