Jane the Actuary General Interest

Covid update: contact tracing, hospitalizations, international comparisons

Yesterday, the state of Illinois announced that suburban Cook County is now subject to covid mitigations, in which, as of tomorrow, indoor dining and bar service will be prohibited and groups at any one gathering will be capped at 25 people.  As the Chicago Tribune reported,

By Wednesday, more than half of Illinois residents will be living under stricter measures meant to slow the latest surge of the coronavirus as suburban Cook County and the Metro East region outside St. Louis join four other regions where the state has shut down indoor dining and bar service and lowered the cap on crowds to 25.

A week ago, only one of the 11 regions in Gov. J.B. Pritzker’s reopening plan was subject to those rules. . . .

Suburban Cook County has had eight consecutive days of test positivity rate increases and seven days of increased hospital admissions. It is the first region to surpass the state-set thresholds for those two metrics at the same time. The other region have triggered tougher rules by reaching an 8% positivity rate threshold for three consecutive days. As of Friday, the rolling seven-day positivity rate for the Cook County suburbs was 7.7%.

With that context, I wanted to write down some of my recent thoughts on developments.

First, why did contact tracing fail?

Contact tracing was, after all, the subject of my May 15 Tribune commentary, in which I observed that the state’s reopening metrics required that for Phase 3, contact tracing would be fully-rolled out, and for Phase 4, contact tracing would be fully scaled up, so that tracing would begin within 24 hours for 90% of new diagnoses in a region.  Turns out, Phase 3 began according to the timeline for all the other metrics, at the end of May, and Phase 4 began at the end of June.  But according to the organization TestAndTrace, as reported at Patch, Illinois has a failing grade in their assessment of Illinois’ contact tracing, due to their very low number of tracers and lack of transparency about their progress.

With respect to Cook County, the Cook County Department of Health announced on June 11 the receipt of a $41 million grant to scale up contact tracing.  But even just last week, Daily Line reporter Alex Nitkin had this to say:

Now, the question above is, I admit, partly rhetorical.  I have my own suppositions as to why, when it really would have made a difference in the late spring/early summer, when cases were down and when Illinoisans were relatively more hunkered down — that is, when a contact tracing effort would have uncovered fewer contacts for any given individual — the state, and Cook County and the city of Chicago as well, fumbled this:  the desire to create union jobs, for one, and the focus on “equity” even when a focus on “low-hanging fruit” might have been more effective.  Various reports in the meantime have described the suspicion with which immigrants and low-income Chicagoans treat contact tracer outreach; focusing resources on parts of the state which would have a greater success rate, as well as more rural areas where there would have been fewer contacts to trace in the first place, would surely have found more success.

In any event, at this point, it is far less effective to contact-trace when cases are becoming increasingly prevalent, as is the case now.

And, in addition, it would appear that the state is misinterpreting the data that it does garner from contact tracing, in any case.

Earlier in the month, Capitol Fax provided a graph produced by the state which was used as justification for its focus on bars and restaurants.  The graph is appallingly difficult to read (and is shown only as a picture, with no data accessible), but claims that, statewide, when contact tracees were asked where they had “visited or worked” within the past 14 days, the second-largest category of response was “restaurant/bar,” with 2300 responses.  The top response was “other,” which includes “vacations, family gatherings, weddings, college parties.”

But this graph is exceedingly unhelpful.  Respondents could give multiple answers, and the graph’s “n” is given as 17,939, but that’s the total number of boxes checked; a true “n” from such a survey would be the number of people surveyed.  We don’t know what percent of tracees visited bars or restaurants, and, more importantly, we don’t know whether people who went to bars/restaurants were disproportionately likely to have been diagnosed with covid.  In other words, to tell us something meaningful, this graph would need to be paired with another one, in which a random sampling of people who matched the demographic characteristics of covid-diagnosed tracing respondents.

