Sieving 2012.

Methods	Design: pilot project to refine intervention protocols and determine preliminary efficacy. Phase 1, random assignment; Phase 2, purposive assignment. Location: Midwestern metropolitan area (USA). Enrollment: Phase I, Sept 1999 to May 2001; Phase II, Oct 2001 to Dec 2002. Sample size calculation (and outcome of focus): not specified
Participants	General with N: 128 girls, sexually active. Source: 3 clinics in a Midwestern metropolitan area Inclusion criteria: clinic visit involving negative pregnancy test, post‐abortion exam, or emergency contraception; age 13 to 14 years; elevated scores on 7‐item screening tool indicating high‐risk sexual behaviors. Exclusion criteria: did not understand consent materials; married, pregnant or had given birth.
Interventions	Study focus: reduce sexual risk behaviors among girls at high risk for early pregnancy Theory or model: social cognitive theory, resilience paradigm, and research on antecedents of sexual risk behaviors among adolescent girls. Treatment: clinic‐based, youth development program provided case management with standardized topic guidelines plus peer leadership program (peer education, which included contraceptive use skills, and service learning) Comparison or control: not specified; presumably received usual clinic services Duration: 18 months
Outcomes	Consistency of dual‐method use (hormonal + condoms) as well as of condom use and of hormonal contraceptive use. Consistency = # months (of past 6) reportedly used method every time had sex with recent partner / # months had sex with that partner; scored on continuum (0 to 1).
Notes	Pilot project for Sieving 2013. Desire to use birth control and beliefs supporting birth control use were not specific to dual‐method use.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Phase 1: Clinics 1 and 2, participants were randomly assigned. Phase 2: Clinic 3, participants were assigned to intervention for 10 months (for sample size to test expanded intervention); for remainder of recruitment, participants were assigned to control.
Allocation concealment (selection bias)	High risk	No mention; presume none since main study (Sieving 2013) did not use any.
Blinding of participants and personnel (performance bias) All outcomes	High risk	None, according to investigator
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Self‐report survey at baseline and at 12 and 18 months; paper‐pencil survey, according to investigator. Investigator communicated that separate groups of research staff were responsible for intervention and evaluation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Intervention group: outcome data for 89% at 12 months and 83% at 18 months. Control group: outcome data 82% at 12 and 18 months.
Selective reporting (reporting bias)	Low risk	Not apparent