Search

Benchmark datasets have a powerful normative influence: by determining how the real world is represented in data, they define which problems will first be solved by algorithms built using the datasets and, by extension, who these algorithms will work for. It is desirable for these datasets to serve four functions: (1) enabling the creation of clinically relevant algorithms; (2) facilitating like-for-like comparison of algorithmic performance; (3) ensuring reproducibility of algorithms; (4) asserting a normative influence on the clinical domains and diversity of patients that will potentially benefit from technological advances. Without benchmark datasets that satisfy these functions, it is impossible to address two perennial concerns of clinicians experienced in computational research: “the data scientists just go where the data is rather than where the needs are,” and, “yes, but will this work for my patients?” If algorithms are to be developed and applied for the care of patients, then it is prudent for the research community to create benchmark datasets proactively, across specialties. As yet, best practice in this area has not been defined. Broadly speaking, efforts will include design of the dataset; compliance and contracting issues relating to the sharing of sensitive data; enabling access and reuse; and planning for translation of algorithms to the clinical environment. If a deliberate and systematic approach is not followed, not only will the considerable benefits of clinical algorithms fail to be realized, but the potential harms may be regressively incurred across existing gradients of social inequity.

The death toll of the COVID-19 pandemic has been unprecedented, due to both the high number of SARS-CoV-2 infections and the seriousness of the disease resulting from these infections. Here, we present mortality rates and case fatality rates for COVID-19 over the past year compared with other historic leading causes of death in the United States. Among the risk categories considered, COVID-19 is the third leading cause of death for individuals 40 years old and over, with an overall annual mortality rate of 325 deaths per 100K individuals, behind only cancer (385 deaths per 100K individuals) and heart disease (412 deaths per 100K individuals). In addition, for individuals 40 years old and over, the case fatality rate for COVID-19 is greater than the case fatality rate for motor vehicle accidents. In particular, for the age group 40-49, the relative case fatality rate of COVID-19 is 1.5 fold (95% CI: [1.3, 1.7]) that of a motor vehicle accident, demonstrating that SARS-CoV-2 infection may be significantly more dangerous than a car crash for this age group. For older adults, COVID-19 is even more dangerous, and the relative case fatality rate of COVID-19 is 29.4 fold (95% CI: [23.2, 35.7]) that of a motor vehicle accident for individuals over 80 years old. On the other hand, motor vehicle accidents have a 4.5 fold (95% CI: [3.9, 5.1]) greater relative case fatality rate compared to COVID-19 for the age group of 20-29 years. These results highlight the severity of the COVID-19 pandemic especially for adults above 40 years of age and underscore the need for large-scale preventative measures to mitigate risks for these populations. Given that FDA-authorized COVID-19 vaccines have now been validated by multiple studies for their outstanding real-world effectiveness and safety, vaccination of all individuals who are over 40 years of age is one of the most pressing public health priorities of our time.

Nearly 150 million doses of FDA-authorized COVID vaccines have been administered in the United States. Sex-based differences of adverse events remain poorly understood, mandating the need for real-world investigation from Electronic Health Records (EHRs) and broader epidemiological data sets. Based on an augmented curation of EHR clinical notes of 31,064 COVID-vaccinated individuals (19,321 females and 11,743 males) in the Mayo Clinic, we find that nausea and vomiting were documented significantly more frequently in females than males after both vaccine doses (nausea: RRDose 1 = 1.67, pDose 1