And, in fact, here’s my transcript of Dr. Ngozi Ezike’s comments at the press briefing last Friday, upon being asked why bars and restaurants are being singled out for closure despite being linked to only 6% of outbreaks (about the 28 minute mark):

Ezike:  An outbreak would be something if somebody works like say at a manufacturing plant and a lot of people work in close proximity and 50 people develop covid in that setting we would have that as a documented outbreak where this person knows that I was working next to this person, this person contracted the virus, a week later so did I, a week later this person did, three days later, so that is like a clearly documented outbreak.  In most cases of covid, the person who has it cannot tell you exactly who they got it from, they cannot say, “oh, I was working in this setting and all these people got it and I got similar symptoms, so in the absence of a documented outbreak all those individuals that did contract the virus the way that you look at where they may have gotten it from is to look at the time at which, the time frame from when you catch the virus to when you show symptoms or to whey you’re diagnosed, in those preceding two weeks,  we ask the cases, where have you been, and all of those places that they list, that they have been in the preceding two weeks are exposure sites.  Any of those places could have been where they contracted the virus.  And time after time, bars and restaurants come up as the number two or the number three place of all of these places frequented, so that’s why we put it as a high because it consistently comes up as a place where people who are infected listed as one of their exposure sites.

What’s remarkable is that Pritzker appears to recognized that this is a poor rationale for closing bars and restaurants, and he jumps in:

Pritzker:  And I would just add that there are literally a dozen studies, many many articles about these bars and restaurants being exposure sites, and the effect of bars and restaurants on the spread of the virus and that is why there is a focus on bars and restaurants.

Ezike’s specialty is pediatrics and her expertise within public health comes from working on health care within juvenile detention centers.  Is it possible that she just doesn’t have the grounding in statistics that’s necessary to grasp these concepts?

Second, why are hospitalizations level in suburban Cook County?

Wirepoints has been tracking key covid data for Illinois as a whole since early on.  When cases rose starting in July, in a way that did not appear to be clearly linked to the ramping up of testing (because at that point the state had already increased testing substantially), I followed their tracking of hospitalizations and deaths and observed that these numbers were holding steady, in a manner that fit the theory that the rise in cases were due to increasing numbers of low-risk young adults becoming diagnosed, or that masks were having the effect of reducing the viral load and thus the severity of the infection.

But that’s no longer the case.  On September 19, hospitalizations stood at 1,417.  They rose gradually, to reach 1,575 on Oct. 3; since then they’ve risen steadily, to 2,605 on October 24.

What’s more, deaths have been increasing during the same time period.  Ranging from the upper teens to the low 20s all summer and early fall, the 7 day average stood at 42 on Oct. 26.  It’s also difficult to make a visual judgement, but there is no apparent lag, as you’d expect there should be, from the start of the increase in hospitalizations to the start of the increase in deaths; these are occurring simultaneously.

Frustratingly, it is not easy, from the information available at the Illinois Department of Health’s website, to look at hospitalizations by region or county.  With a little bit of patience, we can look at admissions for Covid-like illnesses region-by-region:

Region 1, northwest Illinois:  hospitalizations level over the summer, then steadily increase from 4 on September 20 to 14 on October 21.

Region 2, north-central-west:  steady increase in hospitalizations over July (4 on July 3) to early August (12 on August 4), then another small increase in recent weeks (14 on Oct 22).

Region 3, central-west: again, small incease in mid-July, level to mid-October, then increase from 8 on Oct. 12 to 12 on Oct. 18.

Region 4, south- west (St. Louis area): relative peaks in late July and again in late August; decline since then to match the level of June.

Region 5, south:  level/very gradual increase through September; then increase from 4 on Oct. 1 to 8 on Oct. 21.

Region 6, east: same pattern as south, very gradual increase through end of September, then jump: 8 on Oct. 2 to 19 on Oct. 22.

Region 7, far southern suburbs/exurbs: lots of bouncing around: peak in mid-August, decline, then increase from the beginning of October to now (6 to 16).

Region 8, western/far western suburbs: increase in mid-June, level through the end of September, then steep increase since then: 14 on Oct. 5 to 27 on Oct. 23.

