Toward Enhanced Clinical Decision Support for Patients Undergoing a Hip or Knee Replacement: Focus Group and Interview Study With Surgeons.

BACKGROUND: The current assessment of recovery after total hip or knee replacement is largely based on the measurement of health outcomes through self-report and clinical observations at follow-up appointments in clinical settings. Home activity-based monitoring may improve assessment of recovery by enabling the collection of more holistic information on a continuous basis. OBJECTIVE: This study aimed to introduce orthopedic surgeons to time-series analyses of patient activity data generated from a platform of sensors deployed in the homes of patients who have undergone primary total hip or knee replacement and understand the potential role of these data in postoperative clinical decision-making. METHODS: Orthopedic surgeons and registrars were recruited through a combination of convenience and snowball sampling. Inclusion criteria were a minimum required experience in total joint replacement surgery specific to the hip or knee or familiarity with postoperative recovery assessment. Exclusion criteria included a lack of specific experience in the field. Of the 9 approached participants, 6 (67%) orthopedic surgeons and 3 (33%) registrars took part in either 1 of 3 focus groups or 1 of 2 interviews. Data were collected using an action-based approach in which stimulus materials (mock data visualizations) provided imaginative and creative interactions with the data. The data were analyzed using a thematic analysis approach. RESULTS: Each data visualization was presented sequentially followed by a discussion of key illustrative commentary from participants, ending with a summary of key themes emerging across the focus group and interview data set. CONCLUSIONS: The limitations of the evidence are as follows. The data presented are from 1 English hospital. However, all data reflect the views of surgeons following standard national approaches and training. Although convenience sampling was used, participants' background, skills, and experience were considered heterogeneous. Passively collected home monitoring data offered a real opportunity to more objectively characterize patients' recovery from surgery. However, orthopedic surgeons highlighted the considerable difficulty in navigating large amounts of complex data within short medical consultations with patients. Orthopedic surgeons thought that a proposed dashboard presenting information and decision support alerts would fit best with existing clinical workflows. From this, the following guidelines for system design were developed: minimize the risk of misinterpreting data, express a level of confidence in the data, support clinicians in developing relevant skills as time-series data are often unfamiliar, and consider the impact of patient engagement with data in the future. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): RR2-10.1136/bmjopen-2018-021862.

A multi-sensor dataset with annotated activities of daily living recorded in a residential setting.

SPHERE is a large multidisciplinary project to research and develop a sensor network to facilitate home healthcare by activity monitoring, specifically towards activities of daily living. It aims to use the latest technologies in low powered sensors, internet of things, machine learning and automated decision making to provide benefits to patients and clinicians. This dataset comprises data collected from a SPHERE sensor network deployment during a set of experiments conducted in the 'SPHERE House' in Bristol, UK, during 2016, including video tracking, accelerometer and environmental sensor data obtained by volunteers undertaking both scripted and non-scripted activities of daily living in a domestic residence. Trained annotators provided ground-truth labels annotating posture, ambulation, activity and location. This dataset is a valuable resource both within and outside the machine learning community, particularly in developing and evaluating algorithms for identifying activities of daily living from multi-modal sensor data in real-world environments. A subset of this dataset was released as a machine learning competition in association with the European Conference on Machine Learning (ECML-PKDD 2016).

Collecting Food and Drink Intake Data With Voice Input: Development, Usability, and Acceptability Study.

BACKGROUND: Voice-based systems such as Amazon Alexa may be useful for collecting self-reported information in real time from participants of epidemiology studies using verbal input. In epidemiological research studies, self-reported data tend to be collected using short, infrequent questionnaires, in which the items require participants to select from predefined options, which may lead to errors in the information collected and lack of coverage. Voice-based systems give the potential to collect self-reported information "continuously" over several days or weeks. At present, to the best of our knowledge, voice-based systems have not been used or evaluated for collecting epidemiological data. OBJECTIVE: We aimed to demonstrate the technical feasibility of using Alexa to collect information from participants, investigate participant acceptability, and provide an initial evaluation of the validity of the collected data. We used food and drink information as an exemplar. METHODS: We recruited 45 staff members and students at the University of Bristol (United Kingdom). Participants were asked to tell Alexa what they ate or drank for 7 days and to also submit this information using a web-based form. Questionnaires asked for basic demographic information, about their experience during the study, and the acceptability of using Alexa. RESULTS: Of the 37 participants with valid data, most (n=30, 81%) were aged 20 to 39 years and 23 (62%) were female. Across 29 participants with Alexa and web entries corresponding to the same intake event, 60.1% (357/588) of Alexa entries contained the same food and drink information as the corresponding web entry. Most participants reported that Alexa interjected, and this was worse when entering the food and drink information (17/35, 49% of participants said this happened often; 1/35, 3% said this happened always) than when entering the event date and time (6/35, 17% of participants said this happened often; 1/35, 3% said this happened always). Most (28/35, 80%) said they would be happy to use a voice-controlled system for future research. CONCLUSIONS: Although there were some issues interacting with the Alexa skill, largely because of its conversational nature and because Alexa interjected if there was a pause in speech, participants were mostly willing to participate in future research studies using Alexa. More studies are needed, especially to trial less conversational interfaces.

