Training and testing of a gradient boosted machine learning model to predict adverse outcome in patients presenting to emergency departments with suspected covid-19 infection in a middle-income setting.

COVID-19 infection rates remain high in South Africa. Clinical prediction models may be helpful for rapid triage, and supporting clinical decision making, for patients with suspected COVID-19 infection. The Western Cape, South Africa, has integrated electronic health care data facilitating large-scale linked routine datasets. The aim of this study was to develop a machine learning model to predict adverse outcome in patients presenting with suspected COVID-19 suitable for use in a middle-income setting. A retrospective cohort study was conducted using linked, routine data, from patients presenting with suspected COVID-19 infection to public-sector emergency departments (EDs) in the Western Cape, South Africa between 27th August 2020 and 31st October 2021. The primary outcome was death or critical care admission at 30 days. An XGBoost machine learning model was trained and internally tested using split-sample validation. External validation was performed in 3 test cohorts: Western Cape patients presenting during the Omicron COVID-19 wave, a UK cohort during the ancestral COVID-19 wave, and a Sudanese cohort during ancestral and Eta waves. A total of 282,051 cases were included in a complete case training dataset. The prevalence of 30-day adverse outcome was 4.0%. The most important features for predicting adverse outcome were the requirement for supplemental oxygen, peripheral oxygen saturations, level of consciousness and age. Internal validation using split-sample test data revealed excellent discrimination (C-statistic 0.91, 95% CI 0.90 to 0.91) and calibration (CITL of 1.05). The model achieved C-statistics of 0.84 (95% CI 0.84 to 0.85), 0.72 (95% CI 0.71 to 0.73), and 0.62, (95% CI 0.59 to 0.65) in the Omicron, UK, and Sudanese test cohorts. Results were materially unchanged in sensitivity analyses examining missing data. An XGBoost machine learning model achieved good discrimination and calibration in prediction of adverse outcome in patients presenting with suspected COVID19 to Western Cape EDs. Performance was reduced in temporal and geographical external validation.

Beyond acute concussion assessment to office management: a systematic review informing the development of a Sport Concussion Office Assessment Tool (SCOAT6) for adults and children.

OBJECTIVES: To systematically review the scientific literature regarding the assessment of sport-related concussion (SRC) in the subacute phase (3-30 days) and provide recommendations for developing a Sport Concussion Office Assessment Tool (SCOAT6). DATA SOURCES: MEDLINE, Embase, PsycINFO, Cochrane CENTRAL, CINAHL, SPORTDiscus and Web of Science searched from 2001 to 2022. Data extracted included study design, population, definition of SRC diagnosis, outcome measure(s) and results. ELIGIBILITY CRITERIA: (1) Original research, cohort studies, case-control studies, diagnostic accuracy and case series with samples >10; (2) SRC; (3) screening/technology that assessed SRC in the subacute period and (4) low risk of bias (ROB). ROB was performed using adapted Scottish Intercollegiate Guidelines Network criteria. Quality of evidence was evaluated using the Strength of Recommendation Taxonomy classification. RESULTS: Of 9913 studies screened, 127 met inclusion, assessing 12 overlapping domains. Results were summarised narratively. Studies of acceptable (81) or high (2) quality were used to inform the SCOAT6, finding sufficient evidence for including the assessment of autonomic function, dual gait, vestibular ocular motor screening (VOMS) and mental health screening. CONCLUSION: Current SRC tools have limited utility beyond 72 hours. Incorporation of a multimodal clinical assessment in the subacute phase of SRC may include symptom evaluation, orthostatic hypotension screen, verbal neurocognitive tests, cervical spine evaluation, neurological screen, Modified Balance Error Scoring System, single/dual task tandem gait, modified VOMS and provocative exercise tests. Screens for sleep disturbance, anxiety and depression are recommended. Studies to evaluate the psychometric properties, clinical feasibility in different environments and time frames are needed. PROSPERO REGISTRATION NUMBER: CRD42020154787.

Consensus statement on concussion in sport: the 6th International Conference on Concussion in Sport-Amsterdam, October 2022.

