혈액투석 적정성 평가 결과지표의 중증도 보정 모형 개발

Abstract

Hemodialysis is one of the renal replacement therapies for patients with end-stage renal failure, and the number of hemodialysis patients continue to increase with the aging population. Patients with end stage renal failure undergoing hemodialysis have a high risk of anemia due to blood loss in the dialysis tubing and dialyzer and loss its ability to produce the erythropoietin to the production of hemoglobin. Since anemia greatly affects the quality of life of patients, anemia management is important. The Health Insurance Review and Assessment Service has been evaluating the quality of medical care for patients with end stage renal failure undergoing hemodialysis since 2009. Among the quality evaluations of care for patients undergoing hemodialysis, the ratio of patients under Hb 10g/dL, an anemia management measure, is used as a monitoring measure, not a regular measure. However, anemia management is an important measure because it affects the quality of life of patients. Therefore, it is necessary to consider converting the anemia management measure to a regular one. In order to convert to a regular measure, it is essential to establish a validated model for risk adjustment of anemia in patients with end stage renal failure undergoing hemodialysis. The purpose of this study is to develop a model for risk adjustment of anemia in patients with end stage renal failure undergoing hemodialysis. To this end, through literature review and consultation, major variables affecting anemia in patients with chronic renal failure are summarized and applied We wanted to develop the model for validate risk adjust model. Currently, the medical quality evaluation target was expanded to all hemodialysis patients receiving ESA or the evaluation target to all hemodialysis patients. In addition, model discrimination power was analyzed by changing the anemia management measures. The current evaluation index is the percentage of patients whose mean of Hb measured after the first ESA(Erythropoiesis-Stimulating Agents) administration during the evaluation period(6 months) is less than 10 g/dL. In addition, the case where the number of months with Hb<10g/dL was more than 2 months and the number of months with ➁ Hb<10g/dL was more than 3 times during the ➀ evaluation period was also calculated. The analysis data are from 38,794 hemodialysis patients collected for medical quality evaluation, and 37,165(95.8%) patients received ESA during the evaluation period. Since the results of the anemia index are in the form of dichotomous data, multiple logistic regression analysis was used for the statistical method. As a result of the analysis, the model's discrimination performance was higher for all patients than for patients who administrated ESA, as for the measures of anemia management, the average Hb<10g/dL was the highest. The highest C-statistics are the models with an average Hb<10g/dL for all patients 0.636 (95% CI 0.627~0.646) was obtained, but this result is difficult to apply because the C-statistics did not reach above 0.7. In the analysis by medical institution size, the C-statistics of tertiary hospitals were 0.718 (95% CI 0.686~0.749). The highest C-statistics are the models with an average Hb<10g/dL for all patients 0.636(95% CI 0.627~0.646) was obtained, but this result means that it predict anemia poorly because the C-statistics did not reach above 0.7. In the analysis by medical institution size, the C-statistics of tertiary hospitals were 0.718 (95% CI 0.686~0.749). However, only 6.9% of all hemodialysis patients received hemodialysis at tertiary hospitals. We could not develop the model for risk adjusted of anemia. The model developed in this study is not validated because the C-statistic is not 0.7 or higher. However, we found the prevalence of anemia in patients undergoing hemodialysis in Korea and it is possible to identify factors that affect the occurrence of anemia. The reason why validated model for risk adjustment have not been developed is that resistance to erythropoiesis-stimulating agents, which is difficult to reflect as a variable and there may be an absence of an anemia risk index due to concomitant diseases.