Application of a Machine Learning Algorithm to Develop and Validate a Prediction Model for Ambulatory Non-Arrivals.

BACKGROUND: Non-arrivals to scheduled ambulatory visits are common and lead to a discontinuity of care, poor health outcomes, and increased subsequent healthcare utilization. Reducing non-arrivals is important given their association with poorer health outcomes and cost to health systems. OBJECTIVE: To develop and validate a prediction model for ambulatory non-arrivals. DESIGN: Retrospective cohort study. PATIENTS OR SUBJECTS: Patients at an integrated health system who had an outpatient visit scheduled from January 1, 2020, to February 28, 2022. MAIN MEASURES: Non-arrivals to scheduled appointments. KEY RESULTS: There were over 4.3 million ambulatory appointments from 1.2 million adult patients. Patients with appointment non-arrivals were more likely to be single, racial/ethnic minorities, and not having an established primary care provider compared to those who arrived at their appointments. A prediction model using the XGBoost machine learning algorithm had the highest AUC value (0.768 [0.767-0.770]). Using SHAP values, the most impactful features in the model include rescheduled appointments, lead time (number of days from scheduled to appointment date), appointment provider, number of days since last appointment with the same department, and a patient's prior appointment status within the same department. Scheduling visits close to an appointment date is predicted to be less likely to result in a non-arrival. Overall, the prediction model calibrated well for each department, especially over the operationally relevant probability range of 0 to 40%. Departments with fewer observations and lower non-arrival rates generally had a worse calibration. CONCLUSIONS: Using a machine learning algorithm, we developed a prediction model for non-arrivals to scheduled ambulatory appointments usable for all medical specialties. The proposed prediction model can be deployed within an electronic health system or integrated into other dashboards to reduce non-arrivals. Future work will focus on the implementation and application of the model to reduce non-arrivals.

Use of TIMP-2 and IGFBP-7 for prediction of postoperative acute kidney injury after cardiac surgery.

BACKGROUND: The following cell cycle arrest urinary biomarkers, tissue inhibitor of metalloproteinases-2 (TIMP-2) and insulin-like growth factor-binding protein 7 (IGFBP-7), have been used for early detection of acute kidney injury (AKI) in critically ill patients. The purpose of this study is to validate the use of these urinary biomarkers in patients undergoing open heart surgery. MATERIALS AND METHODS: In a single-center prospective observational study, urine samples were collected in 108 consecutive patients who underwent open heart surgery immediately after separation from cardiopulmonary bypass and on postoperative day 1, and were sent for the biomarker [TIMP-2]*[IGFBP7] analysis. Acute kidney injury was defined based on KDIGO criteria, and levels of [TIMP-2]*[IGFBP7] were analyzed for the ability to predict AKI. RESULTS: Of the 108 patients, 19 (17.6%) patients developed postoperative AKI within 48 hours of surgery. At the threshold of > 0.3 (ng/mL)2/1,000, post-cardiopulmonary bypass [TIMP-2]*[IGFBP-7] had a sensitivity of 13% and specificity of 82% for predicting postoperative AKI. Postoperative day-1 [TIMP-2]*[IGFBP-7] had a sensitivity of 47% and a specificity of 59% for predicting postoperative AKI. There were no differences in [TIMP-2]*[IGFBP-7] values at either timepoint between patients who developed postoperative AKI as compared to those who did not. CONCLUSION: Urinary [TIMP-2]*[IGFBP7] was not predictive of the risk of AKI after cardiac surgery in this single-center study population.

An early experience on the effect of solid organ transplant status on hospitalized COVID-19 patients.

