Exploring the Impact of the NULL Class on In-the-Wild Human Activity Recognition.

Monitoring activities of daily living (ADLs) plays an important role in measuring and responding to a person's ability to manage their basic physical needs. Effective recognition systems for monitoring ADLs must successfully recognize naturalistic activities that also realistically occur at infrequent intervals. However, existing systems primarily focus on either recognizing more separable, controlled activity types or are trained on balanced datasets where activities occur more frequently. In our work, we investigate the challenges associated with applying machine learning to an imbalanced dataset collected from a fully in-the-wild environment. This analysis shows that the combination of preprocessing techniques to increase recall and postprocessing techniques to increase precision can result in more desirable models for tasks such as ADL monitoring. In a user-independent evaluation using in-the-wild data, these techniques resulted in a model that achieved an event-based F1-score of over 0.9 for brushing teeth, combing hair, walking, and washing hands. This work tackles fundamental challenges in machine learning that will need to be addressed in order for these systems to be deployed and reliably work in the real world.

Exercise prescription for overhead athletes with shoulder pathology: a systematic review with best evidence synthesis.

OBJECTIVE: To produce a best evidence synthesis of exercise prescription used when treating shoulder pathology in the overhead athlete. DESIGN: A systematic review of exercises used in overhead athletes including case studies and clinical commentaries. DATA SOURCES: MEDLINE, PubMed, SPORTDiscus and CINAHL from database inception through July 8, 2016. METHODS: We examined data from randomised controlled trials and prospective cohort (level I-IV evidence) studies that addressed exercise intervention in the rehabilitation of the overhead athlete with shoulder pathology. Case studies and clinical commentaries (level V evidence) were examined to account for expert opinion-based research. Data were combined using best evidence synthesis and graded (A-F) recommendations (Centre for Evidence-Based Medicine). RESULTS: There were 33 unique exercises in six level I-IV studies that met our inclusion criteria. Most exercises were single-plane, upper extremity exercises performed below 90o of elevation. There were 102 unique exercises in 33 level V studies that met our inclusion criteria. These exercises emphasised plyometrics, kinetic chain and sport-specific training. CONCLUSIONS AND RELEVANCE: Overall, evidence for exercise interventions in overhead athletes with shoulder pathology is dominated by expert opinion (grade D). There is great variability between exercise approaches suggested by experts and those investigated in research studies and the overall level of evidence is low. The strongest available evidence (level B) supports the use of single-plane, open chain upper extremity exercises performed below 90° of elevation and closed chain upper extremity exercises. Clinical expert pieces support a more advanced, global treatment approach consistent with the complex, multidimensional nature of sport.

Meningococcal carriage in periods of high and low invasive meningococcal disease incidence in the UK: comparison of UKMenCar1-4 cross-sectional survey results.

BACKGROUND: The incidence of invasive meningococcal disease in the UK decreased by approximately four times from 1999 to 2014, with reductions in serogroup C and serogroup B disease. Lower serogroup C invasive meningococcal disease incidence was attributable to implementation of the meningococcal serogroup C conjugate vaccine in 1999, through direct and indirect protection, but no vaccine was implemented against serogroup B disease. UK Meningococcal Carriage surveys 1-3 (UKMenCar1-3), conducted in 1999, 2000, and 2001, were essential for understanding the impact of vaccination. To investigate the decline in invasive meningococcal disease incidence, we did a large oropharyngeal carriage survey in 2014-15, immediately before the changes to meningococcal vaccines in the UK national immunisation schedule. METHODS: UKMenCar4 was a cross-sectional survey in adolescents aged 15-19 years who were enrolled from schools and colleges geographically local to one of 11 UK sampling centres between Sept 1, 2014, and March 30, 2015. Participants provided an oropharyngeal swab sample and completed a questionnaire on risk factors for carriage, including social behaviours. Samples were cultured for putative Neisseria spp, which were characterised with serogrouping and whole-genome sequencing. Data from this study were compared with the results from the UKMenCar1-3 surveys (1999-2001). FINDINGS: From the 19 641 participants (11 332 female, 8242 male, 67 not stated) in UKMenCar4 with culturable swabs and completed risk-factor questionnaires, 1420 meningococci were isolated, with a carriage prevalence of 7·23% (95% CI 6·88-7·60). Carriage prevalence was substantially lower in UKMenCar4 than in the previous surveys: carriage prevalence was 16·6% (95% CI 15·89-17·22; 2306/13 901) in UKMenCar1 (1999), 17·6% (17·05-18·22; 2873/16 295) in UKMenCar2 (2000), and 18·7% (18·12-19·27; 3283/17 569) in UKMenCar3 (2001). Carriage prevalence was lower for all serogroups in UKMenCar4 than in UKMenCar1-3, except for serogroup Y, which was unchanged. The prevalence of carriage-promoting social behaviours decreased from 1999 to 2014-15, with individuals reporting regular cigarette smoking decreasing from 2932 (21·5%) of 13 650 to 2202 (11·2%) of 19 641, kissing in the past week from 6127 (44·8%) of 13 679 to 7320 (37·3%) of 19 641, and attendance at pubs and nightclubs in the past week from 8436 (62·1%) of 13 594 to 7662 (39·0%) of 19 641 (all p<0·0001). INTERPRETATION: We show that meningococcal carriage prevalence in adolescents sampled nationally during a low incidence period (2014-15) was less than half of that in an equivalent population during a high incidence period (1999-2001). Disease and carriage caused by serogroup C was well controlled by ongoing vaccination. The prevalence of behaviours associated with carriage declined, suggesting that public health policies aimed at influencing behaviour might have further reduced disease. FUNDING: Wellcome Trust, UK Department of Health, and National Institute for Health Research.

Quasi-experimental studies suggest that lowering air pollution levels benefits infants' and children's health.

Many human studies have shown a correlation between air pollution and poor health in children. This paper focuses on studies that employ quasi-experimental designs to study the effects of air pollution on specific populations in an effort to isolate the causes and minimize the effect of confounding factors. Human studies of many and varied designs have found that exposure to some pollutants, even at levels below regulatory thresholds, adversely affects health. Our review of quasi-experimental studies adds additional support to these findings. Together, the research suggests that lowering the thresholds for acceptable air pollution levels may be a prudent and necessary step toward improving population health, especially among the most vulnerable members of society: infants and children. Policy makers should also consider providing specific information to families and pregnant women about when and where the risk of pollution exposure is highest, so that they can minimize their exposure or avoid it altogether.