Search

Introduction: Artificial intelligence (AI) has the potential to improve healthcare quality when thoughtfully integrated into clinical practice. Current evaluations of AI solutions tend to focus solely on model performance. There is a critical knowledge gap in the assessment of AI–clinician interactions. We systematically reviewed existing literature to identify interaction traits that can be used to assess the quality of AI–clinician interactions. Methods: We performed a systematic review of published studies to June 2022 that reported elements of interactions that impacted the relationship between clinicians and AI-enabled clinical decision support systems. Due to study heterogeneity, we conducted a narrative synthesis of the different interaction traits identified from this review. Two study authors categorised the AI–clinician interaction traits based on their shared constructs independently. After the independent categorisation, both authors engaged in a discussion to finalise the categories. Results: From 34 included studies, we identified 210 interaction traits. The most common interaction traits included usefulness, ease of use, trust, satisfaction, willingness to use and usability. After removing duplicate or redundant traits, 90 unique interaction traits were identified. Unique interaction traits were then classified into seven categories: usability and user experience, system performance, clinician trust and acceptance, impact on patient care, communication, ethical and professional concerns, and clinician engagement and workflow. Discussion: We identified seven categories of interaction traits between clinicians and AI systems. The proposed categories may serve as a foundation for a framework assessing the quality of AI–clinician interactions.

Background Transfemoral carotid artery stenting (TFCAS) carries important perioperative risks. Outcome prediction tools may help guide clinical decision‐making but remain limited. We developed machine learning algorithms that predict 1‐year stroke or death following TFCAS. Methods and Results The VQI (Vascular Quality Initiative) database was used to identify patients who underwent TFCAS for carotid artery stenosis between 2005 and 2024. We identified 112 features from the index hospitalization (82 preoperative [demographic/clinical], 13 intraoperative [procedural], and 17 postoperative [in‐hospital course/complications]). The primary outcome was 1‐year postprocedural stroke or death. The data were divided into training (70%) and test (30%) sets. Six machine learning models were trained using preoperative features with 10‐fold cross‐validation. The primary model evaluation metric was area under the receiver operating characteristic curve. The algorithm with the best performance was further trained using intra‐ and postoperative features. Model robustness was assessed using calibration plots and Brier scores. Overall, 35 214 patients underwent TFCAS during the study period and 3257 (9.2%) developed 1‐year stroke or death. The best preoperative prediction model was extreme gradient boosting, achieving an area under the receiver operating characteristic curve of 0.94 (95% CI, 0.93–0.95). In comparison, logistic regression had an AUROC of 0.65 (95% CI, 0.63–0.67). The extreme gradient boosting model maintained excellent performance at the intra‐ and postoperative stages, with area under the receiver operating characteristic curve values of 0.94 (95% CI, 0.93–0.95) and 0.98 (95% CI, 0.97–0.99), respectively. Calibration plots showed good agreement between predicted/observed event probabilities with Brier scores of 0.11 (preoperative), 0.11 (intraoperative), and 0.09 (postoperative). Conclusions Machine learning can accurately predict 1‐year stroke or death following TFCAS, performing better than logistic regression.

Introduction Genetic testing is used across medical disciplines leading to unprecedented demand for genetic services. This has resulted in excessive waitlists and unsustainable pressure on the standard model of genetic healthcare. Alternative models are needed; e-health tools represent scalable and evidence-based solution. We aim to evaluate the effectiveness of the Genetics Navigator, an interactive patient-centred digital platform that supports the collection of medical and family history, provision of pregenetic and postgenetic counselling and return of genetic testing results across paediatric and adult settings.Methods and analysis We will evaluate the effectiveness of the Genetics Navigator combined with usual care by a genetics clinician (physician or counsellor) to usual care alone in a randomised controlled trial. One hundred and thirty participants (adults patients or parents of paediatric patients) eligible for genetic testing through standard of care will be recruited across Ontario genetics clinics. Participants randomised into the intervention arm will use the Genetics Navigator for pretest and post-test genetic counselling and results disclosure in conjunction with their clinician. Participants randomised into the control arm will receive usual care, that is, clinician-delivered pretest and post-test genetic counselling, and results disclosure. The primary outcome is participant distress 2 weeks after test results disclosure. Secondary outcomes include knowledge, decisional conflict, anxiety, empowerment, quality of life, satisfaction, acceptability, digital health literacy and health resource use. Quantitative data will be analysed using statistical hypothesis tests and regression models. A subset of participants will be interviewed to explore user experience; data will be analysed using interpretive description. A cost-effectiveness analysis will examine the incremental cost of the Navigator compared with usual care per unit reduction in distress or unit improvement in quality of life from public payer and societal perspectives.Ethics and dissemination This study was approved by Clinical Trials Ontario. Results will be shared through stakeholder workshops, national and international conferences and peer-reviewed journals.Trial registration number NCT06455384.

Abstract Lower extremity open revascularization is a treatment option for peripheral artery disease that carries significant peri-operative risks; however, outcome prediction tools remain limited. Using machine learning (ML), we developed automated algorithms that predict 30-day outcomes following lower extremity open revascularization. The National Surgical Quality Improvement Program targeted vascular database was used to identify patients who underwent lower extremity open revascularization for chronic atherosclerotic disease between 2011 and 2021. Input features included 37 pre-operative demographic/clinical variables. The primary outcome was 30-day major adverse limb event (MALE; composite of untreated loss of patency, major reintervention, or major amputation) or death. Our data were split into training (70%) and test (30%) sets. Using tenfold cross-validation, we trained 6 ML models. Overall, 24,309 patients were included. The primary outcome of 30-day MALE or death occurred in 2349 (9.3%) patients. Our best performing prediction model was XGBoost, achieving an area under the receiver operating characteristic curve (95% CI) of 0.93 (0.92–0.94). The calibration plot showed good agreement between predicted and observed event probabilities with a Brier score of 0.08. Our ML algorithm has potential for important utility in guiding risk mitigation strategies for patients being considered for lower extremity open revascularization to improve outcomes.

Background Lower extremity endovascular revascularization for peripheral artery disease carries nonnegligible perioperative risks; however, outcome prediction tools remain limited. Using machine learning, we developed automated algorithms that predict 30‐day outcomes following lower extremity endovascular revascularization. Methods and Results The National Surgical Quality Improvement Program targeted vascular database was used to identify patients who underwent lower extremity endovascular revascularization (angioplasty, stent, or atherectomy) for peripheral artery disease between 2011 and 2021. Input features included 38 preoperative demographic/clinical variables. The primary outcome was 30‐day postprocedural major adverse limb event (composite of major reintervention, untreated loss of patency, or major amputation) or death. Data were split into training (70%) and test (30%) sets. Using 10‐fold cross‐validation, 6 machine learning models were trained using preoperative features. The primary model evaluation metric was area under the receiver operating characteristic curve. Overall, 21 886 patients were included, and 30‐day major adverse limb event/death occurred in 1964 (9.0%) individuals. The best performing model for predicting 30‐day major adverse limb event/death was extreme gradient boosting, achieving an area under the receiver operating characteristic curve of 0.93 (95% CI, 0.92–0.94). In comparison, logistic regression had an area under the receiver operating characteristic curve of 0.72 (95% CI, 0.70–0.74). The calibration plot showed good agreement between predicted and observed event probabilities with a Brier score of 0.09. The top 3 predictive features in our algorithm were (1) chronic limb‐threatening ischemia, (2) tibial intervention, and (3) congestive heart failure. Conclusions Our machine learning models accurately predict 30‐day outcomes following lower extremity endovascular revascularization using preoperative data with good discrimination and calibration. Prospective validation is warranted to assess for generalizability and external validity.

Abstract Background In January 2018, the Government of Ontario, Canada, initiated a universal pharmacare program (OHIP+) for all individuals aged 24 years and younger. In April 2019, the program was amended to cover only children and youth without private insurance. Because benzodiazepines are commonly prescribed to children and youth despite their potential hazards, we examined whether changes in publicly-funded drug coverage influenced benzodiazepine dispensing trends in this demographic. Methods We conducted a population-based natural experiment study of benzodiazepine dispensing to children and youth in Ontario between January 2013 and March 2020. We used interventional autoregressive integrated moving average models to estimate the impact of OHIP + and its subsequent modification on these trends. Results The implementation of OHIP + was associated with an immediate increase in the monthly rate of benzodiazepine dispensing of 12.9 individuals per 100,000 population (95% confidence interval [CI]; 7.5 to 18.3 per 100,000). Benzodiazepine dispensing rates rose from 214.2 to 241.5 per 100,000 from December 2017 to March 2019, a 12.8% (95% CI 9.6–16.0%) increase. In stratified analyses, increases were most pronounced among females, children and youth living in the lowest income neighbourhoods and individuals aged 20 to 24. The April 2019 modification to OHIP + was not associated with changes in monthly benzodiazepine dispensing trends (0.39 individuals per 100,000; 95% CI -1.3 to 2.1 per 100,000). However, rates remained elevated relative to the period preceding OHIP + implementation. Conclusions Implementation of a publicly-funded pharmacare program resulted in more children and youth being prescribed benzodiazepines.

