Machine learning algorithms for predicting determinants of COVID-19 mortality in South Africa

Background The aim of this study is to evaluate the impact of diabetes, hypertension, cardiovascular disease and the use of angiotensin converting enzyme inhibitors/angiotensin II receptor blockers (ACEI/ARB) with severity (invasive mechanical ventilation or intensive care unit admission or O2 saturation < 90%) and mortality of COVID-19 cases. Methods Systematic review of the PubMed, Cochrane Library and SciELO databases was performed to identify relevant articles published from December 2019 to 6th May 2020. Forty articles were included involving 18.012 COVID-19 patients. Results The random-effect meta-analysis showed that diabetes mellitus and hypertension were moderately associated respectively with severity and mortality for COVID-19: Diabetes [OR 2.35 95% CI 1.80–3.06 and OR 2.50 95% CI 1.74–3.59] Hypertension: [OR 2.98 95% CI 2.37–3.75 and OR 2.88 (2.22–3.74)]. Cardiovascular disease was strongly associated with both severity and mortality, respectively [OR 4.02 (2.76–5.86) and OR 6.34 (3.71–10.84)]. On the contrary, the use of ACEI/ARB, was not associate with severity of COVID-19. Conclusion In conclusion, diabetes, hypertension and especially cardiovascular disease, are important risk factors for severity and mortality in COVID-19 infected people and are targets that must be intensively addressed in the management of this infection.

Many countries have seen a two-wave pattern in reported cases of coronavirus disease-19 during the 2020 pandemic, with a first wave during spring followed by the current second wave in late summer and autumn. Empirical data show that the characteristics of the effects of the virus do vary between the two periods. Differences in age range and severity of the disease have been reported, although the comparative characteristics of the two waves still remain largely unknown. Those characteristics are compared in this study using data from two equal periods of 3 and a half months. The first period, between 15th March and 30th June, corresponding to the entire first wave, and the second, between 1st July and 15th October, corresponding to part of the second wave, still present at the time of writing this article. Two hundred and four patients were hospitalized during the first period, and 264 during the second period. Patients in the second wave were younger and the duration of hospitalization and case fatality rate were lower than those in the first wave. In the second wave, there were more children, and pregnant and post-partum women. The most frequent signs and symptoms in both waves were fever, dyspnea, pneumonia, and cough, and the most relevant comorbidities were cardiovascular diseases, type 2 diabetes mellitus, and chronic neurological diseases. Patients from the second wave more frequently presented renal and gastrointestinal symptoms, were more often treated with non-invasive mechanical ventilation and corticoids, and less often with invasive mechanical ventilation, conventional oxygen therapy and anticoagulants. Several differences in mortality risk factors were also observed. These results might help to understand the characteristics of the second wave and the behaviour and danger of SARS-CoV-2 in the Mediterranean area and in Western Europe. Further studies are needed to confirm our findings.