Potential utility of physical function measures to improve the risk prediction of functional disability in community-dwelling older Japanese adults: a prospective study

Mortality and disability risk among older adults unable to complete grip strength and physical performance tests: a population-based cohort study from China

BACKGROUND

The link between low grip strength, diminished physical performance, and adverse health outcomes in older adults has been well-established. However, the impact of older adults who cannot complete these tests on disability and mortality rates remains unexplored without longitudinal study.

METHODS

We collected data from the China Health and Retirement Longitudinal Study (CHARLS). Participants aged 60-101 were enrolled at baseline. We analyzed the prevalence of populations unable to complete handgrip strength (HGS), gait speed (GS), and five times chair stand test (FTCST). Completing risk models were used to estimate the risk of mortality and disability over seven years.

RESULTS

A total of 3,768 participants were included in the analysis. The percentage of older adults unable to complete the GS and FTCST tests increased notably with age, from 2.68 to 8.90% and 2.60-20.42%, respectively. The proportion of older people unable to perform the HGS was relatively stable, ranging from 1.40 to 3.66%. Compared to older adults who can complete these tests, those who cannot perform FTCST face a significantly higher risk of mortality, with 49.1% higher risk [hazard ratio (HR) = 1.491, 95% CI = 1.156, 1.922; subdistribution hazard ratio (SHR) = 1.491, 95%CI = 1.135,1.958)]. Participants who were unable to complete the GS test had a higher risk of developing ADL disability, regardless of whether they were compared to the lowest-performing group (HR = 1.411, 95%CI = 1.037,1.920; SHR = 1.356, 95%CI = 1.030,1.785) or those who can complete the GS (HR = 1.727, 95%CI = 1.302,2.292; SHR = 1.541, 95%CI = 1.196,1.986). No statistically significant difference in the risk of developing ADL disability among older adults who were unable to complete the HGS test compared with either the poorest performing group (HR = 0.982, 95% CI = 0.578, 1.666; SHR = 1.025, 95% CI = 0.639, 1.642) or those who were able to complete the HGS test (HR = 1.008, 95% CI = 0.601, 1.688; SHR = 0.981, 95% CI = 0.619, 1.553). The risk of all-cause mortality was not significantly different for older adults who were unable to complete the HGS test compared to those with the worst performance (HR = 1.196, 95%CI = 0.709-2.020; SHR = 1.196, 95%CI = 0.674, 2.124) or those who were able to complete the test (HR = 1.462, 95%CI = 0.872-2.450; SHR = 1.462, 95%CI = 0.821,2.605).

CONCLUSION

The risks of adverse events faced by older adults unable to complete the tests vary, indicating the necessity for future research to conduct separate analyses on this high-risk population.

Utilization of Japanese long-term care-related data including Kaigo-DB: An analysis of current trends and future directions

Despite high expectations from the government and researchers regarding data utilization, comprehensive analysis of long-term care (LTC)-related data use has been limited. This study reviewed the use of LTC-related data, including Kaigo-DB, in Japan after 2020. There was an increase in studies using LTC-related data in Japan between 2020 and 2021, followed by a stabilization period. The national government provided 13.5% of this data (6.5% from Kaigo-DB), while prefectures and municipalities contributed 85.2%, and facilities provided 1.3%. The linked data used in 90.4% of the studies primarily consisted of original questionnaire or interview surveys (34.6%) and medical claims (34.0%). None of the studies based on Kaigo-DB utilized linked data. In terms of study design, cohort studies were the most common (84.6%), followed by descriptive (5.1%), cross-sectional (3.2%), and case-control studies (1.3%). Among the 138 individual-based analytical descriptive studies, the most frequently used LTC-related data as an exposure was LTC services (26.8%), and the most common data used as an outcome was LTC certification or care need level (43.5%), followed by the independence degree of daily living for the older adults with dementia (18.1%). To enhance the use of LTC-related data, especially the valuable national Kaigo-DB, insights can be gleaned from how researchers effectively utilize municipal and prefectural data. Streamlining access to Kaigo-DB and enabling its linkage with other datasets are promising for future research in this field.