For over two decades, the Concussion in Sport Group has held meetings and developed five international statements on concussion in sport. This 6th statement summarises the processes and outcomes of the 6th International Conference on Concussion in Sport held in Amsterdam on 27-30 October 2022 and should be read in conjunction with the (1) methodology paper that outlines the consensus process in detail and (2) 10 systematic reviews that informed the conference outcomes. Over 3½ years, author groups conducted systematic reviews of predetermined priority topics relevant to concussion in sport. The format of the conference, expert panel meetings and workshops to revise or develop new clinical assessment tools, as described in the methodology paper, evolved from previous consensus meetings with several new components. Apart from this consensus statement, the conference process yielded revised tools including the Concussion Recognition Tool-6 (CRT6) and Sport Concussion Assessment Tool-6 (SCAT6, Child SCAT6), as well as a new tool, the Sport Concussion Office Assessment Tool-6 (SCOAT6, Child SCOAT6). This consensus process also integrated new features including a focus on the para athlete, the athlete's perspective, concussion-specific medical ethics and matters related to both athlete retirement and the potential long-term effects of SRC, including neurodegenerative disease. This statement summarises evidence-informed principles of concussion prevention, assessment and management, and emphasises those areas requiring more research.

The Socio-Demographics and Health Service Use of Opioid Overdose Decedents in Wales: A Cross-Sectional Data Linkage Study.

BACKGROUND: Fatal opioid overdose is a significant public health problem with increasing incidence in developed countries. This study aimed to describe demographic and service user characteristics of decedents of opioid overdose in Wales to identify possible targets for behaviour modification and life-saving interventions. METHODS: A retrospective cross-sectional analysis was conducted of a census sample of opioid overdose-related deaths recorded between January 01, 2012, and October 11, 2018, in Wales. UK Office for National Statistics, Welsh Demographic Service, and National Health Service datasets were linked deterministically. Decedents' circumstances of death, demographic characteristics, residency, and health service use were characterized over 3 years prior to fatal overdose using descriptive statistics. RESULTS: In total, 638 people died of opioid overdose in Wales between January 01, 2012, and October 11, 2018, with an incidence rate of 3.04 per 100,000 people per year. Decedents were predominantly male (73%) and middle aged (median age 50 years). Fatal overdoses predominantly occurred in the community (93%) secondary to heroin (30%) or oxycodone derivative use (34%). In the 3 years prior to death, decedents changed address frequently (53%) but rarely moved far geographically. The majority of decedents had recently visited the emergency department (83%) or were admitted to the hospital (64%) prior to death. Only a minority had visited specialist drug services (32%). CONCLUSIONS: Deaths from opioid overdose typically occur in middle-aged men living peripatetic lifestyles. Victims infrequently visit specialist drug services but often attend emergency medical services. Emergency department-based interventions may therefore be important in prevention of opioid overdose fatalities in the community.

Expansion of cognitive testing for off-field concussion screening in elite rugby players: A cohort study.

OBJECTIVES: Current off-field concussion screening instruments have sub-optimal accuracy and additional testing domains may be necessary to detect the full spectrum of concussion presentations. This study aimed to determine if additional cognitive tests add utility to off-field screening for sport-related concussion. DESIGN: Reproducibility and diagnostic accuracy cohort studies were performed in the 2017 and 2018 seasons of the Super Rugby competition, conducted in Argentina, Australia, Japan, New Zealand, and South Africa. METHODS: Abridged versions of Stroop (score, time), Spatial Memory (score, failed trials), and Trail Making Trial-B (time, errors) cognitive tests, modified for off-field use, were examined. Players performed baseline testing prior to each season. Cases undergoing off-field screening as part of the World Rugby Head Injury Assessment Process underwent evaluation with the same cognitive tests during competition matches. Agreement between repeated pre-season baseline measurements, and the diagnostic accuracy of off-field testing against a clinical reference standard of concussion, was evaluated. RESULTS: Data were available for repeated preseason baseline testing in 644 players, and 100 cases undergoing off-field concussion screening. There was little individual agreement across pre-season baseline assessments for all tests (Lin's correlation and Gwets AC1 coefficients ranging between 0.2 and 0.3). There was significantly worse performance for the time taken to complete the modified Stroop Test in concussed players undergoing off-field screening, compared to non-concussed players (median time 21.1 v 18.4 s, p < 0.01; area under the receiver operating characteristic curve 0.7 (95% CI 0.5-0.8)). Other cognitive measures did not discriminate between injured and un-injured players, with no-statistically significant differences in distribution medians (p = 0.6-0.9) and AUROC values close to 0.5. CONCLUSIONS: The time taken to perform a modified Stroop Test may provide additional diagnostic accuracy if added to current off-field concussion screening tools. Abridged Spatial Memory and Trail Making Trial-B tests did not discriminate between concussed and non-concussed players.

