Many lifestyle intervention trials depend on collecting self-reported outcomes, such as dietary intake, in order to assess the intervention’s effectiveness. Self-reported outcome measures may be subject to measurement error, which could impact estimation of the treatment effect. Methods have been developed to correct for measurement error by using external validation studies, which measure both the self-reported outcome and an accompanying biomarker, to model and account for the measurement error structure. Most validation data, though, are only relevant for the outcome under control conditions. Statistical methods have been developed to use external validation data to correct for the error under control, and then conduct sensitivity analyses around the error of the outcome under treatment to obtain estimates of the corrected average treatment effect. However, an assumption underlying this approach is that the measurement error structure of the outcome is the same in the validation sample as it is in the intervention trial, and that the error correction is therefore transportable to the trial. This may not always be a valid assumption to make.
As part of my dissertation, I propose an approach that adjusts the validation sample to more closely resemble the trial sample and thus leads to more transportable measurement error corrections. I also formally investigate scenarios where bias due to poor transportability may arise.