Abdominal aortic aneurysm (AAA) is a progressive dilatation of the aorta that can lead to aortic rupture. The pathophysiology of the disease is not well characterized but is known to be caused by the general breakdown of the extracellular matrix within the aortic wall. In this comprehensive literature review, all current research on proteins that have been investigated for their potential prognostic capabilities in patients with AAA was included. A total of 45 proteins were found to be potential prognostic biomarkers for AAA, predicting incidence of AAA, AAA rupture, AAA growth, endoleak, and post-surgical mortality. The 45 proteins fell into the following seven general categories based on their primary function: (1) cardiovascular health, (2) hemostasis, (3) transport proteins, (4) inflammation and immunity, (5) kidney function, (6) cellular structure, (7) and hormones and growth factors. This is the most up-to-date literature review on current prognostic markers for AAA and their functions. This review outlines the wide pathophysiological processes that are implicated in AAA disease progression.

BackgroundThe COVID-19 pandemic was associated with increases in the prevalence of depression, anxiety and behavioural problems among children and youth. Less well understood is the influence of the pandemic on antidepressant and antipsychotic use among children. This is important, as it is possible that antidepressants and antipsychotics were used as a “stop-gap” measure to treat mental health symptoms when in-person access to outpatient care and school-based supportive services was disrupted. Furthermore, antipsychotics and antidepressants have been associated with harm in children and youth. We examined trends in dispensing of these medications two years following the pandemic among children 18 years of age and under in Ontario, Canada.MethodsWe conducted a population-based time-series study of antidepressant and antipsychotic medication dispensing to children and adolescents ≤18 years old between September 1, 2014, and March 31, 2022. We measured monthly population-adjusted rates of antidepressant and antipsychotics obtained from the IQVIA Geographic Prescription Monitor (GPM) database. We used structural break analyses to identify the pandemic month(s) when changes in the dispensing of antidepressants and antipsychotics occurred. We used interrupted time series models to quantify changes in dispensing following the structural break and compare observed and expected use of these drugs.ResultsOverall, we found higher-than-expected dispensing of antidepressants and antipsychotics in children and youth. Specifically, we observed an immediate step decrease in antidepressant dispensing associated with a structural break in April 2020 (−55.8 units per 1,000 individuals; 95% confidence intervals [CI] CI: −117.4 to 5.8), followed by an increased monthly trend in the rate of antidepressant dispensing of 13.0 units per 1,000 individuals (95% CI: 10.2–15.9). Antidepressant dispensing was consistently greater than predicted from September 2020 onward. Antipsychotic dispensing increased immediately following a June 2020 structural break (26.4 units per 1,000 individuals; 95% CI: 15.8–36.9) and did not change appreciably thereafter. Antipsychotic dispensing was higher than predicted at all time points from June 2020 onward.ConclusionWe found higher-than-expected dispensing of antidepressants and antipsychotics in children and youth. These increases were sustained through nearly two years of observation and are especially concerning in light of the potential for harm with the long-term use of antipsychotics in children. Further research is required to understand the clinical implications of these findings.

BackgroundWarfarin dosing in cardiac surgery patients is complicated by a heightened sensitivity to the drug, predisposing patients to adverse events. Predictive algorithms are therefore needed to guide warfarin dosing in cardiac surgery patients. ObjectiveThis study aimed to develop and validate an algorithm for predicting the warfarin dose needed to attain a therapeutic international normalized ratio (INR) at the time of discharge in cardiac surgery patients. MethodsWe abstracted variables influencing warfarin dosage from the records of 1031 encounters initiating warfarin between April 1, 2011, and November 29, 2019, at St Michael’s Hospital in Toronto, Ontario, Canada. We compared the performance of penalized linear regression, k-nearest neighbors, random forest regression, gradient boosting, multivariate adaptive regression splines, and an ensemble model combining the predictions of the 5 regression models. We developed and validated separate models for predicting the warfarin dose required for achieving a discharge INR of 2.0-3.0 in patients undergoing all forms of cardiac surgery except mechanical mitral valve replacement and a discharge INR of 2.5-3.5 in patients receiving a mechanical mitral valve replacement. For the former, we selected 80% of encounters (n=780) who had initiated warfarin during their hospital admission and had achieved a target INR of 2.0-3.0 at the time of discharge as the training cohort. Following 10-fold cross-validation, model accuracy was evaluated in a test cohort comprised solely of cardiac surgery patients. For patients requiring a target INR of 2.5-3.5 (n=165), we used leave-p-out cross-validation (p=3 observations) to estimate model performance. For each approach, we determined the mean absolute error (MAE) and the proportion of predictions within 20% of the true warfarin dose. We retrospectively evaluated the best-performing algorithm in clinical practice by comparing the proportion of cardiovascular surgery patients discharged with a therapeutic INR before (April 2011 and July 2019) and following (September 2021 and May 2, 2022) its implementation in routine care. ResultsRandom forest regression was the best-performing model for patients with a target INR of 2.0-3.0, an MAE of 1.13 mg, and 39.5% of predictions of falling within 20% of the actual therapeutic discharge dose. For patients with a target INR of 2.5-3.5, the ensemble model performed best, with an MAE of 1.11 mg and 43.6% of predictions being within 20% of the actual therapeutic discharge dose. The proportion of cardiovascular surgery patients discharged with a therapeutic INR before and following implementation of these algorithms in clinical practice was 47.5% (305/641) and 61.1% (11/18), respectively. ConclusionsMachine learning algorithms based on routinely available clinical data can help guide initial warfarin dosing in cardiac surgery patients and optimize the postsurgical anticoagulation of these patients.

Background Carotid endarterectomy (CEA) is a major vascular operation for stroke prevention that carries significant perioperative risks; however, outcome prediction tools remain limited. The authors developed machine learning algorithms to predict outcomes following CEA. Methods and Results The National Surgical Quality Improvement Program targeted vascular database was used to identify patients who underwent CEA between 2011 and 2021. Input features included 36 preoperative demographic/clinical variables. The primary outcome was 30‐day major adverse cardiovascular events (composite of stroke, myocardial infarction, or death). The data were split into training (70%) and test (30%) sets. Using 10‐fold cross‐validation, 6 machine learning models were trained using preoperative features. The primary metric for evaluating model performance was area under the receiver operating characteristic curve. Model robustness was evaluated with calibration plot and Brier score. Overall, 38 853 patients underwent CEA during the study period. Thirty‐day major adverse cardiovascular events occurred in 1683 (4.3%) patients. The best performing prediction model was XGBoost, achieving an area under the receiver operating characteristic curve of 0.91 (95% CI, 0.90–0.92). In comparison, logistic regression had an area under the receiver operating characteristic curve of 0.62 (95% CI, 0.60–0.64), and existing tools in the literature demonstrate area under the receiver operating characteristic curve values ranging from 0.58 to 0.74. The calibration plot showed good agreement between predicted and observed event probabilities with a Brier score of 0.02. The strongest predictive feature in our algorithm was carotid symptom status. Conclusions The machine learning models accurately predicted 30‐day outcomes following CEA using preoperative data and performed better than existing tools. They have potential for important utility in guiding risk‐mitigation strategies to improve outcomes for patients being considered for CEA.

Abstract Background Population-based research examining geographic variability in psychotropic medication dispensing to children and youth and the sociodemographic correlates of such variation is lacking. Variation in psychotropic use could reflect disparities in access to non-pharmacologic interventions and identify potentially concerning use patterns. Methods We conducted a population-based study of all Ontario residents aged 0 to 24 years who were dispensed a benzodiazepine, stimulant, antipsychotic or antidepressant between January 1, 2018, and December 31, 2018. We conducted small-area variation analyses and identified determinants of dispensing using negative binomial generalized estimating equation models. Results The age- and sex-standardized rate of psychotropic dispensing to children and youth was 76.8 (range 41.7 to 144.4) prescriptions per 1000 population, with large variation in psychotropic dispensing across Ontario’s census divisions. Males had higher antipsychotic [rate ratio (RR) 1.40; 95% confidence interval (CI) 1.36 to 1.44) and stimulant (RR 1.75; 95% CI 1.70 to 1.80) dispensing rates relative to females, with less use of benzodiazepines (RR 0.85; 95% CI 0.83 to 0.88) and antidepressants (RR 0.81; 95% CI 0.80 to 0.82). Lower antipsychotic dispensing was observed in the highest income neighbourhoods (RR 0.72; 95% CI 0.70 to 0.75) relative to the lowest. Benzodiazepine (RR 1.12; 95% CI 1.01 to 1.24) and stimulant (RR 1.11; 95% CI 1.01 to 1.23) dispensing increased with the density of mental health services in census divisions, whereas antipsychotic use decreased (RR 0.82; 95% CI 0.73 to 0.91). The regional density of child and adolescent psychiatrists and developmental pediatricians (RR 1.00; 95% CI 0.99 to 1.01) was not associated with psychotropic dispensing. Conclusion We found significant variation in psychotropic dispensing among young Ontarians. Targeted investment in regions with long wait times for publicly-funded non-pharmacological interventions and novel collaborative service models may minimize variability and promote best practices in using psychotropics among children and youth.