Disability risk prediction model based on machine learning among Chinese healthy older adults: results from the China Health and Retirement Longitudinal Study

Background

Predicting disability risk in healthy older adults in China is essential for timely preventive interventions, improving their quality of life, and providing scientific evidence for disability prevention. Therefore, developing a machine learning model capable of evaluating disability risk based on longitudinal research data is crucial.

Methods

We conducted a prospective cohort study of 2,175 older adults enrolled in the China Health and Retirement Longitudinal Study (CHARLS) between 2015 and 2018 to develop and validate this prediction model. Several machine learning algorithms (logistic regression, k-nearest neighbors, naive Bayes, multilayer perceptron, random forest, and XGBoost) were used to assess the 3-year risk of developing disability. The optimal cutoff points and adjustment parameters are explored in the training set, the prediction accuracy of the models is compared in the testing set, and the best-performing models are further interpreted.

Results

During a 3-year follow-up period, a total of 505 (23.22%) healthy older adult individuals developed disabilities. Among the 43 features examined, the LASSO regression identified 11 features as significant for model establishment. When comparing six different machine learning models on the testing set, the XGBoost model demonstrated the best performance across various evaluation metrics, including the highest area under the ROC curve (0.803), accuracy (0.757), sensitivity (0.790), and F1 score (0.789), while its specificity was 0.712. The decision curve analysis (DCA) indicated showed that XGBoost had the highest net benefit in most of the threshold ranges. Based on the importance of features determined by SHAP (model interpretation method), the top five important features were identified as right-hand grip strength, depressive symptoms, marital status, respiratory function, and age. Moreover, the SHAP summary plot was used to illustrate the positive or negative effects attributed to the features influenced by XGBoost. The SHAP dependence plot explained how individual features affected the output of the predictive model.

Conclusion

Machine learning-based prediction models can accurately evaluate the likelihood of disability in healthy older adults over a period of 3 years. A combination of XGBoost and SHAP can provide clear explanations for personalized risk prediction and offer a more intuitive understanding of the effect of key features in the model.

Machine Learning-Based Prediction of Functional Disability: a Cohort Study of Japanese Older Adults in 2013-2019

BACKGROUND

It is important to identify older adults at high risk of functional disability and to take preventive measures for them at an early stage. To our knowledge, there are no studies that predict functional disability among community-dwelling older adults using machine learning algorithms.

OBJECTIVE

To construct a model that can predict functional disability over 5 years using basic machine learning algorithms.

DESIGN

A cohort study with a mean follow-up of 5.4 years.

PARTICIPANTS

We used data from the Japan Gerontological Evaluation Study, which involved 73,262 people aged  ≥ 65 years who were not certified as requiring long-term care. The baseline survey was conducted in 2013 in 19 municipalities.

MAIN MEASURES

We defined the onset of functional disability as the new certification of needing long-term care that was ascertained by linking participants to public registries of long-term care insurance. All 183 candidate predictors were measured by self-report questionnaires.

KEY RESULTS

During the study period, 16,361 (22.3%) participants experienced the onset of functional disability. Among machine learning-based models, ridge regression (C statistic = 0.818) and gradient boosting (0.817) effectively predicted functional disability. In both models, we identified age, self-rated health, variables related to falls and posture stabilization, and diagnoses of Parkinson's disease and dementia as important features. Additionally, the ridge regression model identified the household characteristics such as the number of members, income, and receiving public assistance as important predictors, while the gradient boosting model selected moderate physical activity and driving. Based on the ridge regression model, we developed a simplified risk score for functional disability, and it also indicated good performance at the cut-off of 6/7 points.

CONCLUSIONS

Machine learning-based models showed effective performance prediction over 5 years. Our findings suggest that measuring and adding the variables identified as important features can improve the prediction of functional disability.

