COVID-19: The World’s Crash Course in Data Analysis and Modeling

Published on
August 25th, 2020
Duration
61 minutes


COVID-19: The World’s Crash Course in Data Analysis and Modeling

The Interview ·
Featuring Christina Ramirez and Michael Green

Published on: August 25th, 2020 • Duration: 61 minutes

For much of the population, analysis and projections of COVID-19 cases, mortality, and the associated policy response have brought forward the importance of statistical analysis and modeling techniques that many have never learned and, at best, probably haven't used since long past college years. This has resulted in heated debate, politicization, and huge differences in policy decisions and attitudes across jurisdictions. In this Interview, Christina Ramirez, professor of biostatistics at the UCLA Fielding School of Public Health, sits down with Michael Green of Logica Capital Advisors to help viewers understand the constantly evolving story that the data is telling. Together, they examine the quality of the available data sets, how the consequences of lockdowns have born themselves out in other data like reports of child abuse, and how data indicates that as our knowledge and experience has grown, so too has our ability to combat the pandemic. Viewers can find more info about Christina Ramirez's work and access her updated data sets and charts here: https://github.com/cmramirez/Covid-plots. Filmed on August 19, 2020.

Comments

Transcript

  • NI
    Nate I.
    31 August 2020 @ 04:37
    I remain unconvinced that there is a lot to be learned from the data because the data quality is extremely poor. We lack agreement about extremely basic parameters like how to characterize death with covid versus death from covid. The test kits are of widely varying quality/accuracy, some kits were contaminated, the people administering the tests were often poorly trained and there is debate among the experts about whether or not the PCR test is even a valid diagnostic tool. This seems hopeless to me no matter how sophisticated the mathematics are that we apply to the data. There is no nice way to say it. The data is just a shit show. About all we can say is that failing to protect nursing homes was a bad idea. Well, I'm not a doctor but I knew that much 40 years ago. It was common sense passed on from my parents not to enter a nursing home if I wasn't feeling well because I could end up killing someone. We also knew early on that the lockdowns were a bad idea. WHO estimated 150 million deaths from starvation alone, which was 10x bigger than the worse case fatality estimates from the virus had we remained open. Lockdowns were very selfish thing to do. Mike did a great interview, but the idea that we get anything from this data seems to me like a fool's errand.
  • MB
    Mike B.
    31 August 2020 @ 04:20
    I love Mike Green but this analysis is flawed, maybe tragically so. They are looking at morbidity but what if morbidity is not the major factor here. What if it's disability? We now have a growing group of people who are referred to as the "Long Haulers". These are people who have caught this thing and still have persistent symptoms across a wide spectrum. Reports are surfacing that even people who were deemed a symptomatic have been shown to have developed significant dis-regulation in bodily functions. Is this just an odd occurrence of sequlae or is this just the tip of the iceberg? I submit that we don't know. Only in the last month have studies been set up to look at this problem. The data sets they are talking about are useless in the assessment of damage or it's severity. We simply don't have anything but reports and anecdotal evidence. I hope that we should start to see some info from these studies in September. Till then, I suggest we hold our horses. Do will really want to risk damaging 2 or 3 generations for the long haul simply because we have failed to control this thing by conventional means even though others have? GIGO indeed, Mike.
  • LC
    Lee C.
