ANNUAL NIGHT VISITING RATES IN 129 GENERAL PRACTICES IN ONE FAMILY HEALTH-SERVICES AUTHORITY - ASSOCIATION WITH PATIENT AND GENERAL-PRACTICE CHARACTERISTICS

Background. Rates of night visiting by general practitioners have incr eased steadily over the last 30 years and vary widely between general practices. Aim. An ecological study was carried out to examine night v isiting rates by general practices in one family health services autho rity, and to determine the extent to which differences in night visiti ng rates between practices could be explained by patient and practice characteristics. Method. The study examined the variation in annual ni ght visiting rates, based on night visit fees claimed between April 19 93 and March 1994, among 129 general practices in Merton, Sutton and W andsworth Family Health Services Authority, London. Results. Practices ' annual night visiting rates varied from three per 1000 to 75 per 100 0 patients. The percentages of the practice population aged under five years and aged five to 14 years were both positively correlated with night visiting rates (r = 0.38 and r = 0.35, respectively), as were va riables associated with social deprivation such as the estimated perce ntage of the practice population living in one-parent households (r = 0.24) and in households where the head of household was classified as unskilled (r = 0.20). The percentage of the practice population report ing chronic illness was also positively associated with night visiting rates (r = 0.26). The percentages of the practice population aged 35 to 44 years and 45 to 54 years were both negatively associated with ni ght visiting rates (r = -0.34 and r = -0.31, respectively) as was the estimated list inflation for a practice (r = -0.31). There was no sign ificant correlation between night visiting rates and the distance of t he main practice surgery from the nearest hospital accident and emerge ncy department. There was also no association between night visiting r ates and permission to use a deputizing service. In a stepwise multipl e regression model, the multiple correlation coefficient was 0.56 with four factors (percentage of the practice population aged under five y ears, percentage aged 35-44 years, percentage who were chronically ill and estimated list inflation) explaining 32% of the variation in nigh t visiting rates. Conclusion. Only about one third of the variation in night visiting rates between practices could be explained by patient and practice variables derived from routine data. Population-based res earch using data collected on individual patients and practices is req uired to improve current understanding of the patient and practice cha racteristics that influence the demand for night visits and of why nig ht visiting rates vary so widely between practices.