Axes of Prognosis: Identifying Subtypes of COVID-19 Outcomes
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (ScreenIT)
Abstract
COVID-19 is a disease with vast impact, yet much remains unclear about patient outcomes. Most approaches to risk prediction of COVID-19 focus on binary or tertiary severity outcomes, despite the heterogeneity of the disease. In this work, we identify heterogeneous subtypes of COVID-19 outcomes by considering ‘axes’ of prognosis. We propose two innovative clustering approaches − ‘Layered Axes’ and ‘Prognosis Space’ – to apply on patients’ outcome data. We then show how these clusters can help predict a patient’s deterioration pathway on their hospital admission, using random forest classification. We illustrate this methodology on a cohort from Wuhan in early 2020. We discover interesting subgroups of poor prognosis, particularly within respiratory patients, and predict respiratory subgroup membership with high accuracy. This work could assist clinicians in identifying appropriate treatments at patients’ hospital admission. Moreover, our method could be used to explore subtypes of ‘long COVID’ and other diseases with heterogeneous outcomes.
Article activity feed
-
SciScore for 10.1101/2021.03.16.21253371: (What is this?)
Please note, not all rigor criteria are appropriate for all manuscripts.
Table 1: Rigor
Institutional Review Board Statement not detected. Randomization Our imputation technique was to randomly choose a value within the normal range for each test result, in order to reduce biasing the results. Blinding not detected. Power Analysis not detected. Sex as a biological variable not detected. Table 2: Resources
No key resources detected.
Results from OddPub: Thank you for sharing your code and data.
Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:Limitations: We highlight the fact that the large majority of our cohort did not suffer from any significant side-effects from COVID-19. In fact, the clusters contain over 2300 patients who can be …
SciScore for 10.1101/2021.03.16.21253371: (What is this?)
Please note, not all rigor criteria are appropriate for all manuscripts.
Table 1: Rigor
Institutional Review Board Statement not detected. Randomization Our imputation technique was to randomly choose a value within the normal range for each test result, in order to reduce biasing the results. Blinding not detected. Power Analysis not detected. Sex as a biological variable not detected. Table 2: Resources
No key resources detected.
Results from OddPub: Thank you for sharing your code and data.
Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:Limitations: We highlight the fact that the large majority of our cohort did not suffer from any significant side-effects from COVID-19. In fact, the clusters contain over 2300 patients who can be deemed to be ‘non-severe’ (clusters 1 and 3 in baseline K-Modes). This leaves only around 500 patients from whom we can derive more ‘interesting’ clusters and build classification models, leading to classifiers built from very small samples. Upsampling the smaller classes when training our models cannot capture diversity in test samples, so does not appear to be a good solution here. Therefore, our approach, especially when using the clusters found across multiple axes, must be tested on larger datasets. Additionally, although we have extrapolated meaning from the clusters found, we cannot truly know which clustering is ‘best’, or even, ‘good’ since they are found in an unsupervised setting. If clustering is not optimal, our classification will likely also be worse - but this is a hard problem to overcome! A larger and more diverse dataset may also help with our confidence in clustering ability and predictions made. Future Work: This work provides a demonstration of our methodology for exploring heterogeneous disease prognosis. Therefore, potential future work is vast. Predominantly, these techniques need to be tested on a larger dataset, particularly with more patients with severe outcomes. This will likely improve the accuracy of clustering on severe patients, and the ability of c...
Results from TrialIdentifier: No clinical trial numbers were referenced.
Results from Barzooka: We did not find any issues relating to the usage of bar graphs.
Results from JetFighter: We did not find any issues relating to colormaps.
Results from rtransparent:- Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
- No funding statement was detected.
- No protocol registration statement was detected.
-
