Table 4.

Multivariable-adjusted linear regression for change in maximum prescription count

	Coefficient (95% CI), p-value
Age group (years)
65-74	reference
75-84	0.04 (0.03, 0.05), p<0.001
85-94	0.04 (0.03, 0.06), p<0.001
95+	-0.03 (-0.07, 0.02), p=0.270
Sex
Male	reference
Female	-0.01 (-0.02, -0.01), p=0.002
Original prescription count
1	0.30 (0.28, 0.32), p<0.001
2-4	0.14 (0.13, 0.15), p<0.001
5-9	reference
10-14	-0.21 (-0.22, -0.19), p<0.001
15-19	-0.42 (-0.47, -0.38), p<0.001
20-24	-0.55 (-0.69, -0.40), p<0.001
Practice region
Scotland	reference
Wales	0.05 (0.03, 0.07), p<0.001
Northern Ireland	0.12 (0.09, 0.14), p<0.001
London	0.07 (0.00, 0.13), p=0.058
Rest of England	0.02 (0.00, 0.05), p=0.031
Consultation type
Face-to-face	reference
Telephone	-0.01 (-0.04, 0.03), p=0.674
Other	-0.02 (-0.04, -0.01), p<0.001
Missing	0.04 (-0.15, 0.22), p=0.715
Staff role
General practitioner	reference
Pharmacist	-0.05 (-0.07, -0.03), p<0.001
Nurse	0.00 (-0.02, 0.02), p=0.863
Other	-0.04 (-0.06, -0.01), p=0.008
Admin	0.02 (-0.01, 0.06), p=0.207
Missing	0.01 (-0.10, 0.12), p=0.804
Townsend Quintile
Quintile 1 (least deprived)	reference
Quintile 2	0.02 (-0.01, 0.06), p=0.208
Quintile 3	0.01 (-0.02, 0.04), p=0.380
Quintile 4	0.02 (-0.02, 0.05), p=0.306
Quintile 5 (most deprived)	0.03 (0.00, 0.07), p=0.056
Intercept	0.06 (0.02, 0.09), p=0.001

The outcome is change in maximum prescription count in the three months after vs three months before a medication review. Negative coefficients indicate a decrease in prescription count after the review, and vice versa. The model is adjusted for the factors shown in the table. CI confidence interval