Are we missing the importance of missing values in HIV prevention randomized clinical trials? Review and recommendations.

摘要: Missing data in HIV prevention trials is a common complication to interpreting outcomes. Even small proportion of missing values randomized can cause bias, inefficiency and loss power. We examined the extent methods which clinical (RCT) have managed values. used database maintained by HIV/AIDS Prevention Research Synthesis (PRS) Project at Centers for Disease Control (CDC) identify related our review. The PRS cumulative was searched on June 15, 2010 all citations that met following criteria were retrieved: All RCTs reported HIV/STD/HBV/HCV behavioral interventions with biological outcome from 2005 present. Out 57 intervention identified, had some level found average per study ranged between 3 97%. Averaging over studies percent 26%. None any assumptions managing their RCTs. Under relaxed discussed below, we expect only 12% report unbiased results. There need more detailed thoughtful consideration problem trials. In current state risk major biases interpretations. Several viable alternatives are available improving internal validity data.