A Systematic Review of Reporting and Handling of Missing Data in Observational Studies Using the UNOS Database.

Abstract

Background: Missing data can undermine a registry's ability to draw valid inferences by decreasing study power and introducing bias. We evaluated how missing data are reported and addressed in observational studies of adult heart transplantation (HT) recipients using the United Network for Organ Sharing (UNOS) database.

Methods: We conducted a systematic literature search of Medline from January 1^st, 2018 through August 22nd, 2023 and included studies that used the UNOS database to evaluate adult (≥18 years) de novo HT recipients. We collected details on the study population, timeframe, primary endpoint, use of missing data, and whether and what methods were used to handle missing data. Approaches were classified as variable selection, complete case analysis (CCA), missing indicator method, single imputation, or multiple imputation.

Results: 229 studies were included from 2018-2023. 67 (29.3%) studies limited their cohorts to those without missing data for the outcome or key variables and 93 (40.6%) reported missing data in their final cohort. 78 (34.1%) studies reported how they handled missing data in their statistical modeling. Of these, CCA was most used (n=41, 52.6%) followed by multiple imputation (n=22, 28.2%), and other methods (n=15, 19.2%) including a combination of missing indicator method (for binary variables) and single imputation. Thirty-one (13.5%) studies reported removing covariates from their analysis because of missing data. Only four (5.1%) studies reported sensitivity analysis to evaluate different missing data approaches.

Conclusions: Merely a third of the identified UNOS database studies reported how they handled missing data in their analysis, with strategies varying. Although no singular approach to handling missing data exists, methods are available that can improve upon the most used approaches by, for example, reducing bias and increasing statistical power. Though these improvements are contingent on the missing data mechanism and the goals of the analysis. Future best practices should include explicit reporting of missingness, detailed methods, and sensitivity checks to maximize the credibility of UNOS database analyses.