We compared the outcome of COVID-19 in immunosuppressed solid organ transplant (SOT) patients to a transplant naïve population. In total, 10 356 adult hospital admissions for COVID-19 from March 1, 2020 to April 27, 2020 were analyzed. Data were collected on demographics, baseline clinical conditions, medications, immunosuppression, and COVID-19 course. Primary outcome was combined death or mechanical ventilation. We assessed the association between primary outcome and prognostic variables using bivariate and multivariate regression models. We also compared the primary endpoint in SOT patients to an age, gender, and comorbidity-matched control group. Bivariate analysis found transplant status, age, gender, race/ethnicity, body mass index, diabetes, hypertension, cardiovascular disease, COPD, and GFR <60 mL/min/1.73 m2 to be significant predictors of combined death or mechanical ventilation. After multivariate logistic regression analysis, SOT status had a trend toward significance (odds ratio [OR] 1.29; 95% CI 0.99-1.69, p = .06). Compared to an age, gender, and comorbidity-matched control group, SOT patients had a higher combined risk of death or mechanical ventilation (OR 1.34; 95% CI 1.03-1.74, p = .027).

Gastrointestinal bleeding in hospitalized COVID-19 patients: a propensity score matched cohort study.

BACKGROUND AND AIMS: Gastrointestinal (GI) bleeding has been observed amongst patients hospitalized with COVID-19. Recently, anticoagulation has shown to decrease mortality, but it is unclear whether this contributes to increased GI bleeding. The aims of this study are: (i) to examine whether there are risk factors for GI bleeding in COVID-19 patients and (ii) to study whether there is a mortality difference between hospitalized patients with COVID-19 with and without GI bleeding. METHODS: This is a propensity score matched case-control study from a large health system in the New York metropolitan area between March 1st and April 27th. COVID-19 patients with GI bleeding were matched 1:1 to COVID-19 patients without bleeding using a propensity score that took into account comorbidities, demographics, GI bleeding risk factors and length of stay. RESULTS: Of 11, 158 hospitalized with COVID-19, 314 patients were identified with GI bleeding. The point prevalence of GI bleeding was 3%. There were no identifiable risk factors for GI bleeding. Use of anticoagulation medication or antiplatelet agents was not associated with increased risk of GI bleeding in COVID-19 patients. For patients who developed a GI bleed during the hospitalization, there was an increased mortality risk in the GI bleeding group (OR 1.58, P = 0.02). CONCLUSION: Use of anticoagulation or antiplatelet agents was not risk factors for GI bleeding in a large cohort of hospitalized COVID-19 patients. Those with GI bleeding during the hospitalization had increased mortality.

Outcomes Among Patients Hospitalized With COVID-19 and Acute Kidney Injury.

RATIONALE & OBJECTIVE: Outcomes of patients hospitalized with coronavirus disease 2019 (COVID-19) and acute kidney injury (AKI) are not well understood. The goal of this study was to investigate the survival and kidney outcomes of these patients. STUDY DESIGN: Retrospective cohort study. SETTING & PARTICIPANTS: Patients (aged≥18 years) hospitalized with COVID-19 at 13 hospitals in metropolitan New York between March 1, 2020, and April 27, 2020, followed up until hospital discharge. EXPOSURE: AKI. OUTCOMES: Primary outcome: in-hospital death. SECONDARY OUTCOMES: requiring dialysis at discharge, recovery of kidney function. ANALYTICAL APPROACH: Univariable and multivariable time-to-event analysis and logistic regression. RESULTS: Among 9,657 patients admitted with COVID-19, the AKI incidence rate was 38.4/1,000 patient-days. Incidence rates of in-hospital death among patients without AKI, with AKI not requiring dialysis (AKI stages 1-3), and with AKI receiving dialysis (AKI 3D) were 10.8, 31.1, and 37.5/1,000 patient-days, respectively. Taking those without AKI as the reference group, we observed greater risks for in-hospital death for patients with AKI 1-3 and AKI 3D (HRs of 5.6 [95% CI, 5.0-6.3] and 11.3 [95% CI, 9.6-13.1], respectively). After adjusting for demographics, comorbid conditions, and illness severity, the risk for death remained higher among those with AKI 1-3 (adjusted HR, 3.4 [95% CI, 3.0-3.9]) and AKI 3D (adjusted HR, 6.4 [95% CI, 5.5-7.6]) compared with those without AKI. Among patients with AKI 1-3 who survived, 74.1% achieved kidney recovery by the time of discharge. Among those with AKI 3D who survived, 30.6% remained on dialysis at discharge, and prehospitalization chronic kidney disease was the only independent risk factor associated with needing dialysis at discharge (adjusted OR, 9.3 [95% CI, 2.3-37.8]). LIMITATIONS: Observational retrospective study, limited to the NY metropolitan area during the peak of the COVID-19 pandemic. CONCLUSIONS: AKI in hospitalized patients with COVID-19 was associated with significant risk for death.

