• Josh Schiffer

March 30, 2020

I am sorry that I have been unable to post over the last few days. I have been busy. My method for the blog is more or less that I read something, write a blurb and move on. As a result, there is no order here at all today. Sorry for that.

I do have an essay that is in my head but not on paper yet, which is essentially an ode to the basic reproductive number in a person (not a population). Hopefully I’ll put it up by Friday.


I want to start by saying how impressed how I have been with the citizens of my city and my country. It is easy to focus on the few who are disobeying social distance rules and placing their own short-term happiness over the well-being of the whole. For the most part, I have seen a stoicism and sense of collective responsibility that is just awesome.

However, the growing tragedy is hard to accept. This video piece from the New York Times with Dr. Ashley Bray (a hero) is just one of dozens of personal anecdotes from the health care front lines from across the globe. For whatever reason, this was the one that broke me. I showed the video to my boys in an attempt to team them the gravity of what we are experiencing, and I lost it. I am in disbelief that this is actually happening.


Herd immunity. I am getting asked about this a lot by friends and family. Recall from my last post, we are still waiting for broad sero-surveys to know whether a larger portion of the population is infected than previously understood. If this is the case, then herd immunity may have set in already and the infection could theoretically run its awful course without the theoretical possibility of a second equally destructive wave.

The prerequisite for herd immunity would be a very high percentage of asymptomatic and / or mildly symptomatic cases. I sense a lot of wishful thinking: friends tell me they had a slight cough or sore throat a few weeks ago, and that must have been corona. Recall that there at least 10 viruses, a few bacteria, and many allergens, that can give these exact symptoms. We are in the midst of an historical epidemic in Seattle and 92-94% of tests that are sent each day return negative. The chance that your symptoms a month ago were COVID related? In my opinion, less than 5%.

Data from the cruise ship now suggests 18% of infected people were asymptomatic. Based on this and some other fuzzier data, my instinct is that there is not enough infection for herd immunity yet. My hope is that I am wrong. Still no answers but the data is definitely coming.


Rapid test. This is a piece of great news. A rapid point of care test will allow a much more rapid flow of patients and also may allow expedited home testing, which in my view is a necessity for initiating rapid self-quarantine and for the early test and treat strategy.


Clinical management. A fantastic document from our local friends at Evergreen Hospital. This post is mostly for MDs but for all readers, gives a scope of the importance of pattern recognition that is critical for internal medicine doctors.


Re-infection. If you see reports of re-infection, be skeptical. Think viral persistence. NPR reported that several people in Wuhan are being quarantined as they have detectable shedding after a period without it:

These cases are unlikely to be reinfection as there is little to no virus circulating in Wuhan right now. Assuming the testing is kosher, SARS CoV-2 never completely left their bodies. There are probably lots of people shedding the virus in a similar way worldwide.

This is interesting. What does it mean? First, this is not viral latency (Latency is when a virus hibernates in a cell, sometimes for years, without replicating. HIV and all of the human herpes viruses have expert strategies to establish latency and survive punishing human immune responses and antiviral therapies.). SARS CoV-2 cannot establish latency: it is an RNA virus and needs to make copies of itself to survive.

SARS CoV-2 is also not surviving outside of cells like a fomite as it has been shown to do on metal or cardboard for a few days. In my opinion, it would be implausible for the virus to survive outside of a cell in the body for weeks.

Instead, in these people in Wuhan, in all likelihood, the virus is still replicating, probably at low levels in some anatomic compartments, for days and weeks after people feel better. Think of it as the last smoldering embers of a fire. This is not so surprising. Zika does the same thing. Fantastic COVID data from Germany & Singapore suggests slow clearance of the virus from the airways and the gut so I even wonder if these people are an anomaly or the norm (e Figure 3A) (Figure 3)

Is persistent shedding important? I can only venture educated guesses. Please do not hold me to these answers. From a health point of view, I tend to doubt that the virus is inflicting further harm. As I will describe eventually in an essay on modeling infections in people, viral load is a good predictor in most cases for the number of infected cells and the extent of tissue damage. The low viral load at late time points suggests that the cast majority of the lung is spared ongoing damage from the virus.