Introduction: Over 400,000 out-of-hospital cardiac arrest (OHCA) occur each year in Canada and the United States with less than 10% survival to hospital discharge. Cardiac arrest is a heterogenous condition and patient outcomes are impacted by a multitude of factors. Prognostication is recommended at 72 hours after return of spontaneous circulation (ROSC), however there may be other factors that could predict patient outcome earlier in the post-arrest period. The objective of our study was to develop and internally validate a novel clinical prediction rule to risk stratify patients early in the post-cardiac arrest period. Methods: We performed a retrospective cohort study of adult (≥18 years) post-cardiac arrest patients between 2010 and 2015 from the Epistry Cardiac Arrest database in Toronto. Our primary analysis used ordinal logistic regression to examine neurologic outcome at discharge using the modified Rankin Scale (mRS). Our secondary analysis used logistic regression for neurologic outcome and survival to hospital discharge. Models were internally validated using bootstrap validation. Results: A total of 3432 patients met our inclusion criteria. Our clinical prediction model was able to predict neurologic outcome on an ordinal scale using our predefined variables with an AUC of 0.89 after internal validation. The predictive performance was maintained when examining neurologic outcome as a binary variable and survival to hospital discharge. Conclusion: We were able to develop a model to accurately risk stratify adult cardiac arrest patients early in the post-cardiac arrest period. Future steps are needed to externally validate this model in other healthcare settings.

Background: Artificial intelligence (AI) and machine learning (ML) are rapidly advancing fields with increasing utility in health care. We conducted a survey to determine the perceptions of Canadian vascular surgeons toward AI/ML. Methods: An online questionnaire was distributed to 162 members of the Canadian Society for Vascular Surgery. Self-reported knowledge, attitudes, and perceptions with respect to potential applications, limitations, and facilitators of AI/ML were assessed. Results: Overall, 50 of the 162 Canadian vascular surgeons (31%) responded to the survey. Most respondents were aged 30 to 59 years (72%), male (80%), and White (67%) and practiced in academic settings (72%). One half of the participants reported that their knowledge of AI/ML was poor or very poor. Most were excited or very excited about AI/ML (66%) and were interested or very interested in learning more about the field (83.7%). The respondents believed that AI/ML would be useful or very useful for diagnosis (62%), prognosis (72%), patient selection (56%), image analysis (64%), intraoperative guidance (52%), research (88%), and education (80%). The limitations that the participants were most concerned about were errors leading to patient harm (42%), bias based on patient demographics (42%), and lack of clinician knowledge and skills in AI/ML (40%). Most were not concerned or were mildly concerned about job replacement (86%). The factors that were most important to encouraging clinicians to use AI/ML models were improvements in efficiency (88%), accurate predictions (84%), and ease of use (84%). The comments from respondents focused on the pressing need for the implementation of AI/ML in vascular surgery owing to the potential to improve care delivery. Conclusions: Canadian vascular surgeons have positive views on AI/ML and believe this technology can be applied to multiple aspects of the specialty to improve patient care, research, and education. Current self-reported knowledge is poor, although interest was expressed in learning more about the field. The facilitators and barriers to the effective use of AI/ML identified in the present study can guide future development of these tools in vascular surgery.

OBJECTIVES:. Hospital early warning systems that use machine learning (ML) to predict clinical deterioration are increasingly being used to aid clinical decision-making. However, it is not known how ML predictions complement physician and nurse judgment. Our objective was to train and validate a ML model to predict patient deterioration and compare model predictions with real-world physician and nurse predictions. DESIGN:. Retrospective and prospective cohort study. SETTING:. Academic tertiary care hospital. PATIENTS:. Adult general internal medicine hospitalizations. MEASUREMENTS AND MAIN RESULTS:. We developed and validated a neural network model to predict in-hospital death and ICU admission in 23,528 hospitalizations between April 2011 and April 2019. We then compared model predictions with 3,374 prospectively collected predictions from nurses, residents, and attending physicians about their own patients in 960 hospitalizations between April 30, and August 28, 2019. ML model predictions achieved clinician-level accuracy for predicting ICU admission or death (ML median F1 score 0.32 [interquartile range (IQR) 0.30-0.34], AUC 0.77 [IQ 0.76-0.78]; clinicians median F1-score 0.33 [IQR 0.30–0.35], AUC 0.64 [IQR 0.63–0.66]). ML predictions were more accurate than clinicians for ICU admission. Of all ICU admissions and deaths, 36% occurred in hospitalizations where the model and clinicians disagreed. Combining human and model predictions detected 49% of clinical deterioration events, improving sensitivity by 16% compared with clinicians alone and 24% compared with the model alone while maintaining a positive predictive value of 33%, thus keeping false alarms at a clinically acceptable level. CONCLUSIONS:. ML models can complement clinician judgment to predict clinical deterioration in hospital. These findings demonstrate important opportunities for human-computer collaboration to improve prognostication and personalized medicine in hospital.

BackgroundIn 2011, the Canadian Alliance for Monitoring Effectiveness and Safety of Antipsychotics in Children (CAMESA) published guidelines for the metabolic monitoring of antipsychotic-treated children and youth. Population-based studies examining adherence to these guidelines are needed to ensure the safe use of antipsychotics in children and youth.MethodsWe conducted a population-based study of all Ontario residents aged 0 to 24 who were newly dispensed an antipsychotic between April 1, 2018, and March 31, 2019. We estimated prevalence ratios (PRs) and 95% confidence intervals (CI) associating sociodemographic characteristics with the receipt of baseline and follow-up (3- and 6-month) laboratory testing using log-Poisson regression models.ResultsOverall, 6,505 of 27,718 (23.5%) children and youth newly dispensed an antipsychotic received at least one guideline-recommended baseline test. Monitoring was more prevalent among individuals aged 10 to 14 years (PR 1.20; 95% CI 1.04 to 1.38), 15 to 19 years (PR 1.60; 95% CI 1.41 to 1.82), and 20 to 24 years (PR 1.71; 95% CI 1.50 to 1.94) compared to children under the age of 10. Baseline monitoring was associated with mental health-related hospitalizations or emergency department visits in the year preceding therapy (PR 1.76; 95% CI 1.65 to 1.87), a prior diagnosis of schizophrenia (PR 1.20; 95% CI 1.14 to 1.26) or diabetes (PR 1.35; 95% CI 1.19 to 1.54), benzodiazepine use (PR 1.13; 95% CI 1.04 to 1.24), and receipt of a prescription from a child and adolescent psychiatrist or developmental pediatrician versus a family physician (PR 1.41; 95% CI 1.34 to 1.48). Conversely, monitoring was less frequent in individuals co-prescribed stimulants (PR 0.83; 95% CI 0.75 to 0.91). The prevalence of any 3- and 6-month follow-up monitoring among children and youth receiving continuous antipsychotic therapy at these time points was 13.0% (1,179 of 9,080) and 11.4% (597 of 5,261), respectively. Correlates of follow-up testing were similar to those of baseline monitoring.ConclusionMost children initiating antipsychotic therapy do not receive guideline-recommended metabolic laboratory monitoring. Further research is needed to understand reasons for poor guideline adherence and the role of clinician training and collaborative service models in promoting best monitoring practices.

Abstract Background The primary aim of this study was to determine if screen use in early childhood is associated with overall vulnerability in school readiness at ages 4 to 6 years, as measured by the Early Development Instrument (EDI). Secondary aims were to: (1) determine if screen use was associated with individual EDI domains scores, and (2) examine the association between screen use and EDI domains scores among a subgroup of high screen users. Methods This prospective cohort study was carried out using data from young children participating in a large primary care practice-based research network in Canada. Logistic regression analyses were run to investigate the association between screen use and overall vulnerability in school readiness. Separate linear regression models examined the relationships between children’s daily screen use and each separate continuous EDI domain. Results A total of 876 Canadian participants participated in this study. Adjusted logistic regression revealed an association between increased screen use and increased vulnerability in school readiness (p = 0.05). Results from adjusted linear regression demonstrated an association between higher screen use and reduced language and cognitive development domain scores (p = 0.004). Among high screen users, adjusted linear regression models revealed associations between increased screen use and reduced language and cognitive development (p = 0.004) and communication skills and general knowledge domain scores (p = 0.042). Conclusions Screen use in early childhood is associated with increased vulnerability in developmental readiness for school, with increased risk for poorer language and cognitive development in kindergarten, especially among high users.

Abstract Machine learning (ML) is a rapidly advancing field with increasing utility in health care. We conducted a systematic review and critical appraisal of ML applications in vascular surgery. MEDLINE, Embase, and Cochrane CENTRAL were searched from inception to March 1, 2021. Study screening, data extraction, and quality assessment were performed by two independent reviewers, with a third author resolving discrepancies. All original studies reporting ML applications in vascular surgery were included. Publication trends, disease conditions, methodologies, and outcomes were summarized. Critical appraisal was conducted using the PROBAST risk-of-bias and TRIPOD reporting adherence tools. We included 212 studies from a pool of 2235 unique articles. ML techniques were used for diagnosis, prognosis, and image segmentation in carotid stenosis, aortic aneurysm/dissection, peripheral artery disease, diabetic foot ulcer, venous disease, and renal artery stenosis. The number of publications on ML in vascular surgery increased from 1 (1991–1996) to 118 (2016–2021). Most studies were retrospective and single center, with no randomized controlled trials. The median area under the receiver operating characteristic curve (AUROC) was 0.88 (range 0.61–1.00), with 79.5% [62/78] studies reporting AUROC ≥ 0.80. Out of 22 studies comparing ML techniques to existing prediction tools, clinicians, or traditional regression models, 20 performed better and 2 performed similarly. Overall, 94.8% (201/212) studies had high risk-of-bias and adherence to reporting standards was poor with a rate of 41.4%. Despite improvements over time, study quality and reporting remain inadequate. Future studies should consider standardized tools such as PROBAST and TRIPOD to improve study quality and clinical applicability.