    25 August 2020 @ 15:07
    Viruses don't have wants and needs. COVID 19 is a 32kb piece of genetic data. Viruses have no arms and no legs and no will. They don't "want" to do anything. The are completely passively integrated into society through the host's behavior and the laws of thermodynamics. Blaming a virus is like blaming a rock. As viruses mutate they respond to the thermodynamics of mixing, entropy and mean reversion. Very virulent viruses kill the host quickly. Less virulent viruses kill the host less quickly and so have a better survival so it is the host's response to the virus's mutation which propagates attenuation. Viruses are controlled by thermodynamic state functions, and as such state functions are independent of path. So all of this hand waving about rates of infection in the end won't stop the infection from proceeding so a state of maximized mixing. If you start in St Louis and drive to Chicago, whether you go 20 mph or 100 mph, take the interstate or take the back roads, your destination is still Chicago. This means all this "science" is just a narrative giving the illusion of control. It may provide marginally fewer deaths, like going 20 instead of 100 may provide marginally fewer deaths, but no cure. Dead people are dead. Once dead Covid has no impact on them. Infected but surviving people are prone to morbidity and that reality is being largely ignored. Post infection morbidity can result in severe long term consequences. It's hard to run a mile or climb stairs if 50% of your lung tissue is scar. Economically the effect will be severe. If you have a junior and senior engineer, both know how to build a bridge only one knows how to build a bridge under budget. so the loss of the wrong one has a much bigger impact than the loss of the other. Near as I can see this is going to last 2.5 years as the virus penetrates humanity and maximal entropy is realized. It's time to stop publishing silly self serving narratives and try to maximize on reality.
    • LS
      Lemony S.
      25 August 2020 @ 18:08
      Yes, the reality being 0.2-0.4% IFR and thus not different than things we have dealt with before. Yet we can stop economic disaster and suffering by loving freedom and life, not fearing death. That fear has hurt and killed far more people than this virus ever will.
    • MD
      Matt D.
      25 August 2020 @ 23:08
      Interesting and original comment.
    • DS
      David S.
      26 August 2020 @ 16:00
      Not sure the thermodynamics I studied in school works here. The existential question is dismissed by just being dead. We could have and should even now handle the economy better. DLS
    • dm
      dude m.
      27 August 2020 @ 00:25
      OMG...reading comments like this is like reading contributions made by DLS. Not applicable and why even bother, unless you are reading your own comments for posterity.
    • CM
      Cory M.
      30 August 2020 @ 23:14
      Michael Pollan (earlier work) makes an interesting argument why plants “use” humans. Virus self-survival makes sense to me in the same way.
  • RY
    Roy Y.
    29 August 2020 @ 21:21
    Thank you Christina & Michael! - such a level headed discussion. Superb.
  • SB
    Scott B.
    29 August 2020 @ 14:41
    Real Vision has to have the most critical viewers; how do you have 28 thumbs down or ~10% of the thumbs up on an interview based on math?! Touch crowd.
  • SB
    Scott B.
    29 August 2020 @ 14:24
    "New York didn't report nursing home COVID deaths if they occurred in the hospital"...really?! In just the last minute of watching we discover the fear-based notion of "excess mortality" is and must be unattributable to covid-19 and that this number excludes suicides and other modes of death. Nursing homes equal 8% of cases and 41% of deaths, but that is likely under-reported because of variations in state laws/regs. Wow! So let's just say half of the deaths are due to nursing homes; we literally shut the world down for ~100k deaths in the US? I love how she mentions the reduced life expectancy when one loses a job...since there are now 30M US citizens in this boat, how does the math equivalate to the deaths caused by covid-19? It would be a long tail of course.
  • wj
    wiktor j.
    29 August 2020 @ 07:58
    If you forget the past you tend to repeat it. We need to find out where this came from as it seems it came from a wuhan lab. The nay sayers still didnt explain about the farin cleavage point. And the aids homology in the gnome. Yet you can recreate this virus in a lab. If this came from a lab we need to hold the responsible to account.
  • TM
    The-First-James M.
    25 August 2020 @ 15:17
    Please please please add this video to your free YouTube channel. It's message needs to get out to the thinking members of the wider general public. I am apalled by how politicised COVID-19 has become given the horrendous damage done to young lives by these lockdowns. I find it incredible that 'only' 49 under 14's make up the US death count of 150,000. True, these deaths are tragic and no parents want to lose their child. However, I suspect the probability of them being hit by a bus is greater.
    • BF
      Bill F.
      25 August 2020 @ 16:33
      i second that recommendation, this is damn near a public service announcement...and,im sure it would make more people become aware of, and want to subscribe to, RV after seeing it.
