Month: June 2020

COVID-19 case count has been increasing significantly in California in June (see figure below), but compared to some of the other hot states like Florida, Texas, and Arizona, its growth has roughly matched the growth in testing capacity and its positivity rate (percent of test results returning positive) has stayed below 7% unlike FL, TX, and AZ whose positivity rates are more than double this.  While hospitalizations have also increased in CA it has not approached full capacity as they have in TX.  Most importantly its death count has declined through May, and even with the recent outbreaks in June it has remained steady rather than surge along with cases.

Death counts usually lag confirmed case counts by 1-2 weeks but we remain optimistic that CA death counts will only increase modestly this summer.  The reasons are fivefold:

1. Expanded testing has turned up milder and asymptomatic cases which while still contagious do not usually result in deaths.
2. Expanded testing has skewed toward younger adults who have case fatality rates much lower than older adults (see figure below) and will lower the Case Fatality Rate (CFR = deaths/cases) for the state as a whole from 6% this spring to below 2% this summer.
3. Earlier diagnosis and treatment could reduce hospital stay times and fatalities as drugs such as Remdesivir and Dexamethasone could be used to slow progression and allow the body’s natural defenses to help patients recover.  The availability of plasma from previously infected patients could also help. (From a forecasting perspective – slowing the progression would also increase the lag time and delay the upturn in death counts.) 
4. Better treatment techniques such as proning (laying the patient on his stomach) delay or eliminate the need for ventilators which are difficult to wean patients off.
5. Government is responsive to changes in data and willing to do such things as mandate mask usage  and delay reopening when infections accelerate. 

Given all this, our forecast for CA is that death counts will edge up in July to around 70 per day from its June average of 61 per day due to the June surge in cases.  CFR will stabilize below 2%, the ratio of the right axis for death counts to the left axis for case counts in the uppermost figure.  The CFR dependence on age will continue to show the steep exponential behavior (seen in many other countries around the globe such as Spain and South Korea) demonstrated in the lower figure with older people (>80yrs old) having a thousand times greater chance of dying from COVID-19 (27%) than younger people (<17yrs old).  Gender and existence of comorbidities will also affect CFR.  COVID-19 remains a difficult beast to tame so continued vigilance on the part of individuals, businesses, and governments is required to prevent new outbreaks.  

New York, New Jersey, and Connecticut have just moved to restrict travel from hot zones around the country.  The states that had the worst Spring outbreak (and whose residents were quarantined by Florida and others) are now requiring Florida and other states’ residents to quarantine for 14 days.  As they might say, “turnabout is fair play”.  A single figure of merit that is not arbitrary or capricious is not easy to agree on, but their definition for current hotness sounds reasonable: more than 10 confirmed COVID-19 cases per 100,000 residents based on a rolling 7-day average.  The table below lists the 8 official ignominious states plus California (CA) which qualified yesterday.  We also include the infection reproduction rate R, and the doubling time for each state as indicators of how fast infections are growing in each state.  The doubling time might also be used as a gating criterion since a state with low current hotness might have a very short doubling time and qualify for the list quickly.  

State	Infections	Infections /million	Deaths	Deaths /million	Fatality (CFR)	R	Doubling (days)	Hotness (7-day)
AZ	59,974	8,240	1,463	201	2.4%	1.3	10	31.7
SC	29,022	5,637	693	135	2.4%	1.2	14	16.9
AR	17,375	5,757	240	80	1.4%	1.1	28	15.9
AL	33,206	6,772	896	183	2.7%	1.1	29	15.4
UT	18,784	5,859	163	51	0.9%	1.2	16	14.5
FL	114,018	5,353	3,327	156	2.9%	1.3	9	14.1
NC	57,183	5,452	1,318	126	2.3%	1.1	29	12.8
TX	137,152	4,730	2,334	80	1.7%	1.3	10	12.0
CA	201,004	5,087	5,809	147	2.9%	1.1	24	11.1

To truly minimize USA reinfections, we had argued that the US needs to update the restricted travel countries list to include many countries that are currently hot and remove countries that are not.  The table below lists some of the potential countries and their 7-day hotness measure.  As you can see NY, NJ, CT could quarantine the entire USA by their gating criteria.  The USA not only qualifies on a rolling 7-day infection basis but it is also growing faster than other highly troubled countries such as Brazil.  Russia and India should be considered for the restricted travel list because both have huge infections that are managed to look good but are probably out of control.  Others, such as countries in Europe should be considered for removal from the CDC restricted travel list.

