Skip to content

How much of public health work “involves not technology but methodicalness and record keeping”?

Palko points us to this interesting point from Josh Marhsall:

I [Marshall] am always struck by, amazed at how much of public health work involves not technology but methodicalness and record keeping. In purely technological terms much of it could have been done 100 years ago or, in outlines at least, 500 years ago. Phones and texts and emails obviously provide a critical new tool, allowing public health officials to remain in regular daily contact with hundreds or thousands of people currently self-quarantined across the country. But at its heart it’s an elementary process: find anyone who has tested positive for the infection, track down everyone with whom they had significant contact and get those people to isolate themselves (usually at home) from everyone else for 14 days.

This seems like a good insight, and perhaps it can be generalized to say that 90% of statistics is measurement, and 90% of what’s left is avoiding and accounting for selection bias. Meanwhile, we spend most of our time talking about the remaining 1%.

But that’s a bit too simplistic on my part. For example, we correct for selection bias and random variation in surveys by using MRP, but for that we need probability models, prior distributions, efficient computation, Stan, etc….

But, to get back to Marshall’s point, the first step is to measure and record.


  1. Bill says:

    Just looking at the excerpt, but I think outsiders would be amazed about how much technical knowledge is involved in accurately collecting and organizing records. As early-career quantitative researcher, I would say that over 90% of my day is spent on data cleaning, validation, and management.

    • David J. Littleboy says:

      This is (part of) the reason “AI” isn’t going to solve all our problems: getting clean data that actually corresponds to the phenomenon of concern is hard. The term “GIGO” is almost as old as computers, but we keep forgetting it.

  2. Yeah, for a non-expert [like myself], I speculated that testing upwards of 70-90 million of the US population back in February, say, would have yielded a more accurate response to COVID-19. Of course it is claimed that we did not have anywhere near the number of tests to use.

    One strategy now is to mass produce portable COVID-19 test kids which can be self administered or administered by a school nurse during school hours and a parent while at home, for example. We could manage the potential for outbreaks. But there is a lag quite often in getting the most streamlined response of a strategy that is welll….streamlined. The brain processing speed sometimes is hampered by biases.

    OK, I want to introduce Dr. Michael Mina, who I would seek to write a blog article for Andrew’s blog. Then we can discuss the pros and cons of such a strategy.

    I would be interested in everyone’s opinion about it. So here goes.

    • Eric says:

      Self-administered or home-administered COVID-19 test kits are an interesting idea, but the devil is in the details. As Bill and Andrew pointed out (above) a lot of effort and thought needs to be put into accurately collecting the data. What would be the outcome of a mis-administered home test?

      The person doing the test might get back a false-negative (because of inadequate sample collection or less-than-ideal sample handling.) They might then think that they’re safe and not be sufficiently vigilant with precautions. If they ultimately become symptomatic, they might either be confused, angry, or dismissive of the test.

      On the other hand, as Andrew pointed out, public health statisticians may be able to “correct” for some errors in testing. This might be able to be handled as yet another source of bias and/or noise. So the overall numbers for the community could be semi-accurate.

      • Clyde Schechter says:

        Well, yes, at the individual level you would prefer the test that is closest to perfectly accurate.

        But at the population level, which is the concern of public health, a somewhat less accurate (lower sensitivity or specificity, or perhaps both) test that can be used on a massive scale will save more lives than a perfect test that cannot be deployed in sufficient numbers. It’s easy enough to set up simple decision analytic models to choose the best test for the population.

        This is an important point that, not just in the case of covid-19, the FDA seems not to get, or to acknowledge only reluctantly in extreme circumstances like a pandemic. They have a long history of rejecting “second best” tests that would actually be better in the real world because they would be more widely accessible.

        • I’ve been told they’re required by law NOT to consider the price of things in their decisions. Not sure if it’s true, but if it is, it’s one of the most serious examples of regulatory capture I’ve ever heard of.

          If you can get a super expensive but just slightly better analysis approved, then no one can ever get their vastly cheaper but just mildly less good thing approved and you make a bank while the world suffers.

          • jim says:

            “they’re required by law NOT to consider the price of things”

            The FDA?

            Of course, that makes sense. It should be up to the manufacturer to determine whether the treatment or whatever can be profitably marketed. But the trade off between test certainty and production volume isn’t necessarily a price effect. Their could be a material availability or manufacturing technology constraint.

            As Clyde describes it, it sounds like the question FDA struggles with is: what is “safe enough” or “good enough quality” for public use? If they always err on the side of higher quality/price, that’s almost always going to mean lower production volume, for whatever reason.

            • No it doesn’t make sense. We’re giving a government enforced monopoly to a company, they should be able to bid on the maximum price they will charge for the product, and the FDA should be able to decide whether the public gets enough benefit from giving up market efficiency at that price. Furthermore competitors should be free to come in and say they will create a thing not quite as good but with a dramatically lower price ceiling, and have that weigh on the FDA’s decision.

              You never want to just give away not only a monopoly on the production of a thing, but also a complete monopoly over any competing product that can’t be shown to be physically better than your product. Imagine if for example we could only buy individually wrapped Harry and David pears because there were regulations keeping people from selling bulk pears because they’re less juicy.

              • jim says:

                I believe Clyde was referring to the general case, not the specific case of COVID.

                However, in the case of COVID, we aren’t providing monopolies as far as I know. We’re funding research for several candidates and there are *many* other candidates that are privately funded and/or funded by other national governments.

            • I believe that video/grand rounds interview with Dr. Michael Mina, I posted to Andrew’s blog, contains an explanation of the FDA’s role in a product like the COVID-19 home saliva kits. Dr. Mina’s explanation was a bit complicated. Apparently such a COVID-19 self-administered home test/kit would not be labeled as a ‘diagnostic’ as are the PCR/antigen tests utilized by the hospitals, physicians, and state health agencies.

              • Joshua says:

                Don’t know what happened to the link I put up. No matter.

                What I found interesting was the distinction he makes between a public health perspective and an individual health perspective. It helped me to btwer understand why there is this illogical disconnect that leads to the lack of a program for mass use oof the quick results testing.

      • Joshua says:

        I recommend the following video out of UCSF for a discussion of mass scale rapid testing. I think it’s insane that we aren’t implementing this. It makes sense at multiple levels (would cost less than stimulus, would help to more quickly normalize the economy, and could effectively create the behaviors that would effectively advance reaching a kind of “herd immunity threshold.)

        Its a good question, how it would mesh with data-collection, but I don’t think movement in that direction should be conditional on how that question is answered.

      • Eric,

        Very good questions. I would strongly recommend viewing Dr. Michael Mina’s Youtube. He does address the prospect of a false-negative. Frequent or daily testing is an answer to acquiring an accurate result. Again, I hope you watch the video.

  3. Lombardi says:

    … depends on what you mean by “public health work” and “public health”.

    Most public-health work indeed involves technology — clean water, sewage & trash disposal, animal/insect control, hygiene facilitation, routine medical care availability, pragmatic medical research etc.

    Deficiencies in above technology result in casualties

    Normal “public health” is government bureaucrats bean counting these casualties and urging corrective actions in the previous technology shortfalls.

  4. I may be oversimplifying. Marshall seems to suggest that the measurement process relating to COVID-19 that has yielded useful information hasn’t required complicated math, despite the availability of sophisticated technological tools and complex algorithms

  5. rm bloom says:

    [For example, Dr. Snow, Cholera and the London wells. Long before the “germ theory” had any substantial support]

Leave a Reply