Can the virus still be transmitted 28 days after infection? Maybe. My best guess is rarely if ever. Viral load is a key determinant of transmission and high amounts of shedding may be needed for most person to person transmissions. Please see Siddhartha Mukherjee’s elegant essay this week in the New Yorker:

The issue of late COVID transmission definitely needs to be studied. Long term studies of shedding in nasal specimens, sputum, urine, semen and stool will happen, probably in Seattle. We are good at this stuff.

Is it possible that this will be a chronic infection is some people that is never cleared? I doubt it, but we definitely cannot rule it out at this point. To quote Tony Fauci, the virus makes the rules.


How to talk about modeling to politicians (and scientists). Mathematical modeling entered the political sphere last week when several US politicians misconstrued model conclusions and misquoted modelers. This is nothing new for our field. Even in my quieter, less sexy, neck of the modeling woods (modeling infection in a person rather than a population), it is common for non-quantitative scientists to reject our work as out of hand if it does not jibe with their opinion, or maybe even worse, to embrace it with unbridled enthusiasm because it supports their work.

My personal philosophy of modeling is that for the most part, we are trying to help policy makers (or in my world, experimentalists and clinical trialists) to make informed decisions. More often than not, we are generating hypotheses rather than testing them. The Imperial college modelers projected the number of infected people and deaths, with and without interventions. Their goal was not to make precise predictions about the future, but rather to give policy makers some estimates to inform policy making.

In the case of social distancing, the data from China

( slide 35)

suggests that modelers were correct. Limiting social contacts can absolutely lower transmission rates and even eliminate a local epidemic from a city. The current rebound of cases in Hong Kong validates their other key prediction, that relaxation of social distancing is bound to be followed by a surge in the local epidemic.

When the epidemic is over and the data is more complete, modelers will be able to reconstruct the precise impact of interventions. Until then, they are simply trying to help make sound policy decisions with incomplete, and sometime inaccurate data. To this end, they bring an invaluable but imperfect set of tools to the table.


I want to touch on this model below which was incidentally quoted in the Washington Post piece above. The model comes from the Institute of Health Metrics and Evaluation (IHME) in Seattle. As a starting point, IHME is a world class organization whose focus is usually on measuring the burden and costs of all diseases. Their work is unique. I recommend their website for fun graphics and informative comparisons of country by country disease burden


However, IHME are not traditionally epidemic modelers. Modelers use differential equations which capture rate of change of variables like number of susceptible people, number of infected people, number of immune people) as a necessary tool to capture the non-linear behavior of these variables during an epidemic.

In his twitter feed, Carl Bergstrom at UW ( points out that that the IHME group does not use this type of model for projecting the future of COVID. Rather it looks they use a method to project what happens next based on the shape of existing data. If true, this is unfortunate because this approach ignores the mechanisms, the physical laws, that underly an epidemic. These rules are not perfectly understood, and modelers’ equations are always slightly wrong. However, in the right hands, the math can be close enough to the epidemiologic truth, that it adequately captures the behavior of a spreading pathogen and helps guide rational policy making.

Having mechanisms in a model’s equations allows one to play by the rules of the virus, to make the best decisions possible. Using a non-mechanistic approach can lead to some wildly wrong projections as Carl shows in his HIV epidemic example in his feed (slide 7).

Here is a list of examples of how incredibly useful modeling has been so far during the COVID-19 pandemic:


Viral genomics. Another really stunning explanation from Trevor on 3/27 giving a country by country breakdown of viral spread across the globe and within communities. Extraordinary.


Mask & hand washing science. In 2008, these authors went through the trouble of doing a meta-analysis of dozens of published studies on hand washing, wearing gloves, wearing masks & wearing gowns to prevent respiratory virus spread. Bless their hearts. Truly. This type of analysis requires a special level of diligence and tolerance for boredom. The including studies took place in a gemish of setting including daycare, homes, hospitals and military sites,

The results are awesome. See Table 6 for the number needed to treat (NNT) to prevent one infection: in other words, the number of people required to do the intervention to prevent one infection. NNT is the greatest clinical tool: I use it constantly when I am seeing patients. For instance, the NNT for people with HIV and antiretroviral therapy is ~1.1 which means that starting antiretroviral therapy is lifesaving in a majority of infected people: this is one of the many reasons that I love HIV medicine.

Other common interventions are far less effective. Sigmoidoscopy has an NNT of 450 to save one life from colon cancer. Statins have a NNT of ~80 for heart attack prevention in people with known heart disease. Before scoffing at these numbers, consider how many lives are saved if 100,000 people with heart disease take a statin: ~1200.

