Survival and Hospitalization Among Dialysis Patients In Taiwan In 2001-2005:Propensity Score Approach

Fenni Widya Sari(Fenni Widya Sari)

指導教授 ： 石曜堂

亞洲大學/健康學院/健康管理研究所/碩士(2012年)

摘要

Background: End-Stage Renal Disease (ESRD) patients in Taiwan are increasing year by year.
This ranked Taiwan as number one in incidence and prevalence in ESRD patients worldwide.
Dialysis is one of the treatment choices for ESRD patients. Since Taiwan has implemented
National Health Insurance (NHI), the utilization of dialysis increased rapidly. Due to the
urgency of a problem, it necessary to investigate survival and hospitalization among dialysis
patients.
Objectives: The purpose of this study is to compare PD and HD in five years survival and
hospitalization within 1st year, 2nd and 3rd year since 2001 to 2005 in Taiwan and which factors
influence them.
Method: Secondary data from the NHRI database was used to analyze the survival and also
hospitalization of dialysis patients. The researcher chose dialysis patients who took dialysis for
more than 3 months. We used the Propensity Score Matching Method to analyze 38023 cases by
using Nearest Neighbor within Caliper (1.0) to get the matched cases between dialysis groups
(PD and HD). We got 2797 pairs to be analyzed by Kaplan Meier as univariate survival
analysis, and Cox Proportional Hazard Model adjusted for age, gender, hospital location,
hospital type, presence of Diabetes Mellitus, and Charlson Comorbidity Index as multivariate
survival analyses and paired t-test and also Multiple Linear Regression for hospitalization.
Result: We found significant results for survival and hospitalization in univariate analysis, thus
we rejected the null hypotheses. There is a difference between PD and HD in survival and
hospitalization. PD patients had better survival than HD patients. From the multivariate analysis,
we found only age, gender, hospital location, dialysis modality, presence Diabetes Mellitus and
comorbid conditions had significant results. PD patients had better survival when they were
younger (less than 65 years-old), female, not having DM, and less comorbidity conditions.
Meanwhile, HD patients had better survival when they were older (more than 65 years-old),
having DM and greater comorbid condition. Age, gender, dialysis modality, presence of DM
and comorbid condition were have relationship to hospitalization within 1st year. Meanwhile,
gender was excluded in association for hospitalization within 2nd and 3rd year. Increasing age or
CCI score, PD group, female, and DM patients were associated with greater hospitalization
within 1st, 2nd and 3rd years.
Conclusion: PD and HD groups have differences in survival and hospitalization adjusted for
age, gender, hospital location, hospital type, presence of DM and Chalson Comorbidity Index.
This factors should be considered when choosing type of dialysis regarding patient’s
characteristics.

關鍵字

ESRD, dialysis, propensity score, surviva analysis, hospitalization.

並列摘要

Background: End-Stage Renal Disease (ESRD) patients in Taiwan are increasing year by year.
This ranked Taiwan as number one in incidence and prevalence in ESRD patients worldwide.
Dialysis is one of the treatment choices for ESRD patients. Since Taiwan has implemented
National Health Insurance (NHI), the utilization of dialysis increased rapidly. Due to the
urgency of a problem, it necessary to investigate survival and hospitalization among dialysis
patients.
Objectives: The purpose of this study is to compare PD and HD in five years survival and
hospitalization within 1st year, 2nd and 3rd year since 2001 to 2005 in Taiwan and which factors
influence them.
Method: Secondary data from the NHRI database was used to analyze the survival and also
hospitalization of dialysis patients. The researcher chose dialysis patients who took dialysis for
more than 3 months. We used the Propensity Score Matching Method to analyze 38023 cases by
using Nearest Neighbor within Caliper (1.0) to get the matched cases between dialysis groups
(PD and HD). We got 2797 pairs to be analyzed by Kaplan Meier as univariate survival
analysis, and Cox Proportional Hazard Model adjusted for age, gender, hospital location,
hospital type, presence of Diabetes Mellitus, and Charlson Comorbidity Index as multivariate
survival analyses and paired t-test and also Multiple Linear Regression for hospitalization.
Result: We found significant results for survival and hospitalization in univariate analysis, thus
we rejected the null hypotheses. There is a difference between PD and HD in survival and
hospitalization. PD patients had better survival than HD patients. From the multivariate analysis,
we found only age, gender, hospital location, dialysis modality, presence Diabetes Mellitus and
comorbid conditions had significant results. PD patients had better survival when they were
younger (less than 65 years-old), female, not having DM, and less comorbidity conditions.
Meanwhile, HD patients had better survival when they were older (more than 65 years-old),
having DM and greater comorbid condition. Age, gender, dialysis modality, presence of DM
and comorbid condition were have relationship to hospitalization within 1st year. Meanwhile,
gender was excluded in association for hospitalization within 2nd and 3rd year. Increasing age or
CCI score, PD group, female, and DM patients were associated with greater hospitalization
within 1st, 2nd and 3rd years.
Conclusion: PD and HD groups have differences in survival and hospitalization adjusted for
age, gender, hospital location, hospital type, presence of DM and Chalson Comorbidity Index.
This factors should be considered when choosing type of dialysis regarding patient’s
characteristics.

並列關鍵字

ESRD, dialysis, propensity score, surviva analysis, hospitalization.

參考文獻

Hemodialysis in a National Cohort of Patients with Chronic Kidney Disease. Ann Intern

Xue, J. L., Chen, S. C., Ebben, J. P., et al. (2002). Peritoneal and hemodialysis: I. Differences in

Beddhu, S., Burns., F., et al. (2000). A simple comorbidity scale predicts clinical outcomes and

patients undergoing peritoneal dialysis. Am J Med, 112, pp 696-701.

Bureau of Health Promotion, Taiwan. (2007). The current situation and prospect of CKD