Supporting clinical decision making in the emergency department for paediatric patients using machine learning: A scoping review protocol.

INTRODUCTION: Machine learning as a clinical decision support system tool has the potential to assist clinicians who must make complex and accurate medical decisions in fast paced environments such as the emergency department. This paper presents a protocol for a scoping review, with the objective of summarising the existing research on machine learning clinical decision support system tools in the emergency department, focusing on models that can be used for paediatric patients, where a knowledge gap exists. MATERIALS AND METHODS: The methodology used will follow the scoping study framework of Arksey and O'Malley, along with other guidelines. Machine learning clinical decision support system tools for any outcome and population (paediatric/adult/mixed) for use in the emergency department will be included. Articles such as grey literature, letters, pre-prints, editorials, scoping/literature/narrative reviews, non-English full text papers, protocols, surveys, abstract or full text not available and models based on synthesised data will be excluded. Articles from the last five years will be included. Four databases will be searched: Medline (EBSCO), CINAHL (EBSCO), EMBASE and Cochrane Central. Independent reviewers will perform the screening in two sequential stages (stage 1: clinician expertise and stage 2: computer science expertise), disagreements will be resolved by discussion. Data relevant to the research question will be collected. Quantitative analysis will be performed to generate the results. DISCUSSION: The study results will summarise the existing research on machine learning clinical decision support tools in the emergency department, focusing on models that can be used for paediatric patients. This holds the promise to identify opportunities to both incorporate models in clinical practice and to develop future models by utilising reviewers from diverse backgrounds and relevant expertise.

The p53 DNA damage response and Fanconi anemia DNA repair pathway protect against acetaldehyde-induced replication stress in esophageal keratinocytes.

Alcohol contributes to cellular accumulation of acetaldehyde, a primary metabolite of alcohol and a major human carcinogen. Acetaldehyde can form DNA adducts and induce interstrand crosslinks (ICLs) that are repaired by the Fanconi anemia DNA repair pathway (FA pathway). Individuals with deficiency in acetaldehyde detoxification or in the FA pathway have an increased risk of squamous-cell carcinomas (SCCs) including those of the esophagus. In a recent report, we described the molecular basis of acetaldehyde-induced DNA damage in esophageal keratinocytes [1]. We demonstrated that, at physiologically relevant concentrations, acetaldehyde induces DNA damage at the DNA replication fork. This resulted in replication stress, leading to activation of the ATR-Chk1-dependent cell cycle checkpoints. We also reported that the p53 DNA damage response is elevated in response to acetaldehyde and that the FA pathway limits acetaldehyde-induced genomic instability. Here, we highlight these findings and present additional results to discuss the role of the FA pathway and p53 DNA damage response in the protection against genomic instability and esophageal carcinogenesis.

Inclusion4EU: Co-Designing a Framework for Inclusive Software Design and Development.

Digital technology is now pervasive, however, not all groups have uniformly benefitted from technological changes and some groups have been left behind or digitally excluded. Comprehensive data from the 2017 Current Population Survey shows that older people and persons with disabilities still lag behind in computer and internet access. Furthermore unique ethical, privacy and safety implications exist for the use of technology for older persons and people with disabilities and careful reflection is required to incorporate these aspects, which are not always part of a traditional software lifecycle. In this paper we present the Inclusion4EU project that aims to co-design a new framework, guidelines and checklists for inclusive software design and development with end-users from excluded categories, academics with expertise in human-computer interaction and industry practitioners from software engineering.

Development of a low-dimensional model to predict admissions from triage at a pediatric emergency department.

Objectives: This study aims to develop and internally validate a low-dimensional model to predict outcomes (admission or discharge) using commonly entered data up to the post-triage process to improve patient flow in the pediatric emergency department (ED). In hospital settings where electronic data are limited, a low-dimensional model with fewer variables may be easier to implement. Methods: This prognostic study included ED attendances in 2017 and 2018. The Cross Industry Standard Process for Data Mining methodology was followed. Eligibility criteria was applied to the data set, splitting into 70% train and 30% test. Sampling techniques were compared. Gradient boosting machine (GBM), logistic regression, and naïve Bayes models were created. Variables of importance were obtained from the model with the highest area under the curve (AUC) and used to create a low-dimensional model. Results: Eligible attendances totaled 72,229 (15% admission rate). The AUC was 0.853 (95% confidence interval [CI], 0.846-0.859) for GBM, 0.845 (95% CI, 0.838-0.852) for logistic regression and 0.813 (95% CI, 0.806-0.821) for naïve Bayes. Important predictors in the GBM model used to create a low-dimensional model were presenting complaint, triage category, referral source, registration month, location type (resuscitation/other), distance traveled, admission history, and weekday (AUC 0.835 [95% CI, 0.829-0.842]). Conclusions: Admission and discharge probability can be predicted early in a pediatric ED using 8 variables. Future work could analyze the false positives and false negatives to gain an understanding of the implementation of these predictions.

