Page 179 - 臺大管理論叢第33卷第1期

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

2002 to December 2003) and has a value of zero otherwise (January 2001 to June 2002).
The coefficient of the CPs represents the changes in the utilization of thyroidectomy
patients in the different time periods. The coefficient of the interaction term, i.e., Treatment
* CPs, represents the changes in the utilization of GS patients relative to the changes for

ENT patients in different time periods and reveals the impact of the CPs on the medical
utilization. The DID approach allows us to account for any secular trend in utilization.
By comparing changes, the observed and unobserved time-invariant, patient-related and

surgeon-associated characteristics that may be correlated with medical expenses can be
controlled. Any remaining significant differences in utilization can be attributed to the CPs.
In this study, patient covariates are presented as means with standard deviations and
frequencies with percentages for continuous and categorical variables, respectively, unless
otherwise stated. The Chi-square test is used to examine the differences in categorical

variables among groups while the Analysis of Variance (ANOVA) is conducted for the
same purpose for the continuous variables. Independent variables include Age, Sex, Major
diagnostic category, Major intervention, CCI scores and surgeon-related characteristics.

Multiple linear regression is used to investigate the effects of these variables on the
hospital charges and LOS. A p-value of less than 0.05 is considered to be statistically
significant.

3. Results

3.1 Descriptive Statistics
We divide the primary diagnoses of patients into the following three categories:

malignant neoplasm of the thyroid gland (ICD-9-CM code 193), benign neoplasm of the
thyroid gland (ICD-9-CM code 226), and other diagnoses of the thyroid gland. Table 1
shows that there are no significant differences among patients in terms of the proportion
of each major diagnostic category (p = 0.465). CCI scores are used to denote the severity
of the disease conditions. Notably, the average CCI scores range from 0.21 to 0.53. The

reasons for the relatively low average CCI scores may be attributed to the fact that most
patients in the current study are young adults and those middle-aged with the average age
ranging from 40 to 55. The statistical tests reveal no significant differences in CCI scores
among GS and ENT patients before and after the CPs’ implementation (p = 0.325). This

171