Papers were independently assessed for quality of reporting by two assessors, one of whom was blind to information which could have introduced systematic bias (e.g. date of publication).
Two authors (KHK and JWK) independently assessed the completeness of reporting in each item.
Two authors (DC and JA) independently analysed the quality of reporting by using a qualitative classification according to risk of bias (high, unclear, or low).
During the first assessment, between March 2003 and May 2003 (Pre-STARD evaluation), two reviewers independently evaluated the quality of reporting in the included articles.
We also showed that we have internal validity due to our good inter-rater reliability correlation between two reviewers who independently assessed the quality of reporting, key methodological items and PICOT framing of the research question.
Based on multivariate analysis, older age and higher level of education were independently associated with increased odds of reporting vaccination in 2009/2010 and 2010/2011.
We observed that HIV-specific traumatic stress was independently associated with increased likelihood of reporting at least weekly use as well as binge use of cocaine/crack.
Increases in HbA1c from 1 to 5 years post-diagnosis were independently associated with increased odds of reporting a negative impact of diabetes on QoL.
Higher positive affect was independently associated with increased likelihood of reporting perfect ART adherence during the past month (AOR = 1.79, P < .05).05
Our findings suggest that both lower general anxiety and higher specific worry were independently associated with the likelihood of reporting behavioural change.