Broad quarantines are crude and often political instruments that should only be used as a last resort.  If a state fails to execute good practices such as masking, social distancing, testing, contact tracing, and selective isolation, it may be the only tool left to control a pandemic.  The CDC has now come out to downplay the efficacy of out-of-state quarantines, but they offered no evidence to the contrary and no alternative policy or procedure to minimize reinfections.  This is one of the dangers of not having a quick and decisive Federal response to the Pandemic in the first place.    

Country	Infections	Infections /million	Deaths	Deaths /million	Fatality (CFR)	R	Doubling (days)	Hotness (7-day)
USA	2,504,298	7,583	126,771	384	5.1%	1.13	20	10.3
Brazil	1,233,147	5,810	55,054	89	4.5%	1.12	22	16.5
Russia	613,994	4,208	8,605	20	1.4%	1.00		5.5
India	491,170	356	15,308	11	3.1%	1.12	22	1.1
Peru	268,602	8,148	8,761	153	3.3%	1.00		12.3
Chile	259,064	13,554	4,903	71	1.9%	1.27	11	26.2

Many people are asking why COVID-19 death counts in the USA are decreasing or stabilizing if the number of confirmed COVID-19 cases are soaring in most of the USA.  The governors of many of the worst-hit states such as Florida, Texas, Arizona, and California are attributing much of the rise in cases to increased testing of milder and asymptomatic cases which tend to not require hospitalization nor result in deaths.  In fact, many of these states did experience dips in Case Fatality Rate (CFR = deaths/cases) as the denominator increased faster than the numerator.  While this was true in May, it is no longer true in June.  The reason is that in many states such as Florida the number of tests performed daily (brown squares against the left axis in the figure below) increased impressively in April and May but has stalled in June to the extent that the percentage of tests returning positive (blue diamonds against right axis) has increased to new highs.

The governors also argue that newer cases tend to skew younger as they are the least likely to wear masks and observe social distancing.  This can be seen in the next graph that shows the distribution of new cases in Florida by age group from 6/19 to 6/23.  While the age distribution cumulative to 6/19 (blue bars) had a median age of 45 yrs, the incremental 15,073 new cases (brown bars) over 4 days had a median age of 35 yrs. 

The younger patients tend to require less hospitalization and die less often.  The following graph shows the very strong CFR dependence on age.  The same age dependence seen in Florida is also seen in California and throughout the world.  Those younger than 30 yrs do get sick but <0.1% will die from COVID-19.  Those older than 50 yrs will have >1.0% chance of dying from COVID-19.  Nearly 1,000 new cases for those older than 50 yrs have been confirmed each day in Florida.  Knowing the age distribution of new cases, we can make a fairly accurate forecast of deaths to come in Florida and the rest of the country over the next two weeks.   

The reason is that there is always a lag of 1-2 weeks between test confirmation and case resolution (death or recovery) so that the sharp increase in case counts seen around June 10^th is just feeding into increased hospitalizations last week and death counts now and will cause them to rise into July.  The projected daily death count (green triangles against the right axis in the figure below) are way below historic averages. The reason is that CFR  peaked near 5% for Florida earlier this year when a lot of those sickened were older males with comorbidities but now the underlying CFR for the state is below 2% as more young and asymptomatic patients have been identified.    

What we see in Florida is what is playing out in dozens of other states around the country as states reopened too quickly without regard for data or common sense.  We should point out again that this Summer disaster need not play out this way.  Masking, social distancing, testing, contact tracing, and selective isolation are proven techniques to limit the spread of COVID-19 and can reduce the number of unnecessary casualties in the USA. 

The number of daily confirmed COVID-19 cases in Florida and several of its major cities such as Miami, Tampa, and Orlando continues to rise exponentially (with Orlando showing a doubling time of 4 days). The latest data shows a surge to record high levels (7-day average) in Orange (brown squares against right axis) and Miami-Dade (blue diamonds against left axis) Counties. 

