Note to self: consult experts, Dworkin case studies and low-COVID19 countries’ playbooks to control a respiratory epidemic

tl;dr: To control a respiratory epidemic, especially if I’m ever in a public health agency, or if a new epidemic is emerging and I have the urge to write [a widely read Medium post](https://medium.com/@tomaspueyo/coronavirus-the-hammer-and-the-dance-be9337092b56), in addition to consulting experienced experts, I would consult Chapters 18, 20 and 8 of Mark Dworkin’s  _Outbreak Investigations Around the World: Case Studies in Infectious Disease Field Epidemiology_ plus the playbooks of countries that have had low cases of COVID-19.

There are many options for controlling a respiratory epidemic. Treatments or vaccines are options when available for the disease, and antimicrobial treatments may also offer the option of antimicrobial prophylaxis. There are additionally the many, _many_ versions of “separating people” and other “non-pharmaceutical” interventions (NPIs), including variations of social distancing, with which many of us have become familiar during COVID19[1]; if controlling a pandemic, I would brainstorm and source many NPIs and review the footnoted ones.

Knowing the various intervention options by name, however, would not be enough, as so clearly illustrated by Mark Dworkin’s _Outbreak Investigations Around the World: Case Studies in Infectious Disease Field Epidemiology_. The book is twenty chapters; each chapter is the story of a particular nonfictional outbreak, written by a public health professional who responded to that outbreak at the time. The chapters I read on respiratory epidemics—18 (“Whipping Whooping Cough in Rock Island County, Illinois”), 20 (“A Mumps Epidemic, Iowa, 2006”) and 8 (“Measles Among Religiously Exempt Persons”)—show that planning in advance of the epidemic (e.g. having relationships with various media and political entities; having flexible stockpiles, manufacturing capacity and personnel for different interventions) can be critical. Moreover, they illustrate that each intervention, along with decisions of whether and how to enforce it, and among whom to intervene, has considerations around complexity, cost, timing and payoff. These factors should affect decisions about scheduling of interventions, including which interventions implement first and to what extent to implement them (my guess is that they should all be implemented as far as they can!). This is why I would refer to these chapters as part of my effort to control a respiratory epidemic. Although I have not had time to find them yet, I would also refer to the playbooks of the countries with low COVID19 incidence.

If you want a sneak peek of the chapters, here are some of the tidbits on complexity, cost and payoff of an intervention that I found most interesting, with direct quotes from Dworkin:

  1. By complexity, I refer to the number of moving parts that must be accounted for.
    1. Anyone implementing an intervention that relies upon diagnostic tests must additionally learn about the sensitivity/specificity of the test, the supply chain of appropriate materials for the test, the time to develop the test, etc., and account for these in implementing that intervention (as illustrated by COVID19). They must think about advance communication to diagnostic labs that their workload for a particular test is about to increase 10x or 100x.
    2. Patricia Quinlisk writes about factors in the vaccine intervention in Chapter 20’s Iowan mumps epidemic: “there were issues of vaccine supply, payment of vaccine purchased, distribution of the vaccine to each county, determination of the amount needed by each county, getting the word out about the program, and efforts to have clinics in the midst of the highest risk groups (such as on college campuses). A special committee meeting of legislators with the Director of the Iowa Department of Public Health was needed to release the emergency funds required to buy the vaccine. (For the best uptake of vaccine, it had to be offered free of charge; thus, public funding for the vaccine had to obtained.) This whole process took about 10 days from the decision to do a statewide mass vaccination program, purchase the vaccine, have it shipped to Iowa, distribute it to the local health departments, and have vaccination clinics. In the end, a total of 37,500 doses of MMR vaccine were purchased and made available. This was no small task… We found it difficult to get college students to go and get vaccinated, and in the end only about 10,000 doses of vaccine were used in Iowa’s 99 counties.” Another factor is that some people may have already been vaccinated; one would have to consider whether it’s worth it to review people’s vaccination records to determine who does not need to be vaccinated again.
  2. By cost, I refer to the amount of work that must be done, as well as monetary cost.
    1. A stay-at-home order is simple to describe and is not costly if not enforced, but can be costly if enforced (i.e. consider the number of police required to keep everyone in their homes!).
    2. Educating the public and particular communities (e.g. healthcare workers) on preventive action seems to be good bang-per-buck to me. There are many channels through which to educate, as in a PR campaign: keeping the public health agency website up to date, getting institutions like colleges to educate their constituents, being in the media in all its channels, etc.
    3. The workload for contact tracing, which is instrumental to many interventions, can increase exponentially, exhausting the personnel of a public health department.
    4. Case finding and confirmation, also instrumental to many interventions and often hand-in-hand with contact tracing, can require diagnosing individuals over the phone one-by-one, which scales linearly with the number of cases. This and contact tracing are instrumental to “respiratory precautions to prevent droplet spread from cases, treatment of cases, and prophylactic treatment of close contacts of the cases.”
    5. Any vaccination, treatment or prophylaxis campaign can require convincing a subject to undergo the medicine and then actually administering it. This can scale linearly with the number of people targeted.
  3. By payoff, I refer to the amount of benefit—the amount of transmission reduced—minus any costs to people’s livelihoods or basic needs.
    1. An unenforced ban/closure has lower cost but potentially lower payoff if people do not comply.
    2. Patricia Quinlisk writes about the decision to not broadly quarantine in Chapter 20’s Iowan mumps epidemic: “Quarantine for the public was not used for several reasons, unlike during a measles outbreak in Iowa 2 years earlier. Those reasons included (1) 20% to 30% of cases are asymptomatic, (2) mumps virus can be spread up to 3 days prior to symptoms, (3) some mumps cases are so mildly ill that they do not seek medical attention thus are not reported to public health, and (4) mumps is generally a mild illness. Isolation and work quarantine (i.e., restrictions on patient contact during the incubation period), however, were used for health care workers because of the risk of spread to vulnerable patients.” Asymptomatic-ness makes quarantine lower payoff per cost inflicted upon people’s freedom.
    3. It may also be useful to read [a National Academics Press review of social distancing’s effectiveness, including the effectiveness of different methods](https://www.dropbox.com/s/47erqziz3ivhr55/Social%20distancing%20effectiveness%20NAP.pdf?dl=0).

Some additional thoughts:

  1. Microbial ecology, e.g. understanding where microbes are and how they’re moving around, may be useful to coming up with ideas for new, more granular NPIs, even in respiratory epidemics.
  2. I have assumed that all the interventions above are worth doing because they follow the logic of separation (even though their payoffs may vary). Perhaps a review of these interventions would show that some are not. Evidence for this would likely come from historical case studies and/or modeling studies.

Footnotes:

  1. Options are included below. Many were sourced from [Hatchet, Mecher and Lipsitch 2007](https://www.dropbox.com/s/49p0oxqplhyxsoo/NPIs_spanish_flu_Hatchett_Lipsitch.pdf?dl=0) and [Market et al. 2007](https://www.dropbox.com/s/g24uaggpvf2pih3/NPIs_Spanish_flu_Markel_Cetron.pdf?dl=0):
    1. isolation (separate/restrict the movement of sick people);
    2. quarantine (separate/restrict the movement of exposed people);
    3. public risk communications and emergency declarations;
    4. closures of schools and businesses;
    5. transportation and public transit restrictions;
    6. public gathering bans, or “outdoor gatherings only”;
    7. staggered business hours;
    8. shelter-in-place (which includes work from home);
    9. notifiability (legally requiring hospitals/physicians to report instances of a disease to a public health agency);
    10. no-crowding rules;
    11. release or no-crowding of students from or in dorms, soldiers from or in barracks or prisoners from or in prisons;
    12. placards on the doors of the sick;
    13. face mask ordinances;
    14. delivery of goods instead of in-person shopping;
    15. hospitals set up to draw out the potentially sick and isolate them;
    16. voluntary quarantine of infected households;
    17. and more? (Brainstorm; there are many variations!)