26 April 2020 Blog Post: COVID-19 Update On Generalizability
I had been blissfully unaware of the “news conference” held by Dr. Dan Erickson of Accelerated Health Care in Bakersfield until a colleague texted me a link to a YouTube video. It has now generated more than 6.5 millions views. I had just finished reading Bill Gates’ 13 page blog post entitled “Pandemic I – The First Modern Pandemic” when the text arrived.
Bill Gates’ paper is thoughtful and composed in its messaging, reflecting the intellectual fortitude of a generational genius who has now devoted his life to understanding global public health challenges. Dr. Erickson’s monologue is riddled with sampling bias, factual errors and sweeping generalizations from somebody with no such acumen or experience. The contrast between the two could not be more stark.
Dr. Erickson begins his press conference with the basic facts of his investigation, which are useful to know and laudable insofar as he took such initiative particularly early in the pandemic. HIs company, Accelerated Urgent Care performed 5,213 COVID-19 tests at its five Bakersfield locations, and, of these, 340 were positive (6.5%). This effort is impressive and far outstrips the pound-for-pound efforts of the UC Health System. UC Irvine for instance has only done 1.987 tests to date.
However, it isn’t long before his conference goes completely off the statistical rails. Dr. Erickson claims that if that percentage of positive cases seen in their Urgent Care were assumed to represent the entire population of Kern County, it would mean about 58,000 people in Kern have had the virus, far more than the nearly 700 official total. But it is impossible that the patients seen at 5 Urgent Care sites represent the underlying population. Who goes to Urgent Care? People that feel unwell. Who doesn’t go to Urgent Care? Everybody else. In fact, outpatient visits to healthcare providers have declined nearly 60% since mid-March and have stayed at that level through April (Dr. Erickson references himself references a ‘dramatic decrease’ in visits to his own facilities). This is akin to estimating the average height of an American adult male by surveying NBA players.
But wait! There’s more. Dr. Erickson then extrapolates his findings to the entire State of California. He overlooks the fact that Kern County comprises only 2.25% of the State’s population, is significantly younger (average 31.5 years versus 36.3 years for California as a whole) and more widely spread geographically (109 people per square mile – Los Angeles County is 2488 people per square mile) on average.
He then doubles down (literally doubles down) on his estimate of actual infection rates stating that the real prevalence of COVID-19 infections in California is actually 12 percent of the population. From where did he get this number? His own data suggested 6.5%. Who knows – and it doesn’t really matter. His misadventures in mathematics aside, he ceremoniously arrives at a COVID-19 mortality rate of 0.03% or 3 deaths per 10,000 infections.
𝗦𝗶𝗴𝗻 𝗨𝗽 𝗳𝗼𝗿 𝗢𝘂𝗿 𝗡𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿
Dig deeper into the health topics you care about most by signing up for our newsletter.
by submitting this form you indicate you have
of Use. Please contact us to for us for more