Table 5.

Table 5 Adjusted Odds Ratio of Inappropriate Prescriptions According to the Method for Validation (Pharmacist or Electronic) and to the Type of Prescriptions Included (First or All)

Variables	Model 1 Pharmacist Validation First Prescriptions	Model 2 Pharmacist Validation All Prescriptions	Model 3 Electronic Validation First Prescriptions Really Given to Patients	Model 4 Electronic Validation All Prescriptions Really Given to Patients
OR (CI 95%)	N = 700	N = 942	N = 673	N = 843
Intervention ∗
Residents	0.69 (0.41–1.15)	0.70 (0.36–1.36)	0.70 (0.42–1.18)	0.75 (0.44–1.30)
Senior physicians	1.88 (0.91–3.89)	1.72 (0.92–3.23)	1.66 (0.82–3.36)	1.79 (0.85–3.75)
Patient's age (yr)	0.98 (0.97–1.00)†	0.97 (0.95–0.99) †	0.98 (0.96–0.99)‡	0.97 (0.95–0.99)‡
Drug category	§	§	§	§
ACEI	1.0	1.0	1.0	1.0
B-adrenergic blocking agent	2.0 (1.0–4.2)	3.4 (1.5–7.5)	3.8 (1.8–7.9)	5.3 (2.5–11.1)
Antibiotic	10.7 (5.5–20.7)	27.2 (10.7–69.3)	16.1 (7.8–33.2)	23.6 (10.3–54.2)
Antigout drug	2.5 (1.1–6.1)	3.3 (1.2–8.6)	3.2 (1.3–7.8)	3.2 (1.3–7.6)
Digoxin	3.2 (1.3–8.3)	7.3 (2.3–23.6)	4.4 (1.6–11.7)	6.4 (2.3–17.7)
Antidiabetes drug	78.0 (15.2–400.0)	527 (56–4785)	—	—

^∗ Due to the significance of interaction term Intervention *Physician's level in Model 1, the odds ratio for the term Intervention was different for residents and senior physicians. The odds ratio of inappropriate prescriptions for senior physicians in intervention versus control periods was obtained by multiplying the odds ratio associated with the variable Intervention and the interaction term Intervention *Physician's level.

^† p < 0.05.

^‡ p < 0.01.

^§ p < 0.001.

Model 1 and Model 3: ORs were estimated in an ordinary multivariable logistic regression model (no patient effect). Model 2 and Model 4: ORs were estimated in mixed multivariable logistic regression model (significant patient effect).

ACEI = angiotensin converting enzyme inhibitor; CI = confidence interval; OR = odds ratio.