Examination of Post-discharge Follow-up Appointment Status and 30-Day Readmission.

BACKGROUND: Post-hospital discharge follow-up appointments are intended to evaluate patients' recovery following a hospitalization, but it is unclear how appointment statuses are associated with readmissions. OBJECTIVE: To examine the association between post-discharge ambulatory follow-up status, (1) having a scheduled appointment and (2) arriving to said appointment, and 30-day readmission. DESIGN AND SETTING: A retrospective cohort study of patients hospitalized at 12 hospitals in an Integrated Delivery Network and their ambulatory appointments in that same network. PATIENTS AND MAIN MEASURES: We included 50,772 patients who had an ambulatory appointment within 18 months of an inpatient admission in 2018. Primary outcome was readmission within 30 days post-discharge. KEY RESULTS: There were 32,108 (63.2%) patients with scheduled follow-up appointments and 18,664 (36.8%) patients with no follow-up; 28,313 (88.2%) patients arrived, 3149 (9.8%) missed, and 646 (2.0%) were readmitted prior to their scheduled appointments. Overall 30-day readmission rate was 7.3%; 6.0% [5.75-6.31] for those who arrived, 8.8% [8.44-9.25] for those without follow-up, and 10.3% [9.28-11.40] for those who missed a scheduled appointment (p < 0.001). After adjusting for covariates, patients who arrived at their appointment in the first week following discharge were significantly less likely to be readmitted than those not having any follow-up scheduled (medical adjusted hazard ratio (aHR) 0.57 [0.47-0.69], p < 0.001; surgical aHR 0.58 [0.44-0.75], p < 0.001) There was an increased risk at weeks 3 and 4 for medical patients who arrived at a follow-up compared to those with no follow-up scheduled (week 3 aHR 1.29 [1.10-1.51], p = 0.001; week 4 aHR 1.46 [1.26-1.70], p < 0.001). CONCLUSIONS: The benefit of patients arriving to their post-discharge appointments compared with patients who missed their follow-up visits or had no follow-up scheduled, is only significant during first week post-discharge, suggesting that coordination within 1 week of discharge is critical in reducing 30-day readmissions.

Incidence of Venous Thromboembolism and Mortality in Patients with Initial Presentation of COVID-19.

Venous thromboembolism (VTE) has emerged as an important issue in patients with COVID-19. The purpose of this study is to identify the incidence of VTE and mortality in COVID-19 patients initially presenting to a large health system. Our retrospective study included adult patients (excluding patients presenting with obstetric/gynecologic conditions) across a multihospital health system in the New York Metropolitan Region from March 1-April 27, 2020. VTE and mortality rates within 8 h of assessment were described. In 10,871 adults with COVID-19, 118 patients (1.09%) were diagnosed with symptomatic VTE (101 pulmonary embolism, 17 deep vein thrombosis events) and 28 patients (0.26%) died during initial assessment. Among these 146 patients, 64.4% were males, 56.8% were 60 years or older, 15.1% had a BMI > 35, and 11.6% were admitted to the intensive care unit. Comorbidities included hypertension (46.6%), diabetes (24.7%), hyperlipidemia (14.4%), chronic lung disease (12.3%), coronary artery disease (11.0%), and prior VTE (7.5%). Key medications included corticosteroids (22.6%), statins (21.2%), antiplatelets (20.6%), and anticoagulants (20.6%). Highest D-Dimer was greater than six times the upper limit of normal in 51.4%. Statin and antiplatelet use were associated with decreased VTE or mortality (each p < 0.01). In COVID-19 patients who initially presented to a large multihospital health system, the overall symptomatic VTE and mortality rate was over 1.0%. Statin and antiplatelet use were associated with decreased VTE or mortality. The potential benefits of antithrombotics in high risk COVID-19 patients during the pre-hospitalization period deserves study.

