Background: To validate Japanese claims-based disease-identifying algorithms for herpes zoster (HZ), Mycobacterium tuberculosis (MTB), nontuberculous mycobacteria infections (NTM), and Pneumocystis jirovecii pneumonia (PJP).
Methods: VALIDATE-J, a multicenter, cross-sectional, retrospective study, reviewed the administrative claims data and medical records from two Japanese hospitals. Claims-based algorithms were developed by experts to identify HZ, MTB, NTM, and PJP cases among patients treated 2012-2016. Diagnosis was confirmed with three gold standard definitions; positive predictive values (PPVs) were calculated for prevalent (regardless of baseline disease-free period) and incident (preceded by a 12-month disease-free period for the target conditions) cases.
Results: Of patients identified using claims-based algorithms, a random sample of 377 cases was included: HZ (n = 95 [55 incident cases]); MTB (n = 100 [58]); NTM (n = 82 [50]); and PJP (n = 100 [84]). PPVs ranged from 67.4-70.5% (HZ), 67.0-90.0% (MTB), 18.3-63.4% (NTM), and 20.0-45.0% (PJP) for prevalent cases, and 69.1-70.9% (HZ), 58.6-87.9% (MTB), 10.0-56.0% (NTM), and 22.6-51.2% (PJP) for incident cases, across definitions. Adding treatment to the algorithms increased PPVs for HZ, with a small increase observed for prevalent cases of NTM.
Conclusions: VALIDATE-J demonstrated moderate to high PPVs for disease-identifying algorithms for HZ and MTB using Japanese claims data.
Keywords: Herpes zoster; Japan; Nontuberculous mycobacteria infections; Pneumocystis jirovecii pneumonia; Positive predictive value; Tuberculosis.
© 2023. BioMed Central Ltd., part of Springer Nature.