N=patients with outcomes recorded plus withdrawn patients | Number of adherent patients/total responses N (%), p | Model* 1 OR (95% CI, p) | Model* 2 (Adjusted) OR (95% CI, p) | Model* 3 (Imputation) OR (95% CI, p) |
---|---|---|---|---|
Adherence NMS (N=378, 126 responses missing) | ||||
Normal practice | 115/190 (60.5) | 1.00 | 1.00 | 1.00 |
NMS | 133/188 (70.7), 0.037 | 1.58 (1.03 to 2.42, 0.037) | 1.67 (1.06 to 2.62, 0.027) | 1.62 (1.04 to 2.53, 0.032) |
Composite NMS (N=443, 61 responses missing)† (adherent+stopped+changed) | ||||
Normal practice | 144/222 (64.9) | 1.00 | 1.00 | 1.00 |
NMS | 165/221 (74.7), 0.025 | 1.60 (1.06 to 2.40, 0.025) | 1.68 (1.09 to 2.58, 0.018) | 1.64 (1.08 to 2.50, 0.021) |
Adherence MMAS-8 (N=267, 237 responses missing) | ||||
Normal practice | 85/143 (59.4) | 1.00 | 1.00 | 1.00 |
NMS | 89/124 (71.8), 0.035 | 1.74 (1.04 to 2.90, 0.036) | 1.88 (1.06 to 3.34, 0.030) | 1.77 (0.96 to 3.28, 0.068) |
Composite MMAS-8 (N=321, 183 responses missing)† (adherent+stopped+changed) | ||||
Normal practice | 108/167 (64.7) | 1.00 | 1.00 | 1.00 |
NMS | 116/154 (75.3), 0.038 | 1.67 (1.03 to 2.71, 0.039) | 1.78 (1.06 to 3.00, 0.029) | 1.81 (1.07 to 3.05, 0.027) |
*Model 1: Simple logistic regression model; Model 2 (Adjusted): Multilevel logistic regression model adjusted for recruiting pharmacy, disease, age, sex and medication count; Model 3 (Imputation): Adjusted logistic regression model incorporating imputation of missing data.
†The difference in numbers of patients with composite and simple adherence outcome is larger for the NMS adherence question, when compared with MMAS questionnaire (65 and 54, respectively). This is because, at 10 weeks, more patients whose medicine was changed responded to the NMS adherence question, compared with the MMAS questionnaire (34 and 23, respectively; 11 patients more for the NMS question).
MMAS-8, Morisky's Medication Adherence Scale 8-item version; N, number; NMS, New Medicine Service.