Algorithmic Fairness and Bias Mitigation for Clinical Machine Learning: Insights from Rapid COVID-19 Diagnosis by Adversarial Learning

Machine learning is becoming increasingly promi- nent in healthcare. Although its benefits are clear, growing attention is being given to how machine learning may exacerbate existing biases and disparities. In this study, we introduce an adversarial training framework that is capable of mitigating biases that may have been acquired through data collection or magnified during model development. For example, if one class is over-presented or errors/inconsistencies in practice are reflected in the training data, then a model can be biased by these. To evaluate our adversarial training framework, we used the statistical definition of equalized odds. We evaluated our model for the task of rapidly predicting COVID-19 for patients presenting to hospital emergency departments, and aimed to mitigate regional (hospital) and ethnic biases present. We trained our framework on a large, real-world COVID-19 dataset and demonstrated that adversarial training demonstrably improves outcome fairness (with respect to equalized odds), while still achieving clinically-effective screening performances (NPV>0.98). We compared our method to the benchmark set by related previous work, and performed prospective and external validation on four independent hospital cohorts. Our method can be generalized to any outcomes, models, and definitions of fairness.

Analysis of Medical Management in Geriatric Patients in the Hospital Emergency Department by Example of Selected Cities with County Status in Poland: A Retrospective Cohort Study

The aim of this study was to analyse medical management in geriatric patients in the Hospital Emergency Departments in the Biała Podlaska County and Chełm County (Poland) between 2016 and 2018 in a group of patients ≥65 years of age. We analysed medical records of 829 patients transported to Hospital Emergency Departments by Medical Emergency Teams. The research was conducted in the period from June 2019 to March 2020. We analysed emergency medical procedure forms and medical records of patients transported to the hospitals. Cardiovascular diseases were diagnosed in 40% of patients. Mortality cases accounted for 3.1% of the 1200 interventions analysed. Ambulance dispatch resulted in the patient being transported to the Hospital Emergency Departments in more than 2/3 of cases. The concordance between the diagnoses made by the Medical Emergency Teams and those made at the Hospital Emergency Departments was confirmed for 78% patients admitted to the department (n = 647), whereas the concordance of classification at the group level was estimated at 71.7% (n = 594). Further in-patient treatment was initiated in some of the patients admitted to the department (n = 385). The mean time of hospital stay was 10.1 days. In conclusion, differences between the initial diagnosis made by the heads of the Medical Emergency Teams and the diagnosis made by the doctor on duty in the Hospital Emergency Departments depended on the chapter of diseases in the ICD-10 classification, but they were acceptable. The majority of the patients were transported to Hospital Emergency Departments. The most common groups of diseases that require Hospital Emergency Departments admission include cardiovascular diseases, injuries due to external causes, and respiratory diseases. A moderate percentage of patients were qualified for further specialist treatment in hospital departments.

Who Is Next: Patient Prioritization Under Emergency Department Blocking

In “Who Is Next: Patient Prioritization Under Emergency Department Blocking,” Li, Sun, and Hong study how physicians and nurses choose the next patient for treatment in hospital emergency departments (EDs). Using data from a tertiary hospital in Alberta, Canada, they conduct an empirical investigation and find that both clinical factors and resource constraints are considered in patient-prioritization decisions. In particular, discharged patients are prioritized when ED beds are increasingly occupied by boarding patients so as to avoid further blocking the ED. A stylized model is developed to explain the rationale behind the prioritization behavior. Using a simulation study, they show such behavior can improve ED operations by reducing the average patient waiting time and length of stay without adding extra capacity, which results in significant cost savings for hospitals.

Fast, Reliable, and Simple Point-of-Care-like Adaptation of RT-qPCR for the Detection of SARS-CoV-2 for Use in Hospital Emergency Departments

During COVID-19 pandemics, the availability of testing has often been a limiting factor during patient admissions into the hospital. To circumvent this problem, we adapted an existing diagnostic assay, Seegene Allplex SARS-CoV-2, into a point-of-care-style direct qPCR (POC dqPCR) assay and implemented it in the Emergency Department of Clinical Hospital Center Rijeka, Croatia. In a 4-month analysis, we tested over 10,000 patients and demonstrated that POC-dqPCR is robust and reliable and can be successfully implemented in emergency departments and similar near-patient settings and can be performed by medical personnel with little prior experience in qPCR.

