Updated: 22nd March: Estimating COVID-19 Case Fatality Rates (CFR) and Infection Rate Fatality (IFR)The Infection Rate Fatality (IFR) differs from the CFR in that aims to estimate the fatality rate in all those with infection: the detected disease (cases) and those with an undetected disease (asymptomatic and not tested group). if tested, this group would be counted as infected and at least temporarily be immune.
Our current best assumption, as of the 22nd March, is the IFR is approximate 0.20% (95% CI, 0.17 to 0.25).*
In the elderly, co-morbidities have a significant impact on the CFR: those with ≥ 3 comorbidities are at much higher risk, particularly those with cardiovascular conditions. Modelling the data on the prevalence of comorbidities is essential to understand the CFR and IFR by age.
In those without pre-existing health conditions, and over 70, the data is reassuring that the IFR will likely not be above 1%. The prevalence of comorbidities is highly age-dependent and is higher in socially deprived.
How do we arrive at this estimated IFR figure?
The current COVID outbreak seems to be following previous pandemics in that initial CFRs start high and then trend downward. In Wuhan, for instance, the CFR has gone down from 17% in the initial phase to near 1% in the late stage. Current testing strategies are not capturing everybody. At least 50% of those on the Diamond Princess was. asymptomatic, who usually wouldn’t get a test.
In South Korea, considerable numbers who tested positive were also asymptomatics. Asymptomatic people and mild cases are likely driving the rapid worldwide spread.
Early IFR rates are subject to selection bias as more severe cases are tested – generally those in the hospital settings or those with more severe symptoms. Mortality in children seems to be near zero (unlike flu) which will drive down the IFR significantly.
In Swine flu, the IFR was fivefold less than the lowest estimate in the 1st ten weeks (0.1%)Therefore, to estimate the IFR, we used the estimate from Germany’s current data 22nd March (93 deaths 23129) cases); CFR 0.40% (95% CI, 0.33% to 0.49%) and halved this for the IFR of 0.20% (95% CI, 0.17% to 0.25%) based on the assumption that half the cases go undetected by testing and none of this group dies. Our assumptions, however, do not account for some exceptional cases, as in Italy, where the population is older, smoking rates are higher, comorbidities may be higher, and antibiotic resistance is the highest in Europe, which all can act to increase the CFR and the subsequent IFR.
Given the estimated mortality in over 80’s is higher (CFR near 15%); there is considerable uncertainty over the IFR rates in this group.
It is currently not clear what the excess mortality is in this group.
It is essential to understand whether the elderly are dying with or from the disease (see the Sarah Newy report).
It is also not clear if the presence of other circulating influenza illnesses acts to increase the CFR (testing for co-pathogens is not occurring)https://www.cebm.net/global-covid-19-ca ... ity-rates/Jason OkeMSc, DPhil
SENIOR STATISTICIAN
I have been a statistician at the Nuffield Department of Primary Care Health Sciences since 2007. In the last five years, my research has been mainly on the diagnosis of cancer in primary care, overdiagnosis and overtreatment. I am interested in methods for evaluating monitoring and screening programmes, test evaluation and diagnostic accuracy.
I am the study statistician for the SCAN pathway and IDEAL (Artificial Intelligence and Big Data for Early Lung Cancer Diagnosis). I am a statistical reviewer for Emergency Medicine, Lancet Diabetes and Endocrinology and Colorectal Disease Journal. I have recently taken on a role as editorial board member for Nature Scientific Reports.
I am currently co-supervising two part-time DPhil students; Ms Ranin Soliman who is studying "Time Trends of Childhood Cancer Outcomes and Resource Use in Egypt" and Dr Hendrijks Dijkstra who is studying "Factors predicting cam morphology in athletes".
I am the co-coordinator of two online modules for the MSc in Evidence Based Health Care; Statistical Computing with R and Stata and the Introduction to Statistics for Health Care Researchers and deputy coordinator for the Year 1 & 2 medical statistics teaching of the Oxford pre-clinical (First BM). With Thomas Fanshawe, I run a one-day workshop on Statistical methods for diagnostic accuracy in medical research.
Carl HeneghanBM, BCH, MA, MRCGP, DPhil
PROFESSOR OF EVIDENCE-BASED MEDICINE
Director of CEBM & Programs in EBHC
Editor in Chief, BMJ EBM
NHS Urgent Care GP
NIHR Senior Investigator
Carl Heneghan is a Clinical Epidemiologist with expertise in Evidence-Based Medicine, Synthesising evidence and informing decision for better healthcare.
He is Director of the NIHR SPCR Evidence Synthesis Working Group a collaboration of nine primary care departments He has published 95 Systematic reviews, and authored high impact work including the Tamiflu systematic reviews.
Hei s a frontline urgent care GP.
His work also includes investigating drug and devices, advising governments on regulatory evidence and working extensively with the public and the media.
He has investigated the evidence for sports drinks, IVF 'Add-on' treatments, metal-hips, screening, surgical mesh, medical devices and hormone pregnancy tests. He has worked in the field of diagnosis for over 15 years and is one of the leads for the Preventing Overdiagnosis Conference.
He is a clinical advisor to two UK All Parliamentary Party Group on Surgical Mesh and Hormone Pregnancy Tests, and A founder of the AllTrials campaign twice voted one of the top 100 NHS clinical leaders by the HSJ. In 2018 he was awarded NIHR Senior Investigator status