Predicting Admissions From a Paediatric Emergency Department - Protocol for Developing and Validating a Low-Dimensional Machine Learning Prediction Model.

Introduction: Patients boarding in the Emergency Department can contribute to overcrowding, leading to longer waiting times and patients leaving without being seen or completing their treatment. The early identification of potential admissions could act as an additional decision support tool to alert clinicians that a patient needs to be reviewed for admission and would also be of benefit to bed managers in advance bed planning for the patient. We aim to create a low-dimensional model predicting admissions early from the paediatric Emergency Department. Methods and Analysis: The methodology Cross Industry Standard Process for Data Mining (CRISP-DM) will be followed. The dataset will comprise of 2 years of data, ~76,000 records. Potential predictors were identified from previous research, comprising of demographics, registration details, triage assessment, hospital usage and past medical history. Fifteen models will be developed comprised of 3 machine learning algorithms (Logistic regression, naïve Bayes and gradient boosting machine) and 5 sampling methods, 4 of which are aimed at addressing class imbalance (undersampling, oversampling, and synthetic oversampling techniques). The variables of importance will then be identified from the optimal model (selected based on the highest Area under the curve) and used to develop an additional low-dimensional model for deployment. Discussion: A low-dimensional model comprised of routinely collected data, captured up to post triage assessment would benefit many hospitals without data rich platforms for the development of models with a high number of predictors. Novel to the planned study is the use of data from the Republic of Ireland and the application of sampling techniques aimed at improving model performance impacted by an imbalance between admissions and discharges in the outcome variable.

Prevalence of depressive symptoms and associated factors in an urban, ophthalmic population.

OBJECTIVE: To determine the prevalence of depressive symptoms in an adult ophthalmic patient population and to delineate correlates. DESIGN: Cross-sectional study. PARTICIPANTS: Adult patients (⩾18 years) were approached in general and sub-specialty cornea, retina, and glaucoma ophthalmic clinics. A total of 367 patients from the four clinics were enrolled. METHODS: Depressive symptoms were assessed using the Patient Health Questionnaire-9. A cut-off score of ⩾10 was used to indicate clinically significant depressive symptoms. Patient Health Questionnaire-9 scores were used to evaluate bivariate relationships between depressive symptoms and distance visual acuity, ocular diagnosis, diabetes status, smoking status, demographic information, and medications. RESULTS: The majority of patients were female (52.9%) and Caucasian (48.6%). The mean age was 52.0 years (standard deviation: 16.7). Clinically significant depressive symptoms were present in 19.9% of patients overall; this rate varied slightly by clinic. Patients with low vision and blindness (visual acuity worse than 20/60) were more likely to have depressive symptoms (odds ratio = 2.82; 95% confidence interval: 1.90-4.21). Smoking and diabetes were also associated with depressive symptoms (odds ratio = 3.11 (2.66-3.64) and 3.42 (1.90-6.16), respectively). CONCLUSION: In a sample of urban ophthalmic adult patients, depressive symptoms were highly associated with low vision, smoking, and diabetes. This information can be used to target interventions to those at greatest risk of depressive symptoms.

Intestinal barrier tightening by a cell-penetrating antibody to Bin1, a candidate target for immunotherapy of ulcerative colitis.

Patients afflicted with ulcerative colitis (UC) are at increased risk of colorectal cancer. While its causes are not fully understood, UC is associated with defects in colonic epithelial barriers that sustain inflammation of the colon mucosa caused by recruitment of lymphocytes and neutrophils into the lamina propria. Based on genetic evidence that attenuation of the bridging integrator 1 (Bin1) gene can limit UC pathogenicity in animals, we have explored Bin1 targeting as a therapeutic option. Early feasibility studies in the dextran sodium sulfate mouse model of experimental colitis showed that administration of a cell-penetrating Bin1 monoclonal antibody (Bin1 mAb 99D) could prevent lesion formation in the colon mucosa in part by preventing rupture of lymphoid follicles. In vivo administration of Bin1 mAb altered tight junction protein expression and cecal barrier function. Strikingly, electrophysiology studies in organ cultures showed that Bin1 mAb could elevate resistance and lower 14 C-mannitol leakage across the cecal mucosa, consistent with a direct strengthening of colonic barrier function. Transcriptomic analyses of colitis tissues highlighted altered expression of genes involved in circadian rhythm, lipid metabolism, and inflammation, with a correction of the alterations by Bin1 mAb treatment to patterns characteristic of normal tissues. Overall, our results suggest that Bin1 mAb protects against UC by directly improving colonic epithelial barrier function to limit gene expression and cytokine programs associated with colonic inflammation.