Performance of a new symptom checker in patient triage: Canadian cohort study

Background Computerized algorithms known as symptom checkers aim to help patients decide what to do should they have a new medical concern. However, despite widespread implementation, most studies on symptom checkers have involved simulated patients. Only limited evidence currently exists about symptom checker safety or accuracy when used by real patients. We developed a new prototype symptom checker and assessed its safety and accuracy in a prospective cohort of patients presenting to primary care and emergency departments with new medical concerns. Method A prospective cohort study was done to assess the prototype’s performance. The cohort consisted of adult patients (≥16 years old) who presented to hospital emergency departments and family physician clinics. Primary outcomes were safety and accuracy of triage recommendations to seek hospital care, seek primary care, or manage symptoms at home. Results Data from 281 hospital patients and 300 clinic patients were collected and analyzed. Sensitivity to emergencies was 100% (10/10 encounters). Sensitivity to urgencies was 90% (73/81) and 97% (34/35) for hospital and primary care patients, respectively. The prototype was significantly more accurate than patients at triage (73% versus 58%, p<0.01). Compliance with triage recommendations in this cohort using this iteration of the symptom checker would have reduced hospital visits by 55% but cause potential harm in 2–3% from delay in care. Interpretation The prototype symptom checker was superior to patients in deciding the most appropriate treatment setting for medical issues. This symptom checker could reduce a significant number of unnecessary hospital visits, with accuracy and safety outcomes comparable to existing data on telephone triage.

Covid-19 and dengue fever combine infection, a dangerous combination for the health system of Pakistan

As the world struggles to deal with the consequences of the COVID-19 pandemic, dengue fever endemic regions are facing the threat of a double pandemic that might totally overwhelm administrations of health care service. In Latin America and Asia, simultaneous epidemics of dengue and COVID-19, as well as possible instances of co-infections, have already begun. Because the healthcare system is already stretched enough, a fatal combination like this may devastate hospital emergency departments and the economy as a whole. Frontline clinicians must use precise epidemiological and contact history-taking techniques, as well as pay close attention to false-positive dengue serology and the possibility of co-infections, to tackle this apparently overwhelming obstacle. Key words: Covid-19; Dengue fever; Co-infection; health system

Emergency Telepsychiatry

Emergency telepsychiatry is the evaluation and treatment of patients experiencing mental health crises using audio/video communications. Patients in hospital emergency departments (EDs) often experience long waits, delayed treatment, and uncomfortable conditions when in-person psychiatric care is not readily available; prolonged lengths of stay in the ED also increase hospital costs. Telepsychiatry can make psychiatric care more accessible and timely by linking mental health specialists in various locations to patients in EDs. Initial data indicate that the financial investment in emergency telepsychiatry is more than offset by savings in ED and hospital costs. Rapid availability of telepsychiatric consultation also improves the quality of care in EDs and leads to more appropriate and timely dispositions of patients.

Characteristics of frequently attending children in hospital emergency departments: a systematic review

ObjectiveTo summarise the literature on frequent attendances to hospital emergency departments (EDs) and describe sociodemographic and clinical characteristics of children who attend EDs frequently.SettingHospital EDs.ParticipantsChildren <21 years, attending hospital EDs frequently.Primary outcome measuresOutcomes measures were defined separately in each study, and were predominantly the number of ED attendances per year.ResultsWe included 21 studies representing 6 513 627 children. Between 0.3% and 75% of all paediatric ED users were frequent users. Most studies defined four or more visits per year as a ‘frequent ED’ usage. Children who were frequent ED users were more likely to be less than 5 years old. In the USA, patients with public insurance were more likely to be frequent attenders. Frequent ED users more likely to be frequent users of primary care and have long-term conditions; the most common diagnoses were infections and gastroenteritis.ConclusionsThe review included a wide range of information across various health systems, however, children who were frequent ED users have some universal characteristics in common. Policies to reduce frequent attendance might usefully focus on preschool children and supporting primary care in responding to primary care oriented conditions.