Table 1 presented the total number of medical malpractice lawsuits, the average number of malpractice lawsuits per million outpatient visits and per million population in each province. The first column showed that there were great disparities of the incidences of medical malpractice lawsuits across different provinces. For example, populous provinces were likely to observe more medical malpractice lawsuits, such as Jiangsu, Shandong and Henan. While, total malpractice lawsuits per million visits were much higher in Jilin, Anhui and Chongqing than other provinces.
Several patterns could be identified from comparisons of total number of malpractice lawsuits with the average number of malpractice lawsuits per million population. Some provinces were characterized with relatively low total malpractice lawsuits and higher population-averaged number of malpractice lawsuits, such as Beijing and Shanghai. This pattern showed that medical malpractice lawsuits arose more often in large cities, indicating a potential migration of cases from other regions. Some provinces were perceived with low total and population-averaged number of malpractice lawsuits, such as Gansu and Hainan. While others were observed with high total and population-averaged number of malpractice lawsuits, such as Jilin and Jiangsu.
Table 2 presented the summary statistics for provincial variables. The average occurrences of malpractice lawsuits were about three times greater in 2013 compared to that in 2012. This dramatic increase might be due to the initiation of “Liability for Medical Malpractice” in 2013. The proportion of private hospitals showed an increasing trend from 33% in 2010 to 54% in 2016. Accordingly, the ratio of outpatient visits and discharges from private hospitals to public hospitals had been rising over the same period. On average, the total number of hospitals has been increased about 25% in 2016. The number of doctors per 1000 people in each province had increased slightly from 1.67 to 2. The proportion of first-class public hospitals remained stable around 8%. Both the coverage of UEBMI and URBMI presented an upward trend, ranging from 16% to 29%. The health care expenditures had been growing rapidly from 783.61 hundred million of Yuan in 2010 to 1,409.77 hundred million of Yuan in 2016. Proportion of people aged 65 and above grew to reach 10.56% in 2016. More than a half (52% in 2010 vs. 59% in 2016) of the population were urban residents. With respect to log of per capita household income, small changes were observed between 2010 and 2016 after adjusting for CPI in each year.
Regression results showed that the over-dispersion parameters were significantly different from 0, indicating that the negative binomial models were preferred over Poisson models. The results of fixed-effects negative binomial models were reported in Table 3 to account for potential endogeneity of the time-invariant unobserved provincial factors as discussed in previous sections. Further, Akaike Information Criterion (AIC) indicated that fixed-effects negative binomial models were superior to random-effects counterpart. Year dummies were added to specify time effects in all provinces. Moreover, an indicator of initiation of “Liability for Medical Malpractice” (e.g. “Indicator of Liability for Medical Malpractice”) were included in the regression models to account for the shifts in legal environment. Indicators of regions were also incorporated into the models.
In Table 3, we see that greater expansion of private hospitals (measured by “Total number of private hospitals”, “Proportion of private hospitals” and “Ratio of discharges from private to public hospitals”) were associated with fewer medical malpractice lawsuits, and this association was statistically significant at the 10% level of significance at least. Specifically, the first column of Table 3 presented the results using total number of private hospitals as an indicator of private hospital expansion. More private hospitals in health care market were associated with less medical malpractice lawsuits. Column (1) presented that an increase of 100 private hospitals in health care market was expected to have a decrease of 21% (=1-exp(-0.23)) in the number of medical malpractice lawsuits holding all other variables constant. Column (2) investigated the proportion of private hospitals and showed a significantly negative correlation with the occurrences of medical malpractice lawsuits in each province. Column (3) presented the results of regressions using ratio of outpatient visits from private hospitals to public hospitals to measure the expansion of private hospitals. The coefficients indicated that there was a positive association between expansion of private hospitals and the ratio of outpatient visits from private to public hospitals, but the correlation was not statistically significant. In column (4), we could see that higher ratio of discharges from private hospitals to public hospitals was negatively and significantly correlated with incidences of medical malpractice lawsuits in each province.
The results in Table 3 also reported other significant variables. The proportion of people aged 65 and above was not significantly associated with the number of malpractice lawsuits. This finding was supported by a previous study, which found that demographic variables were not associated with frequency of medical malpractice claims[25]. While, proportion of females was positively related to the occurrences of medical malpractice lawsuits in three out of four models, and this relationship was significant at 10% level. Higher proportion of urban residents was significantly associated with greater medical malpractice lawsuits. Previous studies suggested that increasing urbanization might contribute to the filed medical malpractice claims because patients in urban hospitals might have greater propensity to litigate a malpractice injuries[25]. More doctors per 1000 people and higher per capita household income were associate with higher amount of medical malpractice lawsuits. Both coverage rate of UEBMI and URBMI were negatively associated with the malpractice lawsuits, but only statistically significant for coefficients for UEBMI. Ramseyer (2009) [34]found that universal health insurance system in Japan prompted the government to cover the rudimentary health services and suppressed the price for sophisticated procedures. Hence, it was possible that the medical malpractices showed a decreasing trend because patients were less likely to sue over primary health care. The positive sign of the Indicator of Liability for Medical Malpractice suggested that more lawsuits were expected to occur after the initiation of Liability for Medical Malpractice in 2013. The coefficients for total health expenditures were negative and significant, implying that health care expenditures were negatively associated with medical malpractice lawsuits. Volume-outcome relationship could be a source of explanations for this finding. According to a systematic review, better outcomes were associated with higher hospital volume[35]. As health care expenditure was usually considered as measurement of health care utilization, it was possible that greater health care services might improve the hospital quality.
To check the robustness of our results, we first utilize the lagged indicators of private hospitals expansion in regression analyses. Table 4 showed the estimation results using one-year-lagged measurement of private hospitals expansion. The directions of the association between the growth of private hospitals and the number of medical malpractice lawsuits were in accordance with our expectation, indicating the growth of private hospitals was negatively associated with the incidences of medical malpractice lawsuits. However, the association was only significant at 10% level in models utilizing proportion of private hospitals and ratio of discharges from private to public hospitals. Further, we restricted the sample between 2013 and 2016 to examine the association because almost all judgement documents became publicly available in 2013.
Table 5 presented the results of regressions using restricted sample. Higher ratio of the discharges was significantly associated with fewer medical malpractice lawsuits. There was no significant association between other indicators of private hospitals expansion and the incidences of medical malpractice lawsuits. In general, our results were robust in different specifications.