Physician characteristics and prescribing for elderly people in New Brunswick: relation to patient outcomes

Abstract

OBJECTIVE: To examine the relation between physician characteristics, prescribing behaviour and patient outcomes. DESIGN: Descriptive study linking four provincial databases. SETTING: New Brunswick. PARTICIPANTS: All 366 general practitioners (GPs) (accounting for 40% of all physicians with a general licence in New Brunswick) who ordered at least 200 prescriptions for elderly beneficiaries of the New Brunswick Prescription Drug Program and saw at least 20 elderly patients in an office setting between Apr. 1, 1990, and Mar. 31, 1991. Physicians with palliative care practices were excluded. OUTCOME MEASURES: GPs' personal, professional and practice characteristics, their prescribing patterns, and mortality, morbidity (number of days in hospital per patient) and hip-fracture rates among their elderly patients. RESULTS: Compared with the GPs who had a lower mortality rate, those with a higher mortality rate prescribed more drugs overall (p < 0.001), specifically antidepressants, bronchodilators, cholesterol-lowering agents, gastrointestinal drugs, neuroleptics and nonsteroidal anti-inflammatory drugs (NSAIDs). They also were more likely to be male (p < 0.01), had larger practices (p < 0.001), saw more patients per day (p < 0.05) and billed more per year (p < 0.001). Compared with the GPs who had a lower morbidity rate, those with a higher morbidity rate prescribed more drugs overall (p < 0.005), specifically bronchodilators, gastrointestinal drugs and NSAIDs. They also were more likely to be younger (p < 0.005) and male (p < 0.01), had fewer years in practice (p < 0.001), saw more patients per day (p < 0.05) and billed more per patient (p < 0.01). The GPs who had a higher hip-fracture rate prescribed more drugs overall (p < 0.001), notably antihypertensives, bronchodilators, cholesterol-lowering agents, gastrointestinal drugs and NSAIDs, than those who had a lower hip-fracture rate. They also had a larger practice (p < 0.001), practised more days per year (p < 0.005), had more patient visits per year (p < 0.05) and billed more per year (p < 0.001). Younger male GPs who practised with relatively more intensity and prescribed more drugs per patient had practices with higher morbidity, mortality and hip-fracture rates among their elderly patients than the other GPs. CONCLUSIONS: There is a significant relation between certain physician characteristics, their prescribing behaviour and patient outcomes. Further study is required to determine what physician characteristics and prescribing behaviours for specific illnesses contribute to patient outcomes. Regional differences should also be examined, as should incentives in this fee-for-service system. Linkage of these types of provincial databases may help in the evaluation of physicians' performance and in the development of strategic interventions and practice guidelines.