    • SB
      Stephen B.
      26 August 2020 @ 00:03
      The question is....If RV put this out on its free service (as i encourage them to do so), will it be deleted by YouTube for "violating its terms of service"? I understand that other academics have tried getting out balanced viewpoints and have been frustrated in that way.
    • MC
      Mark C.
      28 August 2020 @ 21:40
      I agree 100%. It would be a public service.
  • MC
    Mark C.
    28 August 2020 @ 21:33
    Thank you so much to both Christina and Professor Plum. I am showing the last 5 minutes to my kids.
  • SB
    Sunil B. | Contributor
    28 August 2020 @ 04:31
    Math is beautiful language. Enjoyed as usual especially when it so off wicket yet relevant.
  • PO
    Paul O.
    26 August 2020 @ 18:11
    So I thought this interview was a great way to step back from the noise and try to have a balanced discussion on a topic where everyone has an opinion. I am not a statistician so I have a question on the graph that shows Europe v. the US absent NY NJ CT and MA. In her answer Professor Ramirez says "I'm like, well, let's not look at the US as a whole because if you think about it, we're like 50 different countries, at least in terms of population, but also just demographics and everything else." and then goes on to say "I didn't define Europe, I used the European Center for Disease Control definition of Europe because they have a column that-- they call it continent" Why would we treat the US as different countries and eliminate the hardest hit ones, but then treat Europe as a monolith? If the goal is to show how much better we are at treating the disease wouldn't you want to show Europe sans Italy and the US minus the tri-state area? As presented that graph seems to be a statistical version of community adjusted EBITDA to make people in the US feel like we are doing better than everyone else.
    • RM
      Robert M.
      26 August 2020 @ 21:32
      Just stated above. Poor use of statistics to make a point. You can't take out negative info from one data set and leave it in the other data set and expect to make a valid comparison.
    • MG
      Michael G. | Contributor
      26 August 2020 @ 23:10
      Appreciate the question. To be clear, the data is available precisely to allow you to do this analysis yourself. There are no "outlier" equivalents for Europe which is why we demonstrated this dynamic. Europe's "slope line" is unbiased by outliers while the US original slope line is distinctly biased by those states that committed SARS-CoV2 patients to nursing homes. Removing that bias presents a much more accurate picture of the dynamics of the US experience. Again, the point is not political blame, but rather to identify the improving dynamics. This (improving outcomes) is also true of Europe, albeit less so.
    • PO
      Paul O.
      27 August 2020 @ 23:55
      Appreciate the answer and thanks for making the data available.
  • DL
    Dan L.
    27 August 2020 @ 21:20
    I found Michael’s leading questions on incentives for potentially “aggressive coding” to be a bit biased and offensive, frankly. We know there are many deaths that were caused by Covid, but which were not recorded as such, because limited virus tests were needed for the living. I think questions regarding the accuracy of coding are best posed in a non-biased manner, especially if one is trying to get to some objective understanding of the data.
  • OS
    Oliver S.
    26 August 2020 @ 21:21
    The stats are utterly goosed. A COVID death doesn't mean that you have died because of COVID. Even with the goosed stats, this is no worse than seasonal flu (and that according to WHO figures). I guess what is real, and you can adjust investment bias, is the Governments' response. Stats are great, but this is a case of s$&t in, s£$t out.
    • LS
      Lemony S.
      27 August 2020 @ 16:23
      Yes, see my post below. The disease is real but not that meaningful now or historically, and the response is one of creeping marxism and anti liberty central powers.
  • SB
    Stewart B.
    27 August 2020 @ 11:12
    One interesting thing I've noticed is that where there has been a Covid19 outbreak in a largish group of unrelated people (eg a cruise ship or factory), not everyone got Covid19. This supports the idea of some people having some type of immunity or partial immunity to Covid19.
  • JG
    Jeff G.
