Health and Quality of Life Outcomes最新文献

Background: Arthritis is a prevalent chronic disease substantially impacting patients' quality of life (QoL). While identifying key determinants associated with arthritis is critical for targeted interventions, traditional statistical methods often struggle with complex interactions, and existing machine learning (ML) approaches frequently lack the interpretability needed to guide clinical decisions. This study integrates a comprehensive, explainable machine learning (XAI) workflow to identify and interpret key QoL-related predictors of arthritis status in a large national cohort.

Methods: Data were obtained from 15,011 participants aged > 45 years in the 2020 China Health and Retirement Longitudinal Study (CHARLS). We initially selected 55 potential QoL-related predictors spanning demographic, functional, pain, psychosocial, and lifestyle domains. Feature engineering was performed to create aggregate scores, indicators, and binned variables. Missing data were handled using imputation combined with missing indicator variables. A LightGBM-based feature selection process identified 68 key predictors. Nine ML models (including Logistic Regression, RandomForest, GradientBoosting, LightGBM, CatBoost, XGBoost, DecisionTree, NaiveBayes, KNN) were developed using SMOTE-resampled training data, with hyperparameters optimized via Optuna targeting recall. Performance was evaluated on a held-out test set using Area Under the ROC Curve (AUC), Average Precision (AP), Recall, Specificity, Precison, and F1-score. SHapley Additive exPlanations (SHAP) analysis was applied to the best-performing model (GradientBoosting) for interpretation.

Results: Several models achieved strong predictive performance, with GradientBoosting yielding the highest AUC (0.767, 95% CI: 0.752-0.782) and high AP (0.678, 95% CI: 0.655-0.702). SHAP analysis identified multi-site pain burden (particularly knee/leg pain and pain location count), age, self-rated health, sleep quality, functional limitations (ADL counts/scores), and negative affect as the most influential predictors driving arthritis status prediction.

Conclusions: This study successfully applied an XAI pipeline to identify and rank key QoL-related factors predictive of arthritis status in a large Chinese cohort, achieving robust model performance. Pain burden, age, subjective health, sleep, functional status, and psychological well-being are critical determinants. These interpretable findings can inform risk stratification and guide targeted interventions focusing on these key areas to potentially improve arthritis management.

Purpose: Different health-related quality of life scales are suitable for various populations and scenarios; hence, the selection of measurement tools must be based on research objectives and focal points. To facilitate the choice between EQ-5D-5L and SF-6Dv1 in clinical and economic research of myopic patients undergoing corneal refractive surgery, the study aims to assess and compare EQ-5D-5L and SF-6Dv1 among the patients in China.

Methods: A total of 411 participants with a mean age of 25.6 years were recruited. Information collected including demographic characteristics, EQ-5D-5L and SF-6Dv1. The ceiling effect was evaluated by calculating the proportion of participants who reported the best possible health states. Agreement between EQ-5D-5L and SF-6Dv1 health utility scores (HUSs) was tested using intraclass correlation coefficient (ICC) and Bland-Altman plot. Convergent validity of EQ-5D-5L and SF-6Dv1 HUSs was evaluated using the Spearman correlation coefficient (r) with VF-14 as a calibration standard.

Results: Ceiling effects were 75.2% and 40.4% for EQ-5D-5L and SF-6Dv1, respectively. ICC between the two HUSs was 0.611 (average-measure). Bland-Altman plots showed that the agreement varied across different HUS segments. Correlation between EQ-5D-5L HUS and VF-14 (r = 0.363) was higher than that between SF-6Dv1 HUS and VF-14 (r = 0.226). EQ-5D-5L and SF-6Dv1 HUSs can only distinguish one clinical condition (i.e., chronic pain symptom); but they can identify health differences across all self-reported known-groups. EQ-5D-5L HUS had greater sensitivity; but SF-6Dv1 HUS was more responsive at follow-ups.

Conclusions: EQ-5D-5L and SF-6Dv1 HUS cannot be used interchangeably in myopic patients undergoing corneal refractive surgery. It appears that EQ-5D-5L HUS performs better in terms of convergent validity and sensitivity; but SF-6Dv1 HUS is more responsive.

In this study, we propose several methods to account for reporting heterogeneity in self-reported data coming from Visual Analogue Scales (VAS) using corresponding VAS-based anchoring vignettes. Though widely used as a measurement tool in many disciplines, VAS may suffer from reporting heterogeneity. Such reporting heterogeneity and potential solutions to solve this problem in the context of VAS measures have not yet been addressed in the literature. Using VAS-based anchoring vignettes and standard vignettes assumptions, we show how double-index models can be used to address reporting heterogeneity in VAS. We then apply our methods to real data assessing reporting heterogeneity in VAS-measured Quality of Life (QoL) among students in Switzerland. We show that the findings of previous studies showing positive associations between being a female and QoL might be entirely driven by reporting heterogeneity.