External validation of the IMPROVE-DD risk assessment model for venous thromboembolism among inpatients with COVID-19.

There is a need to discriminate which COVID-19 inpatients are at higher risk for venous thromboembolism (VTE) to inform prophylaxis strategies. The IMPROVE-DD VTE risk assessment model (RAM) has previously demonstrated good discrimination in non-COVID populations. We aimed to externally validate the IMPROVE-DD VTE RAM in medical patients hospitalized with COVID-19. This retrospective cohort study evaluated the IMPROVE-DD VTE RAM in adult patients with COVID-19 admitted to one of thirteen Northwell Health hospitals in the New York metropolitan area between March 1, 2020 and April 27, 2020. VTE was defined as new-onset symptomatic deep venous thrombosis or pulmonary embolism. To assess the predictive value of the RAM, the receiver operating characteristic (ROC) curve was plotted and the area under the curve (AUC) was calculated. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated. Of 9407 patients who met study criteria, 274 patients developed VTE with a prevalence of 2.91%. The VTE rate was 0.41% for IMPROVE-DD score 0-1 (low risk), 1.21% for score 2-3 (moderate risk), and 5.30% for score ≥ 4 (high risk). Approximately 45.7% of patients were classified as high VTE risk, 33.3% moderate risk, and 21.0% low risk. Discrimination of low versus moderate-high VTE risk demonstrated sensitivity 0.971, specificity 0.215, PPV 0.036, and NPV 0.996. ROC AUC was 0.703. In this external validation study, the IMPROVE-DD VTE RAM demonstrated very good discrimination to identify hospitalized COVID-19 patients at low, moderate, and high VTE risk.

A Machine Learning Prediction Model of Respiratory Failure Within 48 Hours of Patient Admission for COVID-19: Model Development and Validation.

BACKGROUND: Predicting early respiratory failure due to COVID-19 can help triage patients to higher levels of care, allocate scarce resources, and reduce morbidity and mortality by appropriately monitoring and treating the patients at greatest risk for deterioration. Given the complexity of COVID-19, machine learning approaches may support clinical decision making for patients with this disease. OBJECTIVE: Our objective is to derive a machine learning model that predicts respiratory failure within 48 hours of admission based on data from the emergency department. METHODS: Data were collected from patients with COVID-19 who were admitted to Northwell Health acute care hospitals and were discharged, died, or spent a minimum of 48 hours in the hospital between March 1 and May 11, 2020. Of 11,525 patients, 933 (8.1%) were placed on invasive mechanical ventilation within 48 hours of admission. Variables used by the models included clinical and laboratory data commonly collected in the emergency department. We trained and validated three predictive models (two based on XGBoost and one that used logistic regression) using cross-hospital validation. We compared model performance among all three models as well as an established early warning score (Modified Early Warning Score) using receiver operating characteristic curves, precision-recall curves, and other metrics. RESULTS: The XGBoost model had the highest mean accuracy (0.919; area under the curve=0.77), outperforming the other two models as well as the Modified Early Warning Score. Important predictor variables included the type of oxygen delivery used in the emergency department, patient age, Emergency Severity Index level, respiratory rate, serum lactate, and demographic characteristics. CONCLUSIONS: The XGBoost model had high predictive accuracy, outperforming other early warning scores. The clinical plausibility and predictive ability of XGBoost suggest that the model could be used to predict 48-hour respiratory failure in admitted patients with COVID-19.

Outcomes of patients with end-stage kidney disease hospitalized with COVID-19.

