An machine learning model to predict quality of life subtypes of disabled stroke survivors.

OBJECTIVE: Stroke causes serious physical disability with impaired quality of life (QoL) and heavy burden on health. The goal of this study is to explore the impaired QoL typologies and their predicting factors in physically disabled stroke survivors with machine learning approach. METHODS: Non-negative matrix factorization (NMF) was applied to clustering 308 physically disabled stroke survivors in rural China based on their responses on the short form 36 (SF-36) assessment of quality of life. Principal component analysis (PCA) was conducted to differentiate the subtypes, and the Boruta algorithm was used to identify the variables relevant to the categorization of two subtypes. A gradient boosting machine(GBM) and local interpretable model-agnostic explanation (LIME) algorithms were used to apply to interpret the variables that drove subtype predictions. RESULTS: Two distinct subtypes emerged, characterized by short form 36 (SF-36) domains. The feature difference between worsen QoL subtype and better QoL subtype was as follows: role-emotion (RE), body pain (BP) and general health (GH), but not physical function (PF); the most relevant predictors of worsen QoL subtypes were help from others, followed by opportunities for community activity and rehabilitation needs, rather than disability severity or duration since stroke. INTERPRETATION: The results suggest that the rehabilitation programs should be tailored toward their QoL clustering feature; body pain and emotional-behavioral problems are more crucial than motor deficit; stroke survivors with worsen QoL subtype are most in need of social support, return to community, and rehabilitation.

Factors Influencing Disabled Stroke Survivors' Quality of Life in Rural China: Based on the Structural Characteristics and Psychometric Properties of the SF-36 Assessment.

Many stroke survivors' quality of life is impaired. Few studies of factors influencing their quality of life have been based on the factors tested by the short form 36 instrument. This study did so with 308 physically disabled stroke survivors in rural China. Principal components analysis was applied to refine the dimension structure of the short form 36 assessment, followed by backward multiple linear regression analysis to determine the independent factors influencing quality of life. The structure revealed differed from the generic structure in showing that the mental health and vitality dimensions are not unidimensional. Subjects who reported access to the outdoors as convenient demonstrated better quality of life in all dimensions. Those who exercised regularly achieved better social functioning and negative mental health scores. Other factors influencing a better quality of life in terms of physical functioning were younger age and not being married. Being older and better educated predicted better role-emotion scores. Being female correlated with better social functioning scores, while men scored better on bodily pain. Being less educated predicted higher negative mental health, while being less disabled predicted better physical and social functioning. The results suggest that the SF-36's dimension structure should be re-evaluated before using it to assess stroke survivors.

Predicting Intention to Participate in Community Physical Activities for Adults with Physical Disabilities.

Structural equation modeling was used to derive a relationship predicting the intention to participate in community physical activity among community-dwelling adults with a physical disability in Xiamen, China. The data were collected in a cross-sectional survey. The structural equation modeling combined biomedicine and the theory of planned behavior. It integrated ratings using the rehabilitation set from the international classification of functioning, disability, and health and role-physical scores from the short form 36 health survey questionnaire instrument. The model demonstrated a good ability to predict self-reported participation intentions, explaining 62% of the variance. The standard coefficients showed that activity limitation (27%), role-physical score (21%) and body impairment (14%) were the most influential predictors. ICF-RS ratings and role-physical ratings together can usefully predict physically disabled adults' intention of participating in community physical activities. Suggestions are presented for multidisciplinary intervention and improving this portion of the WHO's classification system.

Analysis of Immune-Stromal Score-Based Gene Signature and Molecular Subtypes in Osteosarcoma: Implications for Prognosis and Tumor Immune Microenvironment.

Objective: Infiltrating immune and stromal cells are essential for osteosarcoma progression. This study set out to analyze immune-stromal score-based gene signature and molecular subtypes in osteosarcoma. Methods: The immune and stromal scores of osteosarcoma specimens from the TARGET cohort were determined by the ESTIMATE algorithm. Then, immune-stromal score-based differentially expressed genes (DEGs) were screened, followed by univariate Cox regression analysis. A LASSO regression analysis was applied for establishing a prognostic model. The predictive efficacy was verified in the GSE21257 dataset. Associations between the risk scores and chemotherapy drug sensitivity, immune/stromal scores, PD-1/PD-L1 expression, immune cell infiltrations were assessed in the TARGET cohort. NMF clustering analysis was employed for characterizing distinct molecular subtypes based on immune-stromal score-based DEGs. Results: High immune/stromal scores exhibited the prolonged survival duration of osteosarcoma patients. Based on 85 prognosis-related stromal-immune score-based DEGs, a nine-gene signature was established. High-risk scores indicated undesirable prognosis of osteosarcoma patients. The AUCs of overall survival were 0.881 and 0.849 in the TARGET cohort and GSE21257 dataset, confirming the well predictive performance of this signature. High-risk patients were more sensitive to doxorubicin and low-risk patients exhibited higher immune/stromal scores, PD-L1 expression, and immune cell infiltrations. Three molecular subtypes were characterized, with distinct clinical outcomes and tumor immune microenvironment. Conclusion: This study developed a robust prognostic gene signature as a risk stratification tool and characterized three distinct molecular subtypes for osteosarcoma patients based on immune-stromal score-based DEGs, which may assist decision-making concerning individualized therapy and follow-up project.