Impact of a machine learning algorithm on time to palliative care in a primary care population: protocol for a stepped-wedge pragmatic randomized trial.

BACKGROUND: As primary care populations age, timely identification of palliative care need is becoming increasingly relevant. Previous studies have targeted particular patient populations with life-limiting disease, but few have focused on patients in a primary care setting. Toward this end, we propose a stepped-wedge pragmatic randomized trial whereby a machine learning algorithm identifies patients empaneled to primary care units at Mayo Clinic (Rochester, Minnesota, United States) with high likelihood of palliative care need. METHODS: 42 care team units in 9 clusters were randomized to 7 wedges, each lasting 42 days. For care teams in treatment wedges, palliative care specialists review identified patients, making recommendations to primary care providers when appropriate. Care teams in control wedges receive palliative care under the standard of care. DISCUSSION: This pragmatic trial therefore integrates machine learning into clinical decision making, instead of simply reporting theoretical predictive performance. Such integration has the possibility to decrease time to palliative care, improving patient quality of life and symptom burden. TRIAL REGISTRATION: Clinicaltrials.gov NCT04604457 , restrospectively registered 10/26/2020. PROTOCOL: v0.5, dated 9/23/2020.

Assessment of Data Quality Variability across Two EHR Systems through a Case Study of Post-Surgical Complications.

Translation of predictive modeling algorithms into routine clinical care workflows faces challenges in the form of varying data quality-related issues caused by the heterogeneity of electronic health record (EHR) systems. To better understand these issues, we retrospectively assessed and compared the variability of data produced from two different EHR systems. We considered three dimensions of data quality in the context of EHR-based predictive modeling for three distinct translational stages: model development (data completeness), model deployment (data variability), and model implementation (data timeliness). The case study was conducted based on predicting post-surgical complications using both structured and unstructured data. Our study discovered a consistent level of data completeness, a high syntactic, and moderate-high semantic variability across two EHR systems, for which the quality of data is context-specific and closely related to the documentation workflow and the functionality of individual EHR systems.

Early Post-Hospitalization Hemoglobin Recovery and Clinical Outcomes in Survivors of Critical Illness: A Population-Based Cohort Study.

Anemia is common during critical illness, is associated with adverse clinical outcomes, and often persists after hospitalization. The goal of this investigation is to assess the relationships between post-hospitalization hemoglobin recovery and clinical outcomes after survival of critical illness. This is a population-based observational study of adults (≥18 years) surviving hospitalization for critical illness between January 1, 2010 and December 31, 2016 in Olmsted County, Minnesota, United States with hemoglobin concentrations and clinical outcomes assessed through one-year post-hospitalization. Multi-state proportional hazards models were utilized to assess the relationships between 1-month post-hospitalization hemoglobin recovery and hospital readmission or death through one-year after discharge. Among 6460 patients that survived hospitalization for critical illness during the study period, 2736 (42%) were alive, not hospitalized, and had available hemoglobin concentrations assessed at 1-month post-index hospitalization. Median (interquartile range) age was 69 (56, 80) years with 54% of male gender. Overall, 86% of patients had anemia at the time of hospital discharge, with median discharge hemoglobin concentrations of 10.2 (9.1, 11.6) g/dL. In adjusted analyses, each 1 g/dL increase in 1-month hemoglobin recovery was associated with decreased instantaneous hazard for hospital readmission (HR 0.87 [95% CI 0.84-0.90]; p < 0.001) and lower mortality (HR 0.82 [95% CI 0.75-0.89]; p < 0.001) through one-year post-hospitalization. The results were consistent in multiple pre-defined sensitivity analyses. Impaired early post-hospitalization hemoglobin recovery is associated with inferior clinical outcomes in the first year of survival after critical illness. Additional investigations are warranted to evaluate these relationships.

Early Detection of Post-Surgical Complications using Time-series Electronic Health Records.

Models predicting health complications are increasingly attempting to reflect the temporally changing nature of patient status. However, both the practice of medicine and electronic health records (EHR) have yet to provide a true longitudinal representation of a patient's medical history as relevant data is often asynchronous and highly missing. To match the stringent requirements of many static time models, time-series data has to be truncated, and missing values in samples have to be filled heuristically. However, these data preprocessing procedures may unconsciously misinterpret real-world data, and eventually lead into failure in practice. In this work, we proposed an augmented gated recurrent unit (GRU), which formulate both missingness and timeline signals into GRU cells. Real patient data of post-operative bleeding (POB) after Colon and Rectal Surgery (CRS) was collected from Mayo Clinic EHR system to evaluate the effectiveness of proposed model. Conventional models were also trained with imputed dataset, in which event missingness or asynchronicity were approximated. The performance of proposed model surpassed current state-of-the-art methods in this POB detection task, indicating our model could be more eligible to handle EHR datasets.

Quantifying the Importance of COVID-19 Vaccination to Our Future Outlook.

Predictive models have played a critical role in local, national, and international response to the COVID-19 pandemic. In the United States, health care systems and governmental agencies have relied on several models, such as the Institute for Health Metrics and Evaluation, Youyang Gu (YYG), Massachusetts Institute of Technology, and Centers for Disease Control and Prevention ensemble, to predict short- and long-term trends in disease activity. The Mayo Clinic Bayesian SIR model, recently made publicly available, has informed Mayo Clinic practice leadership at all sites across the United States and has been shared with Minnesota governmental leadership to help inform critical decisions during the past year. One key to the accuracy of the Mayo Clinic model is its ability to adapt to the constantly changing dynamics of the pandemic and uncertainties of human behavior, such as changes in the rate of contact among the population over time and by geographic location and now new virus variants. The Mayo Clinic model can also be used to forecast COVID-19 trends in different hypothetical worlds in which no vaccine is available, vaccinations are no longer being accepted from this point forward, and 75% of the population is already vaccinated. Surveys indicate that half of American adults are hesitant to receive a COVID-19 vaccine, and lack of understanding of the benefits of vaccination is an important barrier to use. The focus of this paper is to illustrate the stark contrast between these 3 scenarios and to demonstrate, mathematically, the benefit of high vaccine uptake on the future course of the pandemic.

Improving the delivery of palliative care through predictive modeling and healthcare informatics.

OBJECTIVE: Access to palliative care (PC) is important for many patients with uncontrolled symptom burden from serious or complex illness. However, many patients who could benefit from PC do not receive it early enough or at all. We sought to address this problem by building a predictive model into a comprehensive clinical framework with the aims to (i) identify in-hospital patients likely to benefit from a PC consult, and (ii) intervene on such patients by contacting their care team. MATERIALS AND METHODS: Electronic health record data for 68 349 inpatient encounters in 2017 at a large hospital were used to train a model to predict the need for PC consult. This model was published as a web service, connected to institutional data pipelines, and consumed by a downstream display application monitored by the PC team. For those patients that the PC team deems appropriate, a team member then contacts the patient's corresponding care team. RESULTS: Training performance AUC based on a 20% holdout validation set was 0.90. The most influential variables were previous palliative care, hospital unit, Albumin, Troponin, and metastatic cancer. The model has been successfully integrated into the clinical workflow making real-time predictions on hundreds of patients per day. The model had an "in-production" AUC of 0.91. A clinical trial is currently underway to assess the effect on clinical outcomes. CONCLUSIONS: A machine learning model can effectively predict the need for an inpatient PC consult and has been successfully integrated into practice to refer new patients to PC.