Mitigating Machine Learning Bias Between High Income and Low-Middle Income Countries for Enhanced Model Fairness and Generalizability (preprint)

Collaborative efforts in artificial intelligence (AI) are increasingly common between high-income countries (HICs) and low- to middle-income countries (LMICs). Given the resource limitations often encountered by LMICs, collaboration becomes crucial for pooling resources, expertise, and knowledge. Despite the apparent advantages, ensuring the fairness and equity of these collaborative models is essential, especially considering the distinct differences between LMIC and HIC hospitals. In this study, we show that collaborative AI approaches can lead to divergent performance outcomes across HIC and LMIC settings, particularly in the presence of data imbalances. Through a real-world COVID-19 screening case study, we demonstrate that implementing algorithmic-level bias mitigation methods significantly improves outcome fairness between HIC and LMIC sites while maintaining high diagnostic sensitivity. We compare our results against previous benchmarks, utilizing datasets from four independent United Kingdom Hospitals and one Vietnamese hospital, representing HIC and LMIC settings, respectively.

Scalable federated learning for emergency care using low cost microcomputing: Real-world, privacy preserving development and evaluation of a COVID-19 screening test in UK hospitals (preprint)

Background Tackling biases in medical artificial intelligence requires multi-centre collaboration, however, ethical, legal and entrustment considerations may restrict providers' ability to participate. Federated learning (FL) may eliminate the need for data sharing by allowing algorithm development across multiple hospitals without data transfer. Previously, we have shown an AI-driven screening solution for COVID-19 in emergency departments using clinical data routinely available within 1h of arrival to hospital (vital signs & blood tests; CURIAL-Lab). Here, we aimed to extend and federate our COVID-19 screening test, demonstrating development and evaluation of a rapidly scalable and user-friendly FL solution across 4 UK hospital groups. Methods We supplied a Raspberry Pi 4 Model B device, preloaded with our end-to-end FL pipeline, to 4 NHS hospital groups or their locally-linked research university (Oxford University Hospitals/University of Oxford (OUH), University Hospitals Birmingham/University of Birmingham (UHB), Bedfordshire Hospitals (BH) and Portsmouth Hospitals University (PUH) NHS trusts). OUH, PUH and UHB participated in federated training and calibration, training a deep neural network (DNN) and logistic regressor to predict COVID-19 status using clinical data for pre- pandemic (COVID-19-negative) admissions and COVID-19-positive cases from the first wave. We performed federated prospective evaluation at PUH & OUH, and external evaluation at BH, evaluating the resultant global and site-tuned models for admissions to the respective sites during the second pandemic wave. Removable microSD storage was destroyed on study completion. Findings Routinely collected clinical data from a total 130,941 patients (1,772 COVID-19 positive) across three hospital groups were included in federated training. OUH, PUH and BH participated in prospective federated evaluation, with sets comprising 32,986 patient admissions (3,549 positive) during the second pandemic wave. Federated training improved DNN performance by a mean of 27.6% in terms of AUROC when compared to models trained locally, from AUROC of 0.574 & 0.622 at OUH & PUH to 0.872 & 0.876 for the federated global model. Performance improvement was more modest for a logistic regressor with a mean AUROC increase of 13.9%. During federated external evaluation at BH, the global DNN model achieved an AUROC of 0.917 (0.893-0.942), with 89.7% sensitivity (83.6-93.6) and 76.7% specificity (73.9- 79.1). Site-personalisation of the global model did not give a significant improvement in overall performance (AUROC improvement <0.01), suggesting high generalisability. Interpretations We present a rapidly scalable hardware and software FL solution, developing a COVID-19 screening test across four UK hospital groups using inexpensive micro- computing hardware. Federation improved model performance and generalisability, and shows promise as an enabling technology for deep learning in healthcare. Funding University of Oxford Medical & Life Sciences Translational Fund/Wellcome

Estimating the effectiveness of shielding during pregnancy against SARS-CoV-2 in New York City during the first year of the COVID-19 pandemic (preprint)

Pregnant patients have increased morbidity and mortality in the setting of SARS-CoV-2 infection. The exposure of pregnant patients in New York City to SARS-CoV-2 is not well understood due to early lack of access to testing and the presence of asymptomatic COVID-19 infections. Before the availability of vaccinations, preventative (shielding) measures, including but not limited to wearing a mask and quarantining at home to limit contact, were recommended for pregnant patients. Using universal testing data from 2196 patients who gave birth from April through December 2020 from one institution in New York City, and in comparison, with infection data of the general population in New York City, we estimated the exposure and real-world effectiveness of shielding in pregnant patients. Our Bayesian model shows that patients already pregnant at the onset of the pandemic had a 50% decrease in exposure compared to those who became pregnant after the onset of the pandemic and to the general population.