    27 August 2020 @ 03:21
    Medical coding can be quite messy. Subjective and influenced by government rebursement. Great job Michael and Christina! Thanks for your personal story Christina.
  • dm
    dude m.
    27 August 2020 @ 00:34
    Subject matter aside, Michael Green is such an amazing interviewer. And he's so gracious with his guest. This interview seems more of a dance than a static ask-questions-get-answers. And Michael certainly has done his homework. And he really tries to learn the guest story and ads value to the discussion. A superior interview that we've all come to expect and appreciate from RV.
  • RM
    Robert M.
    26 August 2020 @ 21:29
    Taking NYC, NJ out of numbers is apples to oranges comparison to Europe. If you do that, then do the same for European countries like Italy that also got the virus early on and had the same issues as NYC. If you wanted to do that comparison, start with a later date where the initial surge was over in Europe and US and then compare.
    • MG
      Michael G. | Contributor
      26 August 2020 @ 23:12
      Repeat of comment below: Appreciate the question. To be clear, the data is available precisely to allow you to do this analysis yourself. There are no "outlier" equivalents for Europe which is why we demonstrated this dynamic. Europe's "slope line" is unbiased by outliers while the US original slope line is distinctly biased by those states that committed SARS-CoV2 patients to nursing homes. Removing that bias presents a much more accurate picture of the dynamics of the US experience. Again, the point is not political blame, but rather to identify the improving dynamics. This (improving outcomes) is also true of Europe, albeit less so.
    • RM
      Robert M.
      27 August 2020 @ 00:14
      Europe learned the lesson about nursing homes before the US. And they committed the same mistakes. So you have to exclude their data as well. Still disagree with your premise: https://www.marketwatch.com/story/coronavirus-was-devastating-europes-nursing-homes-well-before-the-us-here-are-the-lessons-they-learned-2020-05-07
  • HH
    Heath H.
    26 August 2020 @ 15:44
    I would have liked a lot more statistical content. Covid will pass, but the methods for handling biases, mismatched data, and heteroscedasticity on complex problemswill be useful forever.
    • MG
      Michael G. | Contributor
      26 August 2020 @ 23:22
      Heath, very limited in what we can actually teach on data analysis techniques in this type of forum. Would be interested if others would ACTUALLY be up for a more formal instruction class. Can also suggest some reasonable readings, eg https://www.researchgate.net/publication/308133810_Basic_statistical_tools_in_research_and_data_analysis And one of my favorite books of all time "An Illustrated Guide to Theoretical Ecology" has some used copies for sale on Amazon again... Snap them up while hot https://www.amazon.com/dp/0195085124/ref=cm_sw_em_r_mt_dp_M5UrFbYJBZ4NN
  • LB
    Lukas B.
    26 August 2020 @ 17:49
    I like Mr Green a lot, but he said several times in the interview that he wanted to avoid policy decisions, then posed leading questions about policy. The worst is at 11:40. "In many ways what you are saying is, 'forget what's happened in the past...'" That's not what she said, not even close. That's been his underlying argument at times. It's sloppy.
    • RM
      Robert M.
      26 August 2020 @ 21:40
      Mike was up-to-speed to do this interview, but probably better used elsewhere. Still not sure what the takeaway was from this video. What most people who love to debate the numbers are not realizing, case increases lead to hospitalizations increasing. Hospitalizations increase then lead to death increases. And no matter the disputed cause of death or the rate of death vs the flu, health departments react and shut things down. To me, that is the practical application of Covid case counts. Whether we protect those at high risk or we think young people are low risk, if your community doesn't take the proper masking and social distancing measures, then the cycle restarts where case counts increase, etc and things get shut down again. As someone who produces live events, this cycle keeps people unemployed and will do so until something breaks the case counts to acceptable levels.
  • LS
    Lemony S.