External validation of the Dutch prediction model for prehospital triage of trauma patients in South West region of England, United Kingdom.

IMPORTANCE: This paper investigates the use of a major trauma prediction model in the UK setting. We demonstrate that application of this model could reduce the number of patients with major trauma being incorrectly sent to non-specialist hospitals. However, more research is needed to reduce over-triage and unnecessary transfer to Major Trauma Centres. OBJECTIVE: To externally validate the Dutch prediction model for identifying major trauma in a large unselected prehospital population of injured patients in England. DESIGN: External validation using a retrospective cohort of injured patients who ambulance crews transported to hospitals. SETTING: South West region of England. PARTICIPANTS: All patients ≥16 years with a suspected injury and transported by ambulance in the year from February 1, 2017. EXCLUSION CRITERIA: 1) Patients aged ≤15 years; 2) Non-ambulance attendance at hospital with injuries; 3) Death at the scene and; 4) Patients conveyed by helicopter. This study had a census sample of cases available to us over a one year period. INTERVENTIONS OR EXPOSURES: Tested the accuracy of the prediction model in terms of discrimination, calibration, clinical usefulness, sensitivity and specificity and under- and over triage rates compared to usual triage practices in the South West region. MAIN OUTCOME MEASURE: Major trauma defined as an Injury Severity Score>15. RESULTS: A total of 68799 adult patients were included in the external validation cohort. The median age of patients was 72 (i.q.r. 46-84); 55.5% were female; and 524 (0.8%) had an Injury Severity Score>15. The model achieved good discrimination with a C-Statistic 0.75 (95% CI, 0.73 - 0.78). The maximal specificity of 50% and sensitivity of 83% suggests the model could improve undertriage rates at the expense of increased overtriage rates compared with routine trauma triage methods used in the South West, England. CONCLUSIONS AND RELEVANCE: The Dutch prediction model for identifying major trauma could lower the undertriage rate to 17%, however it would increase the overtriage rate to 50% in this United Kingdom cohort. Further prospective research is needed to determine whether the model can be practically implemented by paramedics and is cost-effective.

Diagnostic Utility of New SCAT5 Neurological Screen Sub-tests.

BACKGROUND: The Sports Concussion Assessment Tool (SCAT) is recommended to screen for concussion following head impact events in elite sport. The most recent 5th edition (SCAT5) included a 'rapid neurological screen' which introduced new subtests examining comprehension, passive neck movement, and diplopia. This study evaluated the additional diagnostic value of these new subtests. METHODS: A prospective cohort study was performed in the Pro14 elite Rugby Union competition between September 2018 and January 2020. The SCAT5 was administered by the team doctor to players undergoing off-field screening for concussion during a medical room assessment. Sensitivity, specificity, false negatives, and positives were examined for SCAT5 comprehension, passive neck movement, and diplopia subtests. The reference standard was a final diagnosis of concussion, established by serial standardised clinical assessments over 48 h. RESULTS: Ninety-three players undergoing off-field screening for concussion were included. Sensitivity and specificity of the comprehension, passive neck movement, and diplopia subtests were 0, 8, 5% and 0, 91, 97%, respectively (concussion prevalence 63%). No players had any abnormality in comprehension. No players had abnormal passive neck movement or diplopia in the absence of abnormalities in other SCAT5 sub-components. CONCLUSIONS: The new SCAT5 neurological screen subtests are normal in the majority of players undergoing off-field concussion screening and appear to lack diagnostic utility over and above other SCAT5 subtests.

Concussion Guidelines in National and International Professional and Elite Sports.

