5. Evidence quality: Sieving 2012¹.

NOS criteria for cohort studies	Met criterion	Support for judgment
Exposed cohort representativeness	✸	Somewhat representative; recruited from adolescent health clinics (reproductive health clinic in middle‐class suburb and primary care clinic in low‐income urban neighborhoods).
Nonexposed cohort selection	✸	Same source as exposed group
Exposure ascertainment: method used	✸	Project staff kept records of participation per activity; level of participation reported per clinic.
Outcome: evidence not present at study start	‐‐‐	Outcomes adjusted for baseline values.
Comparability of groups: design or analysis	‐‐‐	No adjustment for potential confounding factors. Adjusted for baseline measure of outcome and design variation.
Outcome assessment: method used	‐‐‐	Self‐report survey
Follow‐up length	✸	18 months
Follow‐up adequacy	✸	Re‐surveyed: 86% at 12 months; 83% at 18 months.
Quality of evidence	very low	Study met 5 criteria and did not control for confounding.

¹Pilot project for Sieving 2013