Fig. 2

Changes in ADL improvement at discharge and LOS of hospital admission. Interrupted time-series of ADL improvement at hospital discharge and LOS during the pre- and post-introduction periods of the financial incentive. X-axis indicates the ICU admission month, which signifies the elapsed time from the introduction of the financial incentives for each hospital. Y-axis indicates the mean adjusted outcomes in each month. Red lines depict the interrupted time-series regression model describing outcomes before and after the introduction of the incentive. Dashed red lines depict the counterfactual outcomes in the absence of the introduction of the incentive. The box plots show the mean of each outcome by hospital per month. Boxes indicate first quartile, median, and third quartile. Dots denote observations outside the range of 1.5 × interquartile range [IQR]. A light shade of time periods visualizes the time after interruption; the introduction of incentive per each hospital. The outcomes were adjusted for age, sex, body mass index, smoking status, Charlson comorbidity index, Angus organ score, Barthel index at admission, emergency admission, ambulance use, numbers of days hospitalization until ICU admission, drugs and procedures within 2 days of hospital admission, hospital volume, patient-to-nurse ratio, intensivist certified hospital, and dedicated therapist. The main finding is that the financial incentive was associated with a significant slope change in improvement in the Barthel index at discharge and shorter length of stay at the hospital