Cross-sector partnerships are vital for maintaining resilient health systems; however, few studies have sought to empirically assess the barriers and enablers of effective and responsible partnerships during public health emergencies. Through a qualitative, multiple case study, we analyzed 210 documents and conducted 26 interviews with stakeholders in three real-world partnerships between Canadian health organizations and private technology startups during the COVID-19 pandemic. The three partnerships involved: 1) deploying a virtual care platform to care for COVID-19 patients at one hospital, 2) deploying a secure messaging platform for physicians at another hospital, and 3) using data science to support a public health organization. Our results demonstrate that a public health emergency created time and resource pressures throughout a partnership. Given these constraints, early and sustained alignment on the core problem was critical for success. Moreover, governance processes designed for normal operations, such as procurement, were triaged and streamlined. Social learning, or the process of learning from observing others, offset some time and resource pressures. Social learning took many forms ranging from informal conversations between individuals at peer organisations (e.g., hospital chief information officers) to standing meetings at the local university's city-wide COVID-19 response table. We also found that startups' flexibility and understanding of the local context enabled them to play a highly valuable role in emergency response. However, pandemic fueled "hypergrowth" created risks for startups, such as introducing opportunities for deviation away from their core value proposition. Finally, we found each partnership navigated intense workloads, burnout, and personnel turnover through the pandemic. Strong partnerships required healthy, motivated teams. Visibility into and engagement in partnership governance, belief in partnership impact, and strong emotional intelligence in managers promoted team well-being. Taken together, these findings can help to bridge the theory-to-practice gap and guide effective cross-sector partnerships during public health emergencies.

Carotid artery stenosis (CAS), an atherosclerotic disease of the carotid artery, is one of the leading causes of transient ischemic attacks (TIA) and cerebrovascular attacks (CVA). The atherogenic process of CAS affects a wide range of physiological processes, such as inflammation, endothelial cell function, smooth muscle cell migration and many more. The current gold-standard test for CAS is Doppler ultrasound; however, there is yet to be determined a strong, clinically validated biomarker in the blood that can diagnose patients with CAS and/or predict adverse outcomes in such patients. In this comprehensive literature review, we evaluated all of the current research on plasma and serum proteins that are current contenders for biomarkers for CAS. In this literature review, 36 proteins found as potential biomarkers for CAS were categorized in to the following nine categories based on protein function: (1) Inflammation and Immunity, (2) Lipid Metabolism, (3) Haemostasis, (4) Cardiovascular Markers, (5) Markers of Kidney Function, (6) Bone Health, (7) Cellular Structure, (8) Growth Factors, and (9) Hormones. This literature review is the most up-to-date and current comprehensive review of research on biomarkers of CAS, and the only review that demonstrated the several pathways that contribute to the initiation and progression of the disease. With this review, future studies can determine if any new markers, or a panel of the proteins explored in this study, may be contenders as diagnostic or prognostic markers for CAS.

Abstract Machine learning (ML) holds great promise in transforming healthcare. While published studies have shown the utility of ML models in interpreting medical imaging examinations, these are often evaluated under laboratory settings. The importance of real world evaluation is best illustrated by case studies that have documented successes and failures in the translation of these models into clinical environments. A key prerequisite for the clinical adoption of these technologies is demonstrating generalizable ML model performance under real world circumstances. The purpose of this study was to demonstrate that ML model generalizability is achievable in medical imaging with the detection of intracranial hemorrhage (ICH) on non-contrast computed tomography (CT) scans serving as the use case. An ML model was trained using 21,784 scans from the RSNA Intracranial Hemorrhage CT dataset while generalizability was evaluated using an external validation dataset obtained from our busy trauma and neurosurgical center. This real world external validation dataset consisted of every unenhanced head CT scan (n = 5965) performed in our emergency department in 2019 without exclusion. The model demonstrated an AUC of 98.4%, sensitivity of 98.8%, and specificity of 98.0%, on the test dataset. On external validation, the model demonstrated an AUC of 95.4%, sensitivity of 91.3%, and specificity of 94.1%. Evaluating the ML model using a real world external validation dataset that is temporally and geographically distinct from the training dataset indicates that ML generalizability is achievable in medical imaging applications.

BackgroundDeploying safe and effective machine learning models is essential to realize the promise of artificial intelligence for improved healthcare. Yet, there remains a large gap between the number of high-performing ML models trained on healthcare data and the actual deployment of these models. Here, we describe the deployment of CHARTwatch, an artificial intelligence-based early warning system designed to predict patient risk of clinical deterioration.MethodsWe describe the end-to-end infrastructure that was developed to deploy CHARTwatch and outline the process from data extraction to communicating patient risk scores in real-time to physicians and nurses. We then describe the various challenges that were faced in deployment, including technical issues (e.g., unstable database connections), process-related challenges (e.g., changes in how a critical lab is measured), and challenges related to deploying a clinical system in the middle of a pandemic. We report various measures to quantify the success of the deployment: model performance, adherence to workflows, and infrastructure uptime/downtime. Ultimately, success is driven by end-user adoption and impact on relevant clinical outcomes. We assess our deployment process by evaluating how closely we followed existing guidance for good machine learning practice (GMLP) and identify gaps that are not addressed in this guidance.ResultsThe model demonstrated strong and consistent performance in real-time in the first 19 months after deployment (AUC 0.76) as in the silent deployment heldout test data (AUC 0.79). The infrastructure remained online for >99% of time in the first year of deployment. Our deployment adhered to all 10 aspects of GMLP guiding principles. Several steps were crucial for deployment but are not mentioned or are missing details in the GMLP principles, including the need for a silent testing period, the creation of robust downtime protocols, and the importance of end-user engagement. Evaluation for impacts on clinical outcomes and adherence to clinical protocols is underway.ConclusionWe deployed an artificial intelligence-based early warning system to predict clinical deterioration in hospital. Careful attention to data infrastructure, identifying problems in a silent testing period, close monitoring during deployment, and strong engagement with end-users were critical for successful deployment.

BackgroundThe Expanded Disability Status Scale (EDSS) score is a widely used measure to monitor disability progression in people with multiple sclerosis (MS). However, extracting and deriving the EDSS score from unstructured electronic health records can be time-consuming. ObjectiveWe aimed to compare rule-based and deep learning natural language processing algorithms for detecting and predicting the total EDSS score and EDSS functional system subscores from the electronic health records of patients with MS. MethodsWe studied 17,452 electronic health records of 4906 MS patients followed at one of Canada’s largest MS clinics between June 2015 and July 2019. We randomly divided the records into training (80%) and test (20%) data sets, and compared the performance characteristics of 3 natural language processing models. First, we applied a rule-based approach, extracting the EDSS score from sentences containing the keyword “EDSS.” Next, we trained a convolutional neural network (CNN) model to predict the 19 half-step increments of the EDSS score. Finally, we used a combined rule-based–CNN model. For each approach, we determined the accuracy, precision, recall, and F-score compared with the reference standard, which was manually labeled EDSS scores in the clinic database. ResultsOverall, the combined keyword-CNN model demonstrated the best performance, with accuracy, precision, recall, and an F-score of 0.90, 0.83, 0.83, and 0.83 respectively. Respective figures for the rule-based and CNN models individually were 0.57, 0.91, 0.65, and 0.70, and 0.86, 0.70, 0.70, and 0.70. Because of missing data, the model performance for EDSS subscores was lower than that for the total EDSS score. Performance improved when considering notes with known values of the EDSS subscores. ConclusionsA combined keyword-CNN natural language processing model can extract and accurately predict EDSS scores from patient records. This approach can be automated for efficient information extraction in clinical and research settings.

Surgical departments commonly rely on third-party quality improvement registries. As electronic health data become increasingly integrated and accessible within an institution, alternatives to these platforms arise. We present the conceptualization and implementation of an in-house quality improvement platform that provides real-time reports, is less onerous on clinicians and is tailored to an institution’s priorities of care.

Abstract Background High orphan drug prices have gained the attention of payers and policy makers. These prices may reflect the need to recoup the cost of drug development from a small patient pool. However, estimates of the cost of orphan drug development are sparse. Methods Using publicly available data, we estimated the differences in trial characteristics and clinical development costs with 100 orphan and 100 non-orphan drugs. Results We found that the out-of-pocket clinical costs per approved orphan drug to be $166 million and $291 million (2013 USD) per non-orphan drug. The capitalized clinical costs per approved orphan drug and non-orphan drug were estimated to be $291 million and $412 million respectively. When focusing on new molecular entities only, we found that the capitalized clinical cost per approved orphan drug was half that of a non-orphan drug. Conclusions More discussion is needed to better align on which cost components should be included in research and development costs for pharmaceuticals.

