Detecting Potential Overbilling in Medicare Reimbursement via Hours Worked: Reply
回应了Matsumoto指出的数据与编码错误，证明这些错误对原研究的定性结论影响有限，并讨论了服务重复计数问题的量化偏差，强调所提方法对监管者仍有实用价值。
Matsumoto (2020) pointed out data and coding errors in Fang and Gong (2017). We show that these errors have limited impacts: all qualitative findings remain after correcting them. Matsumoto also discussed potential service overcounting in the aggregated utilization data we used to illustrate our method, and then quantified the extent of overcounting with a sample of Medicare claims. We acknowledge the issue but discuss the noise and the bias in his quantification. Overall, our proposed method remains useful, as regulators who are interested in applying the method are unlikely to be subject to the data limitations.