    26 August 2020 @ 19:02
    If you know anything about history and politics (current and historic) you see instantly, if you have a clear mind, that the response to this was and still is born out of political abuse. We have never quarantined healthy people, for any reason, en mass. We have never worried about similar viral diseases (1968), and if we did, they were far more lethal and obviously detrimental than a coronavirus, this or any other. You never had to be a scientist to know this, just had to live life and realize this is all mental and political. You can easily see this now. There is NO EVIDENCE that masks do anything, even N95, at stopping respiratory viruses - let alone faux cloth mask "protection." There is no evidence that lockdowns have done anything. We still are in the whim of the governmental overlords because they have decided that the time has come to institute their plans for the population. These are clearly privacy and economically driven. Conspiracies are real and this one is clear - consolidate the economy. Vaccinate people against their will. They already stole our privacy in so many ways (and in some ways we were willing sheep there too). The realization by the powers that be that the economic system and its debt can't last is obvious. That's WHY. Don't be silly, this is not accidental, just like gain of mutation labs weren't accidental. It is a way to try to get beyond the failures of keynesianism and more, and the cashless society is clearly an objective by demons influencing Summers, Gates, and more names that you don't even know. Wake up, people.
    • OS
      Oliver S.
      26 August 2020 @ 21:18
      Fantastic rant and I totally agree. We need to be clear what a COVID death is. It is a death due to: run over by bus etc, but have COVID; run over by bus etc, and had COVID and recovered; dying with flu like symptoms and not tested for COVID.
  • DG
    Dave G.
    25 August 2020 @ 16:52
    I think we need to not just look at deaths for young people but the lingering health effect that they might have to live with for the rest of their life. It is early in the pandemic so I'm sure data is not that abundant but I wish they would have touch on some of those statistics relevant to people under 50 that survive with lasting effects and what are they.
    • LS
      Lemony S.
      25 August 2020 @ 18:06
      And what are you or is anyone else going to do about some esoteric "lingering effect for the rest of person X life" ... MOAR free? Come on now. Live. Don't fear.
    • mw
      michael w.
      25 August 2020 @ 19:06
      I'd say the lingering health effects will mainly be due to the draconian policies enacted. Depression, anxiety, food insecurity, along with drug and alcohol addictions can follow a person for the rest of their lives
    • DG
      Dave G.
      25 August 2020 @ 23:24
      I for 1 am a survivor of a really bad FLU that I had in 2012. The FLU destroyed my stereocilia in my ears (little hairs) that are part of your bodies balance control. This caused me to have vertigo, bouncing vision and very poor balance. I went thru rehab that lasted 1 year and it is much better but I have lost the input from my ears for balance for life. I now depend on my sight and feet for balance. I don't have the vertigo any more but walking in the dark on rough surfaces is very challenging. I get comments from people thinking I'm drunk cuz I can walk a straight line or I bump into them. Prior to 2012 I never got a flu shot now I get one every year. I feel lucky because I can easily live with these permanent effects. We really don't know the after effects of sever covid 19 yet. Not that you should be scared but you should be aware.
    • DG
      Dave G.
      25 August 2020 @ 23:26
      Can't walk a straight line, lol We need an edit button
    • LS
      Lemony S.
      26 August 2020 @ 16:27
      And I'm glad things worked out as best that they might for Dave, but it proves the point - nothing in life is perfect, and there is nothing he could have done, or others should have done, to stop a single "tragedy" - as if we could stop them all. I'm all for awareness, but usually that term seems to come with taking away other people's liberties. It's not about awareness, it's about holding other people hostage, so that the elites can do what they want, benefit from it, feel more power, etc.
  • GJ
    Gino J.
