Augmentation of Mortality Data Using Disparate Real-World Data Sources: Does More Data = Better Validity?

OBJECTIVES: The ability to include mortality as an outcome in pharmacoepidemiology analyses is key for many therapeutic areas. Identifying reliable sources of death information however can be challenging. This analysis evaluates the incremental value in using disparate real-world data (RWD) to accurately identify the occurrence and date of death.

METHODS: A random sample of 6,000 adults was selected from the MarketScan® by Merative^TM Commercial and Medicare Database (MSN). Death information was obtained from employer data feeds and through a linkage to the Social Security Administration’s Death Master File (DMF). Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were analyzed (%, 95% confidence interval (CI)) using the Centers for Disease Control and Prevention’s National Death Index (NDI) as the ‘gold standard’. Date concordance, based on NDI reported death date, was also evaluated.

RESULTS: Overall, 1,504 and 1,150 deaths were identified in the NDI and MSN+DMF respectively, during either 2019 or 2022. Augmentation of the employer reported death data with DMF increased sensitivity (40.56% to 71.34%) and NPV (83.28% to 91.11%). Specificity and PPV remained relatively unchanged with the addition of DMF data (99.07% to 98.29% and 93.56% to 93.30%, respectively). Concordance around reported date of death remained high with the addition of DMF data, with 97.67% of MSN+DMF death dates +/- 1 day of the NDI reported date.

CONCLUSIONS: Disparate sources of death data often capture deaths occurring in different settings and can be complimentary when trying to capture a comprehensive view of mortality. In this analysis, the integration of multiple sources had a positive impact on the validation metrics for both the presence of death and specific date of death. The ability to deterministically augment real-world data to improve death ascertainment for RWD generation is an important capability.