An Estimated Likelihood Method for Continuous Outcome Regression Models With Outcome-Dependent Sampling

摘要: Many biomedical observational studies attempt to relate a continuous outcome an environmental exposure and other important covariates. If the is easier or cheaper measure relative of interest, then may be observed for every member finite-study population, whereas measurements obtained only relatively small subsample this population. Rather than selecting simple random individuals measurement, investigators enhance study efficiency by allowing selection probabilities depend on outcomes; we refer such sampling schemes as outcome-dependent (ODS). Standard estimation methods that ignore ODS design will yield biased inconsistent parameter estimates. Furthermore, it generally desirable use estimators incorporate all available data analyses restricted subjects with complete information are inefficient. To end, extend estimated likel...