Algorithmic Fairness and Bias Mitigation for Clinical Machine Learning: A New Utility for Deep Reinforcement Learning (preprint)

As machine learning-based models continue to be developed for healthcare applications, greater effort is needed in ensuring that these technologies do not reflect or exacerbate any unwanted or discriminatory biases that may be present in the data. In this study, we introduce a reinforcement learning framework capable of mitigating biases that may have been acquired during data collection. In particular, we evaluated our model for the task of rapidly predicting COVID-19 for patients presenting to hospital emergency departments, and aimed to mitigate any site-specific (hospital) and ethnicity-based biases present in the data. Using a specialized reward function and training procedure, we show that our method achieves clinically-effective screening performances, while significantly improving outcome fairness compared to current benchmarks and state-of-the-art machine learning methods. We performed external validation across three independent hospitals, and additionally tested our method on a patient ICU discharge status task, demonstrating model generalizability.

Machine Learning Generalizability Across Healthcare Settings: Insights from multi-site COVID-19 screening (preprint)

As patient health information is highly regulated due to privacy concerns, the majority of machine learning (ML)-based healthcare studies are unable to test on external patient cohorts, resulting in a gap between locally reported model performance and cross-site generalizability. Different approaches have been introduced for developing models across multiple clinical sites, however no studies have compared methods for translating ready-made models for adoption in new settings. We introduce three methods to do this - (1) applying a ready-made model as-is; (2) readjusting the decision threshold on the output of a ready-made model using site-specific data; and (3) finetuning a ready-made model using site-specific data via transfer learning. Using a case study of COVID-19 diagnosis across four NHS Hospital Trusts, we show that all methods achieve clinically-effective performances (NPV >0.959), with transfer learning achieving the best results (mean AUROCs between 0.870-0.925). Our models demonstrate that site-specific customization improves predictive performance when compared to other ready-made approaches.

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

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.

Privacy-aware Early Detection of COVID-19 through Adversarial Training (preprint)

Early detection of COVID-19 is an ongoing area of research that can help with triage, monitoring and general health assessment of potential patients and may reduce operational strain on hospitals that cope with the coronavirus pandemic. Different machine learning techniques have been used in the literature to detect coronavirus using routine clinical data (blood tests, and vital signs). Data breaches and information leakage when using these models can bring reputational damage and cause legal issues for hospitals. In spite of this, protecting healthcare models against leakage of potentially sensitive information is an understudied research area. In this work, we examine two machine learning approaches, intended to predict a patient's COVID-19 status using routinely collected and readily available clinical data. We employ adversarial training to explore robust deep learning architectures that protect attributes related to demographic information about the patients. The two models we examine in this work are intended to preserve sensitive information against adversarial attacks and information leakage. In a series of experiments using datasets from the Oxford University Hospitals, Bedfordshire Hospitals NHS Foundation Trust, University Hospitals Birmingham NHS Foundation Trust, and Portsmouth Hospitals University NHS Trust we train and test two neural networks that predict PCR test results using information from basic laboratory blood tests, and vital signs performed on a patients' arrival to hospital. We assess the level of privacy each one of the models can provide and show the efficacy and robustness of our proposed architectures against a comparable baseline. One of our main contributions is that we specifically target the development of effective COVID-19 detection models with built-in mechanisms in order to selectively protect sensitive attributes against adversarial attacks.

Real-world evaluation of AI driven COVID-19 triage for emergency admissions: External validation & operational assessment of lab-free and high-throughput screening solutions (preprint)