BackgroundStigma and high-care needs can present barriers to the provision of high-quality primary care for people with opioid use disorder (OUD) and those prescribed opioids for chronic pain. We explored the likelihood of securing a new primary care provider (PCP) among people with varying histories of opioid use who had recently lost access to their PCP.Methods and findingsWe conducted a retrospective cohort study using linked administrative data among residents of Ontario, Canada whose enrolment with a physician practicing in a primary care enrolment model (PEM) was terminated between January 2016 and December 2017. We assigned individuals to 3 groups based upon their opioid use on the date enrolment ended: long-term opioid pain therapy (OPT), opioid agonist therapy (OAT), or no opioid. We fit multivariable models assessing the primary outcome of primary care reattachment within 1 year, adjusting for demographic characteristics, clinical comorbidities, and health services utilization. Secondary outcomes included rates of emergency department (ED) visits and opioid toxicity events. Among 154,970 Ontarians who lost their PCP, 1,727 (1.1%) were OAT recipients, 3,644 (2.4%) were receiving long-term OPT, and 149,599 (96.5%) had no recent prescription opioid exposure. In general, OAT recipients were younger (median age 36) than those receiving long-term OPT (59 years) and those with no recent prescription opioid exposure (44 years). In all exposure groups, the majority of individuals had their enrolment terminated by their physician (range 78.1% to 88.8%). In the primary analysis, as compared to those not receiving opioids, OAT recipients were significantly less likely to find a PCP within 1 year (adjusted hazard ratio [aHR] 0.55, 95% confidence interval [CI] 0.50 to 0.61, p < 0.0001). We observed no significant difference between long-term OPT and opioid unexposed individuals (aHR 0.96; 95% CI 0.92 to 1.01, p = 0.12). In our secondary analysis comparing the period of PCP loss to the year prior, we found that rates of ED visits were elevated among people not receiving opioids (adjusted rate ratio (aRR) 1.20, 95% CI 1.18 to 1.22, p < 0.0001) and people receiving long-term OPT (aRR 1.37, 95% CI 1.28 to 1.48, p < 0.0001). We found no such increase among OAT recipients, and no significant increase in opioid toxicity events in the period following provider loss for any exposure group. The main limitation of our findings relates to their generalizability outside of PEMs and in jurisdictions with different financial incentives incorporated into primary care provision.ConclusionsIn this study, we observed gaps in access to primary care among people who receive prescription opioids, particularly among OAT recipients. Ongoing efforts are needed to address the stigma, discrimination, and financial disincentives that may introduce barriers to the healthcare system, and to facilitate access to high-quality, consistent primary care services for chronic pain patients and those with OUD.

BackgroundAdherence to medicines is low for a variety of reasons, including the cost borne by patients. Some jurisdictions publicly fund medicines for the general population, but many jurisdictions do not, and such policies are contentious. To our knowledge, no trials studying free access to a wide range of medicines have been conducted.Methods and findingsWe randomly assigned 786 primary care patients who reported not taking medicines due to cost between June 1, 2016 and April 28, 2017 to either free distribution of essential medicines (n = 395) or to usual medicine access (n = 391). The trial was conducted in Ontario, Canada, where hospital care and physician services are publicly funded for the general population but medicines are not. The trial population was mostly female (56%), younger than 65 years (83%), white (66%), and had a low income from wages as the primary source (56%). The primary outcome was medicine adherence after 2 years. Secondary outcomes included control of diabetes, blood pressure, and low-density lipoprotein (LDL) cholesterol in patients taking relevant treatments and healthcare costs over 2 years. Adherence to all appropriate prescribed medicines was 38.7% in the free distribution group and 28.6% in the usual access group after 2 years (absolute difference 10.1%; 95% confidence interval (CI) 3.3 to 16.9, p = 0.004). There were no statistically significant differences in control of diabetes (hemoglobin A1c 0.27; 95% CI -0.25 to 0.79, p = 0.302), systolic blood pressure (-3.9; 95% CI -9.9 to 2.2, p = 0.210), or LDL cholesterol (0.26; 95% CI -0.08 to 0.60, p = 0.130) based on available data. Total healthcare costs over 2 years were lower with free distribution (difference in median CAN$1,117; 95% CI CAN$445 to CAN$1,778, p = 0.006). In the free distribution group, 51 participants experienced a serious adverse event, while 68 participants in the usual access group experienced a serious adverse event (p = 0.091). Participants were not blinded, and some outcomes depended on participant reports.ConclusionsIn this study, we observed that free distribution of essential medicines to patients with cost-related nonadherence substantially increased adherence, did not affect surrogate health outcomes, and reduced total healthcare costs over 2 years.Trial registrationClinicalTrials.gov NCT02744963.

BackgroundTuberculosis (TB) is a major cause of death worldwide. TB research draws heavily on clinical cohorts which can be generated using electronic health records (EHR), but granular information extracted from unstructured EHR data is limited. The St. Michael's Hospital TB database (SMH-TB) was established to address gaps in EHR-derived TB clinical cohorts and provide researchers and clinicians with detailed, granular data related to TB management and treatment.MethodsWe collected and validated multiple layers of EHR data from the TB outpatient clinic at St. Michael's Hospital, Toronto, Ontario, Canada to generate the SMH-TB database. SMH-TB contains structured data directly from the EHR, and variables generated using natural language processing (NLP) by extracting relevant information from free-text within clinic, radiology, and other notes. NLP performance was assessed using recall, precision and F1 score averaged across variable labels. We present characteristics of the cohort population using binomial proportions and 95% confidence intervals (CI), with and without adjusting for NLP misclassification errors.ResultsSMH-TB currently contains retrospective patient data spanning 2011 to 2018, for a total of 3298 patients (N = 3237 with at least 1 associated dictation). Performance of TB diagnosis and medication NLP rulesets surpasses 93% in recall, precision and F1 metrics, indicating good generalizability. We estimated 20% (95% CI: 18.4-21.2%) were diagnosed with active TB and 46% (95% CI: 43.8-47.2%) were diagnosed with latent TB. After adjusting for potential misclassification, the proportion of patients diagnosed with active and latent TB was 18% (95% CI: 16.8-19.7%) and 40% (95% CI: 37.8-41.6%) respectively.ConclusionSMH-TB is a unique database that includes a breadth of structured data derived from structured and unstructured EHR data by using NLP rulesets. The data are available for a variety of research applications, such as clinical epidemiology, quality improvement and mathematical modeling studies.

Objective Trauma-related vascular injuries are major contributors to morbidity and mortality worldwide. We conducted a retrospective, population-based, cross-sectional study to examine temporal trends and factors associated with traumatic vascular injury-related in-hospital mortality in Ontario, Canada from 1991 to 2009. Methods We obtained data on Ontario hospital admissions for traumatic vascular injury, including injury mechanism and body region; and patient age, sex, socioeconomic status, and residence from the Canadian Institute for Health Information Discharge Abstract Database and Registered Persons Database from fiscal years 1991 to 2009. We performed time series analysis of vascular injury-related in-hospital mortality rates and multivariable logistic regression analysis to identify significant mortality-associated factors. Results The overall in-hospital mortality rate for trauma-related vascular injury was 5.5%. A slight but non-significant decline in mortality occurred over time. The likelihood of vascular injury-related in-hospital mortality was significantly higher for patients involved in transport-related accidents (odds ratio [OR[=2.21, 95% confidence interval [CI], 1.76–2.76), age ≥65 years (OR = 4.34, 95% CI, 2.25–8.38), or with thoracic (OR = 2.24, 95% CI, 1.56–3.20) or abdominal (OR = 2.45, 95% CI, 1.75–3.42) injuries. Conclusions In-hospital mortality from traumatic vascular injury in Ontario was low and stable from 1991 to 2009.

BackgroundAttention deficit hyperactivity disorder (ADHD) affects approximately 3% of adults globally. Many pharmacologic treatments options exist, yet the comparative benefits and harms of individual treatments are largely unknown. We performed a systematic review and network meta-analysis to assess the relative effects of individual pharmacologic treatments for adults with ADHD.MethodsWe searched English-language published and grey literature sources for randomized clinical trials (RCTs) involving pharmacologic treatment of ADHD in adults (December 2018). The primary outcome was clinical response; secondary outcomes were quality of life, executive function, driving behaviour, withdrawals due to adverse events, treatment discontinuation, serious adverse events, hospitalization, cardiovascular adverse events, and emergency department visits. Data were pooled via pair-wise meta-analyses and Bayesian network meta-analyses. Risk of bias was assessed by use of Cochrane's Risk of Bias tool, and the certainty of the evidence was assessed by use of the GRADE framework.ResultsEighty-one unique trials that reported at least one outcome of interest were included, most of which were at high or unclear risk of at least one important source of bias. Notably, only 5 RCTs were deemed at overall low risk of bias. Included pharmacotherapies were methylphenidate, atomoxetine, dexamfetamine, lisdexamfetamine, guanfacine, bupropion, mixed amphetamine salts, and modafinil. As a class, ADHD pharmacotherapy improved patient- and clinician-reported clinical response compared with placebo (range: 4 to 15 RCTs per outcome); however, these findings were not conserved when the analyses were restricted to studies at low risk of bias, and the certainty of the finding is very low. There were few differences among individual medications, although atomoxetine was associated with improved patient-reported clinical response and quality of life compared with placebo. There was no significant difference in the risk of serious adverse events or treatment discontinuation between ADHD pharmacotherapies and placebo; however, the proportion of participants who withdrew due to adverse events was significantly higher among participants who received any ADHD pharmacotherapy. Few RCTs reported on the occurrence of adverse events over a long treatment duration.ConclusionsOverall, despite a class effect of improving clinical response relative to placebo, there were few differences among the individual ADHD pharmacotherapies, and most studies were at risk of at least one important source of bias. Furthermore, the certainty of the evidence was very low to low for all outcomes, and there was limited reporting of long-term adverse events. As such, the choice between ADHD pharmacotherapies may depend on individual patient considerations, and future studies should assess the long-term effects of individual pharmacotherapies on patient-important outcomes, including quality of life, in robust blinded RCTs.RegistrationPROSPERO no. CRD 42015026049.

