Positive predictive values of nontuberculous mycobacterial diagnosis codes in Denmark: a medical record validation study

Moestrup, Pernille Grand, Barbateskovic, Gabriela, Wejse, Christian, Johansen, Isik Somuncu, Fløe, Andreas, Pedersen, Andreas Arnholdt, Dahl, Victor Naestholt
Mycobacteria · Springer Nature · 2026-11-20 · DOI: 10.1186/s44350-026-00026-9

Аннотация

Background Diagnosis codes are used for disease surveillance and epidemiological research, including nontuberculous mycobacterial (NTM) infections. However, the validity of NTM diagnosis codes in the International Classification of Diseases, Tenth Revision (ICD-10) has not previously been validated. Methods In this retrospective validation study, we reviewed medical records of all adult patients assigned ICD-10 NTM diagnosis codes at Aarhus University Hospital (2019–2023) and Odense University Hospital (2017–2021). We defined positive predictive value (PPV) as the proportion of confirmed NTM cases among all patients assigned at least one diagnosis code. For patients with pulmonary NTM, diagnostic criteria for NTM pulmonary disease were assessed. Results A total of 285 patients with ICD-10 NTM diagnosis codes were identified (median age 69 years; 50.9% female). Coding errors occurred in 9.8% ( n = 28), while 3.9% ( n = 11) had provisional diagnoses that were later dismissed. Among 246 NTM cases, 85.4% ( n = 210) had pulmonary disease, 12.2% ( n = 30) extrapulmonary disease, and 2.4% ( n = 6) disseminated disease. The overall PPV of ICD-10 NTM diagnosis codes was 86.3% (95% CI: 81.8–89.8), reaching 96.5% (95% CI: 92.6–98.4) for pulmonary codes. Among patients with pulmonary involvement, 79.5% ( n = 167) fulfilled guideline criteria, and most patients (73.3%) with a confirmed NTM diagnosis received antibiotic NTM treatment. Conclusion ICD-10 NTM diagnosis codes demonstrated a moderate to high PPV and appeared to be reliable indicators of clinically relevant NTM disease, supported by the high proportion of patients fulfilling diagnostic criteria and receiving relevant treatment. Although coding errors and provisional diagnoses should be considered, ICD-10 codes represent a valuable tool for identifying NTM cases in surveillance and epidemiological research.

Читать у издателя →

Похожие статьи