Orlando, Tampa, and Miami are now more dangerous than any part of Metro-NYC is, and Florida is now hotter than any other state (as indicated by its doubling time of 9.5 days, which is the shortest of any state).  COVID-19 cases surged and will continue to surge higher in Florida for the following reasons:

1. Huge crowds on Florida beaches for Memorial Day weekend, 5/23-25 that showed up as new cases beginning in June
2. Hundreds of thousands watching SpaceX launch attempts on 5/27,5/30 without masks and social distancing on the Space Coast that showed up as new cases in Orlando
3. Ongoing protests against racial injustice soon after George Floyd’s death on 5/25 that are showing up as increased case counts now.
4. Universal Studios reopened on 6/5 and Disney World will reopen on July 11^th .
5. Florida entered Phase 2 reopening on 6/5 (with the exception of Palm Beach, Broward, and Miami-Dade counties that are now in Phase 1).
6. The spread of the Pandemic to Latin America in May that led to the recent travel ban from Brazil that should have also been applied to other countries such as Peru, Chile, Mexico, Ecuador, Colombia, etc.  Hundreds of millions of Latin Americans visit Florida (mostly Miami and Orlando) each year.

The novel coronavirus can take about 5–14 days to incubate before an infected person starts showing symptoms and register as a confirmed case.  So even though case counts stopped declining and started to rise slowly shortly after May 4^th reopening (see figure below) it did not surge until later in May.  Moreover, this resurgence in cases is not confined to the three hot counties of Palm Beach, Broward, and Miami-Dade but it has spread to Hillsborough, Orlando, Pinellas, and the rest of Florida.  The brown squares track the number of daily new cases (left axis) and the blue diamonds track the number of daily new deaths (right axis).  The case fatality rate (CFR = deaths divided by cases) has held near around 4.5% (the ratio of the right axis to the left axis) with deaths lagging case identification by about 2 weeks.  The first broad peak in Florida case counts occurred around 4/12 a broad peak in death count followed on around 4/28.  The death count has averaged near 33 per day for the last 5 weeks, but we expect it to rise slowly in the next few weeks as cases turn into deaths or recoveries. Increased testing should have identified milder and asymptomatic cases associated with younger adults, so the CFR should be lower. 

The recent resurgence in hospitalizations, however, makes us very pessimistic about Florida’s prospects.  The brown squares (7-day averages) below show an increase in new hospital admits over the last two weeks and reached 200 yesterday.  Yesterday 21,559 tests resulted in 17.7% positive cases – a terrible deterioration in the positive test rate from a low of 3% reached on 5/23. 

The recent decision to hold the GOP Convention to Jacksonville will exacerbate the problem for Florida.  The Governor of Florida promised to let the data and science guide reopening so he needs to halt reopening plans to assess the new COVID-19 outbreaks and make sure that testing, contact tracing, and selective isolation can keep up with the recent surge.  The Mayor of Miami has already done so.  If nothing is done statewide, the state may need to lock down again next month to prevent thousands of more Floridians from needlessly dying.   

The number of COVID-19 confirmed cases are rising alarmingly in many parts of the US.  Many of the governors explain away the rise as a result of increased testing.  To some extent this is true.  The number of tests per day has increased from an April plateau around 150,000 tests per day to near 600,000 tests per day (brown squares against the left axis plotted in thousands in the figure below).  This occurred despite the President’s constant criticism of testing for Americans even as he requires everyone around him to be tested multiple times a week.  Some governors base their claim of victory on this improving trend as the average percent of positive results (blue diamonds against the right axis) have fallen below 5% in June for the US as a whole.  Many epidemiologists believe that driving this number below 1% as South Korea and Australia have done will be required to truly control the pandemic. A few states are close to achieving this target but most are not.  

In addition, some governors point to the fact that widespread testing has allowed them to sample more mildly symptomatic and asymptomatic cases, who are generally younger adults.  So all this testing has increased case counts but not hospitalizations and eventually death counts.  They conclude this is all good because this will lower the case fatality rate (CFR) for their state and show that they are managing their epidemic better. 

