| Summary: | 碩士 === 國立陽明大學 === 衛生福利研究所 === 98 === The Bureau of the National Health Insurance (BNHI) has implemented ’The Pay for Performance Program for Hypertensive patients’ in clinics since 2006, it draw financial incentives up to encourage medical provider to provide better quality of health care. Mainly domestic Hypertension related studies were confined to some geographic areas or single institution, and there was no assessment about this program. This study evaluates the effectiveness of this program via avoidable hospitalization for nationally representative samples.
A quasi experimental study to compare the incidence of avoid hospitalization before and after the intervention to control the differences between the two groups before intervention using differences in differences approach was conducted those who joined the program as the experimental group, those who did not as the control group during 2005-2008. Incidences of avoidable hospitalization were observed one year before and/or two years after the intervention program. Based on NHI claim database based on the third-generation Aday & Andersen behavioral model, the independent variable was interaction of year and group, the dependent variables were the hypertension related avoidable hospitalization defined by The Agency for Healthcare Research and Quality (AHRQ). Greedy propensity score matching method was applied to calibrate characteristics of the hypertensive patients (age, gender, the insured amount, having regular source of medical care, severity of illness developed based on Gonella disease staging concept, the co-morbidity index of Clark Chronic disease score in 1995), and hospital characteristics (level, ownership, and location of the hospital)based on logictic regression analysis. Control group was matched to experimental group based on propensity score by ratio a 4 to 1. Data were analyzed using SAS9.2. Descriptive statistics included frequency, percentage and other information about the characteristics of the sample before and after matching; Inferential statistics include: chi-square test, McNemar test, Fisher exact test, and generalized estimating equations (GEE).
There were 1116 persons in experimental and 4464 in control groups of propensity score matching. These two groups didn’t have significantly difference on patients’ and hospital characteristics; GEE analysis showed the likelihood to have avoidable hospitalization of the experimental groups were 3.16 and 2.06 time during the year and one year after the intervention. However the results were not significant.
The results failed to support the hypothesis. This means that this program failed to significantly hospitalization. However, due to low incidence of avoidable hospitalization, small sample sizes, our results should be verified in the future. Limitation of this study include: short duration; lack of clinical examination data and information on attitude and behavior of the providers toward intervention program. Future studies should be conducted by policy makers to further evaluate the effect of this program, so as to decide whether to implement this program continuously in the future.
|
|---|
