Table 1. Unadjusted summary data showing the prevalence of PIPs and estimates of RR with 95% CIs in different subgroups of GP and drug-related characteristics at pre- and post-intervention. The total number of prescriptions pre-intervention was 1 034 033 and in the post-intervention period there were 1 104 391 prescriptions.

	GPs, n	PIPs prevalence, %		Relative riska (95% CI)
		Pre-intervention (n = 23 184)	Post-intervention (n = 19 754)
Overall group	449	2.24	1.79	0.80 (0.78 to 0.81)
Intervention	250	1.32	0.99	0.75 (0.73 to 0.77)
Control	199	0.92	0.80	0.87 (0.84 to 0.89)
Sex
Female	139	0.44	0.37	0.83 (0.80 to 0.87)
Male	310	1.80	1.42	0.79 (0.77 to 0.81)
Practice setting
Rural	208	1.01	0.80	0.79 (0.77 to 0.81)
Urban	241	1.23	0.99	0.80 (0.78 to 0.82)
Type of practice
Single-handed	38	0.25	0.20	0.80 (0.75 to 0.84)
Group practice	411	1.99	1.59	0.80 (0.78 to 0.81)
GP specialist
No	63	0.20	0.18	0.89 (0.84 to 0.95)
Yes	386	2.04	1.61	0.79 (0.77 to 0.80)
Age group, yearsb
27–42	84	0.23	0.22	0.95 (0.90 to 1.01)
43–48	102	0.42	0.32	0.77 (0.74 to 0.81)
49–52	75	0.38	0.29	0.77 (0.73 to 0.80)
53–56	98	0.55	0.44	0.79 (0.77 to 0.83)
57–68	90	0.66	0.51	0.78 (0.75 to 0.80)
Mean PIPs per 100 prescriptions 1-year pre-interventionb
0.5–1.6	89	0.24	0.22	0.91 (0.86 to 0.96)
1.7–2.0	90	0.39	0.33	0.85 (0.81 to 0.89)
2.1 –2.3	90	0.49	0.38	0.77 (0.74 to 0.80)
2.4–2.9	90	0.57	0.44	0.77 (0.74 to 0.80)
3.0–6.3	90	0.55	0.43	0.77 (0.74 to 0.80)
Prescriptions per patient in  the 1 year pre-intervention periodb
1.9–8.6	89	0.21	0.20	0.95 (0.90 to 1.01)
8.7–10.7	90	0.35	0.28	0.79 (0.76 to 0.83)
10.8–12.8	90	0.44	0.35	0.79 (0.76 to 0.83)
12.9–14.9	90	0.56	0.44	0.80 (0.77 to 0.83)
15.0–32.6	90	0.68	0.51	0.76 (0.73 to 0.79)

^aRelative risk indicates how large the prevalence of PIPs at post-intervention was relative to the prevalence at pre-intervention. Relative risk estimates <1 indicate a reduction of PIPs. ^bDivided into quintiles by number of GPs. PIP = potentially inappropriate prescription. RR = relative risk.