Hold onto your hats. Get blown away with the first big seroprevalance report are coming out. And the fatality rate just got bottomed out to influenza levels.https://www.medrxiv.org/content/10.1101/2020.04.14.20062463v1 …
Deze collectie tonen
-
-
Here's what the seroprevalence report means! - The population of Santa Clara county, CA is 1.9M (the heart of Silicon Valley and home to Stanford - Currently, only 1700 people have been diagnosed with
#COVID19 and 66 people have died - That's a crude fatality rate of 3.6% BUT NOWpic.twitter.com/tmqmsGS3vK
18 antwoorden 179 retweets 367 vind-ik-leuksDeze collectie tonen -
Usually tests are limited to people who show symptoms, people who are sick. The seroprevelance tests people randomly to test for possible antibodies: "the population prevalence of COVID-19 in Santa Clara ranged from 2.49% (95CI 1.80-3.17%) to 4.16% (2.58-5.70%)." 2/
8 antwoorden 120 retweets 315 vind-ik-leuksDeze collectie tonen -
This means that the ACTUAL population of people who have been infected with
#COVID19: "estimates represent a range between 48,000 and 81,000 people infected in Santa Clara County by early April, 50-85-fold more than the number of confirmed cases." 3/11 antwoorden 156 retweets 384 vind-ik-leuksDeze collectie tonen -
Als antwoord op @justin_hart
Can you please comment on this study?!
@mlipsitch@K_G_Andersen@MackayIM@BogochIsaac@trvrb@CarlosdelRio7@nataliexdean@SaadOmer3@AdamJKucharski@kakape#COVID191 antwoord 0 retweets 3 vind-ik-leuks -
Als antwoord op @ABsteward @justin_hart en
Assumed test specificity of 100% feels high. A lower specificity alters the adjusted results, dropping seropositivity to closer to 1% (noted by authors in their conclusion). Also, there’s always potential for selection bias (people consent b/c they think they had COVID).
8 antwoorden 7 retweets 87 vind-ik-leuks -
Als antwoord op @nataliexdean @ABsteward en
Natalie E. Dean, PhD heeft geretweet Natalie E. Dean, PhD
I went ahead and made a thread about it, since others seem to have the same questions. These are the main conclusions.https://twitter.com/nataliexdean/status/1251309227760467974?s=20 …
Natalie E. Dean, PhD heeft toegevoegd,
Natalie E. Dean, PhDGeverifieerd account @nataliexdeanSo, the major reasons why I remain skeptical: - Unstable population weighting - Wide bounds after adjusting for clustering - Is test specificity really that high? - Unavoidable potential for consent bias - Is this consistent with other emerging serosurvey data? Fin 10/10Deze collectie tonen3 antwoorden 18 retweets 65 vind-ik-leuks -
Als antwoord op @nataliexdean @ABsteward en
Great thread. IMO the study is a complete failure of Bayesian logic: (1) The CI of their measurements fall entirely within the CI of their specificity; (2) Their conversion to incidence is not appropriate; (3) The only thing they can really conclude is that prevalence is 'low'.
6 antwoorden 15 retweets 90 vind-ik-leuks -
Als antwoord op @K_G_Andersen @nataliexdean en
Folks. Many of the critiques and challenges of the study are acknowledged by the authors in the paper. They account for numerous biases in their analysis. This is one set of data points. Soon: - 10K MLB employees - LA county - Dallas Fortworth area Trends. Hold tight.
1 antwoord 0 retweets 3 vind-ik-leuks -
Als antwoord op @justin_hart @K_G_Andersen en
Also - these are no slouches. Did you read the list of authors signed onto this? Prof. ioannadis basically fathered the modern movement of this domain.
1 antwoord 0 retweets 2 vind-ik-leuks
Als antwoord op @justin_hart @nataliexdean en
Hence the mistakes they make are even more shocking. From experimental design to statistical analysis there is no question this is a flawed study.
07:23 - 18 apr. 2020
0 antwoorden
1 retweet
3 vind-ik-leuks
Het laden lijkt wat langer te duren.
Twitter is mogelijk overbelast of ondervindt een tijdelijke onderbreking. Probeer het opnieuw of bekijk de Twitter-status voor meer informatie.