Efficient and equitable vaccination distribution is a priority for effectively outcompeting the transmission of COVID-19 globally. A recent study from the Centers for Disease Control and Prevention (CDC) identified that US counties with high social vulnerability according to metrics such as poverty, unemployment, low income, and no high school diploma, have significantly lower rates of vaccination compared to the national average1. Here, we build upon this analysis to consider associations between county-level vaccination rates and 68 different demographic, socioeconomic, and environmental factors for 1,510 American counties with over 228 million individuals for which vaccination data was also available. Our analysis reveals that counties with high levels of uninsured individuals have significantly lower COVID-19 vaccination rates (Spearman correlation: −0.264), despite the fact that the CDC has mandated that all COVID-19 vaccines are free and cannot be denied to anyone based upon health insurance coverage or immigration status. Furthermore, we find that the counties with high levels of uninsured individuals tend to have the highest COVID-19 incidence rates in March 2021 relative to December 2020 (Spearman correlation: 0.388). Among the 68 factors analyzed, insurance coverage is the only factor which is highly correlated with both vaccination rate and change in COVID-19 incidence during the vaccination period (|Spearman correlation|> 0.25). We also find that counties with higher percentages of Black and Hispanic individuals have significantly lower vaccination rates (Spearman correlations: −0.128, −0.136) and lesser declines of COVID-incidence rates (Spearman correlations: 0.334, 0.330) during the vaccination period. Surprisingly however, after controlling for race, we find that the association between lack of insurance coverage and vaccination rate as well as COVID-19 incidence rates is largely driven by counties with a majority white population. Among the counties with high proportions of white residents (top 10% decile), the association between insurance coverage and vaccination rate is significant (Spearman correlation: −0.210, p-value: 0.002), but among counties with low proportions of white residents (bottom 10% decile) this association is not significant (Spearman correlation: 0.072, p-value: 0.088). Taken together, this study highlights the fact that intricate socioeconomic factors are correlated not just to COVID-19 vaccination rates, but also to COVID-19 incidence fluctuations, underscoring the need to improve COVID-19 vaccination campaigns in marginalized communities. The strong positive correlation between low levels of health insurance coverage and low vaccination rates is particularly concerning, and calls for improved public health messaging to emphasize the fact that health insurance is not required to be eligible for any of the FDA-authorized COVID-19 vaccines in the United States

Multimorbidity is a defining challenge for health systems and requires coordination of care delivery and care management. Care management is a clinical service designed to remotely engage patients between visits and after discharge in order to support self-management of chronic and emergent conditions, encourage increased use of scheduled care and address the use of unscheduled care. Care management can be provided using digital technology - digital care management. A robust methodology to assess digital care management, or any traditional or digital primary care intervention aimed at longitudinal management of multimorbidity, does not exist outside of randomized controlled trials (RCTs). RCTs are not always generalizable and are also not feasible for most healthcare organizations. We describe here a novel and pragmatic methodology for the evaluation of digital care management that is generalizable to any longitudinal intervention for multimorbidity irrespective of its mode of delivery. This methodology implements propensity matching with bootstrapping to address some of the major challenges in evaluation including identification of robust outcome measures, selection of an appropriate control population, small sample sizes with class imbalances, and limitations of RCTs. We apply this methodology to the evaluation of digital care management at a U.S. payor and demonstrate a 9% reduction in ER utilization, a 17% reduction in inpatient admissions, and a 29% increase in the utilization of preventive medicine services. From these utilization outcomes, we drive forward an estimated cost saving that is specific to a single payor's payment structure for the study time period of $641 per-member-per-month at 3 months. We compare these results to those derived from existing observational approaches, 1:1 and 1:n propensity matching, and discuss the circumstances in which our methodology has advantages over existing techniques. Whilst our methodology focuses on cost and utilization and is applied in the U.S. context, it is applicable to other outcomes such as Patient Reported Outcome Measures (PROMS) or clinical biometrics and can be used in other health system contexts where the challenge of multimorbidity is prevalent.

After one year of the COVID-19 pandemic, over 130 million individuals worldwide have been infected with the novel coronavirus, yet the post-acute sequelae of COVID-19 (PASC), also referred to as the ‘long COVID’ syndrome, remains mostly uncharacterized. We leveraged machine-augmented curation of the physician notes from electronic health records (EHRs) across the multi-state Mayo Clinic health system to retrospectively contrast the occurrence of symptoms and diseases in COVID-19 patients in the post-COVID period relative to the pre-COVID period (n=6,413). Through comparison of the frequency of 10,039 signs and symptoms before and after diagnosis, we identified an increase in hypertensive chronic kidney disease (OR 47.3, 95% CI 23.9-93.6, p=3.50×10−9), thromboembolism (OR 3.84, 95% CI 3.22-4.57, p=1.18×10−4), and hair loss (OR 2.44, 95% CI 2.15-2.76, p=8.46×10−3) in COVID-19 patients three to six months after diagnosis. The sequelae associated with long COVID were notably different among male vs female patients and patients above vs under 55 years old, with the hair loss enrichment found primarily in females and the thromboembolism enrichment in males. These findings compel targeted investigations into what may be persistent dermatologic, cardiovascular, and coagulopathic phenotypes following SARS-CoV-2 infection.