    25 August 2020 @ 20:04
    As a father of two young ladies, I love seeing powerful content promoting women in science or math - Bravo R.V. & Mike for this interview. While I agree with most of the points in this interview, I can't help to think why there was a need to take out the data from the Northeast states to create a framework that Covid-19 is not a concern to most of the U.S. population? I grant you that nursing homes were not protected in early March-May or testing was not available, but how do you explain the rise of U.S. deaths starting July 4th? Roughly 35,000 people have died in less than 8 weeks. See data at: https://www.cdc.gov/nchs/nvss/vsrr/covid19/index.htm And the fact that 49 kids under the age of 14 have passed away, doesn't mean we know everything there is to know about this virus and it's impact on younger people. Most of the kids have been under their parents protection since March. These kids don't live in nursing homes, go to bars, restaurants, schools, attend social or political rallies, so shouldn't we approach the data with caution? Again, I applaud this engaging interview and hope to see similar content in the future.
    • LS
      Lemony S.
      26 August 2020 @ 16:25
      Why do we need to promote women? Why can't we just have a free society? Promoting women to put things above family hasn't worked, Bruno. Teach your daughters the more important things about life, then you won't complain to them about having no grandchildren. You too, like them, can't have it all.
  • RC
    Reese C.
    26 August 2020 @ 06:52
    Mike talks to this lady as if he has all the answers in his head already.
    • BI
      Brett I.
      26 August 2020 @ 08:01
      It’s like the old lawyer’s adage ‘Never ask a question that you don’t know the answer to’. That said, unless there was a serious amount of post editing, he was familiar with her language. Not sure I’d pronounce heteroskedasticity on my first, second or tenth attempt!
    • SS
      Stan S.
      26 August 2020 @ 14:57
      1. He likely does since he was out early pointing out the policy errors in real time. In my view, he has been completely vindicated in his early criticism of the policy panics. 2. It seemed clear to me that these two have been in touch and he likely knew her views and how to best set them up (interviewing skill).
  • PJ
    Peter J.
    26 August 2020 @ 11:23
    Exceptional interview. Has educated me and altered my opinion bias. Great to get objective information upon which we can attain an informed view. Hope this is released to the public.
  • HC
    HIMANSHU C.
    26 August 2020 @ 10:59
    Amazing interview and kinda proves that if we protect the vulnerable specially in the elderly care then we can reduce the covid fatalities without destroying the global economy and mental health of others.
  • HH
    HODL H.
    26 August 2020 @ 04:08
    Great interview, Go Bruins!
  • JR
    James R.
    26 August 2020 @ 03:30
    Please make this public so that I can share it with my in-laws.
  • WR
    William R.
    26 August 2020 @ 02:56
    Powerful content and data, thankful we have so many who have learned so much in a small window of time.
  • SS
    Steven S.
    26 August 2020 @ 01:19
    Excellent discussion. Prof. Ramirez Kitchen is great. Cogent, clear explanations. She's done some exceptional biostats work on HIV viral load, systemic inflammatory responses to viral infections, and very nice work on application of random forests, certain aspects of Machine Learning approaches and Bayesian stats. Her expertise is much needed to make sense of COVID. I also applaud her call for younger folks to get into the field. Thanks!
  • PB
    PHILLIP B.
    26 August 2020 @ 00:52
    No biggie, but when including links, maybe code them to open a new tab, e.g., "https://github.com/cmramirez/Covid-plots" in the abstract.
  • BG
    Bruno G.
    25 August 2020 @ 20:46
    Lock downs preferentially kill children and we have leaders who continue these policies. It’s maddening that MSM does not educate their audience and allow them to make the best choices for their situation.
    • TM
      The-First-James M.
      25 August 2020 @ 21:43
      The MSM have been part of the problem here, as a group, not the solution...
    • BG
      Bruno G.
      26 August 2020 @ 00:19
      Its not my opinion, it what she explains the data shows.
  • RD
    Riki D.
    25 August 2020 @ 23:18
    Interesting story of a child suffering from secondary complications from COVID19. Part of a larger multi trial program to treat inflammation or the over reaction of the body's immune system when confronted with the virus. https://web.musc.edu/about/news-center/2020/08/25/first-in-nation-treatment-for-mis-c-at-musc-shawn-jenkins-childrens-hospital "As doctors worked to save K.J., they suspected he had MIS-C. Multisystem inflammatory syndrome in children is a rare complication of COVID-19 that shows up after the virus has left the body. It’s been described as an immune reaction gone haywire. A test for COVID-19 antibodies came back positive a few days later, helping to confirm their suspicion." Fantastic outcome for the young man and his family!!