Human Activity Recognition Based on Dynamic Active Learning.

Activity of daily living is an important indicator of the health status and functional capabilities of an individual. Activity recognition, which aims at understanding the behavioral patterns of people, has increasingly received attention in recent years. However, there are still a number of challenges confronting the task. First, labelling training data is expensive and time-consuming, leading to limited availability of annotations. Secondly, activities performed by individuals have considerable variability, which renders the generally used supervised learning with a fixed label set unsuitable. To address these issues, we propose a dynamic active learning-based activity recognition method in this work. Different from traditional active learning methods which select samples based on a fixed label set, the proposed method not only selects informative samples from known classes, but also dynamically identifies new activities which are not included in the predefined label set. Starting with a classifier that has access to a limited number of labelled samples, we iteratively extend the training set with informative labels by fully considering the uncertainty, diversity and representativeness of samples, based on which better-informed classifiers can be trained, further reducing the annotation cost. We evaluate the proposed method on two synthetic datasets and two existing benchmark datasets. Experimental results demonstrate that our method not only boosts the activity recognition performance with considerably reduced annotation cost, but also enables adaptive daily activity analysis allowing the presence and detection of novel activities and patterns.

GLU: a software package for analysing continuously measured glucose levels in epidemiology.

Continuous glucose monitors (CGM) record interstitial glucose levels 'continuously', producing a sequence of measurements for each participant (e.g. the average glucose level every 5 min over several days, both day and night). To analyse these data, researchers tend to derive summary variables such as the area under the curve (AUC), to then use in subsequent analyses. To date, a lack of consistency and transparency of precise definitions used for these summary variables has hindered interpretation, replication and comparison of results across studies. We present GLU, an open-source software package for deriving a consistent set of summary variables from CGM data. GLU performs quality control of each CGM sample (e.g. addressing missing data), derives a diverse set of summary variables (e.g. AUC and proportion of time spent in hypo-, normo- and hyper- glycaemic levels) covering six broad domains, and outputs these (with quality control information) to the user. GLU is implemented in R and is available on GitHub at https://github.com/MRCIEU/GLU. Git tag v0.2 corresponds to the version presented here.

Modelling Patient Behaviour Using IoT Sensor Data: a Case Study to Evaluate Techniques for Modelling Domestic Behaviour in Recovery from Total Hip Replacement Surgery.

The UK health service sees around 160,000 total hip or knee replacements every year and this number is expected to rise with an ageing population. Expectations of surgical outcomes are changing alongside demographic trends, whilst aftercare may be fractured as a result of resource limitations. Conventional assessments of health outcomes must evolve to keep up with these changing trends. Health outcomes may be assessed largely by self-report using Patient Reported Outcome Measures (PROMs), such as the Oxford Hip or Oxford Knee Score, in the months up to and following surgery. Though widely used, many PROMs have methodological limitations and there is debate about how to interpret results and definitions of clinically meaningful change. With the development of a home-monitoring system, there is opportunity to characterise the relationship between PROMs and behaviour in a natural setting and to develop methods of passive monitoring of outcome and recovery after surgery. In this paper, we discuss the motivation and technology used in long-term continuous observation of movement, sleep and domestic routine for healthcare applications, such as the HEmiSPHERE project for hip and knee replacement patients. In this case study, we evaluate trends evident in data of two patients, collected over a 3-month observation period post-surgery, by comparison with scores from PROMs for sleep and movement quality, and by comparison with a third control home. We find that accelerometer and indoor localisation data correctly highlight long-term trends in sleep and movement quality and can be used to predict sleep and wake times and measure sleep and wake routine variance over time, whilst indoor localisation provides context for the domestic routine and mobility of the patient. Finally, we discuss a visual method of sharing findings with healthcare professionals.

Activities of Daily Living Ontology for Ubiquitous Systems: Development and Evaluation.

