RESUMEN
Background: Early prediction of oxygen therapy in patients with coronavirus disease 2019 (COVID-19) is vital for triage. Several machine-learning prognostic models for COVID-19 are currently available. However, external validation of these models has rarely been performed. Therefore, most reported predictive performance is optimistic and has a high risk of bias. This study aimed to develop and validate a model that predicts oxygen therapy needs in the early stages of COVID-19 using a sizable multicenter dataset. Methods: This multicenter retrospective study included consecutive COVID-19 hospitalized patients confirmed by a reverse transcription chain reaction in 11 medical institutions in Fukui, Japan. We developed and validated seven machine-learning models (e.g., penalized logistic regression model) using routinely collected data (e.g., demographics, simple blood test). The primary outcome was the need for oxygen therapy (≥1 L/min or SpO2 ≤ 94%) during hospitalization. C-statistics, calibration slope, and association measures (e.g., sensitivity) evaluated the performance of the model using the test set (randomly selected 20% of data for internal validation). Among these seven models, the machine-learning model that showed the best performance was re-evaluated using an external dataset. We compared the model performances using the A-DROP criteria (modified version of CURB-65) as a conventional method. Results: Of the 396 patients with COVID-19 for the model development, 102 patients (26%) required oxygen therapy during hospitalization. For internal validation, machine-learning models, except for the k-point nearest neighbor, had a higher discrimination ability than the A-DORP criteria (P < 0.01). The XGboost had the highest c-statistic in the internal validation (0.92 vs. 0.69 in A-DROP criteria; P < 0.001). For the external validation with 728 temporal independent datasets (106 patients [15%] required oxygen therapy), the XG boost model had a higher c-statistic (0.88 vs. 0.69 in A-DROP criteria; P < 0.001). Conclusions: Machine-learning models demonstrated a more significant performance in predicting the need for oxygen therapy in the early stages of COVID-19.
RESUMEN
INTRODUCTION: Recent advances in troponin sensitivity enabled early and accurate judgement of ruling-out myocardial infarction, especially non-ST elevation myocardial infarction (NSTEMI) in emergency departments (EDs) with development of various prediction-rules and high-sensitive-troponin-based strategies (hs-troponin). Reliance on clinical impression, however, is still common, and it remains unknown which of these strategies is superior. Therefore, our objective in this prospective cohort study is to comprehensively validate the diagnostic accuracy of clinical impression-based strategies, prediction-rules and hs-troponin-based strategies for ruling-out NSTEMIs. METHODS AND ANALYSIS: In total, 1500 consecutive adult patients with symptoms suggestive of acute coronary syndrome will be prospectively recruited from five EDs in two tertiary-level, two secondary-level community hospitals and one university hospital in Japan. The study has begun in July 2018, and recruitment period will be about 1 year. A board-certified emergency physician will complete standardised case report forms, and independently perform a clinical impression-based risk estimation of NSTEMI. Index strategies to be compared will include the clinical impression-based strategy; prediction rules and hs-troponin-based strategies for the following types of troponin (Roche Elecsys hs-troponin T; Abbott ARCHITECT hs-troponin I; Siemens ADVIA Centaur hs-troponin I; Siemens ADVIA Centaur sensitive-troponin I). The reference standard will be the composite of type 1 MI and cardiac death within 30 days after admission to the ED. Outcome measures will be negative predictive value, sensitivity and effectiveness, defined as the proportion of patients categorised as low risk for NSTEMI. We will also evaluate inter-rater reliability of the clinical impression-based risk estimation. ETHICS AND DISSEMINATION: The study is approved by the Ethics Committees of the Kyoto University Graduate School and Faculty of Medicine and of the five hospitals where we will recruit patients. We will disseminate the study results through conference presentations and peer-reviewed journals.