The study determined the policies and procedures for the control and prevention of methicillin-resistant Staphylococcus aureus (MRSA) and its prevalence among nursing and residential homes, and evaluated whether certain home characteristics such as bed size, staffing level, and type of home are related to the prevalence of MRSA. A 21-questionnaire survey, with primarily categorical responses, was mailed to the home managers of all the 121 nursing and residential homes in the district, following which a simple, stratified random sample of 28 (23.14%) homes was taken and all agreeing residents screened from multiple sites for MRSA. Seventy-seven (63.6%) homes returned a completed questionnaire, 13 (46.4%) of whom agreed to participate in the microbiological study. The response rates for returning questionnaires and agreeing to participate in the microbiological study were similar for nursing and residential homes (65% vs. 60%; 67% vs. 40%; P = 0.12; P = 0.62), respectively. Nursing homes had a mean bed size of 30 (95% Confidence Interval (CI) 17-43), not significantly different from residential homes of 23 (95% CI 18-27; P = 0.26). The nursing homes employed a mean of 8.6 (95% CI 4.7-12.5) staff nurses per home; significantly higher than residential homes with a mean of 1.6 (95% CI 0.3-2.8; P = 0.006). No significant differences in mean number of home care assistants employed per home (22.8; 95% CI 12.4-33.13; and 14.4; 95% CI 11.83-16.90; P = 0.098, for nursing and residential homes, respectively) were observed. None of the homes had employed infection control practitioners. Only four (6.8%) of the responding homes stated that MRSA was a problem. Nursing homes were not significantly more likely to have admission policies for colonized person than residential homes (10/13 vs. 40/55, P = 1.00). Of the fifty-five (71.4%) homes who had admission policies, 40 (72.7%) stated that persons colonized/infected with MRSA would not be accepted, while 12 (21.8%) would accept such persons in single-room isolation and/or barrier nursing. Greater proportions of residential homes than nursing homes would not accept admission of persons with documented MRSA colonization (30/35 vs. 4/10, P = 0.007). Four (9.1%) homes (three nursing) had identified a total of five residents colonized/infected with MRSA in 5 years prior to the survey. Two hundred and forty-six residents were screened (552 sites), two (0.81%) of whom were found to be colonized in the nose (one resident) and in the groin (two residents) with MRSA, giving a 2-month weighted point prevalence rate of 0.14% (95% CI 0.01-0.26%). We conclude that in our district the nursing staffing levels and control measures vary widely within these homes, while the prevalence of residents who are colonized/infected with MRSA is lower than in other areas. We suggest that the exclusion admission policy for MRSA positive patients should be abandoned and targeted infection control programmes be instituted.