Abstract Background A randomized controlled trial of adults with empyema recently demonstrated decreased length of stay in hospital in patients treated with intrapleurally administered dornase alfa and fibrinolytics compared to fibrinolytics alone. Whether this treatment strategy is safe and effective in children remains unknown. Methods/design This study protocol is for a superiority, placebo-controlled, parallel-design, multicenter randomized controlled trial. The participants are previously well children admitted to a children’s hospital with a diagnosis of empyema requiring chest tube insertion and fibrinolytics administered intrapleurally. Children will be randomized after the treating physician has decided that pleural drainage is required but prior to chest tube insertion. After chest tube insertion, participants in the treatment group will receive intrapleurally administered tissue plasminogen activator (tPA) 4 mg followed by dornase alfa 5 mg. Participants in the placebo group will receive tPA 4 mg followed by normal saline. Study treatments will be administered once daily for 3 days. All participants, parents or caregivers, clinicians, and research personnel will remain blinded. The primary outcome is length of stay from chest tube insertion to discharge from hospital. Secondary outcomes include time to meeting discharge criteria, chest tube duration, fever duration, need for additional procedures, adverse events, hospital readmission, cost of hospitalization, and mortality. Discussion This multicenter randomized controlled trial will assess the safety, effectiveness, and cost-effectiveness of combined treatment with dornase alfa and fibrinolytics compared to fibrinolytics alone for the treatment of empyema in children. Trial registration ClinicalTrials.gov: NCT01717742 . Registered on 8 October 2012.

Introduction School readiness is a multidimensional construct that includes cognitive, behavioural and emotional aspects of a child’s development. School readiness is strongly associated with a child’s future school success and well-being. The Early Development Instrument (EDI) is a reliable and valid teacher-completed tool for assessing school readiness in children at kindergarten age. A substantial knowledge gap exists in understanding how early child growth, health behaviours, nutrition, cardiometabolic risk and development impact school readiness. The primary objective was to determine if growth patterns, measured by body mass index trajectories in healthy children aged 0–5 years, are associated with school readiness at ages 4–6 years (kindergarten age). Secondary objectives were to determine if other health trajectories, including health behaviours, nutrition, cardiometabolic risk and development, are associated with school readiness at ages 4–6 years. This paper presents the Fit for School Study protocol.Methods and analysis This is an ongoing prospective cohort study. Parents of children enrolled in the The Applied Health Research Group for Kids (TARGet Kids!) practice-based research network are invited to participate in the Fit for School Study. Child growth, health behaviours, nutrition, cardiometabolic risk and development data are collected annually at health supervision visits and linked to EDI data collected by schools. The primary and secondary analyses will use a two-stage process: (1) latent class growth models will be used to first determine trajectory groups, and (2) generalised linear mixed models will be used to examine the relationship between exposures and EDI results.Ethics and dissemination The research ethics boards at The Hospital for Sick Children, Unity Health Toronto and McMaster University approved this study, and research ethics approval was obtained from each school board with a student participating in the study. The findings will be presented locally, nationally and internationally and will be published in peer-reviewed journals.Trial registration number NCT01869530.

Background: Therapeutic inertia (TI) is a common phenomenon among physicians who care for patients with chronic conditions. We evaluated the efficacy of the traffic light system (TLS) educational intervention to reduce TI among neurologists with MS expertise. Methods: In this randomised, controlled trial, 90 neurologists who provide care to MS patients were randomly assigned to the TLS intervention ( n = 45) or to the control group ( n = 45). The educational intervention employed the TLS, a behavioral strategy that facilitates therapeutic choices by facilitating reflective decisions. The TLS consisted in a short, structured, single session intervention of 5-7 min duration. Participants made therapeutic choices of 10 simulated case-scenarios. The primary outcome was a reduction in TI based on a published TI score (case-scenarios in which a participant showed TI divided by the total number of scenarios where TI was possible ranging from 0 to 8). Results: All participants completed the study and were included in the primary analysis. TI was lower in the TLS group (1.47, 95% CI 1.32-1.61) compared to controls (1.93; 95% CI 1.79-2.08). The TLS group had a lower prevalence of TI compared to controls (0.67, 95% CI 0.62-0.71 vs. 0.82, 95% CI 0.78-0.86; p = 0.001). The multivariate analysis, adjusted for age, specialty, years of practice, and risk preference showed a 70% reduction in TI for the TLS intervention compared to controls (OR 0.30; 95% CI 0.10-0.89). Conclusions: In this randomized trial, the TLS strategy decreases the incidence of TI in MS care irrespective of age, expertise, years for training, and risk preference of participants, which would lead to better patient outcomes.

Abstract Background Mental illness is a substantial and rising contributor to the global burden of disease. Access to and utilization of mental health care, however, is limited by structural barriers such as specialist availability, time, out-of-pocket costs, and attitudinal barriers including stigma. Innovative solutions like virtual care are rapidly entering the health care domain. The advancement and adoption of virtual care for mental health, however, often occurs in the absence of rigorous evaluation and adequate planning for sustainability and spread. Methods A pragmatic randomized controlled trial with a nested comparative effectiveness arm, and concurrent realist process evaluation to examine acceptability, effectiveness, and cost-effectiveness of the Big White Wall (BWW) online platform for mental health self-management and peer support among individuals aged 16 and older who are accessing mental health services in Ontario, Canada. Participants will be randomized to 3 months of BWW or treatment as usual. At the end of the 3 months, participants in the intervention group will have the opportunity to opt-in to an intervention extension arm. Those who opt-in will be randomized to receive an additional 3 months of BWW or no additional intervention. The primary outcome is recovery at 3 months as measured by the Recovery Assessment Scale-revised (RAS-r). Secondary outcomes include symptoms of depression and anxiety measured with the Personal Health Questionnaire-9 item (PHQ-9) and the Generalized Anxiety Disorder Questionnaire-7 item (GAD-7) respectively, quality of life measured with the EQ-5D-5L, and community integration assessed with the Community Integration Questionnaire. Cost-effectiveness evaluations will account for the cost of the intervention and direct health care costs. Qualitative interviews with participants and stakeholders will be conducted throughout. Discussion Understanding the impact of virtual strategies, such as BWW, on patient outcomes and experience, and health system costs is essential for informing whether and how health system decision-makers can support these strategies system-wide. This requires clear evidence of effectiveness and an understanding of how the intervention works, for whom, and under what circumstances. This study will produce such effectiveness data for BWW, while simultaneously exploring the characteristics and experiences of users for whom this and similar online interventions could be helpful. Trial registration Clinicaltrials.gov NCT02896894 . Registered on 31 August 2016 (retrospectively registered).

Importance: The prescription of generic (non-proprietary) compared to brand-name drugs is increasing worldwide. In many developing and emerging countries, generics companies market products at similar costs as brand-name competitors benefiting from more flexible compliance rules and regulations for marketing their products in the health system. Together, this phenomenon may influence prescriber's behavior (e.g., maintaining the same treatment despite guideline's recommendations or despite evidence of disease progression).Objectives: To compare the prevalence of therapeutic inertia (TI) between primary prescription of brand-name vs. generic drugs in the management of MS in Argentina.Design: We conducted a population-based online study comprising 117 neurologists with expertise in MS. Participants answered questions regarding their clinical practice, most commonly prescribed disease modifying agents, and therapeutic choices of 10 simulated case-scenarios that assessed TI. Inertia was defined as the lack of treatment initiation or escalation despite evidence of clinical and radiological activity (8 case-scenarios, 720 individual responses). We created the generic-brand name score (GBS) according to the 5 most frequently prescribed generic (n = 16) vs. brand-name (n = 9) drugs for MS, where scores higher than 1 indicated higher prescription of generic drugs and scores lower than 1 indicated higher prescription of brand-name agents. Candidate predictors of prescribing generic drugs included demographic data, MS specialist vs. general neurologist, practice setting, years of practice, volume of MS patients, risk preferences, costs of annual treatment.Participants and setting: population-based prospective study using including neurologists who care for patients with multiple sclerosis across Argentina.Exposure: prescription of generic vs. brand-name MS drugsMain outcome of interest: Therapeutic inertia (TI), defined as lack of treatment escalation when goals are unmet. Secondary outcomes included factors associated with generic drug prescription and costs of MS treatment.Results: Ninety participants completed the study (completion rate 76.9%). TI was observed in 153 (21.3%) of participants' responses. The evaluation of aggregate responses revealed a mean GBS score (SD) of 3.44 (2.1), with 46 (51.1%) participants having a GBS equal to or higher than 1. Older age (OR 1.19; 95% CI 1.00–1.42), being a general neurologist (OR 3.91; 95% CI 1.19–12.8), and being more willing to take risks in multiple domains (SOEP score OR 1.06, 95% CI 1.01–1.12) were associated with higher prescription of generic drugs in MS care. Costs of treatment were not associated with prescribing generic drugs. There was no difference in the annual costs of MS treatment for primary prescribers of brand-name vs. generic drugs (67,500 US$ vs. 67,496 US$; p = 0.99).The evaluation of individual responses revealed that participants with higher prescription of generics—reflected by a higher GBS—had higher incident risk of TI (mean GBS 3.61 for TI vs. 2.96 for no TI; p < 0.001). Multivariate analysis revealed that a prescription of generic agents was associated with an increased incident risk of TI (OR 1.56; 95%CI 1.07–2.29). There was no difference in the annual costs of MS treatment for participants that exhibited TI vs. those without TI (67,426 US$ vs. 67,704 US$; p = 0.66).Conclusions: General neurologist, older age, and willingness to take risks were associated with increased prescription of generic drugs despite similar costs compared to brand-name agents. In our study, the prescription of generic-MS drugs was associated with a higher incident risk of therapeutic inertia.

