Signs and symptoms to determine if a patient presenting in primary care or hospital outpatient settings has COVID-19 disease.

摘要: BACKGROUND: The clinical implications of SARS-CoV-2 infection are highly variable. Some people with remain asymptomatic, whilst the can cause mild to moderate COVID-19 and pneumonia in others. This lead some requiring intensive care support and, cases, death, especially older adults. Symptoms such as fever, cough, or loss smell taste, signs oxygen saturation first most readily available diagnostic information. Such information could be used either rule out COVID-19, select patients for further testing. is an update this review, version which published July 2020. OBJECTIVES: To assess accuracy symptoms determine if a person presenting primary hospital outpatient settings, emergency department dedicated clinics, has COVID-19. SEARCH METHODS: For review iteration we undertook electronic searches up 15 2020 Cochrane Study Register University Bern living search database. In addition, checked repositories publications. We did not apply any language restrictions. SELECTION CRITERIA: Studies were eligible they included clinically suspected recruited known cases controls without when settings. hospitalised only recorded on admission at presentation. including who contracted while admitted eligible. minimum sample size studies was 10 participants. All individual combinations. accepted range reference standards. DATA COLLECTION AND ANALYSIS: Pairs authors independently selected all studies, both title abstract stage full-text stage. They resolved disagreements by discussion third author. Two extracted data assessed risk bias using Quality Assessment tool Diagnostic Accuracy (QUADAS-2) checklist. presented sensitivity specificity paired forest plots, receiver operating characteristic space dumbbell plots. estimated summary parameters bivariate random-effects meta-analysis whenever five more available, heterogeneity across deemed acceptable. MAIN RESULTS: identified 44 26,884 participants total. Prevalence varied from 3% 71% median 21%. There three settings (1824 participants), nine testing centres (10,717 12 performed wards (5061 seven (1048 (3173 setting specified participants). clearly distinguish severe so present results disease severities together. Fifteen had high selection because inclusion depended applicable referral protocols, many under study review. may have influenced those features fever cough. Five imaging, suggesting that population. additional unable bias. makes it very difficult judge validity these studies. applicability improved comparison original A greater proportion where majority assessments take place. However, still none children separately, one focused specifically found 84 symptoms. Results variable Most low specificity. Only cough (25 studies) (7 pooled least 50% but specificities low. Cough 67.4% (95% confidence interval (CI) 59.8% 74.1%) 35.0% CI 28.7% 41.9%). Fever 53.8% 71.7%) 53.3% 78.9%). positive likelihood ratio 1.04 0.97 1.11) 1.65 1.41 1.93). Anosmia alone (11 studies), ageusia (6 anosmia sensitivities below over 90%. 28.0% 17.7% 41.3%) 93.4% 88.3% 96.4%). Ageusia 24.8% 12.4% 43.5%) 91.4% 81.3% 96.3%). 41.0% 27.0% 56.6%) 90.5% 81.2% 95.4%). ratios 4.25 3.17 5.71) 4.31 3.00 6.18) respectively, just our arbitrary definition 'red flag', is, 5. 2.88 2.02 4.09). two combinations different symptoms, mostly combining other These above 80%, cost (< 30%). AUTHORS' CONCLUSIONS: appear poor accuracy, although should interpreted context between Based currently data, neither absence nor presence accurate enough useful red flag given their sensitivities, also identify Prospective unselected population examining evaluate syndromic presentation urgently needed. inform subsequent management decisions.