Table 2.

Decrease in the number of older patients prescribed high doses of BDZ + per quarter per physician and decrease in BDZ + prescribed (DDD) to older patients per quarter per physician (Mean ± standard error) over one year

	Control	Letter	Pharmacist	Physician
Decrease in number patients	0.42 ± 0.1 (44%) **	0.30 ± 0.1 (50%)**	0.27 ± 0.2 (43%)*	0.32 ± 0.2 (54%)**
Decrease in BDZ + prescribed	0.11 ± 0.10 (10%)*	0.18 ± 0.1 (16%)**	0.12 ± 0.1 (11%)*	0.13 ± 0.05 (13%)**

DDD Defined daily dose

^* statistically significant using paired t-test and Wilcoxin signed-rank test at < 0.05 level for change within groups, between group differences were not significant using MANCOVA

^** statistically significant using paired t-test and Wilcoxin signed-rank test at < 0.001 level for change within groups, between group differences were not significant using MANCOVA