Validation of a Composite Real-World Mortality Variable Among Patients With Hematologic Malignancies Treated in the United States.

Purpose: Accurate survival data are critical for high-quality outcomes research. It has been documented that mortality data capture in the real-world setting may be prone to missingness. Our study sought to evaluate the validity of a composite real-world mortality variable compared with the National Death Index (NDI) as the gold standard.

Methods: This was a retrospective, observational research study of patients with hematologic malignancies treated in the United States. Adult patients diagnosed with one of the following cancers between January 1, 2012, and December 31, 2020, were included: AML, chronic lymphocytic leukemia, diffuse large B-cell lymphoma, follicular lymphoma, marginal zone lymphoma, multiple myeloma, and myelodysplastic syndrome. Validation metrics (sensitivity, specificity, positive predictive value [PPV], and negative predictive value [NPV]) and date concordance (exact, ±7, 15, and 30 days) were assessed.

Results: The final study population included N = 21,565 patients across seven cancer types. Validation metrics showed high sensitivity (87.8%), specificity (95.7%), PPV (90.9%), and NPV (94.1%) when comparing the composite real-world mortality variable with the NDI. Exact date concordance was observed in 88.0% of patients, and concordance rates for 7-, 15-, and 30-day intervals were 93.1%, 93.8%, and 94.3%, respectively.

Conclusion: Our study found that a composite mortality variable leveraging multiple data sources yields high validity when compared against the gold-standard NDI. Given evidence highlighting the challenges of mortality data documentation in the real-world setting, the use of a composite mortality variable can provide significant benefits in quality of documentation and research results.