Keder 1998.

Methods	Single site, parallel group RCT Location: Magee‐Womens Hospital clinic in Pittsburgh, PA, USA Time frame: enrolment October 1993 to November 1994 Sample size calculation (and outcome of focus): based on published meta‐analysis of interventions to improve adherence for medical appointments, estimated OR 2.2 for improved follow‐up; assume 50% no‐show, type I error 0.05, and power 90%, 120 needed for each group
Participants	General with N: 250 women attending clinic and willing to use DMPA for contraception. Inclusion criteria: access to telephone; plan to return to same facility for continuing care. Exclusion criteria: currently receiving depot medroxyprogesterone treatment; planning to undergo tubal ligation; immediately postpartum; unwilling to provide back‐up contact information.
Interventions	Intervention: Mail and telephone reminders of next injection 2 weeks before each scheduled injection. Comparison: Written appointment card.
Outcomes	Discontinuation rate at 12 months, rate of on‐time injections (an injection received within 14 weeks of the prior injection was classified as 'on‐time' injection), reason for discontinuation, DMPA‐induced side effects Follow‐up: 12 months
Notes	Trial registration: no details provided Funding: no details Declarations of interest: no details provided
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomization list
Allocation concealment (selection bias)	Unclear risk	No information
Blinding of participants and personnel (performance bias) All outcomes	High risk	Women enrolled were told of their assignment; researchers not blinded
Blinding of outcome assessment (detection bias) All outcomes	High risk	Investigator communicated that she did the analysis and was aware of group assignment
Incomplete outcome data (attrition bias) All outcomes	High risk	Loss to follow‐up: according to data provided by researcher, 36% overall; reminder group, 39%; no reminder group, 33% Report noted 17.6% loss for year‐end interview.
Selective reporting (reporting bias)	Unclear risk	No protocol was available, however, outcomes prespecified in the manuscript were reported on.
Other bias	Low risk	Appeared free of other sources of bias