by Hamilton E. Davis
A Vermont Journal embarked on these Virus Papers less than a month ago on the assumption that my lengthy apprenticeship in health policy might enable me to make a contribution to the process of fending off the Covid-19 threat to our Green Mountain Home. It has been a wild ride, but I think the fundamental issues of epidemic management have been solved. That doesn’t mean the battle is over—it isn’t.
But it seems time to sum up; to lay out the lessons learned and what the run-out looks like over the next few months, beyond which no one has a clue. So, what did the Virus Papers amount to, and where do we stand now?
As of the week of April 20, Governor Phil Scott, to use his own metaphor, is opening the spigot a quarter turn at a time—permitting operations by such relatively manageable categories as small construction crews, lawyers, and accountants. The Scott team has wrapped this move in a blanket of exhortations to Vermonters to maintain their vigilance by social distancing, wearing masks and the like. The Scotties are terrified (not too strong a word) by the risk that lifting the pressure on the brakes too early will generate a resurgence of the virus. That worry is justified, since we don’t fully understand the virus itself, let alone how to manage into the summer and fall.
Still, it seems clear to me that the Scotties have a pretty fat margin of error. The original yield from the epidemiological model was a trio of curves denoting the unconstrained progress of the virus (catastrophic), most likely scenario (manageable by the health care system), and best case scenario (matching South Korea, which would be out of reach). In fact, it now appears that we are on a track (535) below the best-case scenario.
I say appears because the state reporting on the output of our own epi model has shifted away from the backside of the curve and the run-out, which would show where the new hospitalization for new cases approaches zero. The actual data is obviously valuable, but so are the projections of the model; if they are not, we shouldn’t be paying for the damn thing. (We are already in the zone where we are spending money we don’t have). I would be willing to bet that the model shows a statistical return to normal--zero or very near zero on the x-axis for the key metrics--by June 1. (more about this below)
Moreover, the actual data still conflates the presence of the virus in totally contained populations like the inmates of state prisons and nursing homes and the general public. That is ridiculous. The prison inmates are obviously highly vulnerable because they are closely packed in, but they aren’t going anywhere because they are being guarded, literally, by people with guns. The situation with nursing homes is equally clear. So why isn’t the real-world data available to the public? Speaking of which, why has Mike Pieciak, the Commissioner of the Department of Financial Regulation and the chief wrangler of the epi model, been absent from the trice weekly press conferences?
None of this means that Scott should exceed his very cautious “quarter turn of the spigot at a time” approach, but the fact is that we cannot avoid some risk in getting our state economy back in action. And the Scott team, which obviously has the run-out curves from the model, should share them with Vermonters. The Scott team has been doing a good job, but it is the people of the state who are doing a lot of the heavy lifting every day and they should be partners with the Scotties in the effort. The Scotties are sharing some of the output curves from its model, but not the model itself.
The relevance of this distinction can be seen in the two figures below. The first shows what seems to me the most informative curve from the Department of Financial Regulation web site. It shows hospital beds occupied by Covid patients trending toward low single digits over the next several weeks. That is the dashboard reading available to the Governor as he contemplates the next quarter turn of the spigot.
At least part of the risk, however, could be found in Wednesday’s New York Times, which compared the output of several models in use across the country. They included epi versions from MIT, Columbia University, lmperial College (London) and Northeastern University. Our model is Northeastern, and to the inexpert eye (mine) it looks like our model is the most optimistic one in use. Someone should ask Mike Pieciak whether he is concerned about that. The Times survey is national, however, and the actual data from Vermont looks every bit as optimistic as the model is showing…still, worth thinking about.
Credit: New York Times “What 5 Coronavirus Models Say the Next Month Will Look Like” 4/22/20
Further Reflections
Re: contact tracing: It has been a winding path, but we now have a clear picture of where we stand on the one of the most critical tools available to the virus managers. In full-on contact tracing, you interview anybody who has the Covid disease about anyone they have had recent contact with.
And then, you test those contacts, all of them.
That is how the South Koreans stopped the virus in its tracks, and it is what Germany is doing now. In a recent press conference Dr. Mark Levine, the health commissioner, said that in the Vermont effort, they interview the contacts and then recommend that those people self-quarantine for 14 days. That is a good thing, and a clear advance on the pre-tracing situation, but it is not full-on contact tracing.