Ubiquitous eHealth systems based on sensor technologies are seen as key enablers in the effort to reduce the financial impact of an ageing society. At the heart of such systems sit activity recognition algorithms, which need sensor data to reason over, and a ground truth of adequate quality used for training and validation purposes. The large set up costs of such research projects and their complexity limit rapid developments in this area. Therefore, information sharing and reuse, especially in the context of collected datasets, is key in overcoming these barriers. One approach which facilitates this process by reducing ambiguity is the use of ontologies. This article presents a hierarchical ontology for activities of daily living (ADL), together with two use cases of ground truth acquisition in which this ontology has been successfully utilised. Requirements placed on the ontology by ongoing work are discussed.

Presenteeism in a Dutch hand eczema population-a cross-sectional survey.

BACKGROUND: Presenteeism (attending work despite complaints and ill health, which should prompt rest and absence) has been overlooked in the field of hand eczema. OBJECTIVES: To examine the 1-year prevalence of presenteeism related to hand eczema in a population of hand eczema patients who visited a tertiary referral centre. Secondary objectives: to identify intrinsic/extrinsic reasons for presenteeism and to evaluate associated factors. METHODS: This was a cross-sectional questionnaire study. Presenteeism was defined as "going to work despite feeling you should have taken sick leave because of hand eczema". Respondents answered questions about socio-demographic factors, clinical features, occupational characteristics, and hand eczema related to occupational exposure. RESULTS: Forty-one per cent (141/346) of patients who had both worked and had hand eczema during the past 12 months reported presenteeism. The most often reported reasons were: "Because I do not want to give in to my impairment/weakness" (46%) and "Because I enjoy my work" (40%). Presenteeism was associated with: mean hand eczema severity; absenteeism because of hand eczema; improvement of hand eczema when away from work; and high-risk occupations. CONCLUSIONS: In this study, presenteeism was common and predominantly observed in patients with more severe hand eczema and occupational exposure. The most frequently reported reasons for presenteeism were of an intrinsic nature.

The Work Role Functioning Questionnaire v2.0 Showed Consistent Factor Structure Across Six Working Samples.

Objective The Work Role Functioning Questionnaire v2.0 (WRFQ) is an outcome measure linking a persons' health to the ability to meet work demands in the twenty-first century. We aimed to examine the construct validity of the WRFQ in a heterogeneous set of working samples in the Netherlands with mixed clinical conditions and job types to evaluate the comparability of the scale structure. Methods Confirmatory factor and multi-group analyses were conducted in six cross-sectional working samples (total N = 2433) to evaluate and compare a five-factor model structure of the WRFQ (work scheduling demands, output demands, physical demands, mental and social demands, and flexibility demands). Model fit indices were calculated based on RMSEA ≤ 0.08 and CFI ≥ 0.95. After fitting the five-factor model, the multidimensional structure of the instrument was evaluated across samples using a second order factor model. Results The factor structure was robust across samples and a multi-group model had adequate fit (RMSEA = 0.63, CFI = 0.972). In sample specific analyses, minor modifications were necessary in three samples (final RMSEA 0.055-0.080, final CFI between 0.955 and 0.989). Applying the previous first order specifications, a second order factor model had adequate fit in all samples. Conclusion A five-factor model of the WRFQ showed consistent structural validity across samples. A second order factor model showed adequate fit, but the second order factor loadings varied across samples. Therefore subscale scores are recommended to compare across different clinical and working samples.

Physical activity phenotyping with activity bigrams, and their association with BMI.

Background: Analysis of physical activity usually focuses on a small number of summary statistics derived from accelerometer recordings: average counts per minute and the proportion of time spent in moderate-vigorous physical activity or in sedentary behaviour. We show how bigrams, a concept from the field of text mining, can be used to describe how a person's activity levels change across (brief) time points. These variables can, for instance, differentiate between two people spending the same time in moderate activity, where one person often stays in moderate activity from one moment to the next and the other does not. Methods: We use data on 4810 participants of the Avon Longitudinal Study of Parents and Children (ALSPAC). We generate a profile of bigram frequencies for each participant and test the association of each frequency with body mass index (BMI), as an exemplar. Results: We found several associations between changes in bigram frequencies and BMI. For instance, a one standard deviation decrease in the number of adjacent minutes in sedentary then moderate activity (or vice versa), with a corresponding increase in the number of adjacent minutes in moderate then vigorous activity (or vice versa), was associated with a 2.36 kg/m2 lower BMI [95% confidence interval (CI): -3.47, -1.26], after accounting for the time spent in sedentary, low, moderate and vigorous activity. Conclusions: Activity bigrams are novel variables that capture how a person's activity changes from one moment to the next. These variables can be used to investigate how sequential activity patterns associate with other traits.