The Berlin statement on sport-related concussion was published in 2017 using evidence-based recommendations. We aimed to examine (1) the implementation of, distribution and education based on the Berlin recommendations, and the development of sport-specific protocols/guidelines among professional and elite sports, (2) the implementation of guidelines at the community level, (3) translation of guidelines into different languages, and (4) research activities. Senior medical advisers and chief medical officers from Australian Football League, All Japan Judo Federation, British Horseracing Authority, Cricket Australia, Fédération Equestre Internationale, Football Association, Gaelic Athletic Association, International Boxing Association, Irish Horseracing Regulatory Board, Major League Baseball, National Football League, National Hockey League, National Rugby League, and World Rugby completed a questionnaire. The results demonstrated that all 14 sporting organizations have published concussion protocols/guidelines based on the Berlin recommendations, including Recognize, Removal from play, Re-evaluation, Rest, Recovery, and Return to play. There is variable inclusion of Prolonged symptoms. Prevention and Risk reduction and Long-term effects are addressed in the guidelines, rules and regulations, and/or sport-specific research. There is variability in education programs, monitoring compliance with guidelines, and publication in other languages. All sporting bodies are actively involved in concussion research. We conclude that the Berlin recommendations have been included in concussion protocols/guidelines by all the sporting bodies, with consistency in the essential components of the recommendations, whilst also allowing for sport- and regional-specific variations. Education at the elite, community, and junior levels remains an ongoing challenge, and future iterations of guidelines may consider multiple language versions, and community- and junior-level guidelines.

International consensus definitions of video signs of concussion in professional sports.

BACKGROUND: The use of video to assist professional sporting bodies with the diagnosis of sport-related concussion (SRC) has been well established; however, there has been little consistency across sporting codes with regards to which video signs should be used, and the definitions of each of these signs. AIM: The aims of this study were to develop a consensus for the video signs considered to be most useful in the identification of a possible SRC and to develop a consensus definition for each of these video signs across the sporting codes. METHODS: A brief questionnaire was used to assess which video signs were considered to be most useful in the identification of a possible concussion. Consensus was defined as >90% agreement by respondents. Existing definitions of these video signs from individual sports were collated, and individual components of the definitions were assessed and ranked. A modified Delphi approach was then used to create a consensus definition for each of the video signs. RESULTS: Respondents representing seven sporting bodies (Australian Football League, Cricket Australia, Major League Baseball, NFL, NHL, National Rugby League, World Rugby) reached consensus on eight video signs of concussion. Thirteen representatives from the seven professional sports ranked the definition components. Consolidation and refinement of the video signs and their definitions resulted in consensus definitions for six video signs of possible concussion: lying motionless, motor incoordination, impact seizure, tonic posturing, no protective action-floppy and blank/vacant look. CONCLUSIONS: These video signs and definitions have reached international consensus, are indicated for use by professional sporting bodies and will form the basis for further collaborative research.

King-Devick concussion test performs poorly as a screening tool in elite rugby union players: a prospective cohort study of two screening tests versus a clinical reference standard.

BACKGROUND: The King-Devick (KD) test is an objective clinical test of eye movements that has been used to screen for concussion. We characterised the accuracy of the KD test and the World Rugby Head Injury Assessment (HIA-1) screening tools as methods of off-field evaluation for concussion after a suspicious head impact event. METHODS: A prospective cohort study was performed in elite English rugby union competitions between September 2016 and May 2017. The study population comprised consecutive players identified with a head impact event with the potential to result in concussion. The KD test was administered off-field, alongside the World Rugby HIA-1 screening tool, and the results were compared with the preseason baseline. Accuracy was measured against a reference standard of confirmed concussion, based on the clinical judgement of the team doctor after serial assessments. RESULTS: 145 head injury events requiring off-field medical room screening assessments were included in the primary analysis. The KD test demonstrated a sensitivity of 60% (95% CI 49.0 to 70) and a specificity of 39% (95% CI 26 to 54) in identifying players subsequently diagnosed with concussion. Area under the receiver operating characteristic curve for prolonged KD test times was 0.51 (95% CI 0.41 to 0.61). The World Rugby HIA-1 off-field screening tool sensitivity did not differ significantly from the KD test (sensitivity 75%, 95% CI 66 to 83, P=0.08), but specificity was significantly higher (91%, 95% CI 82 to 97, P<0.001). Although combining the KD test and the World Rugby HIA-1 multimodal screening assessment achieved a significantly higher sensitivity of 93% (95% CI 86% to 97%), there was a significantly lower specificity of 33% (95% CI 21% to 48%), compared with the HIA-1 test alone. CONCLUSIONS: The KD test demonstrated limited accuracy as a stand-alone remove-from-play sideline screening test for concussion. As expected with the addition of any parallel test, combination of the KD test with the HIA-1 off-field screening tool provided improved sensitivity in identifying concussion, but at the expense of markedly lower specificity. These results suggest that it is unlikely that the KD test will be incorporated into multimodal off-field screening assessments for concussion at the present time.

