Daily COVID-19 Briefing: Saturday

Top news, reports and insights for today:

1. Daily headline summaries for Saturday:

• Despite widespread belief that the epidemic has moved past New York, over 47,000 people in New York, New Jersey and Connecticut tested positive over the last two weeks, over 13,000 in New York City alone. Essential workers (including medical personnel on the front lines) are still falling ill (New York Times)
• Historian John Barry who wrote the definitive account of the “Spanish” influenza of 1918-19, said that President Trump has ignored the most important lesson of that tragedy: National leaders must tell the truth (CBS News)
• New Harvard working paper reviews data on impact of weather, pollution and sunshine finding that warmer temperatures and moderate outdoor UV exposure may offer a modest reduction in the rate of viral transmission. They warn that weather alone will not be enough to fully contain the epidemic (Harvard University). 
• As U.S. funnels billions toward companies rushing to create a COVID-19 vaccine before the election, Gail Van Norman, a professor of medical ethics at University of Washington said this week that “A new investigational drug that’s going into human trials has a 90% chance of failing” (Vanity Fair)

2. New case update: Which state has the most newly reported daily COVID-19 cases? Maryland
    On Friday, 23,862 new COVID-19 cases were reported, a rise of 1.4% in total cumulative cases, now standing at 1.73 million. Last week, there were 143,328 reported new cases, a 9% increase, but the growth factor for the week was 0.92, indicating that overall, growth in cases is nearly flat. The bottom graph is something new. This figure compares daily new case reports per 100,000 people over the last week by state. This is a useful number to watch because it gives a rate of new cases which evens the playing field among states with very different population sizes. Also, numerous experts have said that new cases should be below 5 per day per 100,000 people before we can say that widespread transmission has stopped. As with other graphs I have made, this one shows states grouped by region in different colors. The vertical axis is daily new cases per 100,000 averaged over the last seven days. Looking at the data this way, the west is doing best, with no state reporting more than 6 daily cases per 100K, and 8 states with fewer than 5. Southern states are mixed with Florida, Kentucky, South Carolina, Texas and West Virginia with sub-5 new cases per day. Alabama, Mississippi and Virginia are all above 9. Four midwestern states saw 11 or more new cases per day per 100K last week: Iowa, Illinois, Minnesota and Nebraska. Only 1 northeastern state meets this benchmark (Vermont), the remaining states have seen 5 or more per 100,000 new cases. The fastest rate of growth in new cases came from my state, Maryland, which had 15.5 new cases per day per 100K. This comes after 2 consecutive days of more than 1,200 new cases on Thursday and Friday. By this metric, Maryland is a clear hotspot with more than 3 times the rate of new daily cases that are safe for reopening.
    The bottom line: Different metrics tell different stories about where the epidemic is going. Despite reassuring signs that things are slowing in the northeast, daily new case rates indicate there is a long way to go before the epidemic is under control in this region. Maryland, DC and Delaware continue to be in the cross-hairs of this disease.

3. Big news: the first evidence that previous exposure to other coronaviruses may offer some partial immunity
   The top story today is one I am particularly excited about. I don’t want to overstate the case, but this may turn out to be an important turning point in solving the case of this disease. First, a little context. Two days ago, David Wallace-Wells, who is perhaps my favorite health journalist, posted a terrific article in New York Magazine, posing a good and simple question: Has the epidemic peaked? In that article, he reminds us that 3 weeks ago, the New York times broke a story about an internal CDC model, predicting 200,000 cases and 3,000 deaths a day by June 1. That’s tomorrow. We are now at about 20,000 cases and 950 deaths a day. The biggest question right now is why those models were so wrong? Even though it’s a positive development, that cases and deaths are not soaring as our models predicted, but it means we have some part of the bigger story wrong. Where is the equivalent of “dark matter” in cosmology that will make the math add up? Over the next week I will be covering several of the hypotheses that have gained traction to explain this, but first I am excited to tell you about one of them.
    A big idea: It’s possible that the rate of infections is slowing because some percentage of the population is partially immune to the new SARS-CoV-2 virus as a result of previous exposure to other coronaviruses. This idea has been sitting in the background waiting for some foundational scientific research start teasing this apart. The first compelling evidence came out in a paper published in the journal Cell on May 14 by Alba Grifoni and colleagues from the Center for Infectious Disease and Vaccine Research at the La Jolla Institute for Immunology in California. They looked at immune system response to SARS-CoV-2 in patients who were known to have had the disease and found more good evidence that the key cells in charge of spotting and attacking the virus were widely circulating in recovering patients, meaning that they were now immune to re-infection. Most interestingly, they also found that several of the primary immune system sentry cells capable of recognizing SARS-CoV-2 were present in about half of the people who were as yet unexposed to the new virus. This may be due to cross-immunity. That means previous exposure to one of the four coronaviruses that cause the common cold may have created a partial immune memory that allows some individuals to recognize and mount a response to the new virus. The study was small, using just 20 previous COVID-19 patients and 20 “unexposed” patients. Much more work needs to be done to determine whether the presence of immune cells that might recognize SARS-CoV-2 translates to actual partial or full immunity to the disease because of a prior common cold. One very cool feature of this study is that they used blood from samples taken years before COVID-19 started to make sure that the unexposed group really had not been infected.
    Bottom line: This is the first well-done peer-reviewed study that demonstrates it’s possible that COVID-19 cases and deaths are less than expected because some fraction of the population may be partially immune. That would be a game changer, because it would mean the threshold for herd immunity may be much closer. This doesn’t prove that half the population is partially immune, but its one piece of a very big and growing puzzle.

