Abstract
Purpose: To perform a cost-utility analysis comparing primary pars-plana vitrectomy (PPV) within 24 hours with primary nonsurgical vitreous tap (or tap and inject [T&I]) for the management of endophthalmitis. Methods: Retrospective cost-utility analysis using decision tree modeling. The Victorian Endophthalmitis Registry was used to model outcome probabilities and costs from a third-party payer perspective. Australian Medicare data were used to calculate costs in a hospital-based setting (Australian dollars [A$]). Cost utility was based on preserved visual acuity and cost per quality-adjusted life year (QALY). Results: The authors identified 206 eyes treated between January 1, 2011, and January 1, 2021; 36 eyes received PPV, and 170 eyes received T&I. Seventeen eyes in the T&I group required delayed PPV. Mean incident ages were 76.29 years (53% female) in the PPV group and 74.28 years (55% female) in the T&I group. Imputed costs were A$1,523 and A$310 for PPV and T&I, with additional per-night admission costs of A$1,177. The mean presenting vs discharge logMAR of endophthalmitis was 2.24 vs 1.25 for the PPV group and 1.88 vs 1.03 for the T&I group. The mean durations of admission were 4.33 nights (PPV) and 4.04 nights (T&I). Total calculated costs per admission were A$6,929.41 and A$5,065.08 for PPV and T&I, respectively. Estimated lifetime QALYs gained were 2.23 (PPV) and 2.45 (T&I). The final costs derived per QALY were A$3,107 (PPV) and A$2,067 (T&I). Conclusions: PPV and T&I are both cost-effective per gained QALY, though the latter provided superior cost utility. A prospective randomized trial is indicated as the 2 groups differed at baseline, with eyes receiving vitrectomy having worse presenting visual acuity and prognosis.