Background: Generic preference-based instruments, such as the Short Form 6-Dimensions (SF-6D) and EuroQol 5-Dimensions (EQ-5D), can generate utility scores that facilitate the estimation of health-related quality of life (HRQoL) which is commonly used in cost-utility analysis. This study investigated the associations between utility scores and potential socio-demographic factors in Chinese patients with dialysis using quantile regression.

Methods: Patients were recruited in a multicenter survey conducted between November 2023 and January 2024 for dialysis patients in China. Patient responses to the SF-6D version 2 (SF-6Dv2) instruments were used to calculate utility scores. The relationships between utility scores and potential socio-demographic factors were examined using both ordinary least squares (OLS) and quantile regression models. The Wald test was employed to test the differences in coefficients across quantiles in quantile regression. Model performance was assessed using 5-fold cross-validation.

Results: A total of 378 patients were included. Age, education level, having a loan due to illness, currently working, monthly income > 8000 RMB and number of comorbidities were associated with utility scores. The quantile regression coefficients and Wald test suggested that the size of the associations between the utility scores and factors varied along with the utility score distribution. Quantile regression yielded more accurate fitted and predicted values compared to OLS regression.

Conclusion: Quantile regression offers a valuable complement in analyzing factors associated with utility scores among Chinese dialysis patients. For policymakers, differentiated nonclinical strategies may be needed to improve HRQoL across varying health states within this population.

Background: Health-related quality of life (HRQoL) is increasingly recognized as a critical indicator of well-being in early childhood, yet its associations with 24-hour movement behaviors-physical activity, screen time, and sleep-remain underexplored in Asian populations. This study aims to examine the association between adherence to the 24-hour movement guidelines (24-MG) and health-related quality of life (HRQOL) among preschoolers residing in major urban areas of Singapore, Japan, and China.

Methods: A cross-sectional survey was conducted among 6,634 children aged 3-6 years across Singapore (n = 3,672), Japan (n = 760), and China (n = 2,202). Movement behaviors were assessed using the validated SMALLQ^® tool, and HRQoL was evaluated using the Pediatric Quality of Life Inventory (PedsQL™). Logistic regression models were applied to determine the associations between different patterns of 24-h MG adherence and HRQoL, adjusting for demographic variables.

Results: Full adherence to all three 24-h MG components was significantly associated with higher physical and psychosocial HRQoL scores in Singapore and China. In Japan, adherence to physical activity guidelines alone is associated with higher HRQoL. Notably, Chinese children had the highest adherence rates across all individual and combined movement behaviors. Conversely, Japanese children had the lowest rates of full adherence and were more likely to fall short of all guidelines.

Conclusions: This study provides novel evidence from three urban Asian contexts that adherence to the 24-hour movement guidelines positively correlates with HRQoL in preschool-aged children. The findings highlight the importance of integrated movement behavior frameworks and support the development of culturally tailored public health policies to improve early childhood well-being.

Background: Australia has a well-established health technology assessment process and there is extensive use of generic health related quality of life (HRQoL) instruments in evidence presented to it. However, there are gaps in tools and evidence available to support evaluation of paediatric health. The aim of this paper is to produce an Australian EQ-5D-Y-3L (Y-3L) value set.

Methods: The methods follow the international Y-3L valuation protocol, but with an expanded design. Data were collected using Composite Time Trade Off (cTTO) and Discrete Choice Experiment (DCE) data from two independent samples of adult members of the Australian general public. In total, 52 Y-3L health states, assigned into four blocks of 14 health states each containing health state 33333, were valued using cTTO. cTTO data were collected via videoconferencing interview and each respondent valued 14 health states. Mean observed cTTO values were adjusted for censoring at -1 using a Tobit model. For the DCE component, 150 latent scale DCE choice pairs were collected via an online survey with each participant completing 15 pairs. DCE data were modelled using a garbage class mixed logit model. Two approaches to anchor DCE data to the Quality Adjusted Life Years (QALYs) scale were explored: anchoring on the value for the worst health state (33333); and mapping DCE data onto the mean cTTO values using all 52 health states. Two evaluation criteria were used to select the final value set: (1) coefficient significance and logical consistency; (2) prediction accuracy of the mean observed cTTO values.

Results: In total, 268 individuals participated in the cTTO interviews, and 1002 completed the DCE. The linear mapping without intercept performed best and was selected as the final value set. Health state values ranged between 0.142 and 1. The relative importance of domains by level 3 coefficients (ordered from most to least important) was: pain/discomfort, then feeling worried, sad or unhappy, usual activities, looking after myself, and mobility.

Conclusion: This study reports an Australian value set for the Y-3L, which enables the calculation of QALYs for use in the economic evaluation of paediatric interventions and can support evidence development and decision making.

Background: The financial burden of cancer care may significantly impair patients' health-related quality of life (HRQoL), yet the extent and nature of this relationship remain underexplored, particularly in low-resource settings. This study aimed to report the HRQoL of patients currently receiving treatment for selected cancers (breast, cervical, lung, oesophageal and stomach) in Nepal. We further investigate the association of out-of-pocket expenditure (OOPE) with HRQoL.

