A Retrospective Review of COVID-19 Severity Risk Factors to Inform Patient Screening at Highland Hospital
Writing sample excerpt · Christine Mineart, MPH
Co-authored IRB protocol, Alameda Health System, April 2020. Excerpted sections below: Introduction (Background, Significance, Study Aim) and Data Analysis.
Introduction
Background
COVID19 has rapidly spread globally and healthcare providers have had to make decisions with limited data. While predictions suggest the initial epidemic of COVID19 may wane in Summer 2020, it is anticipated there will be future outbreaks or continued spread. Therefore, we will need to learn as much as possible from current cases to inform future decisions. The literature describing clinical outcomes from epidemic centers continues to develop. There have been multiple publications describing the risk factors associated with severe COVID19 outcomes in retrospective studies of patient records in an effort to describe this novel coronavirus. As the literature evolves rapidly and health systems gather demographic and severity information from distinct locales and patient populations, it is imperative that Alameda Health System mobilize to gather information, given the distinct demographic make-up and myriad biological and social vulnerabilities of the patient population served by Alameda Health System.
Significance
In light of the COVID19 pandemic, Highland Hospital, like many hospitals, has recently set up a Rule Out COVID19 (ROC) tent at the Emergency Department entrance. This tent was designed to keep as many potential COVID19 positive (COVID+) patients out of the hospital, decreasing the amount of healthcare exposure and personal protection equipment utilization. It has become clear that an informed method for screening patients at the first point of care is needed when testing and screening capacities are scaled. Currently we are making screening decisions for self-quarantine versus hospital admittance based on evolving clinical acumen and minimal data. As a result, patients who eventually develop severe disease may be missed with current protocols. The data of the proposed study will have utility in making evidence based clinical decisions and policies in anticipated future outbreaks or ongoing community spread.
Study Aim
In attempting to create evidence-based, population specific return precautions and stratify low-risk vs. high-risk patients, we propose a retrospective study to analyze risk factors for clinical decompensation in the AHS patient population. In this retrospective study, we plan to review the charts of patients who test positive for COVID19 within AHS. We anticipate the resultant information will provide a greater understanding of clinical and diagnostic predictors of case severity and disease progression to ultimately provide evidence based guidance for symptom follow-up and treatment of cases based on outcome severity risk for our patient population.
The aim of this investigation is to assess characteristics associated with clinical deterioration in COVID19 cases presented at Highland Hospital from January-June 2020. The anticipated practical application of this study is to propose AHS specific correlates for disease severity which may be used to inform clinical decisions including outpatient screening and admission.
Data Analysis
Outcome measures
A retrospective record review will be done by the study personnel to collect data regarding COVID19 infection during the study period. This will include demographics, comorbidities, and past medical history as well as symptom progression while receiving treatment for COVID19 in AHS inpatient and outpatient settings. Data will include outcomes and factors such as intubation, death, ventilator usage time, length of hospital stay, self-quarantine, and medications administered. Data will be analyzed and reported in a de-identified, aggregate form.
Primary Outcome: Severe Cases, defined as positive for any of the following outcomes: ICU admission, death, hospitalization.
Secondary Outcomes: Oxygen supplement, pneumonia, ventilator usage, intubation, multiple organ failure.
Statistical methods
Descriptive analysis of variables will be analyzed as median or number (%). To test for significant associations across primary and secondary outcomes, a chi-square test will be utilized. To explore risk factors associated with outcomes, univariate and multivariate logistic regression models will be used. Analysis will be done in Excel and R.
