ASA Connect

 View Only

A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data

  • 1.  A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data

    Posted 03-23-2020 09:24
    I'd like to know your professional reactions John Ioannidis' article last week.  see link below:


    https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-as-the-coronavirus-pandemic-takes-hold-we-are-making-decisions-without-reliable-data/

    ------------------------------
    David Stokar
    [Principal Statistician]
    ------------------------------


  • 2.  RE: A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data

    Posted 03-24-2020 07:00
    ​It is always important to carefully reflect and Ensure that various, potentially conflicting viewpoints are confronted. That's essential to science. In that sense, there is a lot to say for slow science. That last thing is true in normal times, peace times.

    This is not peace time! We are cutting corners on medicinal development, on vaccine development, etc. Policy decisions are taken that impact states, countries, regions, continents, on partial and variable data. Fortunately, there is a huge body of knowledge regarding the mathematical and statistical dynamics of properties. There is knowledge about SARS, MERS, and, further back in time but extremely relevant, the Spanish flu. Study the Philadelphia versus Saint Louis versus Denver response, for example.

    The only defense we have is our brainpower right now, knowledge and creativity.We should all roll up our sleeves, study, think, look at matters from different angles. There are examples of statisticians in marketing bureaus that are pushing their regular work to the side and start working on the pandemic. The more, the better, as long as it can be organized properly, of course.

    Yes, corners are cut. But if Wald would not have done that in world war II circumstancs, we would not have had sequential trials. If classical clinical trials procedures were followed in the earlier days of the AIDS epidemic, we might not have had cocktail therapy!

    The patients with 80% of their lungs incapacitated are not interested whether the malaria antivarial they are getting has been tested for its effectiveness in corona yes or no - they will be ready to try it!

    The dynamics of this epidemic are well known. After a slow start, like a diesel train, it becomes high speed rail!

    When New York City will be facing North-Italian situations, we will all be happy that people will have been working relentlessly to implement social distancing measures, to work hard towards antivirals, vaccins, and specific/sensitive diagnostics.

    These are war-like times. When they are over, we can read and argue in detail about Ioannidis' article. Being cautious and avoid decisions emotionally taken is a good thing. But other than that, let's get back to work, there is plenty to do.



    ------------------------------
    Geert Molenberghs
    Professor
    Universiteit Hasselt & Katholieke Universiteit Leuven
    ------------------------------



  • 3.  RE: A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data

    Posted 03-24-2020 07:38
    To put it mildly, I am disappointed by Professor Ioannidis article. Of course we are making decisions without reliable data.  We don't have a choice.  Even the highest quality sources of data, such as the Diamond Princess cruise have confidence intervals for important disease parameters that span an order of magnitude.  For instance, efforts to project the infection fatality ratio from the Diamond Princess onto the demographics in China yield IFRs with a 95% confidence interval spanning 0.2%-1.2%.  The implications of being on the lower end are quite different from that of being on the upper end.  We need to do whatever possible to buy time to gather more and better data.

    However, it's quite clear, that taking extreme measures to lower the rate of infection is appropriate.  The hospitalization rate is several times that of the fatality rate, and many more deaths should be expected if hospitals are filled to capacity.  The parameter to consider in evaluating the impact of social distancing and economic disruption, isn't a noisy, and likely biased estimate of the fatality rate from historical data, but the counterfactual: what would be expected absent advanced supportive care.  The lower end of this interval implies a tragedy of immense magnitude.

    ------------------------------
    Andrew McDavid
    Biostatistics and Computational Biology
    University of Rochester
    ------------------------------