The Impact of Adjustment of Reimbursement Rate on Health Care Quality and Utilization, Evidence from Taiwan's Kidney Transplantation

It is a long concern academically and politically that Taiwan has high prevalence rate and incidence rate for End-stage renal diseases (ESRD). This high prevalence rate and high incidence rate are also one of major reasons for the increasing growth of related health expenditure. Among the traditional renal replacement treatments, kidney transplantation is one of methods leading patients to a better quality of life after the treatment. In addition, the overall medical costs are believed to be much lower than those using hemo-dialysis and peritoneal dialysis. However, due to the limited supply of organs and the low reimbursement rate, the average ratio of using renal transplantation in Taiwan is relatively lower than other developed countries. In year 2008, the Bureau of National Health Insurance increased the reimbursement rate for kidney transplantation from 19832 points to 60000 points. In this project we aim to analyze the impact of this adjustment of reimbursement rate on the health care utilization and quality. The data is from National Health Insurance research dataset (year 1996-2013). For the health care utilization, we first estimate the impact of this policy on the probability of using kidney transplantation in the patient level. In addition, we estimate whether the ratio of kidney transplantation of all ESRD patients in a hospital is affected after the policy using a traditional regression technique. The second variable for utilization is the length of stay in a hospital for a given transplantation operation. However, because all patients with kidney transplantation is hospitalized, the data is truncated by default. We therefore use zero-truncated Poisson regression model and zero-truncated negative binomial regression model to tickle this data issue and explore whether the length of stay in hospital for an operation of transplantation is significantly decreased after the policy. Lastly, for the impact of health care quality, two indices are use. One is the outcome (death or not) for patient in a given transplantation. The second one is the one year (and three-year) mortality after the operation. We therefore use logit probability and Cox’s survival model to estimate the mortality rate for a transplantation patients after the increase of reimbursement.

Project ID:PF10308-0244
External Project ID:MOST103-2410-H182-026