Advertisement

Daily COVID-19 Briefing: Thursday

Top news, reports and insights for today:

1. Daily headline summaries for Thursday:

• Optimism about a coronavirus vaccine may be premature. Moderna safety study was only 8 patients and they won’t show anyone the data. Experts warn that despite the hopeful timetable, there is a long way to go before a vaccine is proven to be effective and can be mass produced and distributed (New York Times)
• At least 4 states (Virginia, Texas, Georgia, and Vermont) have combined numbers from viral and antibody tests, providing a misleading picture of testing capacity and hampering our ability to track the epidemic (CNN)
• Brazilian health minister resigns after just weeks on the job after clashing with President Bolsonaro over coronavirus response. Today, Brazil reported its highest daily incidence to date of 19,951 new cases. Brazil now has the third most cases in the world as the epidemic begins to surge in South America. (Aljazeera)
• Researchers at Carnegie Mellon University say that nearly half of the accounts tweeting about coronavirus are ‘bots’ rather than real people. Researchers have identified over 100 false narratives tied to these fake accounts. This is further evidence of the potential role of Chinese and Russian intelligence services in spreading mistrust and misinformation (NPR)

2. U.S. cases are flat, deaths remain volatile
    On Wednesday, the U.S. reported 22,368 new cases (up 1.4%) and 1,528 deaths (+1.8%). The recent trend in overall cases has been flat, while daily deaths have been quite volatile. The 7-day pattern has been fairly steady for cases (growth factor = 0.99) and a modest slowing of deaths (growth factor = 0.89). More regional patterns, as always, tell a better story. Growth in cases has slowed in the northeast, with cumulative cases rising less than 10% in New Jersey, New York and Vermont last week. Maryland saw the largest increase in cases in the region growing 22%. New cases grew by 25% or more in two midwest states (Minnesota and North Dakota) and in North Carolina. In the west, only Arizona saw cases rise by more than 20%.

3. Rhode Island again leading the nation in testing. Overall, testing declines for the first time
    Today’s headlines describe the practice in at least 5 states of combining reports of viral and antibody testing. Experts agree that this is a big problem because the two types of tests do very different things and should be kept separate. Treating antibody tests the same risks exaggerating the testing capacity of the state and doesn’t allow us to tell the fraction of people who have active infection. The latter is the information we need to assess reopening benchmarks and to track the resurgence of cases.
    The figure below shows how each state is doing on testing. The bars show completed tests per 10,000 residents as of yesterday. Rhode Island continues to lead the nation by a wide margin, now having tested more than 11% of Rhode Islanders. The four states with the lowest testing rates are all in the West. Pennsylvania and Maine are the only northeast states that are below the national average of 339 per 10,000. States with patterned bars are likely to be doing less of the most important testing than this graph depicts. The magnitude of the problem is not known. Of particular concern are the three states near the bottom already (Pennsylvania, Texas and Virginia) all of whom are experiencing significant rises in cases in recent weeks.
    Why this matters? States that have moved toward reopening are under the microscope to determine whether infections surge. There is every epidemiological reason to expect they will. The political pressure to blunt the apparent impact of reopening is tremendous. Now is the worst possible time for states to play games with their numbers. Of greater concern is that despite the unanimous opinion of experts that more testing is needed, last week marks the first since the start of the epidemic that the number of COVID-19 tests actually declined (see the lower figure from the CDC website).

Taken from CDC website on May 20: https://www.cdc.gov/coronavirus/2019-ncov/covid-data/covidview/05152020/images/clinical-labs.gif

Daily COVID-19 Briefing: Tuesday

Top news, reports and insights for today:

1. Daily headline summaries for Tuesday:

• Maryland reports largest rise yet in COVID-19 cases 4 days after reopening. The most recent test positivity rate is 25% suggesting testing is still selective; cases no doubt under-estimated (NPR)
• Tracking SARS-CoV-2 using genetic analysis: The virus jumped from animals to humans, then spread through human-to-human contact; circulation started in China as early as October 9, 2019; the virus mutated minimally before March 1; multiple “seeding” events occurred in the U.S. but most infections in Washington State likely came from one individual (Scientific American)
• Wildcat strikes, walkouts and protests erupt across U.S. “essential” industries over low pay, lack of safety and protections for workers (The Guardian)