  • ML
    Michele L.
    25 August 2020 @ 21:35
    Does anyone know what Mike Green's wife does? I've always been curious if she's also in finance bc i'm sure it takes a big brain to keep up with him at the dinner table but when he mentioned her interest in COVID data I wondered if she's in medicine or research. Doesn't matter, I've just been wondering for a while now and thought someone might know since googling 'Jen Green' is unlikely to narrow it down. lol.
  • MA
    Melanie A.
    25 August 2020 @ 21:17
    Great discussion (from a fellow Math girl) - thank you. And hopefully more data will give additional insight in to the “cure” so we can protect the nursing homes/vulnerable and let others lead a more normal (even if it’s a new normal) life.
  • JM
    John M.
    25 August 2020 @ 17:53
    I think it should have been obvious that elderly in residences (or elderly in general) were prime targets for the pandemic. The elderly have weakened immune systems. Don't typical seasonal influenza epidemics disproportionately kill elderly (I recognize that young children with immature immune systems are sometimes targets)? Didn't the 1968-69 pandemic primarily killed seniors?
    • DS
      David S.
      25 August 2020 @ 20:30
      The very fact that we in the US did not place a high priority on nursing homes initially shows a lack of planning and leadership on many levels. This is merely hindsight. Sweden had the same problem. We have had pandemics for centuries. Let us get, keep, and implement reasonable strategies from the beginning of the next epidemic/ pandemic. We live in a global world. The more crowded and interactive the populations the easier any disease spreads. The more developed the internet and social media the faster conspiracy theories spread. We know this is the case. We need to be more objective and proactive about it. DLS
  • DS
    David S.
    25 August 2020 @ 20:18
    Congratulations on an excellent presentation for all. It helped me understand better our and my risks of COVID-19. Objective information gives everyone a better chance for problem solution. The family history was particularly heartwarming. A great way to finish. Thanks to both of you for a most important presentation. DLS
  • JH
    Jesse H.
    25 August 2020 @ 20:04
    Well done, Mike! Couldn’t sit through this one, as her responses were a bit painful for me, as an engineer with a solid grounding in stats. It felt like she was crafting her answers a little too carefully and bureaucratically for my liking. Thanks again. Appreciate the effort to get a statistician on.
  • DS
    David S.
    25 August 2020 @ 19:52
    FYI ; COVID-19 patients could see improved oxygen levels without the help of a ventilator, thanks to one recent Michigan State University College of Nursing graduate. Mariah Foster, a May graduate in the Doctor of Nursing Practice program, as part of her final class project, discovered when COVID-19 patients physically turned over on their stomachs, oxygen levels improved. This finding including both intubated and non-intubated patients and whether nurses or patients themselves assisted with turning over, or “proning.” DLS https://nursing.msu.edu/about-us/news/news-articles/how-tummy-time-helping-covid-19-patients
  • DD
    Dmitry D.
    25 August 2020 @ 10:28
    Great! An engaging, informed discussion without any partisan BS, thanks to Christina and Mike. Sadly, as Mike pointed out, very hard to see discussion such as this (and thus a considered and informed decisions) among the political/bureaucrat class.
    • LS
      Lemony S.
      25 August 2020 @ 18:12
      When you don't have moral leaders (D) or people who have ideas that are good but won't stand up for them ultimately (R), you essentially have bad and worse decisions, and eventually disintegration of a society. Ask Jonathan Haidt, secular multicultural societies don't last. One of the many reasons is that the economics that might hold them together for a while also disintegrates, and then man is show to be the group animal that he is, with clear group differences. It is foolishness and ignorance that disables one's ability to see this, along with the indoctrination of modern schooling, and a dash of affluent society to make it easy for people to not pay attention to what's real, or true.