Given the high risk of infection-related mortality, patients with end-stage kidney disease (ESKD) may be at increased risk with COVID-19. To assess this, we compared outcomes of patients with and without ESKD, hospitalized with COVID-19. This was a retrospective study of patients admitted with COVID-19 from 13 New York hospitals from March 1, 2020, to April 27, 2020, and followed through May 27, 2020. We measured primary outcome (in-hospital death), and secondary outcomes (mechanical ventilation and length of stay). Of 10,482 patients with COVID-19, 419 had ESKD. Patients with ESKD were older, had a greater percentage self-identified as Black, and more comorbid conditions. Patients with ESKD had a higher rate of in-hospital death than those without (31.7% vs 25.4%, odds ratio 1.38, 95% confidence interval 1.12 - 1.70). This increase rate remained after adjusting for demographic and comorbid conditions (adjusted odds ratio 1.37, 1.09 - 1.73). The odds of length of stay of seven or more days was higher in the group with compared to the group without ESKD in both the crude and adjusted analysis (1.62, 1.27 - 2.06; vs 1.57, 1.22 - 2.02, respectively). There was no difference in the odds of mechanical ventilation between the groups. Independent risk factors for in-hospital death for patients with ESKD were increased age, being on a ventilator, lymphopenia, blood urea nitrogen and serum ferritin. Black race was associated with a lower risk of death. Thus, among patients hospitalized with COVID-19, those with ESKD had a higher rate of in-hospital death compared to those without ESKD.

Acute kidney injury in patients hospitalized with COVID-19.

The rate of acute kidney injury (AKI) associated with patients hospitalized with Covid-19, and associated outcomes are not well understood. This study describes the presentation, risk factors and outcomes of AKI in patients hospitalized with Covid-19. We reviewed the health records for all patients hospitalized with Covid-19 between March 1, and April 5, 2020, at 13 academic and community hospitals in metropolitan New York. Patients younger than 18 years of age, with end stage kidney disease or with a kidney transplant were excluded. AKI was defined according to KDIGO criteria. Of 5,449 patients admitted with Covid-19, AKI developed in 1,993 (36.6%). The peak stages of AKI were stage 1 in 46.5%, stage 2 in 22.4% and stage 3 in 31.1%. Of these, 14.3% required renal replacement therapy (RRT). AKI was primarily seen in Covid-19 patients with respiratory failure, with 89.7% of patients on mechanical ventilation developing AKI compared to 21.7% of non-ventilated patients. 276/285 (96.8%) of patients requiring RRT were on ventilators. Of patients who required ventilation and developed AKI, 52.2% had the onset of AKI within 24 hours of intubation. Risk factors for AKI included older age, diabetes mellitus, cardiovascular disease, black race, hypertension and need for ventilation and vasopressor medications. Among patients with AKI, 694 died (35%), 519 (26%) were discharged and 780 (39%) were still hospitalized. AKI occurs frequently among patients with Covid-19 disease. It occurs early and in temporal association with respiratory failure and is associated with a poor prognosis.

COVID-19 in kidney transplant recipients.

There is minimal information on coronavirus disease 2019 (COVID-19) in immunocompromised individuals. We have studied 10 patients treated at 12 adult care hospitals. Ten kidney transplant recipients tested positive for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) by polymerase chain reaction, and 9 were admitted. The median age was 57 (interquartile range [IQR] 47-67), 60% were male, 40% Caucasian, and 30% Black/African American. Median time from transplant to COVID-19 testing was 2822 days (IQR 1272-4592). The most common symptom was fever, followed by cough, myalgia, chills, and fatigue. The most common chest X-ray and computed tomography abnormality was multifocal patchy opacities. Three patients had no abnormal findings. Leukopenia was seen in 20% of patients, and allograft function was stable in 50% of patients. Nine patients were on tacrolimus and a mycophenolic antimetabolite, and 70% were on prednisone. Hospitalized patients had their antimetabolite agent stopped. All hospitalized patients received hydroxychloroquine and azithromycin. Three patients died (30%), and 5 (50%) developed acute kidney injury. Kidney transplant recipients infected with COVID-19 should be monitored closely in the setting of lowered immunosuppression. Most individuals required hospitalization and presenting symptoms were similar to those of nontransplant individuals.