Background There are no recognized pharmacological treatments for abdominal aortic aneurysms (AAA), although statins are suggested to be beneficial. We sought to summarize the literature regarding the effects of statins on human AAA growth, rupture, and 30‐day mortality. Methods and Results We conducted a systematic review and meta‐analysis of randomized and observational studies using the Cochrane CENTRAL database, MEDLINE, and EMBASE up to June 15, 2018. Review, abstraction, and quality assessment were conducted by 2 independent reviewers, and a third author resolved discrepancies. Pooled mean differences and odds ratios with 95% confidence intervals were calculated using random effects models. Heterogeneity was quantified using the I2 statistic, and publication bias was assessed using funnel plots. Our search yielded 911 articles. One case‐control and 21 cohort studies involving 80 428 patients were included. The risk of bias was low to moderate. Statin use was associated with a mean AAA growth rate reduction of 0.82 mm/y (95% confidence interval 0.33, 1.32, P=0.001, I2=86%). Statins were also associated with a lower rupture risk (odds ratio 0.63, 95% confidence interval 0.51, 0.78, P

Introduction The Ontario Narcotics Monitoring System (NMS) captures information on all prescriptions for controlled medications dispensed from outpatient pharmacies in Ontario, Canada, regardless of payer. This system was introduced in 2012, as a strategy to promote appropriate prescribing and dispensing practices. Objectives and Approach We sought to explore the degree to which prescriptions in the NMS can be linked to other health claims databases, and describe the types of medications dispensed between July 2012 and December 2016. We also linked opioid prescriptions to hospitalization and mortality data to examine the relative contributions of prescribed and non-prescribed opioids to opioid toxicity events in 2016. A recent opioid prescription was defined as a prescription with a days’ supply that overlapped the opioid toxicity event. Analyses were stratified by gender and age. Results We examined 1.3 million prescriptions in the NMS during the study period: 72.8% for opioids, 21% for benzodiazepines, 4.4% for stimulants and

Background Statins are commonly used for the prevention of cardiovascular events; however, statins are underutilized in patients with noncoronary atherosclerosis. We sought to establish the rates of statin use in patients with carotid artery disease and to examine the association between statin therapy and outcomes after carotid revascularization. Methods and Results In this population‐level retrospective cohort study, we identified all individuals aged ≥66 years who underwent carotid endarterectomy or stenting in Ontario, Canada (2002–2014). The primary outcome was a composite of 1‐year stroke, myocardial infarction, or death (major adverse cardiac and cerebrovascular events). Five‐year risks were also examined. Adjusted hazard ratios were computed using inverse probability of treatment weighting based on propensity scores. A total of 7893 of 10 723 patients (73.6%) who underwent carotid revascularization were on preprocedural statin therapy; moderate‐ or high‐dose therapy was utilized by 7384 patients (68.9%). The composite rate of 1‐year major adverse cardiac and cerebrovascular events was lower among statin users (adjusted hazard ratio: 0.76; 95% confidence interval, 0.70–0.83). Patients who were on persistent long‐term statin therapy after the carotid procedure continued to experience significantly lower risk of major adverse cardiac and cerebrovascular events at 5 years (adjusted hazard ratio: 0.75, 95% confidence interval, 0.71–0.80). The beneficial associations with statin use were observed regardless of type of carotid revascularization procedure, carotid artery symptom status, or statin dose. Conclusions Continuous statin therapy was associated with a 25% lower risk of long‐term adverse cardiovascular events in patients with significant carotid disease. Along with other supportive evidence, statins should be considered in patients undergoing carotid revascularization, and efforts are required to increase statin use in this undertreated population.

Background: Educational interventions are needed to overcome knowledge-to-action gaps in clinical care. We previously tested the feasibility and potential efficacy of an educational intervention that facilitates treatment decisions in multiple sclerosis care. A demonstration of the usability of such an intervention is crucial prior to demonstration of efficacy in a large trial.Objectives: To evaluate the usability of a novel, pilot-tested intervention aimed at neurologists to improve therapeutic decisions in multiple sclerosis (MS) care.Methods: We surveyed 50 neurologists from Chile, Argentina, and Canada randomized to an educational intervention arm of a pilot feasibility study using the System Usability Score (SUS) to assess the usability of a traffic light system (TLS)-based educational intervention. The TLS facilitates therapeutic decisions, allowing participants to easily recognize high-risk scenarios requiring treatment escalation. The SUS is a validated 10-item questionnaire with five response options. The primary outcome was the average and 95% confidence interval (CI) of the SUS score. Values above 68 are considered highly usable.Results: Of 50 neurologists invited to be part of the study, all completed the SUS scale and the study. For the primary outcome, the average usability score was 74.7 (95%CI 70.1–79.2). There was one outlier with a score of 35. The usability score excluding the outlier was 76.8 (95%CI 72.7–80.8). Multivariate analysis revealed no association between participants' characteristics and the SUS score.Conclusions: Our educational intervention has shown high usability among neurologists. The next step is to evaluate the effectiveness of this educational intervention in facilitating treatment decisions for the management of multiple sclerosis in a large trial.

BackgroundPhysicians often do not initiate or intensify treatments when clearly warranted, a phenomenon known as therapeutic inertia (TI). Limited information is available on educational interventions to ameliorate knowledge-to-action gaps in TI.ObjectivesTo evaluate the feasibility and efficacy of an educational intervention compared to usual care among practicing neurologists caring for patients with multiple sclerosis (MS).MethodsWe conducted a pilot double-blind, parallel-group, randomized clinical trial. Inclusion criteria included neurologists who are actively involved in managing MS patients. Participants were exposed to 20 simulated case-scenarios (10 cases at baseline, and 10 cases post-randomization to usual care vs. educational intervention) of relapsing–remitting MS with moderate or high risk of disease progression. The educational intervention employed a traffic light system (TLS) to facilitate decisions, allowing participants to easily recognize high-risk scenarios requiring treatment escalation. We also measured differences between blocks to invoke decision fatigue. The control group responded as they would do in their usual clinical practice not exposed to the educational intervention. The primary feasibility outcome was the proportion of participants who completed the study and the proportion of participants who correctly identified a high-risk case-scenario with the “red traffic light.” Secondary outcomes included decision fatigue (defined as an increment of TI in the second block of case-scenarios compared to the first block) and the efficacy of the educational intervention measured as a reduction in TI for MS treatment.ResultsOf 30 neurologists invited to be part of the study, the participation rate was 83.3% (n = 25). Of the 25 participants, 14 were randomly assigned to the control group and 11 to the intervention group. TI was present in 72.0% of participants in at least one case scenario. For the primary feasibility outcome, the completion rate of the study was 100% (25/25 participants). Overall, 77.4% of participants correctly identified the “red traffic light” for clinical-scenarios with high risk of disease progression. Similarly, 86.4% of participants correctly identified the “yellow traffic light” for cases that would require a reassessment within 6–12 months. For the secondary fatigue outcome, within-group analysis showed a significant increased prevalence of TI in the second block of case-scenarios (decision fatigue) among participants randomized to the control group (TI pre-intervention 57.1% vs. TI post-intervention 71.4%; p = 0.015), but not in the active group (TI pre-intervention 54.6% vs. TI post-intervention 63.6%; p = 0.14). For the efficacy outcome, we found a non-significant reduction in TI for the targeted intervention compared to controls (22.6 vs. 33.9% post-intervention; OR 0.57; 95% CI 0.26–1.22).ConclusionAn educational intervention applying the TLS is feasible and shows some promising results in the identification of high-risk scenarios to reduce decision fatigue and TI. Larger studies are needed to determine the efficacy of the proposed educational intervention.Clinical Trial Registrationwww.ClinicalTrials.gov, identifier NCT03134794.

Surgeon-dependent factors such as experience and volume are associated with patient outcomes. However, it is unknown whether a surgeon's research productivity could be related to outcomes. The main aim of this study is to investigate the association between the surgeon's academic productivity and clinical outcomes following neurosurgical clipping of ruptured aneurysms.We performed a post-hoc analysis of 3567 patients who underwent clipping of ruptured intracranial aneurysms in the randomized trials of tirilazad mesylate from 1990 to 1997. These trials included 162 centers and 156 surgeons from 21 countries. Primary and secondary outcomes were: Glasgow outcome scale score and mortality, respectively. Total publications, H-index, and graduate degrees were used as academic indicators for each surgeon. The association between outcomes and academic factors were assessed using a hierarchical logistic regression analysis, adjusting for patient covariates.Academic profiles were available for 147 surgeons, treating a total of 3307 patients. Most surgeons were from the USA (62, 42%), Canada (18, 12%), and Germany (15, 10%). On univariate analysis, the H-index correlated with better functional outcomes and lower mortality rates. In the multivariate model, patients under the care of surgeons with higher H-indices demonstrated improved neurological outcomes (p = 0.01) compared to surgeons with lower H-indices, without any significant difference in mortality. None of the other academic indicators were significantly associated with outcomes.Although prognostication following surgery for ruptured intracranial aneurysms primarily depends on clinical and radiological factors, the academic impact of the operating neurosurgeon may explain some heterogeneity in surgical outcomes.