BackgroundUncertainty in patients COVID-19 status contributes to treatment delays, nosocomial transmission, and operational pressures in hospitals. However, typical turnaround times for batch-processed laboratory PCR tests remain 12-24h. Although rapid antigen lateral flow testing (LFD) has been widely adopted in UK emergency care settings, sensitivity is limited. We recently demonstrated that AI-driven triage (CURIAL-1.0) allows high-throughput COVID-19 screening using clinical data routinely available within 1h of arrival to hospital. Here we aimed to determine operational and safety improvements over standard-care, performing external/prospective evaluation across four NHS trusts with updated algorithms optimised for generalisability and speed, and deploying a novel lab-free screening pathway in a UK emergency department. MethodsWe rationalised predictors in CURIAL-1.0 to optimise separately for generalisability and speed, developing CURIAL-Lab with vital signs and routine laboratory blood predictors (FBC, U&E, LFT, CRP) and CURIAL-Rapide with vital signs and FBC alone. Models were calibrated during training to 90% sensitivity and validated externally for unscheduled admissions to Portsmouth University Hospitals, University Hospitals Birmingham and Bedfordshire Hospitals NHS trusts, and prospectively during the second-wave of the UK COVID-19 epidemic at Oxford University Hospitals (OUH). Predictions were generated using first-performed blood tests and vital signs and compared against confirmatory viral nucleic acid testing. Next, we retrospectively evaluated a novel clinical pathway triaging patients to COVID-19-suspected clinical areas where either model prediction or LFD results were positive, comparing sensitivity and NPV with LFD results alone. Lastly, we deployed CURIAL-Rapide alongside an approved point-of-care FBC analyser (OLO; SightDiagnostics, Israel) to provide lab-free COVID-19 screening in the John Radcliffe Hospitals Emergency Department (Oxford, UK), as trust-approved service improvement. Our primary improvement outcome was time-to-result availability; secondary outcomes were sensitivity, specificity, PPV, and NPV assessed against a PCR reference standard. We compared CURIAL-Rapides performance with clinician triage and LFD results within standard-care. Results72,223 patients met eligibility criteria across external and prospective validation sites. Model performance was consistent across trusts (CURIAL-Lab: AUROCs range 0.858-0.881; CURIAL-Rapide 0.836-0.854), with highest sensitivity achieved at Portsmouth University Hospitals (CURIAL-Lab:84.1% [95% Wilsons score CIs 82.5-85.7]; CURIAL-Rapide:83.5% [81.8 - 85.1]) at specificities of 71.3% (95% Wilsons score CIs: 70.9 - 71.8) and 63.6% (63.1 - 64.1). For 3,207 patients receiving LFD-triage within routine care for OUH admissions between December 23, 2021 and March 6, 2021, a combined clinical pathway increased sensitivity from 56.9% for LFDs alone (95% CI 51.7-62.0) to 88.2% with CURIAL-Rapide (84.4-91.1; AUROC 0.919) and 85.6% with CURIAL-Lab (81.6-88.9; AUROC 0.925). 520 patients were prospectively enrolled for point-of-care FBC analysis between February 18, 2021 and May 10, 2021, of whom 436 received confirmatory PCR testing within routine care and 10 (2.3%) tested positive. Median time from patient arrival to availability of CURIAL-Rapide result was 45:00 min (32-64), 16 minutes (26.3%) sooner than LFD results (61:00 min, 37-99; log-rank p<0.0001), and 6:52 h (90.2%) sooner than PCR results (7:37 h, 6:05-15:39; p<0.0001). Sensitivity and specificity of CURIAL-Rapide were 87.5% (52.9-97.8) and 85.4% (81.3-88.7), therefore achieving high NPV (99.7%, 98.2-99.9). CURIAL-Rapide correctly excluded COVID-19 for 58.5% of negative patients who were triaged by a clinician to COVID-19-suspected (amber) areas. ImpactCURIAL-Lab & CURIAL-Rapide are generalisable, high-throughput screening tests for COVID-19, rapidly excluding the illness with higher NPV than LFDs. CURIAL-Rapide can be used in combination with near-patient FBC analysis for rapid, lab-free screening, and may reduce the number of COVID-19-negative patients triaged to enhanced precautions ( amber) clinical areas.

Antibody response to SARS-CoV-2 mRNA vaccines in pregnant women and their neonates (preprint)

