Mask communication redux

The site PreventEpidemics.org, from Resolve to Save Lives and Vital Strategies has added to its resources a Mask Guidance Playbook Promoting mask use is one of the key interventions governments, communities, businesses, and other organizations can implement to control COVID-19. The Playbook includes communication strategies, monitoring approaches, and more. This is a step in the right direction, relating to my own post about the need for more effective public health messaging, What do you mean communication is the biggest challenge?

Public health PSAs

In my post What do you mean communication is the biggest challenge?, I observed social contexts antithetical to effective communication directed towards serving public health objectives, contexts like distrust of science, proliferation of misinformation like conspiracy theories, and politicization of health policy. I asserted that effective health messaging must establish credibility and foster a receptive environment. I suggested leveraging techniques of public relations and advertising towards those ends in order to achieve desirable outcomes of healthy choices made by informed communities.

Morphing gamblers

A piece of cake statistics isn’t. Not for me. My first class was hard enough. I persevered and did well with it, but I really broke a sweat. Then there was my second class, and the textbook, Grinstead and Snell’s Introduction to Probability, really took the starch out of me. This dense, proof-heavy text, barren of concrete examples (or so it seemed to me) gave me the glicks. Gambling The history of statistics is bound closely to the world of gambling.

Crappy hospitalization data

The crappy hospitalization data is why I prefer to review status by state based on excess deaths. Hospitalization Data Reported by the HHS vs. the States: Jumps, Drops, and Other Unexplained Phenomena. Analysis about how crappy the hospitalization data is and suggestions about why. On the other hand, vital statistics reporting is long established, very complete, and very accurate. If you look at all-cause mortality, you take out of the equation questions of coincidence of morbidity and viral positivity.

Endline link bucket

My daily review (blogs, news, papers, plots) exposed me to a few things that I looked up, and this is a capture of some links. Thomas Basbøll will like this post (analogy between common—indeed, inevitable—mistakes in drawing, and inevitable mistakes in statistical reasoning).. From Andrew Gelman. Examples of summarizing data. The Millennium Villages Project: a retrospective, observational, endline evaluation. Andrew’s example of how to do it. Worth studying for the research approach and statistical analysis.

Rapid test advocacy kit

Rapid testing for SARS-CoV-2 would change the game, but your help is needed to make these tests available. Current tests described as “rapid” still require equipment and the results aren’t as rapid as possible. The sticking point is around FDA requirements about test sensitivity. The most rapid tests aren’t authorized yet because they are less sensitive than clinical tests. However, although clinical-grade tests are necessary to diagnose cases for treatment, disease surveillance using rapid tests can greatly deter community spread even if they are somewhat less sensitive.

Things I heard about herd

A couple months ago, I did a back-of-the-envelope calculation of the human cost of herd immunity to SARS-CoV-2. I estimated the U.S. would not see herd immunity until it reached another 195 million cases and, grimly, the deaths that would accompany them. I believed herd immunity was impossible. Shortly thereafter, I saw an article from Johns Hopkins University that did the calculation the same way I did and came up with a similar estimate of 200 million cases.

Are cases important?

News reports in the U.S. of coronavirus test positivity do not necessarily interpret the metrics consistent with the public health objectives for collection of these data. Other metrics easily reviewed on data reporting websites can better serve the typical news viewer. Why do we track a disease or, in public health terms, conduct surveillance? How do we define a case of a disease, and to what ends? Surveillance can have many objectives, and the objectives can differ by disease.

Hash map instead of ladders

Cool tip from Adrian Olszewski about avoiding the “anti-pattern” of a ladder of if then else statements. Posting here to remind me to try that out when I have a chance.