In a multivariable analysis, central nervous system drugs were associated with more off-label use than were cardiovascular drugs (26.3% vs 3.3%; adjusted OR [AOR], 9.91; 95% CI, 9.07-10.84), and formulary-restricted drugs had lower off-label use (2.9%; AOR, 1.01; 95% CI, 0.87-1.16). Drugs with 3 or 4 approved indications had lower off-label use compared with drugs with 1 or 2 approved indications (6.7% vs 15.7%; AOR, 0.44; 95% CI, 0.41-0.48). In addition, drugs with 5 to 7 and those with 8 or more approved indications had lower off-label use: 9.6% (AOR, 0.62; 95% CI, 0.57-0.67) and 9.7% (AOR, 0.32; 95% CI, 0.28-0.37), respectively. Drugs approved after 1995 had lower off-label use than did drugs approved before 1981 (8.0% vs 17.0%; AOR, 0.46; 95% CI, 0.42-0.50); drugs approved between 1981 and 1995 also had lower off-label use than those approved before 1981 (8.4%; AOR, 0.48; 95% CI, 0.43-0.55). Women received more off-label drugs compared with men (11.8% vs 9.7%; AOR, 1.06; 95% CI, 1.03-1.09). Patients with a Charlson Comorbidity Index score of 1 or higher had lower off-label use than did those with a Charlson Comorbidity Index score of 0 (9.6% vs 11.7%; AOR, 0.94; 95% CI, 0.91-0.97). Physicians with higher scores on evidence-based practice were less likely to prescribe off-label. A 5-point increase in the physicians' evidence score on the Evidence-Practicality-Conformity Scale decreased the risk of off-label prescribing by 7% (AOR, 0.93; 95% CI, 0.88-0.99). Patient age, physician sex, and physician graduation year were not associated with off-label use. When the analysis was restricted to off-label prescribing without strong evidence, there were notable differences (Table 4). The AOR for the central nervous system, anti-infective, ear-nose-throat, and antineoplastic drug classes increased by more than 2-fold owing to small percentages of off-label use with strong scientific support in these classes and a large percentage of strong scientific support in the cardiovascular (reference) group. Older drugs and drugs with 1 or 2 approved treatment indications still had the highest risk for off-label use; however, the risk was attenuated. Physicians who graduated in the 1980s and those who graduated in the 1990s-2000s prescribed off-label without scientific evidence more frequently than did the 1960-1970 graduates. In addition, the physician evidence-based practice score had a stronger effect on off-label prescribing without scientific evidence, with a 5-point increase in physicians' evidence scale decreasing off-label prescribing without scientific evidence by 10% (AOR, 0.90; 95% CI, 0.85-0.96).