Table 1.
Meta-analysis plan by type of community-based condom distribution intervention, outcome, population.
Fig 1.
Searching and screening of scientific records for systematic review of community-based condom distribution interventions in the United States (search January 1, 1986 to April 17, 2017).
*Databases searched: SCOPUS, PubMed, PsycINFO, Cochrane Central Register of Controlled Trials. †We separately analyzed and reported this group of studies. They initially met our broad inclusion criteria, but were implemented at the individual context level and were limited in terms of frequency and/or duration of access to condoms (e.g., participants could take as many condoms as they wanted, but only at motivational sessions or when they made contact with a street outreach worker). See S5 File for details. ‡Given the differences between youth populations in school and high-risk adult populations in community settings, we decided while screening of studies was in progress to disseminate findings of school-based studies separately.
Table 2.
Characteristics of community-based condom distribution programs in the United States included in systematic review, by intervention categorya.
Table 3.
Summary of evidence for the effectiveness of community-based condom distribution interventions by intervention type, outcome, and population type in the United Statesa.
Fig 2.
Pooled effect measures and risk of bias for the effect of “Ongoing” community-based condom distribution interventions (compared to no condom distribution) for sexual risk behaviors in the United States.
Legend: CI, Confidence interval; IV, Inverse variance. Size of red square on the forest plots represents IV weights. See S6 for further details on risk ratio calculations. Risk of bias legend: (A) Incomplete outcome data (attrition bias); (B) Selective reporting (reporting bias); (C) Other bias; (D) Failure to develop and apply appropriate eligibility criteria; (E) Flawed measurement of exposure and/or outcome; (F) Failure to control for confounders; (G) Too-short or incomplete length of follow-up;— = high risk of bias; + = low risk of bias;? = unclear risk of bias.
Fig 3.
Pooled effect measures and risk of bias for the effect of “Ongoing-plus” community-based condom distribution interventions (compared to no condom distribution) for sexual risk behaviors in the United States.
Legend: CI, Confidence interval; IV, Inverse variance. Size of red square on the forest plots represents IV weights. See S6 for further details on risk ratio calculations. Risk of bias legend: (A) Incomplete outcome data (attrition bias); (B) Selective reporting (reporting bias); (C) Other bias; (D) Failure to develop and apply appropriate eligibility criteria; (E) Flawed measurement of exposure and/or outcome; (F) Failure to control for confounders; (G) Too-short or incomplete length of follow-up;— = high risk of bias; + = low risk of bias;? = unclear risk of bias.