Region 9, northeast Illinois (Lake and McHenry counties/far north suburban Chicago): increase in June, level through August, drop through mid-Sept. and level to the end of September, then increasing from 7 on Oct. 5, to 13 on Oct. 23.

Region 10, suburban Cook County: level-ish through the end of August, a small drop through the end of September, then a jump from Oct. 2, at 23 to Oct, 22, at 49.

and Region 11, Chicago: level through July and August, drop in September to a low of 21 on Oct. 3, then up to 41 on Oct. 23.

(Note that the Wirepoints numbers are total hospitalizations; these are admissions on any given day.)

But, again, here’s suburban Cook County according to the IDPH dashboard:

Covid-like admissions, October 27 data. https://www.dph.illinois.gov/regionmetrics?regionID=10.

But at the same time, the Cook County Department of Public Health‘s own website’s reported hospitalizations have been level, showing no change other than a drop-off for the past week due presumably to lags in data reporting.  (Note: as of today, they have removed the data on hospitalizations; I have requested an explanation.)

What’s going on?  You’ll have to trust me that the CCDPH data was level, because it’s been removed, but is the fact that the state includes “covid-like illnesses” regardless of whether a patient has covid, causing an increase in the numbers?  I can well understand using this broader definition back when testing was difficult, but covid tests are no routine for anyone who enters a hospital even for unrelated reasons.

What’s more, here’s the equivalent graph for region 11, Chicago:

But here are the hospitalizations as reported by the city of Chicago (as downloaded here on October 27):

— and this, despite rising case numbers:

Ugh.

Again, are hospitalizations due to covid really on the uptick?  Or is it due to “covid-like illnesses”?  Or — benefit of the doubt here — is there something faulty about the “covid hospitalizations” figure even after tests have become available without practical limits in terms of hospital access?

Third, what about Europe?

Biden, and Trump opponents generally speaking, are willing to say that a considerable number of America’s covid infections and deaths can be blamed on Trump’s poor management of the pandemic, and it’s easy to point to countries which have had extremely low infection rates — Japan, Taiwan, South Korea.  It’s also easy to point to stunningly foolish things Trump has done and said, and the whole mask debacle, well, it’s been a debacle — insisting in March that masks were useless only to later on conclude they weren’t, but stoking substantial suspicion due to that prior insistence.

But claims that the US has been singularly incompetent in managing the pandemic are falling apart.

According to the Financial Times‘ website, measured on a cases-per-million basis, averaged over 7 days, the European Union’s rate equaled that of the US on October 11 or thereabouts.  Now it’s rate is substantially higher, at 284.2 cases per million, compared to 200.7 in the US (as of Oct. 22 and 23, respectively).  In fact, very few countries within Europe are lower than the US, and many of those which are, are seeing steady increases.  Even Germany, lauded for its successful handling of the pandemic early on, is now seeing a surge in cases, with a 14 day change of 191% for new cases and 198% for deaths, compared to an increase of 40% and 14% for cases and deaths, respectively, for the US, according to the New York Times.  (Remember, to increase by 191% is not to double, but to triple, that is, for the new case rate to be 3 times that of the original rate.)  That means that seems quite likely indeed for Germany to reach our level of cases relative to the population by Election Day, which would be ironic when Joe Biden claims that he would have had German-levels of success in avoiding infections in the U.S.

Why are cases spiking in Europe?

Does that mean that there’s nothing, really, that can be done but hope that a vaccine and/or an antibody treatment is approved, manufactured, and distributed?  What does an observation of the case increases in Europe mean for our decision-making about whether to shut down restaurants or merely restrict their capacity, or even to re-institute lockdowns?

This is where I end, as I don’t have answers to these questions, and I’ve achieved my short-term goal of writing now new information I’ve learned and issues I wanted to share.

coronavirus

Jane the Actuary General Interest

Contact Tracing is an Urgent Task. So Why Is the State Failing at It?

Illinois state capitol; public domain

Back a month ago, I wrote a commentary in the Chicago Tribune in which I criticized the governor’s complete lack of communication (and seeming lack of plans) regarding contact tracing, despite the mandate that to move to Phase 3 contact tracing must be implemented, and to move to Phase 4, contact tracing must be fully scaled-up (90% of new diagnoses).