BACKGROUND/OBJECTIVE: Nunavut has the highest hospitalization rates for respiratory syncytial virus (RSV) worldwide, with rates of 166 per 1000 live births per year

Background . Inuit infants who reside in the Nunavut (NU) regions of Arctic Canada have extremely high rates of lower respiratory tract infections (LRTIs) associated with significant health expenditures, but the costs in other regions of Arctic Canada have not been documented. Objective . This prospective surveillance compares, across most of Arctic Canada, the rates and costs associated with LRTI admissions in infants less than 1 year of age, and the days of hospitalization and costs adjusted per live birth. Design . This was a hospital-based surveillance of LRTI admissions of infants less than 1 year of age, residing in Northwest Territories (NT), the 3 regions of Nunavut (NU); [Kitikmeot (KT), Kivalliq (KQ) and Qikiqtani (QI)] and Nunavik (NK) from 1 January 2009 to 30 June 2010. Costs were obtained from the territorial or regional governments and hospitals, and included transportation, hospital stay, physician fees and accommodation costs. The rates of LRTI hospitalizations, days of hospitalization and associated costs were calculated per live birth in each of the 5 regions. Results . There were 513 LRTI admissions during the study period. For NT, KT, KQ, QI and NK, the rates of LRTI hospitalization per 100 live births were 38, 389, 230, 202 and 445, respectively. The total days of LRTI admission per live birth were 0.25, 3.3, 2.6, 1.7 and 3 for the above regions. The average cost per live birth for LRTI admission for these regions was $1,412, $22,375, $14,608, $8,254 and $10,333. The total cost for LRTI was $1,498,232 in NT, $15,662,968 in NU and $3,874,881 in NK. Medical transportation contributed to a significant proportion of the costs. Conclusion . LRTI admission rates in NU and Nunavik are much higher than that in NT and remain among the highest rates globally. The costs of these admissions are exceptionally high due to the combination of very high rates of admission, very expensive medical evacuations and prolonged hospitalizations. Decreasing the rates of LRTI in this population could result in substantial health savings.

To examine the impact of national clinical practice guidelines and provincial drug policy interventions on prevalence of high-dose opioid prescribing and rates of hospitalization for opioid toxicity.Interventional time-series analysis.Ontario, Canada, from 2003 to 2014.Ontario Drug Benefit (ODB) beneficiaries aged 15 to 64 years from 2003 to 2014.Publication of Canadian clinical practice guidelines for use of opioids in chronic non-cancer pain (May 2010) and implementation of Ontario's Narcotics Safety and Awareness Act (NSAA; November 2011).Three outcomes were explored: the rate of opioid use among ODB beneficiaries, the prevalence of opioid prescriptions exceeding 200 mg and 400 mg morphine equivalents per day, and rates of opioid-related emergency department visits and hospital admissions.Over the 12 year study period, the rate of opioid use declined 15.2%, from 2764 to 2342 users per 10,000 ODB eligible persons. The rate of opioid use was significantly impacted by the Canadian clinical practice guidelines (p-value = .03) which led to a decline in use, but no impact was observed by the enactment of the NSAA (p-value = .43). Among opioid users, the prevalence of high-dose prescribing doubled (from 4.2% to 8.7%) over the study period. By 2014, 40.9% of recipients of long-acting opioids exceeded daily doses of 200 mg morphine or equivalent, including 55.8% of long-acting oxycodone users and 76.3% of transdermal fentanyl users. Moreover, in the last period, 18.7% of long-acting opioid users exceeded daily doses of 400 mg morphine or equivalent. Rates of opioid-related emergency department visits and hospital admissions increased 55.0% over the study period from 9.0 to 14.0 per 10,000 ODB beneficiaries from 2003 to 2013. This rate was not significantly impacted by the Canadian clinical practice guidelines (p-value = .68) or enactment of the NSAA (p-value = .59).Although the Canadian clinical practice guidelines for use of opioids in chronic non-cancer pain led to a decline in opioid prescribing rates among ODB beneficiaries these guidelines and subsequent Ontario legislation did not result in a significant change in rates of opioid-related hospitalizations. Given the prevalence of high dose opioid prescribing in this population, this suggests that improved strategies and programs for the safe prescribing of long-acting opioids are needed.

PURPOSE: To examine the association between the provision of professional pharmacy services (PPS) and patient complexity as determined by the number of distinct medications dispensed in Ontario. METHODS: We conducted a cross-sectional study among all individuals dispensed one or more medications under the Ontario Public Drug Program (OPDP) between April 1st, 2012 and March 31st, 2013. We compared characteristics of patients receiving ­1 or more PPS to those receiving no PPS. To assess the relationship between patient complexity (as measured by the number of chronic medications dispensed) and receipt of PPS, we reported the number and proportion of patients eligible for Ontario Drug Benefits (ODB) who received a PPS within each patient complexity group, and compared these proportions using the Cochran-Armitage test. RESULTS: Over the 1-year study period, 27.1% (N = 799,674 of 2,946,183) of ODB beneficiaries received at least one professional pharmacy service. Among these services, more than two-thirds of the patients received a MedsCheck service (N=511,490; 64.0%). Overall, individuals who received a PPS tended to be older, more likely to reside in a long-term care (LTC) facility, have multiple comorbidities, and were more likely to have been prescribed 9 or more medications in the past year. As patient complexity increased, the proportion of ODB beneficiaries who received PPS also increased; 3.0% of individuals prescribed between 1 and 2 medications in the past year received PPS, while 53.6% of those treated with 13 or more medications received PPS (p

Vitamin D is associated with established cardiovascular risk factors such as low density lipoprotein (LDL) in adults. It is unknown whether these associations are present in early childhood. To determine whether serum 25-hydroxyvitamin D (25(OH)D) is associated with serum non-high density lipoprotein (non-HDL) cholesterol during early childhood we conducted a cross-sectional study of children aged 1 to 5 years. Healthy children were recruited through the TARGet Kids! practice based research network from 2008-2011 (n=1,961). The associations between 25(OH)D and non-fasting non-HDL cholesterol (the primary endpoint), total cholesterol, triglycerides, HDL, and low density lipoprotein (LDL) cholesterol, were evaluated using multiple linear regression adjusted for age, sex, skin pigmentation, milk intake, vitamin D supplementation, season, body mass index, outdoor play, and screen time. Each 10 nmol/L increase in 25(OH)D was associated with a decrease in non-HDL cholesterol concentration of -0.89 mg/dl (95% CI: -1.16,-0.50), total cholesterol of -1.08 mg/dl (95%CI: -1.49,-0.70), and triglycerides of -2.34 mg/dl (95%CI: -3.23,-1.45). The associations between 25(OH)D and LDL and HDL were not statistically significant. 25(OH)D concentrations were inversely associated with circulating lipids in early childhood, suggesting that vitamin D exposure in early life may be an early modifiable risk factor for cardiovascular disease.

BackgroundDespite a lack of data describing the long-term efficacy and safety of testosterone replacement therapy (TRT), prescribing of testosterone to older men has increased with the availability of topical formulations. The magnitude of this increase and the impact of formulary restrictions on testosterone prescribing are poorly characterized.MethodsWe conducted a time series analysis using the linked health administrative records of men aged 66 years or older in Ontario, Canada between January 1, 1997 and March 31, 2012. We used interventional autoregressive integrated moving average models to examine the impact of a restrictive drug reimbursement policy on testosterone prescribing and examined the demographic profile of men initiating testosterone in the final 2 years of the study period.ResultsA total of 28,477 men were dispensed testosterone over the study period. Overall testosterone prescribing declined 27.9% in the 6 months following the implementation of the restriction policy (9.5 to 6.9 men per 1000 eligible; pConclusionGovernment-imposed restrictions did not influence long-term prescribing of testosterone to older men. By 2012, approximately 1 in every 90 men aged 66 or older was being treated with TRT, most with topical formulations.

BACKGROUND: Carpal tunnel release (CTR) is among the most common hand surgeries, although little is known about its pattern. In this study, we aimed to investigate temporal trends, age and gender variation and current practice patterns in CTR surgeries. METHODS: We conducted a population-based time series analysis among over 13 million residents of Ontario, who underwent operative management for carpal tunnel syndrome (CTS) from April 1, 1992 to March 31, 2010 using administrative claims data. RESULTS: The primary analysis revealed a fairly stable procedure rate of approximately 10 patients per 10,000 population per year receiving CTRs without any significant, consistent temporal trend (p = 0.94). Secondary analyses revealed different trends in procedure rates according to age. The annual procedure rate among those age >75 years increased from 22 per 10,000 population at the beginning of the study period to over 26 patients per 10,000 population (p

BACKGROUND: Diverticular disease is one of the most common gastrointestinal conditions affecting the Canadian population, yet very little is known about its epidemiology.