Zhu 2009.

Methods	Design: cluster RCT Location: Beijing, Shanghai, and Zhengzhou (China); 8 abortion clinics (hospitals) per city Time frame: May to November 2006 Sample size estimation and outcome of focus: based on simulations for varying rates of contraceptive use (0%, 5%, or 10%), needed ≥ 60 women; 80% power when effect size > 10% in contraceptive use. For 25% loss to follow‐up, needed 1800 participants; 900 each arm Cluster randomized trial: 24 hospitals (8 per city) matched in pairs (characteristics of clinics, mainly abortion in 2005; "no substantial difference" in characteristics of women); randomly assigned to 1 of 2 intervention packages 8 pairs of hospitals in analysis: 2 Beijing; 4 Shanghai; 2 Zhengzhou 5 hospitals did not follow randomization protocol; 4 pairs of hospitals ineligible for analysis
Participants	General with N: 16 hospitals; 2336 women who requested abortion in 2‐month period Inclusion criteria: < 25 years old; first trimester of pregnancy; seeking abortion Exclusion criteria: no information
Interventions	Study focus: improve contraceptive use and reduce repeat abortion rate Theory or model: none apparent; interventions adapted for cultural and socioeconomic appropriateness Provider training for both groups; service guidelines and training module Intervention (special or comprehensive): provider training 2 days; group education for women; individual counseling including information on contraceptive methods and recommendation of most suitable methods; free provision of contraceptives (condoms, OCs, IUD, implant); involvement of male partner; and referral to existing family planning services. Comparison (standard or essential): provider training 1 day; group education for women; and referral to existing family planning services.
Outcomes	Primary (questionnaire): contraceptive use; use of more effective contraceptive methods (condoms, OCs, IUDs and implants); regular intake of OCs among OC users; consistent and correct use of condoms among condom users; pregnancy; repeat induced abortion during follow‐up plus unwanted pregnancies (counted once) Secondary: knowledge of contraception (13 items); postabortion family planning services received; patient satisfaction with clinic services Follow‐up: 6 months (mostly by phone)
Notes	Data collected before randomization of hospitals and after intervention implementation. Analysis: Investigators used conditional logistic regression to account for hospital matching and robust variance estimates to account for baseline differences. Final models adjusted for potential confounding factors showed significant effects. Investigators estimated relative effects using an interaction term to account for baseline differences.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization via coin tossing by "neutral person" not involved in study at 1 research center 5 hospitals did not follow randomization protocol, therefore 4 pairs of hospitals ineligible for analysis 8 pairs of hospitals in analysis: Beijing 2; Shanghai 4; Zhengzhou 2
Allocation concealment (selection bias)	Unclear risk	Cluster randomization with matching of hospitals as above; no further information
Blinding of participants and personnel (performance bias) All outcomes	High risk	Presume no blinding of participant or provider; not feasible due to type of intervention
Blinding of outcome assessment (detection bias) All outcomes	High risk	Interviewers not blinded to intervention group
Incomplete outcome data (attrition bias) All outcomes	High risk	Loss to follow‐up: special intervention 44% (473/1065); comparison 40% (372/927)

DMPA: depot medroxyprogesterone acetate
 IUD: intrauterine device
 LARC: long‐acting reversible contraception
 OC: oral contraceptive
 OR: odds ratio
 RCT: randomized controlled trial
 STI: sexually transmitted infection