Reductions in hospitalisations and emergency department visits with early antibiotic initiation in nontuberculous mycobacterial lung disease

Background While antibiotics are recommended for treatment of nontuberculous mycobacterial lung disease (NTMLD), the impact of early antibiotic initiation on healthcare resource utilisation is unclear. This study compared healthcare resource utilisation with early versus delayed antibiotic initiation in NTMLD. Methods A retrospective, claims database study (Merative MarketScan) of patients diagnosed with NTMLD between 1 July 2015 and 30 June 2019. Patients were divided into early antibiotic initiation, i.e. <= 3 months after the first medical claim for NTMLD (index date), and delayed antibiotic initiation groups. Hospitalisations and outpatient visits during a 2-year post-index period were compared to baseline per treatment group; a difference-in-difference analysis compared early and delayed antibiotic initiation groups adjusting for confounding. Results Out of 481 NTMLD treated patients, 364 (76%) and 117 (24%) comprised the early and delayed antibiotic initiation groups, respectively. The early antibiotic initiation group showed significant reductions from baseline in hospitalisations (all-cause, respiratory) and emergency department (ED) visits at follow-up. A significant increase from baseline in mean number of hospitalisations per patient was observed in the delayed antibiotic initiation group in year 1 post-index. Compared to delayed antibiotic initiation, the early antibiotic initiation group showed significantly greater reductions in all-cause hospitalisations in years 1 and 2 post-index (relative risk 0.62 (95% CI 0.41-0.95) and 0.62 (95% CI 0.39-0.98), respectively), and in respiratory-related hospitalisations. Conclusions The early antibiotic initiation group showed significant reductions from baseline in hospitalisations and ED visits over time. Compared to delayed antibiotic initiation, early antibiotic initiation was associated with significantly greater reductions in hospitalisations.