Background The COVID-19 pandemic has profoundly impacted some of the most vulnerable populations in low-resource settings (LRS) across the globe. These settings tend to have underdeveloped healthcare systems that are exceptionally vulnerable to the strain of an outbreak such as SARS-CoV-2. LRS-based clinicians are in need of effective and contextually appropriate triage and assessment tools that have been purpose-designed and validated to aid in evaluating the severity of potential COVID-19 patients. In the context of the COVID-19 crisis, a low-input severity scoring tool could be a cornerstone of ensuring timely access to appropriate care and justified use of critically limited resources. Machine learning was used on data from a retrospective cohort of Sudanese COVID-19 patients to derive a contextually appropriate mortality scale for COVID-19, the African Federation for Emergency Medicine COVID-19 Mortality Scale (AFEM-CMS) model. This MSc aimed to validate the AFEM-CMS, to assist frontline providers in rapidly predicting severe COVID-19 disease in LRS emergency units (EUs) in Sudan. Methods A retrospective quantitative analysis of data collected on adult patients aged 18 years and older screened as potentially positive for COVID-19 was undertaken to validate the AFEMCMS in the same Sudanese setting from which it was derived. Data for this study were collected retrospectively by non-clinical personnel from four government referral hospitals in Sudan's Khartoum State from 01 September 2020 and 31 January 2021. This study's primary outcome was in-hospital mortality due to SARS-CoV-2 infection. A set of predictor variables was collected for all patients based on the requisite inputs for the AFEMCMS tool. The predictor variables comprise demographic and historical data (age and sex), the number of existing comorbidities a patient has on presentation, and a number of clinical inputs (GCS, systolic blood pressure, respiratory rate, heart rate, and pulse oximetry). The AFEM-CMS was validated using C-index measurements (area under the receiver operator curve (AUROC)) in the validation dataset. All analyses were performed in R (version 4.1.0, © The R Foundation) with the dplyr, finalfit, glmnet, mice, pROC, rmda, and tidyverse packages. 4 Missing datapoints were managed using multiple imputation by chained equations (MICE), which imputed values for predictor variables with less than 33% of data points missing. Ethical approvals for this study were obtained from the University of Cape Town and the Sudanese Ministry of Health. Results In this study, the AFEM-CMS was validated against a 936-patient cohort, all of whom All of these included cases met the WHO definitions for suspected, probable, or confirmed SARSCoV-2 infection. Similar to initial derivation outcomes, the tool was found to have reasonable discriminatory power in identifying those at greatest risk of death from COVID-19: The model including pulse oximetry had a C-statistic of 0.732 (95% CI: 0.687-0.777) and the model excluding pulse oximetry had a C-statistic of 0.696 (0.645-0.747). Conclusions This dissertation establishes what is, to our knowledge, the validation of the first COVID-19 mortality prediction tool intentionally designed for frontline providers in LRS. The validation of the AFEM-CMS highlights the feasibility and potential impact of real-time development of clinical tools to improve patient care, even in times of surge in LRS. This study is just one of hundreds of efforts across all resource levels suggesting that rapid use of machine learning methodologies holds promise in improving responses to pandemics and other emergencies. It is our hope that, in future health crises, LRS-based clinicians and researchers can refer to these techniques to inform contextually and situationally appropriate clinical tools and reduce morbidity and mortality.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:uct/oai:localhost:11427/38134 |
| Date | 19 July 2023 |
| Creators | Omer, Yasein |
| Contributors | Wallis, Lee, Pigoga Jennife |
| Publisher | Faculty of Health Sciences, Division of General Surgery |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Master Thesis, Masters, MSc |
| Format | application/pdf |
Page generated in 0.0017 seconds