Clinical variables influencing treatment decisions for agitated dementia patients: survey of physician judgments.

Objective: To better understand primary treatment recommendations and the variables that might influence treatment decisions of physicians who treat agitated dementia patients.

Design: A written cross-sectional survey of three physician groups (geriatric psychiatrists, primary care physicians, and neurologists) who typically treat agitated dementia patients in community settings. We used a written clinical vignette describing a home-bound, agitated dementia patient to ask respondents to provide information regarding their primary treatment recommendation and to estimate the degree to which clinically relevant variables might influence their treatment recommendation. Using principal component analysis, the original set of clinical variables was collapsed into a smaller set of composite factors that better defined the fundamental constructs of the variables that influenced decision making. Analyses compared primary treatment recommendations and factors influencing treatment recommendations by physician groups.

Participants: The pool of survey respondents consisted of a random selection of 207 primary care physicians from western North Carolina, 147 geriatric psychiatrists obtained from the roster of the 1991 American Association for Geriatric Psychiatry, and 120 neurologists obtained from the roster of the American Board of Medical Specialties. The response rate was 65% for geriatric psychiatrists, 38% for primary care physicians, and 33% for neurologists.

Results: Differences in primary treatment recommendations by physician group were not found. Physicians, regardless of specialty, recommended neuroleptic medications as their primary intervention. When medication classes were collapsed into a single category, medications as a primary intervention exceeded 55% for all physician groups. Twenty-two percent of all respondents recommended psychosocial interventions as primary treatment strategies. The principal component analysis of clinical variables influencing treatment recommendations solved for five components that accounted for 64% of the variance. Comparing the five components by specialty groups failed to find significant differences, except for Factor 5, the "Hassle Factor." Primary care physicians were more likely to indicate that this component influenced their decision making than were the other physician groups.

Conclusions: The findings indicate that physicians, regardless of specialty, are likely to use medication and to weight clinically relevant information in a similar fashion when managing agitated dementia patients.