Page 189 - 臺大管理論叢第33卷第1期

NTU Management Review Vol. 33 No. 1 Apr. 2023

result in relation to surgical types in this study may also result from the method of
classifying surgery codes adopted by the NHIA. The NHIA classifies thyroidectomy-
related surgeries as DRG 290, which is defined only by the codes for major surgical
interventions rather than by a major diagnostic category, without consideration of

comorbidities or complications. Accordingly, all such surgeries are granted the same
payment system.
This study demonstrates the beneficial effects of CPs on LOS and total hospital

charges. The findings are expected to serve as a reference for the NHIA and medical
institutions to manage thyroidectomy costs more effectively. In addition, the results of this
study may also serve as a valuable benchmark for medical resource distribution and the
quality control of healthcare under the current self-management project conducted by the
NHIA, where the healthcare provider’s performance such as the efficiency of healthcare

resource utilization and the service items or quantities are evaluated to determine the total
healthcare service points and the growth rate of yearly service quantity that the NHIA will
offer to the individual hospital for the next contractual year.

However, there are several limitations that should be considered. First, this study
uses claims data and is therefore associated with the limitations of the claims information.
Some patients’ characteristics such as socioeconomic status, and different stages of the
disease that may also influence medical care utilization are not coded on the hospital’s
claim form. Nonetheless, this study provides some preliminary findings on the effect

of CPs’ implementation on healthcare utilization by patients receiving a thyroidectomy
performed by different divisions and surgeons and should generate some insights regarding
the effect of the CPs. Further research is necessary to elucidate the effect of CPs on

healthcare utilization while controlling for the patient’s socioeconomic status and different
stages of the disease.
The second limitation is the extent to which the assumptions underlying the DID
method have been met. The DID method relies on the assumption that the selection of
treatment group is not correlated with the observed or unobserved factors (confounders),

which are associated with the outcome (Jones and Rice, 2009). Wherever this assumption
is untenable, the inference will be contaminated with selection bias due to a failure to
control for unobserved or unobservable characteristics (Lin and Hsu, 2014; Song, Safran,

and Chernew, 2019). In this study, although we have a strong belief that the changes in

181