Using MOOCs to Promote Digital Accessibility and Universal Design, the MOOCAP Experience.

The recently completed Massive Open Online Course for Accessibility Partnership project (MOOCAP), had the twin aims of establishing a strategic partnership around the promotion of Universal Design and Accessibility for ICT professionals and of developing a suite of Open Educational resources (OERs) in this domain. MOOCAP's eight university partners from Germany, Norway, Greece, Ireland, the UK and Austria have a significant history in developing and providing courses in the domains of Universal Design and Accessibility, as well as leading research and advocacy roles within Europe. The MOOCAP project consisted of two phases: the development of an introductory MOOC on Digital Accessibility and the delivery of set of online courses with more in-depth and focused learning topics. During the lifetime of the project over 10,000 students signed up for these courses. This paper reflects on the challenges of creating and delivering MOOCs, especially in topics around Digital Accessibility and Universal Design. It considers the outcomes, impacts and legacies of the project. Based on our experiences of integrating these materials into our courses and on feedback and project evaluations, this paper will assess the potential of MOOCs to promote Universal Design for ICT and other professionals, while pointing up the possible trials and opportunities of such activities.

Adherence to Follow-up Recommendations Among Individuals in the Philadelphia Glaucoma Detection and Treatment Project.

PURPOSE: To evaluate rates of adherence to free follow-up eye exam appointments among participants in the Philadelphia Glaucoma Detection and Treatment Project. PATIENTS AND METHODS: Ophthalmologists and testing equipment were brought directly to participants at risk for glaucoma at 43 community sites in Philadelphia. Those diagnosed with glaucoma-related pathology were recommended to return for follow-up to be reexamined on site. Rates of adherence and clinical and demographic risk factors for adherence were evaluated. RESULTS: Five hundred thirty-one participants were diagnosed with glaucoma-related conditions and recommended to attend community-based follow-up exams. Follow-up adherence rate was 61.2% (n=325/531). Significant factors associated with greater eye exam appointment adherence, based on our univariable analysis, included final diagnosis of glaucoma (risk ratio [RR]=1.33; 95% confidence interval [CI], 1.13-1.57), male sex (RR=1.19; 95% CI, 1.04-1.36), white race (RR=1.26; 95% CI, 1.08-1.48), age (RR=1.17; 95% CI, 1.00-1.37) recommendation for glaucoma medication (RR=1.52; 95% CI, 1.35-1.71), recommendation for laser peripheral iridotomy (RR=1.18; 95% CI, 1.02-1.35), diagnosis of age-related macular degeneration (RR=1.42; 95% CI, 1.13-1.77) and an increased intraocular pressure (>22 mm Hg in the worse eye) (RR=1.23; 95% CI, 1.06-1.42). On the basis of our multivariable model, diagnosis, sex, and recommended glaucoma medications were significantly associated with follow-up adherence. CONCLUSIONS: This study demonstrates that individuals living in underserved urban communities would take advantage of free eye exams in community sites and return for follow-up eye exams in these same settings. Future studies could investigate interventions to improve eye exam appointment adherence in community-based settings to detect glaucoma-eye conditions.

The Host Microbiome Regulates and Maintains Human Health: A Primer and Perspective for Non-Microbiologists.

Humans consider themselves discrete autonomous organisms, but recent research is rapidly strengthening the appreciation that associated microorganisms make essential contributions to human health and well being. Each person is inhabited and also surrounded by his/her own signature microbial cloud. A low diversity of microorganisms is associated with a plethora of diseases, including allergy, diabetes, obesity, arthritis, inflammatory bowel diseases, and even neuropsychiatric disorders. Thus, an interaction of microorganisms with the host immune system is required for a healthy body. Exposure to microorganisms from the moment we are born and appropriate microbiome assembly during childhood are essential for establishing an active immune system necessary to prevent disease later in life. Exposure to microorganisms educates the immune system, induces adaptive immunity, and initiates memory B and T cells that are essential to combat various pathogens. The correct microbial-based education of immune cells may be critical in preventing the development of autoimmune diseases and cancer. This review provides a broad overview of the importance of the host microbiome and accumulating knowledge of how it regulates and maintains a healthy human system. Cancer Res; 77(8); 1783-812. ©2017 AACR.

Management of acute traumatic intracranial haematoma in rural and remote areas of Australia.

