What influences physician practice behavior? An interview study of physicians who received consultative geriatric assessment recommendations
R. C. Maly, A. F. Abrahamse, S. H. Hirsch, J. C. Frank and D. B. Reuben
Division of Family Medicine, University of California-Los Angeles School of Medicine, USA.
BACKGROUND: Comprehensive geriatric assessment (CGA) in outpatient settings
has not been shown to be as effective in reducing mortality and improving
health as in hospital settings; this difference has been attributed in part
to a lack of direct control over recommendation implementation. OBJECTIVE:
To identify inhibiting and facilitating factors in physicians' compliance
with consultative CGA recommendations, so that the effectiveness of
outpatient CGA might be improved. METHODS: A 49-item questionnaire was
administered via the telephone to 87 eligible community primary care
physicians in Los Angeles, Calif, whose patients had received consultative
outpatient CGAs as part of a study of CGA (response rate, 96%). The
questionnaire assessed physician compliance with CGA recommendations,
reasons for implementing or not implementing the recommendations, and
specific physician attitudes, perceptions, and characteristics. The focus
of the interview was the CGA recommendation that was determined to be the
"most important" by the evaluating geriatrician. Recommendations addressed
geriatric syndromes, general medical problems, or psychiatric conditions.
RESULTS: Of the 87 physician respondents, 62 (71%) implemented the most
important recommendation. In multivariate analysis, 4 variables were
predictive of physician compliance: (1) a patient's request that the
recommendation be implemented (odds ratio [OR], 10.8; 95% confidence
interval [CI], 1.9-61.3; P = .007); (2) perceived legal liability resulting
from nonimplementation of the recommendation (OR, 10.8; 95% CI, 1.1-108.2;
P = .04); (3) female physician gender (OR, 9.6; 95% CI, 1.4-67.9; P = .04);
and (4) perceived cost-effectiveness of the recommendation (OR, 7.0; 95%
CI, 1.6-30.5; P = .01). CONCLUSIONS: Patient behavior, which may be
modifiable, was among the strongest determinants of physician compliance
with recommended care. Specifically, when patients requested that a
recommendation be implemented, physicians were highly likely to comply.
Changing patient behavior within the physician-patient relationship as a
way of effecting desired changes in physician health care practices merits
further attention.