Pregnant women were excluded from initial clinical trials for COVID-19 vaccines1-2, thus the immunologic response to vaccination in pregnancy and the transplacental transfer of maternal antibodies are just beginning to be studied4-5. Methods: Between January 28 and March 31, 2021, we studied 122 pregnant women and their neonates at time of birth. All women had received one or both doses of a messenger RNA (mRNA)-based COVID-19 vaccine. Fifty-five women received only one dose of the vaccine and 67 women received both doses of the vaccine by time of giving birth. Eighty-five women received the Pfizer-BioNTech vaccine, while 37 women received the Moderna vaccine. All women tested negative for SARS-CoV-2 infection using reverse-transcriptase PCR on nasopharyngeal swabs, and none reported any COVID-19 symptoms at the time of admission for birth. Semi-quantitative testing for antibodies against S-Receptor Binding Domain (RBD) (ET HealthCare)3 was performed on sera of maternal peripheral blood and neonatal cord blood at the time of delivery to identify antibodies mounted against the vaccine. All women tested negative for antibodies against the Nucleocapsid Protein (NP) antigen (Roche Diagnostics EUA) to ensure that the antibodies detected were not produced in response to past SARS-CoV-2 infection. Relationship between IgG antibody levels over time was studied using ANOVA with Tukey posthoc. Relationship between maternal and neonatal IgG levels was studied using Pearson correlation analysis and linear regression on log2-scaled serological values. Relationship between IgG placental transfer ratio (neonatal/maternal) vs. time was studied using Pearson correlation analysis and linear regression on log2-scaled serological values and days. Serology levels represented as log2+1. Statistical analysis was performed using R 3.6.3, RStudio 1.1.463. The study was approved by the Weill Cornell Medicine institutional review board. Results: Pregnant women vaccinated with mRNA-based COVID-19 vaccines during pregnancy and tested at time of birth had detectable immunoglobulin (Ig)G and IgM response. Eighty-seven women tested at birth produced only an IgG response, and 19 women produced both an IgM and IgG response. Sixteen women tested at birth had no detectable antibody response, and they were all within four weeks after vaccination dose 1 (Figure 1A). There was an increase over time in the number of women that mounted an antibody response, as well as the number of women that demonstrated passive immunity to their neonates (Figure 1A). All women and their neonates, except for one neonate, had detectable IgG antibodies by 4 weeks after maternal first dose of vaccination (Figure 1A). 43.6% (24/55) of neonates born to women that received only one vaccine dose had detectable IgG, while 98.5% (65/67) of neonates born to women that received both vaccine doses had detectable IgG. The IgG levels in pregnant women increased weekly from two weeks after first vaccine dose (p=0.0047;0.019), as well as between the first and second weeks after the second vaccine dose (p=2e-07) (Figure 1B). Maternal IgG levels were linearly associated with neonatal IgG levels (R=0.89, p<2.2e-16) (Figure 2A). Placental transfer ratio correlated with the weeks that elapsed since maternal second dose of vaccine (R=0.8, p=2.6e-15) (Figure 2B). Discussion: mRNA-based COVID-19 vaccines in pregnant women lead to maternal antibody production as early as 5 days after the first vaccination dose, and passive immunity to the neonate as early as 16 days after the first vaccination dose. The increasing levels of maternal IgG over time, and the increasing placental IgG transfer ratio over time suggest that timing between vaccination and birth may be an important factor to consider in the vaccination strategies of pregnant women. Further studies are needed to understand the factors that influence transplacental transfer of IgG antibody, as well as the protective nature of these antibodies.

Seroprevalence of Anti-SARS-CoV-2 Antibodies in a Cohort of New York City Metro Blood Donors using Multiple SARS-CoV-2 Serological Assays: Implications for Controlling the Epidemic and Reopening. (preprint)

Projections of the stage of the Severe Acute Respiratory Syndrome-Coronavirus-2 (SARS-CoV-2) pandemic and local, regional and national public health policies designed to limit the spread of the epidemic as well as reopen cities and states, are best informed by reproducible, high throughput, and statically credible antibody (Ab) assays. To date, a myriad of Ab tests, both available and authorized for emergency use by the FDA, has led to confusion rather than insight per se. The present study reports the results of a rapid, point-in-time 1,000-person cohort study using serial blood donors in the New York City metropolitan area (NYC) using multiple serological tests, including enzyme-linked immunosorbent assays (ELISAs) and high throughput serological assays (HTSAs). These were then tested and associated with assays for neutralizing Ab (NAb). Of the 1,000 NYC blood donor samples in late June and early July 2020, 12.1% and 10.9% were seropositive using the Ortho Total Ig and the Abbott IgG HTSA assays, respectively. These serological assays correlated with neutralization activity specific to SARS-CoV-2. The data reported herein suggest that seroconversion in this population occurred in approximately 1 in 8 blood donors from the beginning of the pandemic in NYC (considered March 1, 2020). These findings deviate with an earlier seroprevalence study in NYC showing 13.7% positivity. Collectively however, these data demonstrate that a low number of individuals have serologic evidence of infection during this first wave and suggest that the notion of herd immunity at rates of ~60% or higher are not near. Furthermore, the data presented herein show that the nature of the Ab-based immunity is not invariably associated with the development of NAb. While the blood donor population may not mimic precisely the NYC population as a whole, rapid assessment of seroprevalence in this cohort and serial reassessment could aid public health decision making.