No abstract

The field of fall risk testing using wearable sensors is bustling with activity. In this Letter, the authors review publications which incorporated features extracted from sensor signals into statistical models intended to estimate fall risk or predict falls in older people. A review of these studies raises concerns that this body of literature is presenting over-optimistic results in light of small sample sizes, questionable modelling decisions and problematic validation methodologies (e.g. inherent problems with the overly-popular cross-validation technique, lack of external validation). There seem to be substantial issues in the feature selection process, whereby researchers select features before modelling begins based on their relation to the target, and either perform no validation or test the models on the same data used for their training. This, together with potential issues related to the large number of features and their correlations, inevitably leads to models with inflated accuracy that are unlikely to maintain their reported performance during everyday use in relevant populations. Indeed, the availability of rich sensor data and many analytical options provides intellectual and creative freedom for researchers, but should be treated with caution, and such pitfalls must be avoided if we desire to create generalisable prognostic tools of any clinical value.

Introduction This was a pilot study to examine the effects of home telemonitoring (TM) of patients with severe chronic obstructive pulmonary disease (COPD). Methods A randomised controlled 12-month trial of 42 patients with severe COPD was conducted. Home TM of oximetry, temperature, pulse, electrocardiogram, blood pressure, spirometry, and weight with telephone support and home visits was tested against a control group receiving only identical telephone support and home visits. Results The results suggest that TM had a reduction in COPD-related admissions, emergency department presentations, and hospital bed days. TM also seemed to increase the interval between COPD-related exacerbations requiring a hospital visit and prolonged the time to the first admission. The interval between hospital visits was significantly different between the study arms, while the other findings did not reach significance and only suggest a trend. There was a reduction in hospital admission costs. TM was adopted well by most patients and eventually, also by the nursing staff, though it did not seem to change patients' psychological well-being. Discussion Ability to draw firm conclusions is limited due to the small sample size. However the trends of reducing hospital visits warrant a larger study of a similar design. When designing such a trial, one should consider the potential impact of the high quality of care already made available to this patient cohort.

Identification of older people most at risk of falling may facilitate early preventative intervention to reduce the likelihood of falls occurring. While many clinical fall risk assessment techniques exist, they often require subjective assessor interpretation, or are not appropriate for unsupervised screening of larger populations owing to a number of issues including safety, ability to reliably perform the assessment, and requirements for unwieldy apparatus. Researchers have more recently attempted to address some of these deficits by instrumenting new or existing physical fall risk assessments with wearable motion sensors to make such assessments more objective, quicker to administer, and potentially more appropriate for deployment for unsupervised use in the community. The objective of this paper is to discuss various practical questions involving sensor-based fall risk assessment (SFRA). Many of the issues discussed contribute to answering the important question of whether SFRA should or can be used in either a supervised or an unsupervised manner, and what possible deployment scenarios exist for it.

In developed countries, chronic disease now accounts for more than 75% of health care expenditure and nearly an equivalent percentage of disease-related deaths [1]. The burden of chronic disease (often, but not exclusively, associated with ageing) includes congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), hypertension and diabetes. Over the past several decades there has been an epidemiological shift in disease burden from acute to chronic diseases that has rendered acute care models of health service delivery inadequate to address population health needs.

Falls are a common and serious problem faced by older populations. There is a growing interest in estimating the risk of falling for older people using body-worn sensors and simple movement tasks, allowing appropriate fall prevention programs to be administered in a timely manner to the high risk population. This study investigated the capability and validity of using a waist-mounted triaxial accelerometer (TA) and a directed routine (DR) that includes three movement tasks to discriminate between fallers and non-fallers and between multiple fallers and non-multiple fallers. Data were collected from 98 subjects who were stratified into two separate groups, one for model training and the other for model validation. Logistic regression models were constructed using the TA features from the entire DR and from each single DR task, and were validated using unseen data. The best models were obtained using features from the alternate step test to classify between fallers and non-fallers with κ = 0.34-0.41, sensitivity = 68%-71% and specificity = 63%-73%. However, the overall validation performances were poor. The study emphasizes the importance of independent validation in fall prediction studies.

Chronic obstructive pulmonary disease (COPD) is responsible for significant morbidity and mortality worldwide. Recent clinical research has indicated a strong association between physiological homeostasis and the onset of COPD exacerbation. Thus the analysis of these variables may yield a means of predicting a COPD exacerbation in the near future. However, the accuracy of existing prediction methods based on statistical analysis of periodic snapshots of physiological variables is still far from satisfactory, due to lack of integration of long-term and interactive effects of the physiological variables. Therefore, developing a relatively accurate method for predicting COPD exacerbation is an outstanding challenge. In this paper, a regression-based machine learning technique was developed, using trend pattern variables extracted from COPD patients' longitudinal physiological records, to classify subjects into "low-risk" and "high-risk" categories, indicating their risk of suffering a COPD exacerbation event. Experimental results from cross validation assessment of the classifier model show an average accuracy of 79.27% using this method.