| Study | Potential confounders controlled for in the study | |||||
| Age | Sex | Race/ethnicity | HIV risk | Insurance status | Case mix | |
| Chu 2010 (14) | Yes | Yes | Yes | Yes | Yes | Yes |
| Gardner 2002 (21) | Yes | NA | No | Yes | Yes | Yes |
| Horberg 2012 (15) | Yes | Yes | Yes | Yes | NA | Yes |
| Landon 2003 (16) | Yes | Yes | Yes | Yes | Yes | Yes |
| Landon 2005 (18) | Yes | Yes | No | No | NA | Yes |
| Page 2003 (17) | NA―same baseline characteristics for patients in different intervention arms | |||||
| Schneider 2008 (19) | Yes | Yes | Yes | No | Yes | Yes |
| Turner 2001 (20) | Yes | NA | Yes | No | NA | Yes |
| Wilson 2005 (22) | Yes | Yes | No | Yes | NA | Yes |