Skip to main content
. 2012 Apr 5;13:30. doi: 10.1186/1471-2296-13-30

Table 4.

Subgroup analysis of the intervention effect on the percentage of days adherence to prescribed medication

Sub-group variable Sub-group category (L = Low or H = High) N per sub group Intervention arm Standard care arm Intervention- Control (SE)* Difference (H-L) (95% C.I.) P interaction
HbA1c % (mmol/mol)
7.5 – 7.9 (58-63)
85
80.4 (24.5)
62.6 (34.2)
17.7 (6.4)
 
 
 
8.0 - 12.4 (64-112)
108
74.8 (27.7)
72.8 (27.8)
2.0 (5.3)
-15.8 (-32.9, 1.2)
0.07
Age (years)
37 – 64
94
75.0 (25.61)
72.3 (27.3)
2.7 (5.6)
 
 
 
65 and over
100
80.0 (27.1)
66.4 (33.2)
13.5 (6.0)
10.8 (-5.4, 27.2)
0.19
Total number of medications
0 – 5
89
80.2 (23.3)
73.0 (27.0)
7.2 (5.3)
 
 
 
6 and over
105
75.1 (28.5)
65 .2 (33.7)
10.0 (6.1)
2.8 (-13.5, 19.1)
0.74
Self reported adherence at baseline
0-23
102
75.0 (26.4)
70.9 (27.2)
4.1 (5.5)
 
 
 
24 or 25
70
85.9 (21.3)
68.0 (33.9)
17.9 (6.8)
13.8 (-3.2, 30.0)
0.11
Prior randomisation to medication monitoring device
No
102
76.6 (28.7)
71.0 (29.3)
5.6 (5.8)
 
 
 
Yes
92
78.2 (23.7)
66.65 (32.6)
11.6 (5.9)
6.0 (-10.5, 22.7)
0.48
Gender
Male
127
77.0 (28.3)
66.4 (31.5)
10.6 (5.3)
 
 
  Female 67 78.0 (22.9) 74.8 (28.8) 3.3 (6.4) -7.3 (-24.3, 9.6) 0.39

Effect estimates, confidence intervals and p-values from non-parametric bootstrap percentile method (50,000 replications).

* Estimated intervention effects are positive within each subgroup.