A Clinician's Guide to Running Custom Machine-Learning Models in an Electronic Health Record Environment.

We recently brought an internally developed machine-learning model for predicting which patients in the emergency department would require hospital admission into the live electronic health record environment. Doing so involved navigating several engineering challenges that required the expertise of multiple parties across our institution. Our team of physician data scientists developed, validated, and implemented the model. We recognize a broad interest and need to adopt machine-learning models into clinical practice and seek to share our experience to enable other clinician-led initiatives. This Brief Report covers the entire model deployment process, starting once a team has trained and validated a model they wish to deploy in live clinical operations.

Assessing the Generalizability of a Clinical Machine Learning Model Across Multiple Emergency Departments.

Objective: To assess the generalizability of a clinical machine learning algorithm across multiple emergency departments (EDs). Patients and Methods: We obtained data on all ED visits at our health care system's largest ED from May 5, 2018, to December 31, 2019. We also obtained data from 3 satellite EDs and 1 distant-hub ED from May 1, 2018, to December 31, 2018. A gradient-boosted machine model was trained on pooled data from the included EDs. To prevent the effect of differing training set sizes, the data were randomly downsampled to match those of our smallest ED. A second model was trained on this downsampled, pooled data. The model's performance was compared using area under the receiver operating characteristic (AUC). Finally, site-specific models were trained and tested across all the sites, and the importance of features was examined to understand the reasons for differing generalizability. Results: The training data sets contained 1918-64,161 ED visits. The AUC for the pooled model ranged from 0.84 to 0.94 across the sites; the performance decreased slightly when Ns were downsampled to match those of our smallest ED site. When site-specific models were trained and tested across all the sites, the AUCs ranged more widely from 0.71 to 0.93. Within a single ED site, the performance of the 5 site-specific models was most variable for our largest and smallest EDs. Finally, when the importance of features was examined, several features were common to all site-specific models; however, the weight of these features differed. Conclusion: A machine learning model for predicting hospital admission from the ED will generalize fairly well within the health care system but will still have significant differences in AUC performance across sites because of site-specific factors.

Better clinical outcomes in hospitalized COVID-19 minority patients with accompanying gastrointestinal symptoms.

BACKGROUND AND AIMS: The novel coronavirus (SARS-CoV-2) is highly contagious pathogen that primarily causes respiratory illnesses. Howerver, multiple gastrointestinal (GI) symptoms have been reported in Coronavirus Disease of 2019 (COVID-19). We conducted a retrospective cohort study of inpatients with COVID-19 at the George Washington University Hospital (GWUH) to assess the prevalence of GI symptoms and their association with clinical outcomes. METHODS: We reviewed the charts of 401 adults admitted to GWUH with positive SARS-CoV-2 tests from February 24 to May 21, 2020, ultimately including 382 inpatients. RESULTS: 87% of our cohort was African American or Latinx. 59% of patients reported at least one GI symptom, with diarrhea being the most common (29%). Patients with GI symptoms were slightly younger (58 +/- 15.8 vs. 65 +/- 16.9, p = 0.0005), have higher body mass index (31.5 +/- Standard Deviation of 8.7 vs. 28 +/- 8.2, p = 0.0001), and more likely to be Latinx (34 vs. 27, p = 0.01). Patients who presented with abdominal pain, nausea, vomiting, or diarrhea had significantly lower rates of death during hospitalization compared to those who did not present those symptoms (Odds Ratio 0.48, 95% Confidence Interval 0.28-0.8, p = 0.004). CONCLUSIONS: Our study suggests that GI symptoms portend a less-severe clinical course of COVID-19 which may reflect a different disease phenotype and lower overall immune response. Additional research should focus on more robust symptom reporting and longer follow-up.

Systematic Review Estimating the Burden of Dementia in the Latin America and Caribbean Region: A Bayesian Approach.

Background: The global burden of dementia has increasingly shifted to low- and middle-income regions that lack essential data for monitoring epidemiological progression, and policy and planning support. Drawing upon data that have emerged since the last known estimates published in 2015, this study aims to update dementia estimates in the Latin America and Caribbean (LAC) region for the years 2020, 2030, and 2050 through the application of a recently validated Bayesian approach for disease estimates useful when data sources are scarce. Methods: A comprehensive parallel systematic review of PubMed, EMBASE, PsycINFO, Global Health, and LILACS was conducted to identify prospective population-based epidemiological studies on dementia published in English from 2013 to 2018 in LAC. English and non-English data cited by a recent review on dementia estimates in LAC were also examined for additional data. A Bayesian normal-normal hierarchical model (NNHM) was developed to estimate age-specific and age-adjusted dementia prevalence in people aged 60+. Using age-specific population projections from the UN, the total number of people affected by dementia for the years 2020, 2030, and 2050 were estimated. Results: 1,414 studies were identified, of which only 7 met the inclusion criteria. The studies had 7,684 participants and 1,191 dementia cases. The age-standardized prevalence of all forms of dementia in LAC was 8% (95% CI: 5-11.5%) in people aged 60+. The estimated prevalence varied with age, increasing from 2.5% (95% CI: 0.08-4.0%) in the 60-69 age group, to 9.4% (95% CI: 5.4-13.2%) in the 70-79 age group and 28.9% (95% CI: 20.3-37.2%) in the ≥80 age group. The number of people age 60 and older living with dementia in LAC in 2020 was estimated at 6.86 (95% CI: 4.3-9.8) million, 9.94 (95% CI: 6.16-14.15) million in 2030, and 19.33 (95% CI: 12.3-13.6) million in 2050. Conclusion: We project an upward disease trajectory for dementia in LAC countries. The projection is likely an underestimation of the true dementia burden given the underrepresentation of rural and socio-economically deprived populations. More research is urgently needed to improve the accuracy of disease estimates, guide clinicians to improve evaluations for earlier recognition of dementia, and support the development of effective policies for improving dementia prevention, diagnosis and clinical management in LAC's diverse and aging communities.

