WHO/HIV/2013.69 © World Health Organization 2013
This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS 1
GRADE table: Electronic reminders for promoting adherence to ART among people living with HIV – Randomized Controlled Trials
Author(s): Horvath T Date: 2012-10-08
Question: Should mobile phone text messages be used for promoting adherence to antiretroviral therapy in patients with HIV infection?
Settings: Brazil, Cameroon, Guatemala, Kenya, Peru
Bibliography: Lester 2010, Pop-Eleches 2011, Curioso 2012, da Costa 2012, Ikeda 2012, Mbuagbaw 2012
Quality assessment No. of patients Effect
Quality Importance
No. of
studies Design Risk of
bias Inconsistency Indirectness Imprecision Other considerations
Mobile phone
text messages Control Relative
(95% CI) Absolute
Viral load suppression at 52 weeks: any text messages versus standard care
2 randomized
trials
no serious risk of bias
no serious inconsistency
no serious indirectness
no serious imprecision
none 246/391
(62.9%)
165/373 (44.2%)
RR 0.67 (0.57 to 0.78)
146 fewer per 1000 (from 97 fewer to 190
fewer)
⊕⊕⊕⊕
HIGH
CRITICAL
Viral load suppression at 52 weeks: weekly text messages versus standard care
1 randomized
trials
no serious risk of bias
no serious inconsistency
no serious indirectness
no serious imprecision
none 156/273
(57.1%)
128/265 (48.3%)
RR 0.83 (0.69 to 0.99)
82 fewer per 1000 (from 5 fewer to 150 fewer)
⊕⊕⊕⊕
HIGH
CRITICAL
Viral load suppression at 52 weeks: daily text messages versus standard care
1 randomized
trials
no serious risk of bias
no serious inconsistency
no serious indirectness
serious1 none 90/118
(76.3%)
37/108 (34.3%)
RR 0.36 (0.25 to 0.51)
219 fewer per 1000 (from 168 fewer to 257
fewer)
⊕⊕⊕O MODERATE
CRITICAL
Viral load suppression at 30 weeks: daily text messages versus standard care
1 randomized trials
no serious risk of bias
no serious inconsistency
no serious indirectness
serious1 none 74/118
(62.7%)
32/108 (29.6%)
RR 0.53 (0.41 to 0.69)
139 fewer per 1000 (from 92 fewer to 175
fewer)
⊕⊕⊕O MODERATE
CRITICAL
ART adherence at 48–52 weeks: any text messages versus standard care (overall)
WHO/HIV/2013.69 © World Health Organization 2013
This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS 2
3 randomized
trials
no serious risk of bias
no serious inconsistency
no serious indirectness
no serious imprecision
none 340/649
(52.4%)
201/491 (40.9%)
RR 0.79 (0.71 to 0.88)
86 fewer per 1000 (from 49 fewer to 119 fewer)
⊕⊕⊕⊕
HIGH
CRITICAL
ART adherence at 48–52 weeks: weekly text messages versus standard care (overall)
2 randomized
trials
no serious risk of bias
no serious inconsistency
no serious indirectness
no serious imprecision
none 285/494
(57.7%)
188/404 (46.5%)
RR 0.78 (0.68 to 0.89)
102 fewer per 1000 (from 51 fewer to 149
fewer)
⊕⊕⊕⊕
HIGH
CRITICAL
ART adherence at 48–52 weeks: short weekly messages versus standard care
2 randomized trials
no serious risk of bias
no serious inconsistency
no serious indirectness
no serious imprecision
none 207/346
(59.8%)
188/404 (46.5%)
RR 0.77 (0.66 to 0.9)
107 fewer per 1000 (from 47 fewer to 158
fewer)
⊕⊕⊕⊕
HIGH
CRITICAL
ART adherence at 48 weeks: long weekly messages versus standard care
1 randomized
trials
no serious risk of bias
no serious inconsistency
no serious indirectness
very serious2 none 39/74
(52.7%)
56/139 (40.3%)
RR 0.79 (0.6 to 1.04)
85 fewer per 1000 (from 161 fewer to 16 more)
⊕⊕OO LOW
CRITICAL
ART adherence at 48 weeks: short daily messages versus standard care
1 randomized
trials
no serious risk of bias
no serious inconsistency
no serious indirectness
very serious2 none 28/70
(40%)
56/139 (40.3%)
RR 1 (0.79 to 1.27)
0 fewer per 1000 (from 85 fewer to 109 more)
⊕⊕OO LOW
CRITICAL
ART adherence at 48 weeks: long daily messages versus standard care
1 randomized
trials
no serious risk of bias
no serious inconsistency
no serious indirectness
very serious2 none 30/72
(41.7%)
56/139 (40.3%)
RR 0.98 (0.77 to 1.24)
8 fewer per 1000 (from 93 fewer to 97 more)
⊕⊕OO LOW
CRITICAL
ART adherence at 26 weeks: short weekly text messages versus standard care
1 randomized
trials
no serious risk of bias
no serious inconsistency
no serious indirectness
serious1 none 80/101
(79.2%)
78/99 (78.8%)
RR 0.98 (0.57 to 1.68)
16 fewer per 1000 (from 339 fewer to 536 more)
⊕⊕⊕O MODERATE
CRITICAL
ART adherence at 52 weeks: flexible number and interval of messages versus standard care
WHO/HIV/2013.69 © World Health Organization 2013
This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS 3
1 randomized
trials
no serious risk of bias
no serious inconsistency
serious3 very serious2 none 36/87
(41.4%)
13/87 (14.9%)
RR 0.69 (0.57 to 0.84)
46 fewer per 1000 (from 24 fewer to 64 fewer)
⊕OOO VERY LOW
CRITICAL
ART adherence at 16 weeks: text messages five times per week versus standard care
1 randomized
trials
no serious risk of bias
no serious inconsistency
no serious indirectness
very serious2 none 6/8
(75%)
6/13 (46.2%)
RR 0.46 (0.13 to 1.71)
249 fewer per 1000 (from 402 fewer to 328
more)
⊕⊕OO LOW
CRITICAL
Pharmacy refill data at 26 weeks: short weekly messages versus standard care (better indicated by lower values)
1 randomized trials
no serious risk of bias
no serious inconsistency
no serious indirectness
serious1 none 101 99 - Mean difference 0.1
higher (0.29 lower to 0.49 higher)
⊕⊕⊕O MODERATE
IMPORTANT
1 Few events.
2 Very few events.
3 Self-reported adherence.