International study of video review of concussion in professional sports.

BACKGROUND: Video review has become an important tool in professional sporting codes to help sideline identification and management of players with a potential concussion. AIM: To assess current practices related to video review of concussion in professional sports internationally, and compare protocols and diagnostic criteria used to identify and manage potential concussions. METHODS: Current concussion management guidelines from professional national and international sporting codes were reviewed. Specific criteria and definitions of video signs associated with concussion were compared between codes. Rules and regulations adopted across the codes for processes around video review were also assessed. RESULTS: Six sports with specific diagnostic criteria and definitions for signs of concussion identified on video review participated in this study (Australian football, American football, world rugby, cricket, rugby league and ice hockey). Video signs common to all sports include lying motionless/loss of responsiveness and motor incoordination. The video signs considered by the majority of sports as most predictive of a diagnosis of concussion include motor incoordination, impact seizure, tonic posturing and lying motionless. Regulatory requirements, sideline availability of video, medical expertise of video reviewers and use of spotters differ across sports and geographical boundaries. By and large, these differences reflect a pragmatic approach from each sport, with limited underlying research and development of the video review process in some instances. CONCLUSIONS: The use of video analysis in assisting medical staff with the diagnosis or identification of potential concussion is well established across different sports internationally. The diagnostic criteria used and the expertise of the video review personnel are not clearly established, and research efforts would benefit from a collaborative harmonisation across sporting codes.

Implementation of the 2017 Berlin Concussion in Sport Group Consensus Statement in contact and collision sports: a joint position statement from 11 national and international sports organisations.

The 2017 Berlin Concussion in Sport Group Consensus Statement provides a global summary of best practice in concussion prevention, diagnosis and management, underpinned by systematic reviews and expert consensus. Due to their different settings and rules, individual sports need to adapt concussion guidelines according to their specific regulatory environment. At the same time, consistent application of the Berlin Consensus Statement's themes across sporting codes is likely to facilitate superior and uniform diagnosis and management, improve concussion education and highlight collaborative research opportunities. This document summarises the approaches discussed by medical representatives from the governing bodies of 10 different contact and collision sports in Dublin, Ireland in July 2017. Those sports are: American football, Australian football, basketball, cricket, equestrian sports, football/soccer, ice hockey, rugby league, rugby union and skiing. This document had been endorsed by 11 sport governing bodies/national federations at the time of being published.

A video analysis of head injuries satisfying the criteria for a head injury assessment in professional Rugby Union: a prospective cohort study.

OBJECTIVES: Concussion is the most common match injury in professional Rugby Union, accounting for 25% of match injuries. The primary prevention of head injuries requires that the injury mechanism be known so that interventions can be targeted to specifically overall incidence by focusing on characteristics with the greatest propensity to cause a head injury. METHODS: 611 head injury assessment (HIA) events in professional Rugby Union over a 3-year period were analysed, with specific reference to match events, position, time and nature of head contact. RESULTS: 464 (76%) of HIA events occur during tackles, with the tackler experiencing a significantly greater propensity for an HIA than the ball carrier (1.40 HIAs/1000 tackles for the tackler vs 0.54 HIAs/1000 tackles for the ball carrier, incidence rate ratio (IRR) 2.59). Propensity was significantly greater for backline players than forwards (IRR 1.54, 95% CI 1.28 to 1.84), but did not increase over the course of the match. Head to head contact accounted for the most tackler HIAs, with the greatest propensity. CONCLUSIONS: By virtue of its high propensity and frequency, the tackle should be the focus for interventions that may include law change and technique education. A specific investigation of the characteristics of the tackle is warranted to refine the approach to preventative strategies.