Worldwide Survey of COVID-19-Associated Arrhythmias.

[Figure: see text].

The effect of anticoagulation on clinical outcomes in novel Coronavirus (COVID-19) pneumonia in a U.S. cohort.

BACKGROUND: COVID-19 infection is associated with D-dimer elevations, high rates of thrombus formation, and poor clinical outcomes. We sought to determine if empiric therapeutic anticoagulation (AC) affected survival in COVID-19 patients compared to standard prophylactic AC. METHODS: Retrospective analysis of 402 COVID-19 patients hospitalized between March 15 and May 31, 2020 was performed. Clinical outcomes were compared between 152 patients treated with therapeutic AC to 250 patients on prophylactic AC. An elastic net logistic regression was designed to first identify the important variables affecting mortality. These variables were then included as covariates to AC in standard multivariate logistic regression models studying the effect of AC on death. Nonparametric survival analysis was conducted, and Kaplan Meier curves were constructed. RESULTS: Increased mortality was associated with therapeutic AC [OR 3.42 (2.06, 5.67)]. The log-rank test was statistically significant at p = 0.001 showing higher mortality for patients treated with therapeutic AC compared to prophylactic AC. Subset analysis of critically ill and intubated patients had similar survival curves regardless of AC dose. The log-rank test was not significant even with Prentice modification. For non-ICU patients, the log rank test favoring prophylactic AC disappeared when the analysis was stratified by D-dimer level less or greater than 3 µg/mL. Approximately 9% of patients receiving therapeutic AC experienced clinically significant bleeding or thrombocytopenia, versus 3% in those receiving prophylactic AC. CONCLUSIONS: In our cohort, therapeutic anticoagulation provided no mortality benefit over thromboprophylaxis, independent of co-morbidities or disease severity. More adverse events were observed with therapeutic AC.

Systematic review estimating the burden of dementia in the WHO Southeast Asia Region using Bayesian and frequentist approaches.

BACKGROUND: Rapid increase in life expectancy in low- and middle-income countries including the World Health Organization's Southeast Asia Region (SEAR) has resulted in an increase in the global burden of dementia, which is expected to become a leading cause of morbidity. Accurate burden estimates are key for informing policy and planning. Given the paucity of data, estimates were developed using both a Bayesian methodology and as well as a traditional frequentist approach to gain better insights into methodological approaches for disease burden estimates. METHODS: Seven databases were searched for studies published between 2010-2018 regarding dementia prevalence in SEAR, generating 8 relevant articles. A random-effects model (REM) and a Bayesian normal-normal hierarchical model (NNHM) were used to obtain the pooled prevalence estimate of dementia for people aged 60 and above in SEAR. The latter model was also developed to estimate age-specific dementia prevalence. Using UN population estimates for SEAR, total and age-specific projections of the burden of dementia in 2015, 2020 and 2030 were calculated. RESULTS: The prevalence of dementia in SEAR was found to be 3% (95% confidence interval (CI) = 2-6%) in those above age 60 based on REM, and 3.1% (95% credible interval = 1.5-5.0%) based on the NNHM. The estimated prevalence varies with age, increasing from 1.6% (95% credible interval = 0.8-2.5%) in people aged 60-69 to 12.4% (95% credible interval = 5.6-20%) in people above the age of 80. The risk of developing dementia increased exponentially with age. The number of people living with dementia in SEAR in 2015 was estimated at 5.51 million (95% credible interval = 2.66-8.82), with projections of 6.66 million (95% credible interval = 3.21-10.7) in 2020 and 9.6 million (95% credible interval = 4.62-15.36) in 2030. CONCLUSION: The burden of dementia in SEAR is substantial and will continue to increase rapidly by 2030. The lack of research focusing on dementia in SEAR points to a significant under-recognition of this disease. The projected rise in dementia cases in the future should prompt urgent governmental response to address this growing public health issue. We also argue that given the overall paucity of data for the region, the Bayesian approach offers a promising methodology for improved estimates of disease prevalence and burden and should continue to be explored.

The association between biomarkers and clinical outcomes in novel coronavirus pneumonia in a US cohort.

Aim: To describe the association between D-dimer, CRP, IL-6, ferritin, LDH and the clinical outcomes in a cohort of 299 COVID-19 patients treated on the inpatient medical service at a university hospital in the District of Columbia (DC, USA). Methodology & results: In this retrospective study, we included all laboratory confirmed COVID-19 adults admitted to the inpatient medicine service at the George Washington University Hospital between 12 March 2020 and 9 May 2020. We analyzed the association of biomarkers on intensive care unit transfer, intubation and mortality. Threshold values for all biomarkers were found to be statistically significant and independently associated with higher odds of clinical deterioration and death. Conclusion: Laboratory markers of inflammation and coagulopathy can help clinicians identify patients who are at high risk for clinical deterioration in COVID-19.