Stress testing: National patterns and predictors of test ordering

Objective To determine predictors of ordering of exercise stress tests. Background Because exercise stress testing is routinely used and widely ava ilable and may have an effect on subsequent evaluation of and therapy for h eart disease, understanding current patterns of ordering exercise stress re sts may have important implications for national health care costs. We hypo thesized that factors other than clinical condition exert an influence on o rdering of exercise stress tests. Methods Data from the 1991 and 1992 National Ambulatory Medical Core Survey s conducted by the National Center for Health Statistics were analyzed by m eans of multivariate logistic regression. Results In an estimated 1.12 billion adult visits to office-based physician s in the United Stares (95% confidence interval [CI], 1.07-1.16 billion), 6 .2 million (95% CI, 4.8-7.6 million) exercise stress tests were ordered. Af ter adjustment for clinical and nonclinical variables associated with the o ffice visit, cardiologists were 3.7 (95% CI, 2.7-5.1) times more likely to order exercise stress tests than were internists, who were more likely to o rder an exercise stress test than were family and general practitioners (0. 5, 95% CI, 0.3-0.7). Nonclinical factors associated with increased ordering of exercise stress tests included male sex (odds ratio 2.5; 95% CI, 2.0-3. 2), white race (odds ratio 1.6; 95% CI, 1.1-2.3), new referral status (odds ratio 3.8; 95% CI, 2.5-5.8), and private insurance (odds ratio 1.4; 95% CI , 1.1-1.8). Medicare recipients were about half (95% CI, 0.4-0.9) as likely as other patients to have an exercise stress test ordered. Conclusions Factors other than clinical condition exert an influence on ord ering of exercise stress tests and may represent modifiable elements associ ated with appropriate practice.