Pop‐Eleches 2011

Methods	Randomized controlled trial
Participants	Four hundred and thirty‐one adult patients who had initiated ART within 3 months at a rural clinic in Kenya. Seven hundred and thirty‐five patients were approached and 720 (97.9%) were enrolled. Analyses are restricted to 431 participants enrolled before 31 January 2008.
Interventions	Participants were randomly assigned to one of four intervention groups or to the control group that would receive no text messages. One‐third of the sample was allocated to the control group, and the remaining two‐thirds of the sample were allocated evenly to each of the four intervention groups. The text messages were provided in English, Dhoulou or Kiswahili (as appropriate). The content of the short message was "This is your reminder." The content of the long message was "This is your reminder. Be strong and courageous. We care about you."
Outcomes	The primary outcome was whether adherence exceeded 90% during each 12‐week period of analysis and the 48‐week study period. Outcomes are broken out by which patients got the short daily, short weekly, long daily, long weekly messages. Overall (all lengths, all timing) data are also provided. Adherence >90% at 48 weeks: (Number of events per arm is back‐calculated from proportional figures provided by the authors. These proportional figures are given below in parentheses.) Any text messages: 136/289 (0.47) Short weekly messages: 39/73 (0.53) Long weekly messages: 39/74 (0.53) Short daily: 28/70 (0.40) Long daily: 30/72 (0.42) Standard care: 56/139 (0.40)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation schedule was prepared in advance of enrollment by the investigators. A sequence of random numbers between 0 and 1 was generated, and four equal intervals between 0 and 2/3 corresponded to the four intervention groups, whereas the value interval from 2/3 to 1 corresponded to the control group.
Allocation concealment (selection bias)	Unclear risk	One‐third of the sample was allocated to the control group, and the remaining two‐thirds of the sample was allocated evenly to each of the four intervention groups. Further detail was not provided.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	No blinding, but outcome and outcome measurement unlikely to be influenced by lack of blinding.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Unclear, but outcome and outcome measurement unlikely to be influenced by lack of blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All patients and outcomes were accounted for.
Selective reporting (reporting bias)	Unclear risk	No evidence of selective reporting. The trial is registered at ClinicalTrials.gov (NCT01058694).
Other bias	Low risk