| Methods |
RCT. Cluster randomization by date of discharge from postpartum care, done with random number generator by separate researcher so clinic staff could not predict day's assignment. Data analyzed by individual, re‐evaluated accounting for cluster sampling, no substantial differences. Researchers conducting baseline interviews not masked to group assignment, blinded personnel conducted follow‐up, data entry, and analysis. 12 mo follow‐up. |
| Participants |
370 post‐partum, low income, racially diverse English‐ or Spanish‐speaking women at public inner‐city hospital in San Francisco. Excluded if major contraindications to estrogen use, post‐partum tubal ligation or partner with vasectomy, employees of Labor and Delivery at the hospital, enrolled in another study, or difficult to reach for follow‐up (lack of a phone, psychiatric disorder, untreated substance abuse, plans for relocation). |
| Interventions |
Intervention group received one course of Yuzpe regimen (eight tabs 0.15 mg levonorgestrel plus 30 µg ethinyl estradiol), educational session, verbal and written instructions. Additional pills available on request. Control group received routine counseling, which may or may not have included a discussion of EC. |
| Outcomes |
Pregnancy rates, use of EC, frequency of unprotected intercourse, change in contraceptive methods, EC knowledge. Except for pregnancy rates, most outcomes can only be included for six‐month follow‐up data, as they were reported separately for the six months prior to the six and twelve‐month follow‐up visits. |
| Notes |
Large loss to follow‐up (31% at 12 months). All self‐reported data. Powered to detect difference in EC use. |
| Risk of bias |
| Bias |
Authors' judgement |
Support for judgement |
| Allocation concealment (selection bias) |
High risk |
|
