Exploring the influence of a financial incentive scheme on early mobilization and rehabilitation in ICU patients: an interrupted time-series analysis

Abstract

Abstract Background Clinical guidelines recommend early mobilization and rehabilitation (EMR) for patients who are critically ill. However, various barriers impede its implementation in real-world clinical settings. In 2018, the Japanese universal healthcare coverage system announced a unique financial incentive scheme to facilitate EMR for patients in intensive care units (ICU). This study evaluated whether such an incentive improved patients’ activities of daily living (ADL) and reduced their hospital length of stay (LOS). Methods Using the national inpatient database in Japan, we identified patients admitted to the ICU, who stayed over 48 hours between April 2017 and March 2019. The financial incentive required medical institutions to form a multidisciplinary team approach for EMR, development and periodic review of the standardized rehabilitation protocol, starting rehabilitation within 2 days of ICU admission. The incentive amounted to 34.6 United States Dollars per patient per day with limit 14 days, structured as a per diem payment. Hospitals were not mandated to provide detailed information on individual rehabilitation for government, and the insurer made payments directly to the hospitals based on their claims. Exposure was the introduction of the financial incentive defined as the first day of claim by each hospital. We conducted an interrupted time-series analysis to assess the impact of the financial incentive scheme. Multivariable radon-effects regression and Tobit regression analysis were performed with random intercept for the hospital of admission. Results A total of 33,568 patients were deemed eligible. We confirmed that the basic assumption of ITS was fulfilled. The financial incentive was associated with an improvement in the Barthel index at discharge (0.44 points change in trend per month; 95% confidence interval = 0.20–0.68) and shorter hospital LOS (− 0.66 days change in trend per month; 95% confidence interval = − 0.88 – -0.44). The sensitivity and subgroup analyses showed consistent results. Conclusions The study suggests a potential association between the financial incentive for EMR in ICU patients and improved outcomes. This incentive scheme may provide a unique solution to EMR barrier in practice, however, caution is warranted in interpreting these findings due to recent changes in ICU care practices.