  • RG
    Razmig G.
    25 August 2020 @ 17:57
    "They wouldn't have died if they hadn't sneezed" impressive!
  • BG
    Bryce G.
    25 August 2020 @ 17:18
    Congrats guys, you earned my plus subscription. I was admittedly a little salty about you holding out on us essential members with the September stuff, but this interview of reminded me of the quality of information Realvision provides. Good information is priceless.
  • AC
    Andree C.
    25 August 2020 @ 16:52
    Great content. Too many “um’s”
  • DJ
    Dan J.
    25 August 2020 @ 16:32
    wow a balanced review of the pandemic shocking
  • LK
    Lalith K.
    25 August 2020 @ 14:02
    Several interesting insights into the interrogation and analysis of the data - thank you. The title is a bit deceptive, however, because more time seemed to be spent on biology than statistics. There were also a few issues that seemed to have been skirted over: 1.) The distribution of European data points doesn't look like it lends itself at all well to a linear regression model. 2.) This is may sound controversial, but it is a sincere question: could the lower mortality decreasing over time partly be attributed to increased testing as time has progressed? (This in sort of hinted at around minute 38 but not really discussed.) I have a relative in the London, who displayed classic covid-19 symptoms after contact with what was later a confirmed case. At the time (mid-March), tests were only available for people admitted to hospitals, so she, and many other potential cases, never had the opportunity to be tested during the first peak of the European outbreak. Maybe the same could be argued for the tri-state area, but perhaps not for states where infections peaked during the summer months? 3.) Statistical inference from random surveillance testing (using RT-PCR, not just serology that has the problems discussed in the interview) indicates that confirmed cases only comprise a small proportion of total cases, and this fraction appears to vary between regions. Is this also be a serious limitation of the data? 4.) Obviously there isn't time to mention every salient point, but it was a little disappointing that there was no discussion of the non-linear effects of systems being overwhelmed. This was widely reported in Italy, probably contributed to the decision to move elderly patients out of hospital and into care homes in the UK, and is something Nassim Taleb frequently mentions. 5.) The virologist Vincent Rancaniello of Columbia University and TWiV fame (he would make a good RV guest) often states that selection pressure that drives evolution is primarily dependent on the ability to transmit. Assuming I have understood correctly, his argument goes something like this: given that SARS-CoV-2 has always been able to be spread prior to the onset of symptoms and virus shedding declines by orders of magnitude prior to patients becoming severely ill, there is little reason to expect viral attenuation. This makes intuitive sense to me since it's not as if HIV or several other infectious diseases that are routinely vaccinated against have become less pathogenic over time. Given that SARS-CoV-2 very rarely causes severe disease in young children, the possibility of it becoming a benign illness over the next few decades, like the other cold coronaviruses, does seem plausible if childhood exposure confers long-lasting resistance to serious infection later in life.
  • SC
    Sejong C.
    25 August 2020 @ 13:51
    Wonderful. More data scientists should go into this field. Really appreciate the codes and data on Github.
  • AM
    Alexander M.
    25 August 2020 @ 13:32
    #underlyingdynamicsgang
  • CT
    Chris T.
    25 August 2020 @ 11:59
    The challenge for her data is the counting of deaths as they discussed. UK, Spain and Italy counting those deaths with, not from covid and in the UK where 95% of deaths had pre-conditions and >11k less pneumonia deaths this year (vs 5 yr avg) you really have to question the covid death count. ONS and NHS data provide this in excel format. Given UK's Sage have already admitted they have miscounted hospitalisations the UK's quality of data is very questionable. Further the easy answer of excess deaths, doesn't take into account the impact of hospitals and doctors surgeries being essentially "closed". Given median age globally of covid victims is over age 80 (and above life expectancy), that age group of population likely has a heavy reliance on healthcare, when that is absent to the same extent as prior to lockdown there is likely to be a large impact on health. Model only as good as the numbers you put in to it and may take many years to get the real data. And then you have Sweden.