Methods for Using Data Abstracted from Medical Charts to Impute Longitudinal Missing Data in a Clinical Trial

摘要: Abstract Objective To describe a method for imputing missing follow-up blood pressure data in clinical hypertension trial using pressures abstracted from medical charts. Methods We tested two-step method. In the first, longitudinal mixed-effects model was estimated on second, patient-specific fitted values this at were used to impute trial. Simulations that imposed alternative mechanisms observed compare approach imputation approaches do not incorporate Results For are random, incorporating chart-based models leads estimates of trial-based unbiased and have lower mean squared deviation than imputed without data. ameliorates but does eliminate inherent bias. Conclusions Incorporating chart into an algorithm via use is efficient way randomized