Objective  To examine the effect of long-term care insurance on health condition, medical expenditure and family care of urban elderly in China for providing evidence to the improvement of long-term care insurance system and health condition of elderly population.

  Methods  The data on 2 278 urban residents aged 60 years and above were from three waves of China Health and Retirement Longitudinal Study (CHARLS) conducted in 2013, 2015, and 2018 across China. Of the urban elderly selected, 798 from 11 pilot cities of implementing long-term care insurance for elderly residents were assigned into a study group and 1 480 from non-pilot cities were regarded as the controls. Difference-in-differences (DID) method was adopted to analyze the effect of long-term care insurance on health condition, medical expenditure, family care of urban elderly and to assess the heterogeneity between the young-old and the older-old residents. Propensity score matching-difference-in-differences (PSM-DID) method was used in robustness test.

  Results  Compared to those of the controls, significantly higher self-rated health score (2.92 ± 0.87 vs. 2.62 ± 0.95, P = 0.04 ) and lower Center of Epidemiological Studies Depression Scale-10 (CESD-10) score (9.93 ± 8.05 vs. 10.24 ± 7.08, P = 0.04) were observed in the study group, with a DID of 0.30 (P < 0.01) for self-rated health score and – 0.32 (P = 0.02) for mental health between the two groups, indicating a positive effect of long-term care insurance on the physical and mental health of the elderly; while, the monthly medical expenses of the study group was significantly higher than that of control group (449.08 ± 1024.42 vs. 220.43 ± 605.62 RMB yuan, P = 0.03). The elderly of the study group received significantly lower monthly number of days (3.47) or hours (41.47) of home-based care than those of the control group (both P < 0.05). Heterogeneity analysis results showed that long-term care insurance had a significant effect on the health condition of both the young-old and the older-old persons to a similar extent. The effect of long-term care insurance on monthly medical expense was significant only for the young-old elderly; the effect of long-term care insurance on monthly number of days or hours of home-based care was much stronger for the older-old persons.

  Conclusion  In general, long-term care insurance could improve the health condition and reduced the time of home-based care for urban elderly, but the role of reducing the actual medical expenses is limited.