OBJECTIVES: To investigate the impact of no cost sharing on paediatric care on usage and health outcomes, and whether the effect varies by household income levels. DESIGN: Regression discontinuity design. SETTING: Nationwide medical claims database in Japan. PARTICIPANTS: Children aged younger than 20 years from April 2018 to March 2022. EXPOSURE: Co-insurance rate that increases sharply from 0% to 30% at a certain age threshold (the threshold age varies between 6 and 20 years depending on region). PRIMARY OUTCOME MEASURES: The outpatient care usage (outpatient visit days and healthcare spending for outpatient care) and inpatient care (experience of any hospitalisation and healthcare spending for inpatient care). RESULTS: Of 244 549 children, 49 556 participants were in the bandwidth and thus included in our analyses. Results from the regression discontinuity analysis indicate that no cost sharing was associated with a significant increase in the number of outpatient visit days (+5.26 days; 95% CI, +4.89 to +5.82; p<0.01; estimated arc price elasticity, -0.45) and in outpatient healthcare spending (+US$369; 95% CI, +US$344 to +US$406; p<0.01; arc price elasticity, -0.55). We found no evidence that no cost sharing was associated with changes in inpatient care usage. Notably, the effect of no cost-sharing policy on outpatient healthcare usage was larger among children from high-income households (visit days +5.96 days; 95% CI, +4.88 to +7.64, spending +US$511; 95% CI, +US$440 to +US$627) compared with children from low-income households (visit days +2.64 days; 95% CI, +1.54 to +4.23, spending +US$154; 95% CI, +US$80 to +US$249). CONCLUSIONS: No cost sharing for paediatric care was associated with a greater usage of outpatient care services, but did not affect inpatient care usage. The study found that this effect was more pronounced among children from high-income households, indicating that the no cost sharing disproportionately benefits high-income households and may contribute to larger disparities.