Effects of visual restoration on gait performance and kinematics of lower extremities in patients with age-related cataract

BACKGROUND

Visual inputs are critical for locomotor navigation and sensorimotor integration in the elderly; however, the mechanism needs to be explored intensively. The present study assessed the gait pattern after cataract surgery to investigate the effects of visual restoration on locomotion.

METHODS

The prospective study recruited 32 patients (70.1 ± 5.2 years) with bilateral age-related cataracts in the Department of Ophthalmology at Peking University Third Hospital from October 2016 to December 2019. The temporal-spatial gait parameters and kinematic parameters were measured by the Footscan system and inertial measurement units. Paired t -test was employed to compare data normally distributed and Wilcoxon rank-sum test for non-normally distributed.

RESULTS

After visual restoration, the walking speed increased by 9.3% (1.19 ± 0.40 m/s vs. 1.09 ± 0.34 m/s, P =0.008) and exhibited an efficient gait pattern with significant decrease in gait cycle (1.02 ± 0.08 s vs. 1.04 ± 0.07 s, P =0.012), stance time (0.66 ± 0.06 s vs. 0.68 ± 0.06 s, P =0.045), and single support time (0.36 ± 0.03 s vs. 0.37 ± 0.02 s, P =0.011). High amplitude of joint motion was detected in the sagittal plane in the left hip (37.6° ± 5.3° vs. 35.5° ± 6.2°, P =0.014), left thigh (38.0° ± 5.2° vs. 36.4° ± 5.8°, P =0.026), left shank (71.9° ± 5.7° vs. 70.1° ± 5.6°, P =0.031), and right knee (59.1° ± 4.8° vs. 56.4° ± 4.8°, P =0.001). The motor symmetry of thigh improved from 8.35 ± 5.30% to 6.30 ± 4.73% ( P =0.042).

CONCLUSIONS

The accelerated gait in response to visual restoration is characterized by decreased stance time and increased range of joint motion. Training programs for improving muscle strength of lower extremities might be helpful to facilitate the adaptation to these changes in gait.

Novel physical performance-based models for activities of daily living disability prediction among Chinese older community population: a nationally representative survey in China

Background

Physical performances including upper and lower limb functions have predictive roles in activities of daily living (ADL) disability, but they have rarely been incorporated into prediction models. This study primarily aimed to develop and validate novel physical performance-based models for ADL disability among Chinese older adults. Comparisons of predictive performance across multiple models were performed, and model simplification was further explored.

Methods

Data were obtained from the China Health and Retirement Longitudinal Study in the 2011 and 2015 waves, containing 2192 older adults over 60 years old. Our models were constructed by logistic regression analysis, using a backward stepwise selection. Model performance was internally validated by discrimination, calibration, and clinical utility. Integrated Discrimination Improvement (IDI) and Net Reclassification Improvement (NRI) were used to assess the incremental benefit of the extended models. Moreover, nomograms were built for visualization.

Results

We selected gender, age, smoking, self-report health condition, BMI, depressive symptoms, and cognitive function into the fundamental model (Model 1). Based on Model 1, five novel prediction models were constructed by adding handgrip strength (Model 2), Short Physical Performance Battery (SPPB) (Model 3), gait speed (Model 4), handgrip strength plus SPPB (Model 5), and handgrip strength plus gait speed (Model 6), respectively. Significant improvement in predictive values were observed for all five novel models compared with Model 1 (C-index = 0.693). The lower limb model (Model 3 SPPB model: C-index = 0.731) may play a key role in the prediction of ADL disability, reflecting a comparable predictive value to the comprehensive models combining both upper and lower limbs (Model 5 handgrip strength + SPPB model: C-index = 0.732). When we simplified the lower limb models by replacing SPPB with gait speed, the predictive values attenuated slightly (C-index: Model 3 vs Model 4: 0.731 vs 0.714; Model 5 vs Model 6: 0.732 vs 0.718), but still better than the upper limb model (Model 2 handgrip strength model: C-index = 0.701).

Conclusions

Physical performance-based models, especially lower limb model, provided improved prediction for ADL disability among Chinese older adults, which may help guide the targeted intervention.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-022-02905-y.