| 摘要: | 台灣人罹患糖尿病的情況日益嚴重，本文將透過研究縣市的生活型態和社經差異，從中找尋影響糖尿病就診率的因子，藉此提供政策制定的參考依據。
被解釋變數採用的是「二十大死因全民健康保險利用資料」，並自行計算出糖尿病就診率。資料年度為民國九十三年到民國一零三年，共十一個年份；橫斷面為台灣本島之十九個縣市的資料，以下再分兩組性別及四組年齡層。關注的解釋變數為各縣市家戶所得、人均生產總值、大專生比例、菸酒消費比例，而以人口密度、老化指標、醫療資源、高血壓人口等做為控制變數。本研究使用追蹤資料並加入性別、年齡的虛擬變數。當解釋變數具有內生性時，則採用兩階段最小平方法估計。控制了性別、年齡、控制變數之後，觀察生活型態和社經差異對於糖尿病就診率的影響。
由迴歸結果得知，社經地位越高的地區，糖尿病就診率越低。而越趨向不健康的生活型態，如抽菸、飲酒比例越高，則該地區的糖尿病罹病率越高。性別及年齡是影響糖尿病就診率的決定性因素；女性風險高於男性。另外年齡越高，就診率越大，且高峰會落在六十五到八十四歲。由於抽菸與飲酒比例都顯著影響糖尿病就診率，因此建議政府可以將抽菸和飲酒的外部性內部化，嘗試對菸、酒課更重的稅。
;More and more Taiwanese are suffering from diabetes and it’s getting serious over time. Therefore, this study focus on how lifestyle and socioeconomic position affect the treatment rate of diabetes. By identifying the risky areas, we provide suggestions for policy makers.
The independent variable, treatment rate of diabetes, is in the form of panel data and is calculated based on the data from National Health Insurance. The scale of this research was from 2004 to 2014, and the cross-section is based on 19 cities of Taiwan and then the data are allocated into 2 gender groups and 4 age groups. The explanatory variables include GDP per capita, the proportion of college students, tobacco and alcohol consumption in respect of different city, and we set population density, medical expenses, aging index and high blood pressure as control variables. The method of this study is panel data using gender and age as dummy variables. When there is endogeneity in explanatory variables, the two-stage least squares method will be used. Controlling gender, age, and control variables, we observe the effect of socioeconomic position and lifestyle on the treatment rate of diabetes.
According to our results, the area with higher socioeconomic position tends to have low treatment rate in diabetes. Also, the area with a negative lifestyle is likely to have a higher proportion of tobacco and alcohol consumption tend to have higher treatment rates. Gender and age are two crucial factors that explain the treatment rate of diabetes. Generally, female had a higher risk of diabetes than male. In addition, the higher the age is, the higher the treatment rate will be, and the peak falls in the range of 65 to 84 years old. Due to the significant effect of tobacco and alcohol consumption, this study suggests government internalize the externalities of smoking and drinking, by adding taxes on tobacco and alcohol. |
