Skip to main content
. 2017 Jan 20;7(1):e010568. doi: 10.1136/bmjopen-2015-010568

Table 7.

Risk factors for possible hospital-acquired suspected ADRs in 194 of 762 inpatients, Uganda, 2014

Possible ADR, n (%)
Crude analysis
Adjusted analysis
Characteristics Yes No OR 95% CI for OR p Value β-adj OR 95% CI for OR p Value
Gender
 Male 51 (22) 177 (78) 1.0 1.01
 Female 143 (27) 391 (73) 1.3 0.88 to 1.83 0.201 0.011 1.00 0.67 to 1.53 0.959
Age: mean (SD) 33.5 (13.6) 35.3 (15.1) 1.0 0.98 to 1.00 0.151 −0.0006 1.00 0.98 to 1.01 0.933
Number of conventional medicines
 Five or less 39 (16) 205 (84) 1.0 1.00
 Six or more 155 (30) 363 (70) 2.2 1.52 to 3.32 <0.001 1.001 2.72 1.79 to 4.13 <0.001
Use of herbal medicines in the 4 weeks prior hospitalisation
 No 127 (23) 428 (77) 1.0 1.00
 Yes 67 (32) 140 (68) 1.6 1.13 to 2.29 0.008 0.519 1.68 1.16 to 2.43 0.006
HIV-positive
 No/unknown 138 (26) 392 (74) 1.0 1.00
 Yes 56 (24) 176 (76) 0.9 0.63 to 1.29 0.580 −0.589 0.55 0.31 to 0.99 0.045
Charlson's comorbidity index
 Two or less 147 (26) 420 (74) 1.0 1.00
 Three or more 47 (24) 148 (76) 0.9 0.62 to 1.32 0.614 −0.613 0.54 0.24 to 1.24 0.146
Interaction: Charlson's index≥3 & HIV+status
 No interaction 157 (25) 472 (75) 1.0 1.00
 Interaction 37 (28) 96 (72) 1.2 0.76 to 1.76 0.492 1.235 3.44 1.22 to 9.71 0.020
Hospitalisation in past 3 months
 No 125 (24) 407 (76) 1.0 1.00
 Yes 69 (30) 161 (70) 1.4 0.99 to 1.97 0.059 0.448 1.57 1.09 to 2.26 0.016
Gynaecological ward
 No 128 (22) 443 (78) 1.0 1.00
 Yes 66 (35) 125 (65) 1.8 1.28 to 2.61 0.001 0.771 2.16 1.36 to 3.44 0.001
Intercept −2.19 0.11 0.06 to 0.23

Except for gender (male coded 1 and female coded 2), indicator variables were coded 0 for baseline, 1 otherwise; regression χ2=53.99 on 9 df (p<0.001); β-adj=adjusted regression coefficient.