Smart homes, private homes? An empirical study of technology researchers' perceptions of ethical issues in developing smart-home health technologies.

BACKGROUND: Smart-home technologies, comprising environmental sensors, wearables and video are attracting interest in home healthcare delivery. Development of such technology is usually justified on the basis of the technology's potential to increase the autonomy of people living with long-term conditions. Studies of the ethics of smart-homes raise concerns about privacy, consent, social isolation and equity of access. Few studies have investigated the ethical perspectives of smart-home engineers themselves. By exploring the views of engineering researchers in a large smart-home project, we sought to contribute to dialogue between ethics and the engineering community. METHODS: Either face-to-face or using Skype, we conducted in-depth qualitative interviews with 20 early- and mid-career smart-home researchers from a multi-centre smart-home project, who were asked to describe their own experience and to reflect more broadly about ethical considerations that relate to smart-home design. With participants' consent, interviews were audio-recorded, transcribed and analysed using a thematic approach. RESULTS: Two overarching themes emerged: in 'Privacy', researchers indicated that they paid close attention to negative consequences of potential unauthorised information sharing in their current work. However, when discussing broader issues in smart-home design beyond the confines of their immediate project, researchers considered physical privacy to a lesser extent, even though physical privacy may manifest in emotive concerns about being watched or monitored. In 'Choice', researchers indicated they often saw provision of choice to end-users as a solution to ethical dilemmas. While researchers indicated that choices of end-users may need to be restricted for technological reasons, ethical standpoints that restrict choice were usually assumed and embedded in design. CONCLUSIONS: The tractability of informational privacy may explain the greater attention that is paid to it. However, concerns about physical privacy may reduce acceptability of smart-home technologies to future end-users. While attention to choice suggests links with privacy, this may misidentify the sources of privacy and risk unjustly burdening end-users with problems that they cannot resolve. Separating considerations of choice and privacy may result in more satisfactory treatment of both. Finally, through our engagement with researchers as participants this study demonstrates the relevance of (bio)ethics as a critical partner to smart-home engineering.

Machine learning to assist risk-of-bias assessments in systematic reviews.

BACKGROUND: Risk-of-bias assessments are now a standard component of systematic reviews. At present, reviewers need to manually identify relevant parts of research articles for a set of methodological elements that affect the risk of bias, in order to make a risk-of-bias judgement for each of these elements. We investigate the use of text mining methods to automate risk-of-bias assessments in systematic reviews. We aim to identify relevant sentences within the text of included articles, to rank articles by risk of bias and to reduce the number of risk-of-bias assessments that the reviewers need to perform by hand. METHODS: We use supervised machine learning to train two types of models, for each of the three risk-of-bias properties of sequence generation, allocation concealment and blinding. The first model predicts whether a sentence in a research article contains relevant information. The second model predicts a risk-of-bias value for each research article. We use logistic regression, where each independent variable is the frequency of a word in a sentence or article, respectively. RESULTS: We found that sentences can be successfully ranked by relevance with area under the receiver operating characteristic (ROC) curve (AUC) > 0.98. Articles can be ranked by risk of bias with AUC > 0.72. We estimate that more than 33% of articles can be assessed by just one reviewer, where two reviewers are normally required. CONCLUSIONS: We show that text mining can be used to assist risk-of-bias assessments.

MR-PheWAS: hypothesis prioritization among potential causal effects of body mass index on many outcomes, using Mendelian randomization.

Observational cohort studies can provide rich datasets with a diverse range of phenotypic variables. However, hypothesis-driven epidemiological analyses by definition only test particular hypotheses chosen by researchers. Furthermore, observational analyses may not provide robust evidence of causality, as they are susceptible to confounding, reverse causation and measurement error. Using body mass index (BMI) as an exemplar, we demonstrate a novel extension to the phenome-wide association study (pheWAS) approach, using automated screening with genotypic instruments to screen for causal associations amongst any number of phenotypic outcomes. We used a sample of 8,121 children from the ALSPAC dataset, and tested the linear association of a BMI-associated allele score with 172 phenotypic outcomes (with variable sample sizes). We also performed an instrumental variable analysis to estimate the causal effect of BMI on each phenotype. We found 21 of the 172 outcomes were associated with the allele score at an unadjusted p < 0.05 threshold, and use Bonferroni corrections, permutation testing and estimates of the false discovery rate to consider the strength of results given the number of tests performed. The most strongly associated outcomes included leptin, lipid profile, and blood pressure. We also found novel evidence of effects of BMI on a global self-worth score.