Image captured from the Guardian website: https://www.theguardian.com/world/2020/may/19/strikes-erupt-us-essential-workers-demand-better-protection-amid-pandemic

2. U.S. COVID-19 cases remain flat, deaths dip
    On Monday, the U.S. reported 21,287 new cases, a rise of 1.4%. The 7-day moving average has been declining generally, but flattening in recent days. Last week, 159,270 cases were reported nationwide, a 12% rise. The 1-week growth factor was <1 (0.96) as there were 166,000 new cases the week before. The U.S. is poised to pass 1.5 million cases tomorrow, and continues to have the most cases of any nation, with more than 1 million more cases than Russia (now at #2) and Spain (at #3). States with more than 25% case growth in the last week include Arizona (+25%), Minnesota (+39%), North Dakota (+27%), North Carolina (26%) and Maine (+29%).
    On Monday, there were 792 COVID-19 deaths reported, a rise of less than 1% and the second day of fewer than 800 reported fatalities. Last week, 9,484 Americans died of COVID-19, a rise of 13%, compared to 12,125 the week before (growth factor=0.78). The declining trend in deaths is a positive development, remembering that we would anticipate a rebound in reported deaths tomorrow given the weekend reporting lag. States reporting 25% or more increase in total deaths include Arizona (+28%), New Mexico (30%), Iowa (+31%), Nebraska (+25%), South Dakota (+29%), Delaware (+32%), and New Hampshire (+29%).
    What this means? Because deaths lag behind cases, and cases lag behind infections, it is plausible that the dip in deaths will be temporary. In my opinion, we are now waiting for signs of the conveyor-belt of infections to become visible over the next few weeks as the impact of reopening moves like a ‘pig in a python’.

3. From anosmia to ‘COVID toes’: What to do with the strange mix of COVID-19 symptoms
    Yesterday, there was a nice article in Scientific American summarizing the challenges we face in understanding the often baffling array of symptoms that have popped up during this pandemic. I have been yelling at anyone who would listen since the beginning about the importance of getting the case definition right. Until we have tens of millions more tests, we should be using clinical check-lists to screen everyone. That requires doing a lot of important work to determine how symptoms cluster together and how they map to test results. That itself is a big discussion for another time. In the meantime, lets consider why this has been such a challenge for us disease detectives.
    Any attempt to characterize the symptoms and signs of disease requires balancing two key features of any case definition: sensitivity and specificity. It’s a balancing act because we have competing needs: define the disease broadly enough that it captures all true positive cases, but narrowly enough that it captures all true negative cases. The first part is intuitive (what set of symptoms, when endorsed, identifies everyone who really has COVID-19). The second part is trickier and less intuitive (which symptoms, when endorsed, avoid identifying people with a similar respiratory infection as having COVID-19 when they don’t). The former is the essence of sensitivity, the latter is specificity. The balancing act gets especially tricky when the most common symptoms (fever, cough, fatigue, headache) are also signs of other diseases (ergo low specificity). That’s why my ears always perk up when a rare and mysterious symptom comes onto the radar. Examples include anosmia (loss of sense of smell) or chilblains (sore, red, swollen toes). From a measurement point of view, these symptoms are as good as gold because while they are rare, they are they can be used to differentiate COVID-19 infection from other illnesses. Problem is that requiring rare symptoms lowers our sensitivity and we miss true positives. We know a great deal about how to handle these measurement challenges, its simply a matter of gathering the right data and doing the leg work. To my knowledge, this hasn’t been done yet and is sorely needed.
    Another challenge is that the symptom profiles are clouded by the shear scale of the pandemic. With 5 million cases of a disease nobody had heard of 6 months ago, it is hard to separate the diagnostic wheat from the chaff. When huge numbers of people get a new disease, super rare symptoms will occur and get noticed. Some of these will be noise and some will be signal. There is an art to telling the difference.
    Another challenge is that there are symptoms that occur because of the pathogen itself, and those that occur as a result of host characteristics related to individual differences in immune response. For example, some of the sickest people end up in respiratory failure due to cytokine storm, an exaggerated out-of-control response by the body to a novel viral invader. Separating indicators of differential host susceptibility and symptoms of infection is a key challenge – and its often not a clear line of separation.
    A final challenge is more sociological than scientific. During a fast moving and scary epidemic, facts and opinions go viral, become memes and persist in the collective mindset, even when they no longer serve a purpose. Both the WHO and the CDC were stuck on an initial case definition from a single brief report from China based on a just 41 very sick ICU patients (Huang, Lancet). Long after we had learned that up to 30% of COVID patients don’t present with a fever, that fatigue and malaise were more common than shortness of breath, and that gastrointestinal complaints were important additions to the mix, it took weeks or months before officials moved past the “big 3” (fever, cough and SOB).
    Bottom line: A more complete and nuanced understanding of the diverse symptom profiles of this novel disease are emerging. It’s a fundamental task of the disease detective to gather the right clues and examine them through the right lens. That work still lies ahead. How can I help?