Open AccessUtility Estimation of Health Status of Cancer Patients by Mapping for Cost-Utility Analysis
Author(s) -
Ali Reza Mortezaei,
Mohammad Taher Rajabi,
Masoumeh Erfani Khanghahi,
Hosein Ameri
Publication year - 2019
Publication title -
rāhburdhā-yi mudīriyyat dar niẓām-i salāmat
Language(s) - English
Resource type - Journals
eISSN - 2538-1563
pISSN - 2476-6879
DOI - 10.18502/mshsj.v4i1.1089
Subject(s) - mean squared error , statistics , breast cancer , goodness of fit , medicine , mathematics , cancer , ordinary least squares , quality of life (healthcare) , quality adjusted life year , econometrics , cost effectiveness , nursing
Background: It is important to obtain accurate information about the preferences of people for measuring quality-adjusted life years (QALYs), because it is necessary for cost-utility analysis. In this regard, mapping is a method to access this information. Therefore, the purpose of this study was to map Functional Assessment of Cancer Therapy – General (FACT-G) onto Short Form Six Dimension (SF-6D) in breast cancer patients to provide appropriate conditions for a detailed cost-utility analysis.
Methods: This descriptive analytical study was conducted on 416 patients with breast cancer. The SF-6D and FACT-G questionnaires were completed for patients selected by consecutive sampling from the Imam Khomeini Cancer Institute in Tehran in 2018. The Ordinary Least Squares model was used to estimate the value of utility and the models' goodness of fit was evaluated using R2. In addition, models' predictive performance was assessed by Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Minimal Important Difference (MID) conducted between the observed and predicted SF-6D values. Models were validated using a 10-fold cross validation method.
Results: Given the criteria of goodness of fit, model 2 was the best (R2 = 41.19 %). Moreover, findings of the predictive performance of models showed that model 2 was the best (MAE = 0.06943, RMSE = 0.09031, and MID = 0.0003).
Conclusions: Findings showed that the developed algorithm had a good predictive ability. So, it can enable the policymakers and researchers to convert cancer-specific health-related quality of life instruments to preference-based instruments.