Background: How hospital clinicians facilitate breastfeeding in the first 48–72 h is critical to breastfeeding exclusivity and duration. Mothers who discharge hospital directly breastfeeding are more likely to continue exclusively breastfeeding at 3-months. Objective: To assess the impact of facility-wide implementation of a physiological breastfeeding method (the Thompson method) on direct breastfeeding at hospital discharge and exclusive breastfeeding at 3-months of age. Design: Multi-method design using interrupted time series analysis and surveys. Setting(s): An Australian tertiary maternity hospital. Participants: 13,667 mother-baby pairs (interrupted time series analysis) and 495 postnatal mothers (surveys). Methods: The Thompson method includes cradle position and hold, alignment of mouth-to-nipple, baby-led connection and seal, maternal fine-tuning for symmetry, and leisurely duration. We used a large pre-post implementation dataset and conducted interrupted time series analysis using a 24-month baseline period (January 2016 – December 2017); and a 15-month post-implementation period (April 2018 – June 2019). We recruited a sub-sample of women to complete surveys at hospital discharge and 3-months postpartum. Surveys were primarily used to measure impact of Thompson method on exclusive breastfeeding at 3-months, compared with a baseline survey conducted in same setting. Results: Following implementation of the Thompson method, the declining trend in direct breastfeeding at hospital discharge was significantly averted by 0.39% each month relative to baseline (95% CI: 0.03% to 0.76%; p = 0.037). While the 3-month exclusive breastfeeding rate in the Thompson group was 3 percentage points higher than the baseline group; this result did not reach statistical significance. However, a subgroup analysis of women who discharged hospital exclusively breastfeeding revealed the relative odds of exclusive breastfeeding at 3-months in the Thompson group was 0.25 (95% CI: 0.17 to 0.38; p < 0.001), significantly better than the baseline group (Z = 3.23, p < 0.01) where the relative odds was only 0.07 (95% CI: 0.03 to 0.19; p < 0.001). Conclusions: Implementation of the Thompson method for well mother-baby pairs improved direct breastfeeding trends at hospital discharge. For women who discharged hospital exclusively breastfeeding, exposure to the Thompson method reduced the risk of exclusive breastfeeding discontinuation by 3-months. The positive impact of the method was potentially confounded by partial implementation and a parallel rise in birth interventions which undermine breastfeeding. We recommend strategies to strengthen clinician buy-in to the method, and future research using a cluster randomised trial design. Tweetable abstract: Facility-wide implementation of the Thompson method improves direct breastfeeding at hospital discharge and predicts breastfeeding exclusivity at 3-months.