Motivated by the fact that there may be inaccuracies in features and labels of training data, we apply robust optimization techniques to study in a principled way the uncertainty in data features and labels in classification problems and obtain robust formulations for the three most widely used classification methods: support vector machines, logistic regression, and decision trees. We show that adding robustness does not materially change the complexity of the problem and that all robust counterparts can be solved in practical computational times. We demonstrate the advantage of these robust formulations over regularized and nominal methods in synthetic data experiments, and we show that our robust classification methods offer improved out-of-sample accuracy. Furthermore, we run large-scale computational experiments across a sample of 75 data sets from the University of California Irvine Machine Learning Repository and show that adding robustness to any of the three nonregularized classification methods improves the accuracy in the majority of the data sets. We observe the most significant gains for robust classification methods on high-dimensional and difficult classification problems, with an average improvement in out-of-sample accuracy of robust versus nominal problems of 5.3% for support vector machines, 4.0% for logistic regression, and 1.3% for decision trees.

BACKGROUND. Clinical data to support the use of bamlanivimab for the treatment of outpatients with mild to moderate coronavirus disease-19 (COVID-19) are needed. METHODS. 2335 Patients who received single-dose bamlanivimab infusion between November 12, 2020, and February 17, 2021, were compared with a propensity-matched control of 2335 untreated patients with mild to moderate COVID-19 at Mayo Clinic facilities across 4 states. The primary outcome was the rate of hospitalization at days 14, 21, and 28. RESULTS. The median age of the population was 63 years; 47.3% of the bamlanivimab-treated cohort were 65 years or more; 49.3% were female and 50.7% were male. High-risk characteristics included hypertension (54.2%), BMI greater than or equal to 35 (32.4%), diabetes mellitus (26.5%), chronic lung disease (25.1%), malignancy (16.6%), and renal disease (14.5%). Patients who received bamlanivimab had lower all-cause hospitalization rates at days 14 (1.5% vs. 3.5%; risk ratio [RR], 0.41), 21 (1.9% vs. 3.9%; RR, 0.49), and 28 (2.5% vs. 3.9%; RR, 0.63). Secondary exploratory outcomes included lower intensive care unit (ICU) admission rates at days 14 (0.14% vs. 1%; RR, 0.14), 21 (0.25% vs.1%; RR, 0.25), and 28 (0.56% vs.1.1%; RR. 0.51) and lower all-cause mortality at days 14 (0% vs. 0.33%), 21 (0.05% vs. 0.4%; RR,0.13), and 28 (0.11% vs. 0.44%; RR, 0.26). Adverse events were uncommon with bamlanivimab, occurring in 19 of 2355 patients, and were most commonly fever (n = 6), nausea (n = 5), and lightheadedness (n = 3). CONCLUSIONS. Among high-risk patients with mild to moderate COVID-19, treatment with bamlanivimab was associated with a statistically significant lower rate of hospitalization, ICU admission, and mortality compared with usual care. FUNDING. Mayo Clinic.

To assess the association of COVID-19 vaccines and non-COVID-19 vaccines with cerebral venous sinus thrombosis (CVST). We retrospectively analyzed a cohort of 771,805 vaccination events across 266,094 patients in the Mayo Clinic Health System between 01/01/2017 and 03/15/2021. The primary outcome was a positive diagnosis of CVST, identified either by the presence of a corresponding ICD code or by an NLP algorithm which detected positive diagnosis of CVST within free-text clinical notes. For each vaccine we calculated the relative risk by dividing the incidence of CVST in the 30 days following vaccination to that in the 30 days preceding vaccination. We identified vaccination events for all FDA-approved COVID-19 vaccines including Pfizer-BioNTech (n = 94,818 doses), Moderna (n = 36,350 doses) and Johnson & Johnson - J&J (n = 1,745 doses). We also identified vaccinations events for 10 common FDA-approved non-COVID-19 vaccines (n = 771,805 doses). There was no statistically significant difference in the incidence rate of CVST in 30-days before and after vaccination for any vaccine in this population. We further found the baseline CVST incidence in the study population between 2017 and 2021 to be 45 to 98 per million patient years. This real-world evidence-based study finds that CVST is rare and is not significantly associated with COVID-19 vaccination in our patient cohort. Limitations include the rarity of CVST in our dataset, a relatively small number of J&J COVID-19 vaccination events, and the use of a population drawn from recipients of a SARS-CoV-2 PCR test in a single health system.

