Skip to main content
. 2021 May 26;16(5):e0252161. doi: 10.1371/journal.pone.0252161

Table 3. Subgroup analysis of included studies.

Study Characteristics Subgroup analysis Meta Regression
Studies (n) Pooled estimate (%) (95% CI) I2 (%) Mean Difference P-Value
Age groups
 Paediatric 2 1.01 (0.29, 1.73) 71.8 refa ref
 Adults (excl. elderly) 12 13.60 (10.79, 16.42) 99.7 14.56 (-6.60, 35.80) 0.170
 Elderly 10 28.43 (18.65, 38.21) 99.2 27.71 (5.86, 49.56) 0.015
 All age groups 9 4.53 (4.04, 5.03) 99.6 5.16 (-16.63, 26.95) 0.632
Setting
 General practice 10 5.57 (4.95, 6.19) 99.8 4.96 (-10.19, 20.11) 0.508
 General internal medicine 7 19.33 (10.83, 27.84) 98.6 18.29 (-0.60, 37.18) 0.057
 Community pharmacy 5 6.93 (4.10, 9.75) 98.3 6.19 (-13.45, 25.84) 0.523
 Home setting 4 23.34 (17.60, 29.08) 89.6 23.55 (3.01, 44.07) 0.026
 Long-term care facilitiesb 4 42.22 (17.57, 66.88) 98.9 39.75 (21.94, 57.55) 0.000
 Paediatric practice 2 1.01 (0.29, 1.73) 71.8 Ref ref
 Multidisciplinary 1 - - - -
Methods
 Medical record/notes/medication review and patient survey 16 19.92 (16.11, 23.73) 99.1 23.69 (6.49, 40.88) 0.009
 Medical record review 3 14.90 (0.29, 29.50) 98.3 13.81 (0.66, 26.96) 0.040
 Trigger-based medical record review 3 14.32 (3.80, 24.84) 97.9 10.69 (-7.94, 29.33) 0.249
 Spontaneous/solicited reporting 5 4.74 (0.66, 8.82) 98.3 1.06 (-14.99, 17.11) 0.893
 Administrative database screening 5 3.19 (2.53, 3.85) 99.9 Ref ref
 Combined 1 - - - -
Definitions
 WHO 9 3.38 (2.81, 3.96) 99.5 Ref ref
 Bates et al 6 13.05 (9.37, 16.73) 98.3 -9.92 (-26.98, 7.15) 0.245
 NA/Otherc 18 11.42 (9.23, 13.60) 99.7 -10.86 (-24.10, 2.37) 0.104
Study Quality
 Low risk of bias 12 20.37 (16.89, 23.85) 99.5 9.47 (-2.09, 21.02) 0.105
 High risk of bias 21 6.40 (5.89, 6.91) 99.6 Ref ref
Sample Size
 ≤1000 19 22.49 (18.58, 26.39) 99.3 20.57 (6.35, 34.79) 0.006
 1001–10,000 9 7.67 (4.85, 10.50) 99.2 5.11 (-10.25, 20.47) 0.502
 >10,000 5 2.96 (2.28, 3.64) 99.9 Ref ref

aReference value

bInclude one studies combining nursing home and home nursing care

cInclude three studies that used definition by Parry et al and Britt et al, the remaining studies did not specify the definition used.