P1453THE DEVELOPMENT OF A NEW PATIENT-REPORTED OUTCOME MEASURE TO EVALUATE THE QUALITY OF DIALYSIS THERAPY

Abstract

Abstract Background and Aims There is a growing interest in Patient-Reported Outcome Measures (PROMs) as evaluation tools of medical care globally, including dialysis therapy. We developed our original PROM for patients undergoing maintenance hemodialysis (MHD), which was named the Ai-POD (Ai means love in Japanese, POD stands for Patients Oriented Dialysis). The Ai-POD is composed of 20 items relating to physical symptoms (joint pain, itching, frustration, weakness, shortness of breath, constipation, difficulty getting to sleep, asleep condition), complaints during dialysis therapy (pain of cannulation, headache, drop of blood-pressure, leg crumps, fatigue after dialysis session), dietary conditions (appetite, food enjoyment, sore or dry mouth), mental health (depressive feeling, life enjoyment) and comprehensive life satisfaction. All items are rated on a Likert scale from 0 (best) to 4 (worst) except mental health. The mental health questions are yes-no closed questions converting 0 (excellent) or 4 (bad). The sum of all 20 components ranges from 0 to 80, denoting increasing of severity. We have been collecting the data of Ai -POD from all patients in our facilities for over ten years. We examined the associations between the score of Ai-POD and prospective mortality in our patients undergoing hemodialysis therapy to assess the usefulness of the Ai-POD score. Method Three hundred ninety-six patients (262 men, 134 women; age 66 ± 12years) on MHD therapy for at least three months were evaluated with Ai-POD in October 2013 and followed up for three years. We categorized the patients into the four groups according to the quartile of the Ai-POD score and compared anthropometric measurements and laboratory measurements among the groups. The four groups are as below: Grop1 (Score: 0 ∼ 10), Group 2 (11 ∼18), Group 3 (19 ∼29), and Group 4 (29 ∼). A Cox regression analysis was used to evaluate the association between the severity of the Ai-POD score and mortality risk adjusting with several factors. Kaplan-Meire analysis and Log-rank test assessed the mortality rate for comparison among the four groups. Results The average of Ai-POD score of all patients was 19.6 ± 11.7, following a normal distribution. Among the four quartile groups, we found no differences in age, dialysis vintage and diabetes ratio, but found the statistic difference in gender ratio and serum-albumin. Compared with Group 1, the multivariable-adjusted hazard ratio for mortality was significantly increased to 2.51 (95% CI 1.03 – 7.02) and 3.00 (95% CI 1.25 – 8.46) in Group 3 and Group 4, respectively. Kaplan-Meier analysis showed higher mortality in Group 3 and 4 (16.7% and 19.1% respectively) compared with Group 1 and 2 (5.6% and 6.5% respectively) significantly (p=0.0033) (Figure). Sub-analysis with each Ai-POD item indicated that frustration, appetite, food enjoyment, and overall life satisfaction were strong independent predictors of prognosis. Conclusion The score of PROM is a significant prognostic factor, suggesting the Ai-POD score could be one of the useful tools to estimate the quality of dialysis therapy. The strategies to improve the score in each item could enhance the quality of dialysis therapy.