Case reports of patients infected with COVID-19 and influenza virus (“flurona”) have raised questions around the prevalence and severity of coinfection. Using data from HHS Protect Public Data Hub, NCBI Virus, and CDC FluView, we analyzed trends in SARS-CoV-2 and influenza hospitalized coinfection cases and strain prevalences. We also characterized coinfection cases across the Mayo Clinic Enterprise from January 2020 to April 2022. We compared expected and observed coinfection case counts across different waves of the pandemic and assessed symptoms and outcomes of coinfection and COVID-19 monoinfection cases after propensity score matching on clinically relevant baseline characteristics. From both the Mayo Clinic and nationwide datasets, the observed coinfection rate for SARS-CoV-2 and influenza has been higher during the Omicron era (2021 December 14 to 2022 April 2) compared to previous waves, but no higher than expected assuming infection rates are independent. At the Mayo Clinic, only 120 coinfection cases were observed among 197,364 SARS-CoV-2 cases. Coinfected patients were relatively young (mean age: 26.7 years) and had fewer serious comorbidities compared to monoinfected patients. While there were no significant differences in 30-day hospitalization, ICU admission, or mortality rates between coinfected and matched COVID-19 monoinfection cases, coinfection cases reported higher rates of symptoms including congestion, cough, fever/chills, headache, myalgia/arthralgia, pharyngitis, and rhinitis. While most coinfection cases observed at the Mayo Clinic occurred among relatively healthy individuals, further observation is needed to assess outcomes among subpopulations with risk factors for severe COVID-19 such as older age, obesity, and immunocompromised status.

In light of the massive and rapid vaccination campaign against COVID-19, continuous real-world effectiveness and safety assessment of the FDA-authorized vaccines is critical to amplify transparency, build public trust, and ultimately improve overall health outcomes. In this study, we leveraged large-scale longitudinal curation of electronic health records (EHRs) from the multi-state Mayo Clinic health system (MN, AZ, FL, WN, IA). We compared the infection rate of 2,195 individuals who received a single dose of the Ad26.COV2.S vaccine from Johnson & Johnson (J&J) to the infection rate of 21,950 unvaccinated, propensity-matched individuals between February 27th and April 14th 2021. Of the 1,779 vaccinated individuals with at least two weeks of follow-up, only 3 (0.17%) tested positive for SARS-CoV-2 15 days or more after vaccination compared to 128 of 17,744 (0.72%) unvaccinated individuals (4.34 fold reduction rate). This corresponds to a vaccine effectiveness of 76.7% (95% CI: 30.3-95.3%) in preventing SARS-CoV-2 infection with onset at least two weeks after vaccination. This data is consistent with the clinical trial-reported efficacy of Ad26.COV2.S in preventing moderate to severe COVID-19 with onset at least 14 days after vaccine administration (66.9%; 95% CI: 59.0-73.4%). Due to the recent authorization of the Ad26.COV2.S vaccine, there are not yet enough hospitalizations, ICU admissions, or deaths within this cohort to robustly assess the effect of vaccination on COVID-19 severity, but these outcomes will be continually assessed in near-real-time with our platform. Collectively, this study provides further evidence that a single dose of Ad26.COV2.S is highly effective in preventing SARS-CoV-2 infection and reaffirms the urgent need to continue mass vaccination efforts globally.

Equitable vaccination distribution is a priority for outcompeting the transmission of COVID-19. Here, the impact of demographic, socioeconomic, and environmental factors on county-level vaccination rates and COVID-19 incidence changes is assessed. In particular, using data from 3142 US counties with over 328 million individuals, correlations were computed between cumulative vaccination rate and change in COVID-19 incidence from 1 December 2020 to 6 June 2021, with 44 different demographic, environmental, and socioeconomic factors. This correlation analysis was also performed using multivariate linear regression to adjust for age as a potential confounding variable. These correlation analyses demonstrated that counties with high levels of uninsured individuals have significantly lower COVID-19 vaccination rates (Spearman correlation: −0.460, p-value: <0.001). In addition, severe housing problems and high housing costs were strongly correlated with increased COVID-19 incidence (Spearman correlations: 0.335, 0.314, p-values: <0.001, <0.001). This study shows that socioeconomic factors are strongly correlated to both COVID-19 vaccination rates and incidence rates, underscoring the need to improve COVID-19 vaccination campaigns in marginalized communities.