Identifying employees at risk for job loss during sick leave.

PURPOSE: To examine the associations between medical, work-related, organizational and sociodemographic factors and job loss during sick leave in a Dutch population of 4132 employees on sick leave. METHODS: Data were assessed by occupational health physicians (OHPs) on sociodemographic, medical, work-related and organizational factors. Odds ratios for job loss were calculated in logistic regression models. RESULTS: Job loss during sick leave is associated with mental disorder, a history of sick leave due to these disorders, lack of co-worker and supervisor support, job insecurity and working as a civil servant or a teacher. Associations vary for gender and for company size. CONCLUSIONS: Job loss during sick leave is associated with medical, work-related, organizational and socio-demographic factors. The findings of this study might help the OHP or other health professionals involved in the management of employees on sick leave to identify those employees who are at risk for job loss during sick leave, and might help policymakers to decide on priorities in prevention and treatment. Future studies should have a longitudinal, prospective design and include information about the type of contract, possible causes for job loss, severity and treatment of the disorder causing the sick leave. IMPLICATIONS FOR REHABILITATION: The labor market moves to more and more flexible and temporary contracts. This leads to more precarious types of employment. The risk of job loss during sick leave is associated with medical, work-related, organizational and sociodemographic factors. Occupational health physicians and other professionals in the field of work rehabilitation should be aware of these associations to prevent job loss due to these factors.

Caveats and pitfalls of ROC analysis in clinical microarray research (and how to avoid them).

The receiver operating characteristic (ROC) has emerged as the gold standard for assessing and comparing the performance of classifiers in a wide range of disciplines including the life sciences. ROC curves are frequently summarized in a single scalar, the area under the curve (AUC). This article discusses the caveats and pitfalls of ROC analysis in clinical microarray research, particularly in relation to (i) the interpretation of AUC (especially a value close to 0.5); (ii) model comparisons based on AUC; (iii) the differences between ranking and classification; (iv) effects due to multiple hypotheses testing; (v) the importance of confidence intervals for AUC; and (vi) the choice of the appropriate performance metric. With a discussion of illustrative examples and concrete real-world studies, this article highlights critical misconceptions that can profoundly impact the conclusions about the observed performance.

Factors associated with first return to work and sick leave durations in workers with common mental disorders.

BACKGROUND: Associations are examined between socio-demographic, medical, work-related and organizational factors and the moment of first return to work (RTW) (within or after 6 weeks of sick leave) and total sick leave duration in sick leave spells due to common mental disorders. METHODS: Data are derived from a Dutch database, build to provide reference data for sick leave duration for various medical conditions. The cases in this study were entered in 2004 and 2005 by specially trained occupational health physicians, based on the physician's assessment of medical and other factors. Odds ratios for first RTW and sick leave durations are calculated in logistic regression models. RESULTS: Burnout, depression and anxiety disorder are associated with longer sick leave duration. Similar, but weaker associations were found for female sex, being a teacher, small company size and moderate or high psychosocial hazard. Distress is associated with shorter sick leave duration. Medical factors, psychosocial hazard and company size are also and analogously associated with first RTW. Part-time work is associated with delayed first RTW. The strength of the associations varies for various factors and for different sick leave durations. CONCLUSION: The medical diagnosis has a strong relation with the moment of first RTW and the duration of sick leave spells in mental disorders, but the influence of demographic and work-related factors should not be neglected.

Determinants of sick-leave duration: a tool for managers?

AIMS: To provide managers with tools to manage episodes of sick-leave of their employees, the influence of factors such as age, gender, duration of tenure, working full-time or part-time, cause and history of sick-leave, salary and education on sick-leave duration was studied. METHOD: In a cross-sectional study, data derived from the 2005 sick-leave files of a Dutch university were examined. Odds ratios of the single risk factors were calculated for short spells (or=91 days) of sick-leave. Next, these factors were studied in multiple regression models. RESULTS: Age, gender, duration of employment, cause and history of sick-leave, salary and membership of scientific staff, studied as single factors, have a significant influence on sick-leave duration. In multiple models, this influence remains for gender, salary, age, and history and cause of sick-leave. Only in medium or long spells and regarding the risk for a long or an extended spell do the predictive values of models consisting of psychological factors, work-related factors, salary and gender become reasonable. CONCLUSIONS: The predictive value of the risk factors used in this study is limited, and varies with the duration of the sick-leave spell. Only the risk for an extended spell of sick-leave as compared to a medium or long spell is reasonably predicted. Factors contributing to this risk may be used as tools in decision-making.