Age-groups tied to different frequent COVID-19 symptoms
Researchers found that different age-groups had different associations with various COVID-19 symptoms, according to a Lancet Digital Health study late last week. The study was evaluating an artificial intelligence (AI) model for early COVID-19 detection (1 to 3 days of symptom onset).
The researchers trained their AI model on 182,991 UK people who self-reported symptoms in the ZOE COVID Symptom Study app from Apr 20 to Oct 15, 2020, and then they tested it on 15,409 UK app participants from Oct 16 to Nov 30, 2020. Of the 18 symptoms included in the study, the researchers say the ones most associated with early COVID-19 infection were, in order from most to least relevant, loss of smell, chest pain, persistent cough, shortness of breath, abdominal pain, blisters on the feet, eye soreness, and unusual muscle pain.
The highest area under the receiver operating characteristic curve (AUC) the model achieved was 0.80, with better sensitivity and specificity for older age-groups. AUC is one measure of the model’s usefulness from 0 to 1.
Participants from 16 to 39 years of age showed higher COVID-19 associations with loss of smell, chest pain, and abdominal pain. Loss of smell, however, had less significance in those aged 60 to 79. For those 80 and above, it was not the most common indicator. Instead, chest pain, diarrhea, and sore throat were. Overall, men were more likely to report shortness of breath, fatigue, chills, and shivers, while women were more likely to report loss of smell, chest pain, and a persistent cough.
“Currently, in the UK, only a few symptoms are used to recommend self-isolation and further testing,” said first author Liane dos Santos Canas, PhD, in a King’s College London press release. “Using a larger number of symptoms and only after a few days of being unwell, using AI, we can better detect COVID-19 positive cases. We hope such a method is used to encourage more people to get tested as early as possible to minimise the risk of spread.”
Jul 29 Lancet Digit Health study
Jul 30 King’s College London press release
COVID vs excess deaths vary by country
The 2020 residual mortality rates (RMRs) were half positive and half negative across 35 countries, suggesting that COVID-19 deaths and excess mortality deaths were not comparable, according to a study today in Epidemiology & Infection.
The researchers looked at all deaths from 2015 to 2020 as well as COVID-attributed deaths in 2020, and they included covariates such as economic health, demographics, and government response stringency index. Taiwan and Canada had the highest positive RMR (ie, fewer excess deaths than reported COVID-19 deaths) and Bulgaria and Poland had the most negative (ie, more excess deaths than reported COVID-19 deaths). The researchers add that these countries had more than double the RMR magnitude than the next closest countries.
Overall, the correlation coefficient between excess mortality and COVID-19 mortality was 0.796 (95% confidence interval [CI], 0.629 to 0.892), 0.867 between excess mortality and RMR (95% CI, 0.750 to 0.931), and 0.388 between COVID-19 mortality and RMR (95% CI, 0.062 to 0.638).
RMR had no significant covariates, which the researchers say may be due to how both excess deaths and COVID-19 deaths were related to them. Plus, they write, their study did not look at disruption to health services or under- and over-reporting of COVID-19 deaths.
“In accordance with this survey, we found a high correlation coefficient between excess mortality and excess non–COVID-19 mortality ie RMR, suggesting that health system disruption, reduced health care demand, and health care resources diverted to COVID-19 patients could be some important determinants of RMR,” add the researchers.
“Further data on cause-specific mortality is required to determine the extent to which residual mortality represents non–COVID-19 deaths and to explain differences between countries. Time series modelling of timing and stringency of lockdown measures on COVID-19 deaths are also necessary to assess the impact these have had in different countries,” they conclude.
Aug 2 Epidemiol Infect study