COVID-19 in Hospitalized Patients on Chronic Peritoneal Dialysis: A Case Series.

BACKGROUND: The COVID-19 pandemic has affected the end-stage kidney disease (ESKD) population, with high mortality rates reported among patients on hemodialysis. However, the degree to which it has affected the peritoneal dialysis (PD) population in the United States has not yet been elucidated. In this report, we describe the clinical characteristics, presentations, clinical course, and outcomes of ESKD patients on PD hospitalized with COVID-19. METHODS: We describe the characteristics, presentation, and outcomes of adult ESKD patients on chronic PD hospitalized with CO-VID-19 in our 13 major hospitals in the NY health system using descriptive statistical analysis. RESULTS: Of 419 hospitalized patients with ESKD, 11 were on chronic PD therapy (2.6%). Among those 11, 3 patients required mechanical ventilation, 2 of whom died. Of the entire cohort, 9 of the 11 patients (82%) were discharged alive. While fever was a common presentation, more than half of our patients also presented with diarrhea. Interestingly, 3 patients were diagnosed with culture-negative peritonitis during their hospitalization. Seven patients reported positive SARS-CoV-2 exposure from a member of their household. CONCLUSION: Hospitalized patients on PD with COVID-19 had a relatively mild course, and majority of them were discharged home.

Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the New York City Area.

Importance: There is limited information describing the presenting characteristics and outcomes of US patients requiring hospitalization for coronavirus disease 2019 (COVID-19). Objective: To describe the clinical characteristics and outcomes of patients with COVID-19 hospitalized in a US health care system. Design, Setting, and Participants: Case series of patients with COVID-19 admitted to 12 hospitals in New York City, Long Island, and Westchester County, New York, within the Northwell Health system. The study included all sequentially hospitalized patients between March 1, 2020, and April 4, 2020, inclusive of these dates. Exposures: Confirmed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection by positive result on polymerase chain reaction testing of a nasopharyngeal sample among patients requiring admission. Main Outcomes and Measures: Clinical outcomes during hospitalization, such as invasive mechanical ventilation, kidney replacement therapy, and death. Demographics, baseline comorbidities, presenting vital signs, and test results were also collected. Results: A total of 5700 patients were included (median age, 63 years [interquartile range {IQR}, 52-75; range, 0-107 years]; 39.7% female). The most common comorbidities were hypertension (3026; 56.6%), obesity (1737; 41.7%), and diabetes (1808; 33.8%). At triage, 30.7% of patients were febrile, 17.3% had a respiratory rate greater than 24 breaths/min, and 27.8% received supplemental oxygen. The rate of respiratory virus co-infection was 2.1%. Outcomes were assessed for 2634 patients who were discharged or had died at the study end point. During hospitalization, 373 patients (14.2%) (median age, 68 years [IQR, 56-78]; 33.5% female) were treated in the intensive care unit care, 320 (12.2%) received invasive mechanical ventilation, 81 (3.2%) were treated with kidney replacement therapy, and 553 (21%) died. As of April 4, 2020, for patients requiring mechanical ventilation (n = 1151, 20.2%), 38 (3.3%) were discharged alive, 282 (24.5%) died, and 831 (72.2%) remained in hospital. The median postdischarge follow-up time was 4.4 days (IQR, 2.2-9.3). A total of 45 patients (2.2%) were readmitted during the study period. The median time to readmission was 3 days (IQR, 1.0-4.5) for readmitted patients. Among the 3066 patients who remained hospitalized at the final study follow-up date (median age, 65 years [IQR, 54-75]), the median follow-up at time of censoring was 4.5 days (IQR, 2.4-8.1). Conclusions and Relevance: This case series provides characteristics and early outcomes of sequentially hospitalized patients with confirmed COVID-19 in the New York City area.