INTRODUCTION: Outcomes are reviewed with traumatic brain injury (TBI) involving intracranial haematomata (ICH) with patients in rural and remote areas of South Australia and adjacent states. Patients were referred to the Royal Adelaide Hospital (RAH), a level 1 trauma centre with a major neurosurgical service. METHOD: From 2000 to 2013, 1107 multiple trauma cases included 162 with severe TBI. Local medical officers (LMOs) phoned a specific number to access resuscitative and neurosurgical advice. Onsite neurosurgical support was provided when requested. Specialist teams later retrieved these patients to RAH. Locations were coded according to the Accessibility/Remoteness Index of Australia (ARIA+). Injuries were coded using ICD 9 and 10. RESULTS: General surgeon LMOs drained nine clinically progressive ICHs. Neurosurgical attendance was provided in four instances. Eight patients survived. The remaining 153 patients had other injuries involving thoraco-abdominal organs, spine, pelvis and limbs. The overall mortality was 30%. Twenty-six had ICHs requiring surgical drainage later at the RAH, with 46% mortality. Average Injury Severity Score was 30 (range 9-66). Male/female ratio was 76/24. Motor vehicular accidents predominated (60%), followed by falls (26%) and assaults (10%). Those under 30 years were overrepresented. Patients were transported a mean distance of 283 km (maximum distance 2600 km). CONCLUSION: LMOs in remote locations may consider immediate drainage of deteriorating traumatic ICH. Adequate support from a distant major trauma centre can help achieve acceptable outcomes. Effective communications are vital. The Royal Australasian College of Surgeons and Neurosurgical Society of Australasia guidelines based on the Early Management of Severe Trauma protocols can assist LMOs in making the decision to undertake emergency craniotomy.

Tetanus and the evolution of intensive care in Australia.

A review of two series of patients with tetanus from the Royal Adelaide Hospital provides a historical perspective on the evolution of intensive care in Australia. Nine consecutive severe cases presenting in 1957 constituted one of the first series published. Four patients died. The second series of 38 severe cases, among a total of 56 cases presenting between 1967 and 1985, included two deaths, comparing favourably with survival in other contemporary series. The specialty of intensive care evolved considerably during this time. Neuromuscular blockade introduced in the first series produced radical changes in management. Supportive measures that were not then widely practised, involving intermittent positive pressure ventilation, were used in the second series for up to 46 days and evolved into standard ICU practice. The option of using a tank respirator was rejected. Older patients were susceptible to complications commonly related to respiratory, cardiovascular and diabetic comorbidities, but most returned to their previous lifestyle. Severe tetanus often resulted from mild injuries in patients who were incompletely immunised. Four patients developed tetanus following surgical procedures. The use of nitrous oxide in the first series was abandoned owing to adverse effects on bone marrow function. Complications reported in early literature, such as fractures and myositis ossificans, presumably related to unrelieved spasm, are no longer seen. Clinicians are now likely to see the condition only if working with counter-disaster teams overseas.

An electroelution apparatus for sequential transfer of sodium dodecyl sulfate-proteins into agarose and mass spectrometric identification of Li-Na-dodecyl sulfate-proteins from solubilized agarose.

Separated protein bands are sequentially electrophoresed into low melting agarose plugs distributed in an apparatus of original design along the surface of a plastic drum. The rotation of the drum is synchronized to migration of electrophoretic bands to receive each band individually. Agarose plugs are dissolved enzymatically for transfer into the mass spectrometer. One microL of the agarose solution containing 1 pmol of each of three lithium and natrium salts of dodecyl sulfate (Li-Na-DS)-proteins were applied to matrix-assisted laser desorption/ionization-mass spectrometry (MALDI-MS) without any prepurification. It yields a signal indistinguishable from that obtained in the absence of agarose.

Mass spectrometry after capture and small-volume elution of analyte from a surface plasmon resonance biosensor.

The identification of binding partners of proteins by mass spectrometry following specific capture on a biosensor surface is a promising tool for proteomics research and the identification and characterization of protein-protein interactions. Previous approaches include the direct ionization of analyte from the biosensor chip on a matrix assisted laser desorption ionization time-of-flight mass spectrometer (MALDI-TOFMS) apparatus and the on-chip digestion followed by elution, chromatographic concentration of the fragments, and electrospray mass spectrometry. In the present paper, using the small-volume microfluidic sample manipulation technique with oscillatory flow reported recently (Abrantes et al. Anal. Chem. 2001, 73, 2828-2835), analyte is shown to be eluted from the sensor surface into a small volume of buffer that promotes dissociation from the capture surface and delivery to the mass spectrometer. Both the incubation of the sensor surface with the sample and the recovery of analyte can be achieved with a few microliters and conducted until steady-state is attained. Because the procedure is non-destructive for the sensor surface, multiple cycles of capture and elution allow the transfer and concentration of analyte into the elution buffer. The eluted analyte can be studied directly by MALDI-TOFMS, or subjected to proteolytic digestion for protein identification. Transfer into the elution buffer and MALDI-TOFMS detection was achieved from 5 microL of starting samples containing <50 fmol of analyte. Examples are presented for the specific detection and recovery of a protein from a complex mixture of cytosolic proteins.