Anyway, somewhat to my surprise, for masks, gowns, gloves, and handwashing the NNTs are really low! They are all great strategies.

With this, I have a confession, during the next pandemic or bad flu year, and assuming, and hoping, that we have plenty of masks to go around in hospitals, I am going to wear a mask when in crowded public spaces, and on airplanes. I am convinced enough by the data above. And I don’t mind looking like a dufus.


Mass spreading events. I found this story to be terribly dispiriting, that one of the world’s truly great athletic tournaments, which my boys and I follow together closely, appeared to accelerate the awful outbreak in Italy.

Same story with Mardi Gras. It is just brutal to consider the ramifications. Will we ever feel completely safe about mass events again? If so, when?

My dear friend, college roommate and drummer extraordinaire John Colpitts (aka Kid Millions) interviewed me for the piece below. John spoke with artists from around the world and discussed how they will survive in the current environment. It will really be a challenge to decide when musicians and athletes can reconvene is front of small, medium and large crowds. I think advanced analytics, sort of like weather forecasting, could be a tool to help with this.

John mentions a story at the start of the epidemic where I calculated in my head that it was probably okay for Chihana and I to go to one last Sounders game at the end of February. We were excited to spend a night with some great friends, before the shit hit the fan in our city (as I knew it would). In retrospect, I made the wrong decision and I regret it. The game should have been cancelled.


Broadly neutralized antibodies. There is a race across the world to identify antibodies from patients that prevent the virus from entering cells, mass producing them, and then infuse them either prophylactically as a preventative measure, or as a treatment.

“BNAbs” could have great relevance for the pandemic. Below is a website for the AMP studies: 4 continents, >40,000 antibody or placebo infusions, 4200 participants, to test the role of BNabs for preventing HIV. Enrollment completed. Where there is a will….


Beautiful mortality data from Korea. This is from YaeJean Kim, an extremely talented scientist and physician who worked with us at the Hutch for a while. She was instrumental in managing the MERS outbreak in Korea and now COVID. This paper is awesome. The low death rate is remarkable. The clear data presentation is commednable. Figure 4 teaches me that early test and treat, if it works, will need to be implemented pretty quickly. Humbling.


Hydroxychloroquine. This is a randomized clinical trial of the drug. Not a shining success. However, the sample size is tiny, so it is definitely worth giving it a shot in a bigger trial -------------------------

Washington state. Both the canary in the coal mine perspective at hospitals, and also the incidence data which I continue to plot everyday shows evidence that physical distancing is having an effect. We are treading water with weights on our ankles, and it is early, but it is so great to see this. Remember also that flattening the curve means that we are in this for a while.

I hold the “COVID pager” for UWMC and I do not get the sense that the hospital is overwhelmed as of yet. However, the situation is heartbreaking. In particular, the lack of hospital visitors except under exceptional circumstances is clearly straining for families, patients and health care workers. ----------------------

What went wrong in the US. I have some expertise regarding viruses, epidemics and the like. However, I am as in the dark as the next guy when it comes to why our country failed in this time of need. It feels like Katrina writ large, which I also did not really understand. This piece from Fareed Zakaria helped but I don’t see a solution. Perhaps, we can import public health leaders from Germany, South Korea, Singapore, Vietnam and Taiwan and just promise to do exactly as they say, and to hire as many people as they want. -------------------------

Cocooning the vulnerable. This piece highlights why relaxing physical distancing and cocooning the elderly would be disastrous in this country. Due to heart disease and diabetes, we are uniquely vulnerable to high morbidity COVID. I completely agree. Left out is the fact that cocooning the elderly may not even protect the elderly from the virus.


Sweden. I am watching this. This policy seems a bit loose to me. On the other hand, the Swedes are good at infectious diseases. I have no idea what to expect.


Super spreader events. From Rich Read at the LA Times, some circumstantial evidence of aerosol rather than droplet spread. Very tragic story and informative piece. I keep harking back to the Biogen outbreak in Boston (which seems like a century ago) mentioned in my earlier posts. I still feel like 70 people getting infected at a brief conference does not sound totally like a typical respiratory virus. I am not sure flu would to that.


And a much safer choir practice. This is beautiful.


© 2020 by Joshua T. Schiffer | Proudly created with

  • Facebook Clean Grey
  • Twitter Clean Grey
  • LinkedIn Clean Grey