Comparison of Machine Learning Techniques for Mortality Prediction in a Prospective Cohort of Older Adults.

As global demographics change, ageing is a global phenomenon which is increasingly of interest in our modern and rapidly changing society. Thus, the application of proper prognostic indices in clinical decisions regarding mortality prediction has assumed a significant importance for personalized risk management (i.e., identifying patients who are at high or low risk of death) and to help ensure effective healthcare services to patients. Consequently, prognostic modelling expressed as all-cause mortality prediction is an important step for effective patient management. Machine learning has the potential to transform prognostic modelling. In this paper, results on the development of machine learning models for all-cause mortality prediction in a cohort of healthy older adults are reported. The models are based on features covering anthropometric variables, physical and lab examinations, questionnaires, and lifestyles, as well as wearable data collected in free-living settings, obtained for the "Healthy Ageing Initiative" study conducted on 2291 recruited participants. Several machine learning techniques including feature engineering, feature selection, data augmentation and resampling were investigated for this purpose. A detailed empirical comparison of the impact of the different techniques is presented and discussed. The achieved performances were also compared with a standard epidemiological model. This investigation showed that, for the dataset under consideration, the best results were achieved with Random UnderSampling in conjunction with Random Forest (either with or without probability calibration). However, while including probability calibration slightly reduced the average performance, it increased the model robustness, as indicated by the lower 95% confidence intervals. The analysis showed that machine learning models could provide comparable results to standard epidemiological models while being completely data-driven and disease-agnostic, thus demonstrating the opportunity for building machine learning models on health records data for research and clinical practice. However, further testing is required to significantly improve the model performance and its robustness.

Investigation of the analysis of wearable data for cancer-specific mortality prediction in older adults.

Cancer is an aggressive disease which imparts a tremendous socio-economic burden on the international community. Early detection is an important aspect in improving survival rates for cancer sufferers; however, very few studies have investigated the possibility of predicting which people have the highest risk to develop this disease, even years before the traditional symptoms first occur. In this paper, a dataset from a longitudinal study which was collected among 2291 70-year olds in Sweden has been analyzed to investigate the possibility for predicting 2-7 year cancer-specific mortality. A tailored ensemble model has been developed to tackle this highly imbalanced dataset. The performance with different feature subsets has been investigated to evaluate the impact that heterogeneous data sources may have on the overall model. While a full-features model shows an Area Under the ROC Curve (AUC-ROC) of 0.882, a feature subset which only includes demographics, self-report health and lifestyle data, and wearable dataset collected in free-living environments presents similar performance (AUC-ROC: 0.857). This analysis confirms the importance of wearable technology for providing unbiased health markers and suggests its possible use in the accurate prediction of 2-7 year cancer-related mortality in older adults.