Importance: Over the past year, the COVID-19 pandemic has resulted in the deaths of more than 550,000 individuals in the United States. It is important for us to understand the risks of COVID-19 in relation to other leading causes of death in order to inform public policy decisions and individual behaviors to help mitigate the highest risks. Objective: Contrast COVID-19 with historic leading causes of death. Design: We present mortality rates and case fatality rates for COVID-19 from March 2020 to March 2021 compared with other historic leading causes of death in the United States. For each risk category and 10-year age groups (0-9, 10-19, …, 70-79, ≥ 80), we report the overall mortality and case fatality rates along with the relative risks compared to COVID-19 as a baseline. There were 16 other leading causes of death considered in this analysis, including: cancer, various subtypes of cancer (brain / nervous system, breast, lung / bronchus, pancreatic, prostate, leukemia, lymphoma), cerebrovascular disease, drug overdoses, firearm homicide, heart diseases, influenza, motor vehicle accidents (overall and without restraint), and suicide. Setting: This is a population-based study. Participants: N/A. Exposures: N/A. Main Outcomes and Measures: Overall mortality (number of deaths / population) and case fatality rates (number of deaths / number of cases). Results: COVID-19 is the third leading cause of death for individuals over 40 years old, with an annual mortality rate of 325 deaths/100K, behind only cancer (385 deaths/100K) and heart disease (412 deaths/100K). In addition, for individuals over 40 years old, the case fatality rate for COVID-19 is greater than the case fatality rate for motor vehicle accidents. For the age group 40-49 years old, the relative case fatality rate for COVID-19 is 1.5 fold (95% CI: [1.3,1.7]) that of motor vehicle accidents, and this relative risk increases to 29.4 fold (95% CI: [23.2, 35.7]) for individuals over 80 years old. Conclusions and Relevance: These results highlight the severity of the COVID-19 pandemic especially for adults over 40 years old and underscore the need for large-scale preventative measures to mitigate risks for these populations. Funding Statement: This work was supported by Nference, Inc. Declaration of Interests: None reported.

Background: Equitable vaccination distribution is a priority for outcompeting the transmission of COVID-19. A recent CDC study found that US counties with high social vulnerability – according to metrics such as poverty, unemployment and low income – have significantly lower rates of vaccination compared to the national average. We build upon this analysis to assess the impact of demographic, socioeconomic, and environmental factors on county-level vaccination rates and COVID-19 incidence changes. Methods: Using data from 3,142 US counties (with over 328 million individuals), we computed Spearman rank-order correlations between the outcome variables (cumulative vaccination rate, change in COVID-19 incidence) and 68 demographic, environmental, and socioeconomic factors. We also performed analysis using multivariate linear regression to adjust for age as a potential confounding variable. Finally, we performed fold change analysis on counties stratified by vaccination rates to compute relative risks along with Fisher test p-value associated with these features. Findings: Our analysis reveals that counties with high levels of uninsured individuals have significantly lower COVID-19 vaccination rates (Spearman correlation: -0.463). We separately assessed factors associated with COVID-19 incidence rates and found that the percent of a county’s population with severe housing problems and high housing costs were strongly correlated with COVID-19 incidence (Spearman correlations: 0.337, 0.314, p-values

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: &lt, 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: &lt, 0.001, &lt, 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.

Digital care management programs can reduce health care costs and improve quality of care. However, it is unclear how to target patients who are most likely to benefit from these programs ex ante, a shortcoming of current "risk score"-based approaches across many interventions. This study explores a framework to define impactability by using machine learning (ML) models to identify those patients most likely to benefit from a digital health intervention for care management. Anonymized insurance claims data were used from a commercially insured population across several US states and combined with inferred sociodemographic data. The approach involves creating 2 models and the comparative analysis of the methodologies and performances therein. The authors first train a cost prediction model to calculate the differences in predicted (without intervention) versus actual (with onboarding onto digital health platform) health care expenditures for patients (N