Developing a COVID-19 WHO Clinical Progression Scale inpatient database from electronic health record data.

OBJECTIVE: There is a need for a systematic method to implement the World Health Organization's Clinical Progression Scale (WHO-CPS), an ordinal clinical severity score for coronavirus disease 2019 patients, to electronic health record (EHR) data. We discuss our process of developing guiding principles mapping EHR data to WHO-CPS scores across multiple institutions. MATERIALS AND METHODS: Using WHO-CPS as a guideline, we developed the technical blueprint to map EHR data to ordinal clinical severity scores. We applied our approach to data from 2 medical centers. RESULTS: Our method was able to classify clinical severity for 100% of patient days for 2756 patient encounters across 2 institutions. DISCUSSION: Implementing new clinical scales can be challenging; strong understanding of health system data architecture was integral to meet the clinical intentions of the WHO-CPS. CONCLUSION: We describe a detailed blueprint for how to apply the WHO-CPS scale to patient data from the EHR.

Early identification of patients admitted to hospital for covid-19 at risk of clinical deterioration: model development and multisite external validation study.

OBJECTIVE: To create and validate a simple and transferable machine learning model from electronic health record data to accurately predict clinical deterioration in patients with covid-19 across institutions, through use of a novel paradigm for model development and code sharing. DESIGN: Retrospective cohort study. SETTING: One US hospital during 2015-21 was used for model training and internal validation. External validation was conducted on patients admitted to hospital with covid-19 at 12 other US medical centers during 2020-21. PARTICIPANTS: 33 119 adults (≥18 years) admitted to hospital with respiratory distress or covid-19. MAIN OUTCOME MEASURES: An ensemble of linear models was trained on the development cohort to predict a composite outcome of clinical deterioration within the first five days of hospital admission, defined as in-hospital mortality or any of three treatments indicating severe illness: mechanical ventilation, heated high flow nasal cannula, or intravenous vasopressors. The model was based on nine clinical and personal characteristic variables selected from 2686 variables available in the electronic health record. Internal and external validation performance was measured using the area under the receiver operating characteristic curve (AUROC) and the expected calibration error-the difference between predicted risk and actual risk. Potential bed day savings were estimated by calculating how many bed days hospitals could save per patient if low risk patients identified by the model were discharged early. RESULTS: 9291 covid-19 related hospital admissions at 13 medical centers were used for model validation, of which 1510 (16.3%) were related to the primary outcome. When the model was applied to the internal validation cohort, it achieved an AUROC of 0.80 (95% confidence interval 0.77 to 0.84) and an expected calibration error of 0.01 (95% confidence interval 0.00 to 0.02). Performance was consistent when validated in the 12 external medical centers (AUROC range 0.77-0.84), across subgroups of sex, age, race, and ethnicity (AUROC range 0.78-0.84), and across quarters (AUROC range 0.73-0.83). Using the model to triage low risk patients could potentially save up to 7.8 bed days per patient resulting from early discharge. CONCLUSION: A model to predict clinical deterioration was developed rapidly in response to the covid-19 pandemic at a single hospital, was applied externally without the sharing of data, and performed well across multiple medical centers, patient subgroups, and time periods, showing its potential as a tool for use in optimizing healthcare resources.

Characterizing styles of clinical note production and relationship to clinical work hours among first-year residents.

OBJECTIVE: To characterize variation in clinical documentation production patterns, how this variation relates to individual resident behavior preferences, and how these choices relate to work hours. MATERIALS AND METHODS: We used unsupervised machine learning with clinical note metadata for 1265 progress notes written for 279 patient encounters by 50 first-year residents on the Hospital Medicine service in 2018 to uncover distinct note-level and user-level production patterns. We examined average and 95% confidence intervals of median user daily work hours measured from audit log data for each user-level production pattern. RESULTS: Our analysis revealed 10 distinct note-level and 5 distinct user-level production patterns (user styles). Note production patterns varied in when writing occurred and in how dispersed writing was through the day. User styles varied in which note production pattern(s) dominated. We observed suggestive trends in work hours for different user styles: residents who preferred producing notes in dispersed sessions had higher median daily hours worked while residents who preferred producing notes in the morning or in a single uninterrupted session had lower median daily hours worked. DISCUSSION: These relationships suggest that note writing behaviors should be further investigated to understand what practices could be targeted to reduce documentation burden and derivative outcomes such as resident work hour violations. CONCLUSION: Clinical note documentation is a time-consuming activity for physicians; we identify substantial variation in how first-year residents choose to do this work and suggestive trends between user preferences and work hours.

Assessing the effects of pharmacological agents on respiratory dynamics using time-series modeling.

Developing quantitative descriptions of how stimulant and depressant drugs affect the respiratory system is an important focus in medical research. Respiratory variables-respiratory rate, tidal volume, and end tidal carbon dioxide-have prominent temporal dynamics that make it inappropriate to use standard hypothesis-testing methods that assume independent observations to assess the effects of these pharmacological agents. We present a polynomial signal plus autoregressive noise model for analysis of continuously recorded respiratory variables. We use a cyclic descent algorithm to maximize the conditional log likelihood of the parameters and the corrected Akaike's information criterion to choose simultaneously the orders of the polynomial and the autoregressive models. In an analysis of respiratory rates recorded from anesthetized rats before and after administration of the respiratory stimulant methylphenidate, we use the model to construct within-animal z-tests of the drug effect that take account of the time-varying nature of the mean respiratory rate and the serial dependence in rate measurements. We correct for the effect of model lack-of-fit on our inferences by also computing bootstrap confidence intervals for the average difference in respiratory rate pre- and postmethylphenidate treatment. Our time-series modeling quantifies within each animal the substantial increase in mean respiratory rate and respiratory dynamics following methylphenidate administration. This paradigm can be readily adapted to analyze the dynamics of other respiratory variables before and after pharmacologic treatments.

Correcting for serial dependence in studies of respiratory dynamics.

Understanding the physiological impact of drug treatments on patients is important in assessing their performance and determining possible side effects. While this effect might be best determined in individual subjects, conventional methods assess treatment performance by averaging a physiological measure of interest before and after drug administration for n subjects. Summarizing large numbers of time-series observations in two means for each subject in this way results in significant information loss. Treatment effect can instead be analyzed in individual subjects. Because serial dependence of observations from the same animal must then be considered, methods that assume independence of observations, such as the t-test and z-test, cannot be used. We address this issue in the case of respiratory data collected from anesthetized rats that were injected with a dopamine agonist. In order to accurately assess treatment effect in time-series data, we begin by formulating a method of conditional likelihood maximization to estimate the parameters of a first-order autoregressive (AR) process. We show that treatment effect of a dopamine agonist can be determined while incorporating serial effect into the analysis. In addition, while maximum likelihood estimators of a large sample with independent observations may converge to an asymptotically normal distribution, this result of large sample theory may not hold when observations are serially dependent. In this case, a parametric bootstrap comparison can be used to approximate an appropriate measure of uncertainty.