Unfortunately, this approach makes it extremely difficult for older and more vulnerable Americans to survive.  The strongest driver of COVID-19 mortality is age with the same phenomena seen in every country.  In California those over 80 yrs old are over 1000 times more likely to die (36%) than those younger than 17 yrs old (< 0.03%).  So many young people flaunting the rules of masking and social distancing (encouraged by the President and his allies) which is causing a higher percentage of the population to become infectious.  More vulnerable Americans will also get COVID-19 and many of them will die.  In fact, hospitalizations are now edging up in Florida, Arizona, Texas, California, and many other states.  This laissez-faire attitude was also adopted by Sweden and Brazil where effective lockdowns were never imposed.  Now more Swedes are realizing that their government has failed to protect its most vulnerable.  Mahatma Gandhi suggested that “the true measure of any society can be found in how it treats its most vulnerable members.”  Currently, this country is not adequately protecting those who are most harmed by the pandemic, including sick, old, poor, and black Americans.  Worse, this President is actively working to disenfranchise the most vulnerable, by continuing to fight against mail-in voting, their only means to vote safely in a Pandemic.

We pointed out earlier that if one looks at the US outside of New York (NY) and New Jersey (NJ) (or more precisely outside of Metro-NYC), the country’s COVID-19 pandemic has not peaked but rather has plateaued since mid-April.  Some states outside of NY and NJ appear to have also peaked but the 3 biggest states in the US, California, Texas, and Florida — containing 25% of the country’s population, has not really peaked and are experiencing continued increases in daily confirmed cases.  The figure below shows the daily case count for these 3 states:

California (CA — blue diamonds in figure above with 7-day rolling average), after plateauing in early April, has seen rising case count since mid-April.  Part of this rise has been due to increased testing but while testing has definitely improved it cannot account for all the increases in infections.  Some of it has been due to a series of politically motivated miscalculations undertaken to reopen the state since early May and testing and contact tracing missteps.  In addition, since Memorial Day, the massive crowds at the beaches and at the demonstrations against racial injustice and police brutality have just begun to feed into higher infection rates this month and into Summer.

Texas (TX – red squares) after initial infection plateaued in mid-April has experienced waves of additional infections ever since Phase 1 reopening began on May 1^st.  Governor Abbott just widened the reopening to Phase 3 on June 3 with little regard for the alarming trends in infections.  Texas has just reported its highest rate of hospitalization rates since the Coronavirus began spreading but officials have no plans to slow reopening.  Yesterday, 2166 COVID-19 patients were hospitalized and Texas ICUs are reportedly at 88% capacity.  So while its Deaths per million population and Case Fatality Rate (CFR in table below) has been relatively benign due to its relatively young population (median age of 34.8 yrs), it should not want to see its ICU flooded with more cases.

Florida (FL – green triangles) showed some evidence of peaking in case count in early April that led to halving of infections by early May.  However, as soon as the state reopened on May 2^nd the case counts leveled off and by the end of May begin to rise again for a wide variety of reasons.  More recently the rise has begun to look exponential with an effective reproduction rate of R=1.18 and a doubling time of 13 days (see table below).  This does not bode well for the GOP convention now scheduled for August in Jacksonville.  If they do not enforce testing, masking, and social distancing it could spell disaster for the GOP, city, state, and country. 

State	Infections	Infections /million	Deaths	Deaths /million	Fatality (CFR)	% Test Positive	Median Age	R	Doubling time(days)
CA	147,132	3,724	4,988	126	3.4%	5.5%	36.8	1.08	33
TX	85,641	2,954	1,966	68	2.3%	6.8%	34.8	1.14	20
FL	70,971	3,332	2,877	135	4.1%	5.4%	42.2	1.18	13

While CA, TX and FL had the 3 highest single day totals yesterday, North Carolina, Arizona, and Arkansas also saw large increases (~0.02% of population) recently.  Other states such as Oregon and Utah have prudently paused reopening plans as they investigate recent surges.  Those who are not pausing are inviting disaster because the slower one reacts, the more drastic the measures one needs to apply to stop any outbreak. 

After a slow start, the US has made good progress in the last couple of months against the COVID-19 pandemic.  Confirmed case counts per day have declined to 20,000 after peaking near 35,000 in early April (brown squares against the left axis in the figure below).  

