OBJECTIVE: To assess the consistency of maternal near-miss incidence and mortality index between two definitions across 104 facilities in Tanzania and Uganda. METHODS: Based on WHO guidance, cross-sectional near-miss data were collected in Tanzania (July 2015 to October 2016) and Uganda (June 2016 to September 2017). Prepartum hemorrhage and abortion were included as additional screening events and the number of blood units transfused was recorded. Near-miss incidence and mortality index were determined by using two near-miss definitions: the WHO standard definition, and a modified definition including women receiving at least 1 unit of blood. A sensitivity analysis excluded the additional screening events. RESULTS: Near-miss incidence differed between Tanzania and Uganda (1.79 and 4.00, respectively, per 100 deliveries) when estimated by the standard definition, but was similar (5.24 and 4.94, respectively) by the modified definition. The mortality index was higher in Tanzania than in Uganda when estimated by the standard definition (8.56% vs 3.54%), but was similar by the modified definition (3.10% vs 2.89%). CONCLUSION: The modified definition provided a more consistent estimate of near-miss incidence and mortality index. Lowering the threshold for units of blood transfusion might improve comparability between settings, but more research is needed.