| Methods |
RCT (by colored condom chosen from age‐stratified bucket). Correspondence with primary author indicated that participants could not see inside bucket before choosing and were unaware of the color assignments. Two colors, distributed 50:50, inside each bucket. Most clinic staff unlikely to have been able to decipher the color code, method unlikely to have affected randomization. Six mo follow‐up. |
| Participants |
301 sexually‐active adolescents, aged 15‐20 yrs, in Southwestern Pennsylvania, primarily minority and low‐income. Excluded if using IUD, implant, injectable, if living in foster care or group home, or if had other characteristics which could threaten follow‐up. |
| Interventions |
From study start until April 2000, intervention group received one course Yuzpe regimen 200 mcg ethinyl estradiol plus 2 mg norgestrel, plus an extra dose in case of vomiting, in addition to diphenhydramine. After April 2000, when levonorgestrel only regimens became standard of care, a levonorgestrel‐only regimen was used (two tabs of levonorgestrel 0.75 mg). Participants could obtain two additional courses over six mo period by request, regardless of whether unprotected intercourse had occurred. Participants also received counseling and EC info. Control group received EC on request at the clinic and EC info. |
| Outcomes |
Pregnancy and STI rates (specific STIs not specified), use of EC, interval between unprotected intercourse and EC use, frequency of unprotected intercourse, condom use. |
| Notes |
Large loss to follow‐up (26% at six mo ‐ for reasons other than pregnancy), and loss to follow‐up differential by treatment group (33% in advance provision group, 19% in control group). Not powered to detect differences in pregnancy or STI rates. |
| Risk of bias |
| Bias |
Authors' judgement |
Support for judgement |
| Allocation concealment (selection bias) |
Low risk |
|