Over the same time period, the number of tests conducted per day increased from 100,000 to near 500,000 now (brown squares against the left axis in the figure below).  The percent of tests that yielded positive results has declined from 24% to below 5% over the same time period (blue diamonds plotted against the right axis in the figure below).  At this rate, the country as a whole may finally control the virus by August.

However, the country needs to remain vigilant as the virus can and has the ability to re-emerge in many hotspots.  Much of the national progress over the last two months has been due to the ability of the governors to control the global epicenter that emerged in March in metro-NYC.  The daily case count after removing those from New York and New Jersey shows that for the rest of the country the infection has not peaked but just plateaued – thus still a dangerous mix of smoldering fires and emerging hotspots (see figure below where the smoothed line is the 7-day rolling average).  

For the 3 largest states in the US: California, Texas, and Florida that account for near 25% of the US population the virus is increasing with R-values greater than 1.  Actually according to one modeler, CMMID,  R, reproduction number, for the US as a whole, has now re-emerged above 1 after all the states reopened to some extent in May.  Some of this increase in daily case count was expected due to increased testing revealing more of the mildly symptomatic and asymptomatic cases, and it has positively improved the US case fatality rate from above 6% to below 6% (see the first figure above where the blue diamonds are plotted against the right axis scaled to 6% of the case count). It is nevertheless, concerning that the infection is still growing in large parts of the US with the US reopening more every day.  Recent massive demonstrations against racial injustice are sure to exacerbate the pandemic. All demonstrators should get tested and testing and contact tracing capacity must increase further in the US to give us a chance to contain the pandemic.

We have been tracking COVID-19 cases in Florida with cautious optimism ever since the Governor initiated Phase 1 reopening on May 4^th.  Daily confirmed cases had been tracking down before that date and leveled off after that date (see figure below).  We had assumed that increased testing accounted for most of the increased cases that shows as the gap between the projected downward blue trend line versus the actual 7-day moving average (solid brown line) after reopening.  Phase 1 reopening appeared to be more or less on track.  However, in recent days the confirmed case count has surged anew to a record high weekly total.

What is causing this resurgence of infections?

1. Huge crowds on Florida beaches for Memorial Day weekend, 5/23-25 that showed up as new cases beginning 6/3
2. Hundreds of thousands watching SpaceX launch attempts on 5/27,5/30 without masks and social distancing on the Space Coast that is showing up as new cases now -especially in Orlando
3. Ongoing protests against racial injustice soon after George Floyd’s death on 5/25 that could show up as increased case counts in coming weeks.
4. Reopening of Universal Studios on 6/5 and Disney World on 7/11 that could lead to new cases 10 days from now and later this summer (although both nominally adhere to CDC guidelines).
5. Florida reopening going to Phase 2 on 6/5 (with the exception of Palm Beach, Broward, and Miami-Dade counties) that could result in surges later this month.
6. The spread of the Pandemic to Latin America in May that led to the recent travel ban from Brazil that should have also been applied to other countries such as Peru, Chile, Mexico, Ecuador, Colombia, etc.  Hundreds of millions of Latin Americans enter Florida each year.

The novel coronavirus can take about 5–14 days to incubate before an infected person starts showing symptoms (if at all) and register as confirmed cases.  So even though case counts stopped declining and started to rise slowly shortly after the May 4^th reopening it did not surge until this week.  Moreover, this resurgence in cases is not just confined to the three hot counties of Palm Beach, Broward, and Miami-Dade but is spread throughout Florida.  The figure below shows the daily case count in Orange County that encompasses the city of Orlando.  The same pattern of declining cases in April, mild rise in May, and resurgence in June is observed.  

Governor DeSantis’s move to Phase 2 reopening on June 5^th was disappointing especially after he had promised in early May not to reopen further unless the data permitted it.  Clearly, the data shows a resurgence that has not even reflected the full impact of recent massive protests and the reopening of Universal and Disney that are already baked into Florida’s COVID-19 cases for this summer.  DeSantis would have a better chance of relocating the August GOP Convention to Orlando Florida safely if he delayed Phase 2 reopening until testing and contact tracing caught up with the recent surge in COVID-19 infections.