Methods: A cross-sectional survey was conducted from April to May 2024 among 353 patients undergoing cancer treatment in two tertiary cancer hospitals in Nepal. We used the European Quality of Life 5 dimensions (EQ-5D-5L) and the European Quality of Life Visual Analogue Scale (EQ-VAS) to obtain their HRQoL. Similarly, we collected sociodemographic and treatment-related data, including OOPE and patient satisfaction. We used the ordinary least squares estimation with robust standard errors to identify the association between OOPE and HRQoL (EQ-5D-5L index score).

Results: The mean (SD) EQ-5D-5L index score was 0.39 (0.42), and the mean (SD) EQ-VAS score was 56.65 (21.71). Anxiety/depression and pain/discomfort were the most common reported problems (> 90% of the study participants), whereas pain/discomfort had the greatest disability weight (0.17). In the regression analysis, the logarithms of OOPE (β = -0.086; 95% CI: -0.132 to -0.040) was significantly associated with a lower EQ-5D-5L index score. Other significant covariates included being currently not married (β = -0.149; 95% CI: -0.274 to -0.024), having stage IV cancer during diagnosis (β = -0.212; 95% CI: -0.364 to -0.061) and patient satisfaction score (β = 0.015; 95% CI: 0.001 to 0.030).

Conclusions: The study revealed a moderate quality of life among patients currently receiving cancer treatment in Nepal. Higher OOPE along with stage IV cancer during diagnosis and being currently not married were associated with lower HRQoL, whereas higher patient satisfaction score was associated with higher HRQoL. These insights might be helpful for providing targeted interventions such as emphasizing early diagnosis and management and focusing on patient satisfaction and those at financial risk to improve the HRQoL of people with cancer.

Background: Measuring health-related quality of life (HRQoL) in dialysis patients is essential for clinical assessment and economic evaluation. Despite the emergence and increasing use of updated instruments, evidence comparing their performance in Chinese dialysis patients remains limited. The aim of this study was to compare the measurement properties of the EQ-5D-5L and the SF-6Dv2 instruments in Chinese patients on dialysis and to provide a reference for future utility scale choice for Chinese dialysis patients.

Methods: Data were obtained using Wen Juan Xing questionnaire from dialysis patients during November 2023 to January 2024 in hospitals in China. The questionnaire included the EQ-5D-5L, SF-6Dv2, the kidney disease quality of life instrument (KDQOL-36), clinical and socio-demographic characteristics. The agreement of utility scores was assessed using intra class correlation coefficients (ICC) and Bland-Altman plots. The construct validity was evaluated using Spearman's correlation coefficient. The known group validity was evaluated by comparing the scores among patients with different health states, and sensitivity was compared using relative efficiency and the effect size.

Results: A total of 378 patients (male, 49.5%; mean age, 49.1 years) were included in this study. The ICC between EQ-5D-5L and SF-6Dv2 utility values was 0.639. The correlation between the two scales was strong (0.767). Both scales showed known groups validity, although the SF-6Dv2 was more sensitive. The differences in the SF-6Dv2 scores for patients in better and worse health state were greater than those measured by the EQ-5D-5L scores.

Conclusions: Both EQ-5D-5L and SF-6Dv2 instruments are valid for dialysis patients. However, the two scales cannot be used interchangeably, and it appears that the SF-6Dv2 was more sensitive in capturing health state differences for dialysis patients in China.

Background: To measure and value the health-related quality of life (HRQOL) of Chinese Multiple sclerosis (MS) patients, and to explore the influencing factors.

Methods: A national online survey amongst the patients was conducted in July 2022. Information on patient demographics, clinical characteristics (MS subtype and disability status), and the HRQOL measured by EQ-5D-5L were collected. Disability status was classified based on the self-reported Expanded Disability Status Scale (EDSS). Both the Tobit and logistic regression models were adopted to assess the impact of clinical characteristics on HRQOL with the adjustment of demographic characteristics.

Results: A total of 733 eligible patients (mean age 34.31 years; 67.7% females) were included in the analysis. The mean EQ-5D-5L health utility score (HUS) and VAS score were 0.73(SD:0.29) and 68.7(SD:23.9), respectively. MS patients reported a high prevalence of health problems in all the EQ-5D dimensions, ranging from 25.6% (self-care) to 74.9% (anxiety/depression). According to Tobit regression analyses, secondary progressive MS was associated with lower VAS compared to relapsing-remitting MS. Worse disability status measured by EDSS was related to the lower HUS and VAS scores. Demographic characteristics including younger age, higher education level, and employment were also identified as protective factors of the patients' HRQOL.

Conclusions: This study assessed HRQOL of Chinese MS patients using the EQ-5D-5L. Our finding revealed that MS subtypes and disability status are the most important factors affecting their HRQOL.