In the meantime, the governor has shifted to statements that contact tracing is already underway at local Departments of Public Health, and has shifted to speaking of a 60% objective (e.g., on May 18 and May 29) as well as a doublespeak rewriting of objectives as reaching 90% of the 60% target (I can no longer find this cite), and relabeling the entire project as “‘a goal’ rather than a requirement” (according to a May 26 Tribune report).  However, the Restore Illinois official requirement remains unchanged.

I’ve become resigned to the fact that this is how politics works, that rather than announcing a change that involves an admission of failure and invites demands for other changes, it’s simply memory-holed.  And my anger has shifted from the lack of communication to the lack of urgency in the actions of the governor, the mayor, and the Cook County Board President.

With respect to the last of these, an article on June 11 at the Chicago Tribune was the first reporting on the Cook County Department of Public Health’s actions —  even when I looked just a few days prior there was no information available on the DPH website; now, the website announces that

CCDPH anticipates starting our first group of contact tracers by early August. Contact tracers will be brought on in groups of 50-100. CCDPH will have a full team by the fall.

Again, remember that this is supposed to be in place in order to move to Phase 4, which is otherwise being targeted for just two weeks from now.

Why is this taking so long?

In part, it appears to be the fault of the Illinois Department of Public Health taking nearly three months to allocate funding from the CARES Act, which passed in March.  But it appears, from the Tribune reporting of Preckwinkle’s statements, that the delay is because the county simply does not recognize the urgency of getting the program in place as soon as possible, and is instead using the program to promote social justice objectives even at the cost of delayed implementation.

Preckwinkle said the efforts, funded with a grant from the Illinois Department of Public Health, would focus extensively on disproportionately affected groups that have “experienced systemic racism,” including African Americans and Latinos, both in terms of tracing and hiring of new contact tracers. The program also will be bilingual so hundreds of thousands of Spanish-speaking residents are not left out.

“This grant is so important for those who have been most impacted by COVID-19,” said Dr. Kiran Joshi, one of two senior medical officers running the county Department of Public Health, who said blacks in the county have been affected at three times the rate of whites and Latinos at four times the rate. “We intend to hire suburban Cook County residents for these jobs who are culturally competent, multilingual and have great communication skills.”

The county, however, will take several months to ramp up the program, even though many social-distancing restrictions have been lifted by the state and there’s concern that a future surge could occur soon because of recent crowded conditions during protests over the Minneapolis police killing of George Floyd.

Now, I well understand the importance of hiring tracers who can gain the trust of the tracees, because a program in which individuals are contacted but refuse assistance to enable them to isolate and refuse to provide information about their contacts because they don’t trust the tracer and can’t be persuaded that the greater good of their community warrants these actions, is fairly useless.  But Preckwinkle’s statement goes beyond this acknowledgement to a desire to use the program to advance broader social goals.  And that’s wrong — the top priority should be speed, regardless of whether goals of equal opportunity or extra assistance to underrepresented groups must be sacrificed.

In fact, like it or not, it is likely that a focus should be on disproportionately less affected communities, as the low-hanging fruit, with far more payoff in terms of the effectiveness of the effort.  It seems to me even more the case state-wide, that nipping in the bud an incipient outbreak in a community that’s otherwise been uneffected would be more successful than the greater challenge of dense urban areas with a pre-existing substantial prevalence.

And it’s not just suburban Cook County — in Chicago itself, the process of hiring contact tracers is set to take much longer than it should, due to a process of first identifying an organization with which to contract out the primary organization of the effort, and then distributing funds to

at least 30 neighborhood-based organizations located within, or primarily serving residents of, communities of high economic hardship

which would work at

recruiting, hiring and supporting a workforce of 600 contact tracers, supervisors and referral coordinators to support an operation that has the capacity to trace 4,500 new contacts per day

with an objective of hiring 150 by August 1, and 300 by September 15.

