Quantitative and Qualitative Evaluation of Provider Use of a Novel Machine Learning Model for Favorable Outcome Prediction

Predictive models may be particularly beneficial to clinicians when they face uncertainty and seek to develop a mental model of disease progression, but we know little about the post-implementation effects of predictive models on clinicians' experience of their work. Combining survey and interview methods, we found that providers using a predictive algorithm reported being significantly less uncertain and better able to anticipate, plan and prepare for patient discharge than non-users. The tool helped hospitalists form and develop confidence in their mental models of a novel disease (Covid-19). Yet providers' attention to the predictive tool declined as their confidence in their own mental models grew. Predictive algorithms that not only offer data but also provide feedback on decisions, thus supporting providers' motivation for continuous learning, hold promise for more sustained provider attention and cognition augmentation.

Treating COVID-19 With Hydroxychloroquine (TEACH): A Multicenter, Double-Blind Randomized Controlled Trial in Hospitalized Patients.

BACKGROUND: Effective therapies to combat coronavirus 2019 (COVID-19) are urgently needed. Hydroxychloroquine (HCQ) has in vitro antiviral activity against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), but the clinical benefit of HCQ in treating COVID-19 is unclear. Randomized controlled trials are needed to determine the safety and efficacy of HCQ for the treatment of hospitalized patients with COVID-19. METHODS: We conducted a multicenter, double-blind randomized clinical trial of HCQ among patients hospitalized with laboratory-confirmed COVID-19. Subjects were randomized in a 1:1 ratio to HCQ or placebo for 5 days and followed for 30 days. The primary efficacy outcome was a severe disease progression composite end point (death, intensive care unit admission, mechanical ventilation, extracorporeal membrane oxygenation, and/or vasopressor use) at day 14. RESULTS: A total of 128 patients were included in the intention-to-treat analysis. Baseline demographic, clinical, and laboratory characteristics were similar between the HCQ (n = 67) and placebo (n = 61) arms. At day 14, 11 (16.4%) subjects assigned to HCQ and 6 (9.8%) subjects assigned to placebo met the severe disease progression end point, but this did not achieve statistical significance (P = .350). There were no significant differences in COVID-19 clinical scores, number of oxygen-free days, SARS-CoV-2 clearance, or adverse events between HCQ and placebo. HCQ was associated with a slight increase in mean corrected QT interval, an increased D-dimer, and a trend toward an increased length of stay. CONCLUSIONS: In hospitalized patients with COVID-19, our data suggest that HCQ does not prevent severe outcomes or improve clinical scores. However, our conclusions are limited by a relatively small sample size, and larger randomized controlled trials or pooled analyses are needed.

Evaluating the Effect of a COVID-19 Predictive Model to Facilitate Discharge: A Randomized Controlled Trial.

BACKGROUND: We previously developed and validated a predictive model to help clinicians identify hospitalized adults with coronavirus disease 2019 (COVID-19) who may be ready for discharge given their low risk of adverse events. Whether this algorithm can prompt more timely discharge for stable patients in practice is unknown. OBJECTIVES: The aim of the study is to estimate the effect of displaying risk scores on length of stay (LOS). METHODS: We integrated model output into the electronic health record (EHR) at four hospitals in one health system by displaying a green/orange/red score indicating low/moderate/high-risk in a patient list column and a larger COVID-19 summary report visible for each patient. Display of the score was pseudo-randomized 1:1 into intervention and control arms using a patient identifier passed to the model execution code. Intervention effect was assessed by comparing LOS between intervention and control groups. Adverse safety outcomes of death, hospice, and re-presentation were tested separately and as a composite indicator. We tracked adoption and sustained use through daily counts of score displays. RESULTS: Enrolling 1,010 patients from May 15, 2020 to December 7, 2020, the trial found no detectable difference in LOS. The intervention had no impact on safety indicators of death, hospice or re-presentation after discharge. The scores were displayed consistently throughout the study period but the study lacks a causally linked process measure of provider actions based on the score. Secondary analysis revealed complex dynamics in LOS temporally, by primary symptom, and hospital location. CONCLUSION: An AI-based COVID-19 risk score displayed passively to clinicians during routine care of hospitalized adults with COVID-19 was safe but had no detectable impact on LOS. Health technology challenges such as insufficient adoption, nonuniform use, and provider trust compounded with temporal factors of the COVID-19 pandemic may have contributed to the null result. TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT04570488.

Giving Your Electronic Health Record a Checkup After COVID-19: A Practical Framework for Reviewing Clinical Decision Support in Light of the Telemedicine Expansion.

BACKGROUND: The transformation of health care during COVID-19, with the rapid expansion of telemedicine visits, presents new challenges to chronic care and preventive health providers. Clinical decision support (CDS) is critically important to chronic care providers, and CDS malfunction is common during times of change. It is essential to regularly reassess an organization's ambulatory CDS program to maintain care quality. This is especially true after an immense change, like the COVID-19 telemedicine expansion. OBJECTIVE: Our objective is to reassess the ambulatory CDS program at a large academic medical center in light of telemedicine's expansion in response to the COVID-19 pandemic. METHODS: Our clinical informatics team devised a practical framework for an intrapandemic ambulatory CDS assessment focused on the impact of the telemedicine expansion. This assessment began with a quantitative analysis comparing CDS alert performance in the context of in-person and telemedicine visits. Board-certified physician informaticists then completed a formal workflow review of alerts with inferior performance in telemedicine visits. Informaticists then reported on themes and optimization opportunities through the existing CDS governance structure. RESULTS: Our assessment revealed that 10 of our top 40 alerts by volume were not firing as expected in telemedicine visits. In 3 of the top 5 alerts, providers were significantly less likely to take action in telemedicine when compared to office visits. Cumulatively, alerts in telemedicine encounters had an action taken rate of 5.3% (3257/64,938) compared to 8.3% (19,427/233,636) for office visits. Observations from a clinical informaticist workflow review included the following: (1) Telemedicine visits have different workflows than office visits. Some alerts developed for the office were not appearing at the optimal time in the telemedicine workflow. (2) Missing clinical data is a common reason for the decreased alert firing seen in telemedicine visits. (3) Remote patient monitoring and patient-reported clinical data entered through the portal could replace data collection usually completed in the office by a medical assistant or registered nurse. CONCLUSIONS: In a large academic medical center at the pandemic epicenter, an intrapandemic ambulatory CDS assessment revealed clinically significant CDS malfunctions that highlight the importance of reassessing ambulatory CDS performance after the telemedicine expansion.

A validated, real-time prediction model for favorable outcomes in hospitalized COVID-19 patients.

The COVID-19 pandemic has challenged front-line clinical decision-making, leading to numerous published prognostic tools. However, few models have been prospectively validated and none report implementation in practice. Here, we use 3345 retrospective and 474 prospective hospitalizations to develop and validate a parsimonious model to identify patients with favorable outcomes within 96 h of a prediction, based on real-time lab values, vital signs, and oxygen support variables. In retrospective and prospective validation, the model achieves high average precision (88.6% 95% CI: [88.4-88.7] and 90.8% [90.8-90.8]) and discrimination (95.1% [95.1-95.2] and 86.8% [86.8-86.9]) respectively. We implemented and integrated the model into the EHR, achieving a positive predictive value of 93.3% with 41% sensitivity. Preliminary results suggest clinicians are adopting these scores into their clinical workflows.