And, again, quite apart from the appropriateness of prioritizing workers from low-income communities for city jobs, in general, contact tracing is not just a city jobs program.  It is an urgent task.  The work of hiring tracers should have been started months ago, not months in the future.

What’s more, even this plan is being criticized by Chicago activists, who want the hiring to be done within the Chicago Department of Public Health itself, rather than being outsourced, and who are treating this as a matter of shoring up governmental institutions.

Also joining the group were current and former union officials who have an interest in seeing the ranks of public workers expand. They included Tony Johnston, president of the Cook County College Teachers Union, who said city community colleges should be training new contact tracers, and Matt Brandon, former secretary-treasurer of International Service Employees Union Local 73 and current president of Communities Organized to Win.

Contact tracing is not a jobs program.  It is not a stimulus program.  It is not an economic rebuilding program for poor communities.  It is certainly not a program for building up a unionized workforce.  And city, county, and state government officials who treat it as such, rather than ramping up tracing as quickly as possible, during this limited window of opportunity of lowered infection rates due to lockdowns and warmer weather, are failing the people they serve.

Jane the Actuary General Interest

Has Pritzker Abandoned Contact Tracing as a “Restore Illinois” Requirement?

Illinois state capitol; public domain

It’s right there in black and white:  contact tracing is a key part of Illinois Gov. JB Pritzker’s “Restore Illinois” plan.  To move from Phase 2 to Phase 3, permitting the opening of child care, retail, and gatherings of 10 or fewer people, requires the beginnings of “contact tracing and monitoring within 24 hours of diagnosis.”  To move from Phase 3 to Phase 4, permitting the opening of restaurants, personal care services, health clubs, and schools, as well as gatherings of 50 or fewer people, requires fully scaled-up contract tracing, that is, “for more than 90% of cases in region.”

But as I wrote last Saturday at the Chicago Tribune, however crucial contact tracing is, the state has provided virtually no information on its timing or its progress in implementing the program.

Only just today did the Department of Public Health provide a press release on the topic (can I take credit for this?), informing residents that county public health departments will actually be running the initiative, with funding and technical support from the state, and with Partners in Health in an advisory role.  Two specific counties will be “immediately” piloting the program.  The governor further stated at today’s (Monday’s) press briefing that at present 29% of diagnoses are “engaged in a tracing process” and “that’s a number we want to push as high as possible, to the industry standard of over 60%.”

Despite this, last week Pritzker announced that “all regions across the state are now on track to meet the metrics needed to move into the next phase of reopening.”

How does this make sense?  With only 11 days until the first possible “Phase 3” date, and with only a 2-county pilot program in place, how can the state be on track to meet its Phase 3 contact tracing requirement?

And how does a verbal target of aiming for “the industry standard of 60%” match up with the Phase 4 requirement of 90%?

What’s more, the state provides regular updates to metrics in the areas of testing and hospital admissions and resources, but no updates on contact tracing.

It’s as if they’ve forgotten about these requirements.

Has the state abandoned them, that is, continuing to strive for additional capacity but no longer requiring implementation/scaling to move to the next phase?

And, if so, why is the state not revising its plan, but instead simply treating them as if they don’t exist?

My guess:  the governor knows there is tremendous pressure to revise other components of the plan:  the inclusion of very geographically distinct counties adjacent to “collar counties” in the same region as Chicago, the continued closure of restaurants until Phase 4, the limitation on gatherings to 50 persons regardless of the capacity of a given facility, and so on.  Were he to revise the contact tracing component, he would further increase calls for revisions of other sorts.  So long as no one with any particularly strong voice or much political power calls him on this, he continues to be enabled to insist that his plan is unchangeable, set in stone, rather than risking opening it up to the sort of negotiation which he insists is impossible because he is guided solely by “science” and “data”.

Now, this is an admittedly cynical answer, but I can’t make sense of this any other way.  And, much as I hate for it to be true, as it implicates a wide range of bureaucrats as well in this convenient omission, it does, at the same time, offer some firmer reason to believe that, however painfully delayed Phases 3 and 4 are implemented, however many restaurants and other small business will shut down, it will at least not be delayed even further.