The obvious hole in the donut is the asymptomatic carrier, not to mention the inevitable leakage in self-quarantining: “Well, I had to get some milk and my car is out of gas…” The asymptomatic carrier is the core of the core of the problem going forward. In a recent report, The New York Times said the reason the federal Centers for Disease Control shifted its position on the wearing of masks by the public was the asymptomatic carrier. The problem, of course, is that testing all the contacts would require a dramatic increase in the use of the testing reagent, which Vermont doesn’t have. New York Governor Andrew Cuomo made the point last Friday: “You have to develop a testing capacity that does not now exist.” So, Vermont is doing the best it can with what it has available. But, contra Doctor Levine, it is not full contact tracing, and, contra Mike Smith, the testing program is not “robust.” Were it so…
The Dartmouth-Hitchcock conundrum: When the Scotties first laid out the medical capacity available to deal with the virus, they toted up the number of hospital beds available, the number of ICUs and the number of ventilators. They did not, however, include the resources potentially available at Dartmouth, which is just across the Connecticut River in New Hampshire. There were two reasons for that. One was uncertainty about whether the State of New Hampshire, in the face of an emergency, might force D-H to limit their capacity to New Hampshire residents. A second, I believe, was that the Scotties were trying to build in all the redundancy and fat margin capacity they could to aid both in the management of the virus, and the state of public morale in Vermont.
Now that we are on the downhill side of the curve, it is critical that the Scotties figure out and regularize the role of D-H in the Vermont delivery system. The whole eastern part of the state, and a not insignificant part of the rest of the state, utilizes D-H, mostly for high end, tertiary services. Forty percent of Dartmouth’s patient traffic consists of Vermonters. Yet, we know almost nothing about it. That is a hole in the management of the health reform project; but it is also a vital element in post-virus planning. To that end, we need to see and understand the flow of Covid patients into D-H over the last month. How many Vermonters went there for Covid treatment? Was it stratified—that is, did more serious patients bypass local hospitals in Vermont to get to D-H? What was the D-H ventilator capacity, and what percentage of that was used by Vermont patients.? ICUs used?
A related issue: We’ve had relatively little actual medical reporting on the virus. There has been some of that nationally, but it was thin there also. The disease attacks the lungs, but in an idiosyncratic way. And it’s tricky—it can also affect other organs; a victim may need kidney dialysis, for example, and it apparently can cause heart problems. And sneaky: I have seen stories about patients failing to understand the seriousness of their symptoms because the oxygen levels in their blood had dropped, without other symptoms being evident.
We haven’t seen what percentage of confirmed cases need a ventilator; and it would be good to understand the qualitative outcome for ventilator patients. The national press and the web have carried stories about patients surviving the ventilator, but at a catastrophic cost in quality of life.
Government: To an unusual degree, the Covid crisis has given Vermonters an intimate view of their own state government. That sense was amplified when the Vermont State College debacle broke into public view a few days ago…So, my general view is that Phil Scott has done an excellent job steering the ship through the virus challenge; I think the same is true of his team, not withstanding my caveats about whether the Scotties could bring their game up even more.
On a longer reach, however, I think that the combination of the virus management with the uproar about closing the two northern state colleges has illuminated a pattern in the performance of Phil Scott as Governor. In brief, it looks like this: Scott’s performance differs markedly depending on the type of problem he has to confront. On clear-cut, dramatic public issues, like the virus and the earlier problem about the threat of a possible shooting in a Vermont school, Scott stands right up—calm, firm, strong, gutsy. His performance on the guns issue was striking: he abandoned an earlier hands-off posture, advocated an important gun control, and then went out on the state house entrance to sign the bill, facing a very hostile crowd of gun advocates.
On the other hand, Scott’s reaction to the state college was the kind of mushy, evasive posture that marked his performance on earlier dust-ups, like his battle with the legislature over health insurance for school teachers. In the current state college issue, Scott said he didn’t support the proposal by Jeb Spaulding to close three northern institutions, but that he didn’t have any proposal of his own. The governor got a break in the form of a very short public notice on the issue by Spaulding, but that was just luck. A three-day notice. The horror!
In fact, it wasn’t a three-day notice; it was a three-year notice. The state college mess has been building for years; 25 years ago the faculty got paid less than high school teachers. So, even before the virus took over center stage, the Governor and his team had to know that the fiscal crisis for the colleges was cresting; and that nobody in town had any answer to it. The brute force answer: levy the taxes necessary to pay to keep the northern colleges running in a collapsing educational market. Scott contemplates that the way he would blowing a head gasket and all four tires when driving his race car at Thunder Road. Of course the usual subjects in the legislature went into full on hyper-ventilating mode…puffing themselves up, calling in the cameras.
The reality, however, is that nobody has the faintest clue how to make the state colleges dilemma go away, and after wasting a lot of time, and probably a lot of money, they’ll end up pretty close to the Spaulding plan. The lone note of sanity that I’ve seen was a comment to VTDigger by State Sen. Philip Baruth of Chittenden County:
“If anybody believes that we’re going to move this decision into the Legislature to vote to keep these campuses open and get away without a price tag of hundreds of millions of dollars you’re kidding yourself,” Baruth said.
The broader point is that Vermont lacks the capacity to deal with truly complex problems, like a declining rural population, the attendant drain on elementary and secondary education, the need to recast the community hospital system, the lack of modern telecommunications in rural areas, and, of course, the state colleges. On all of these hairballs, neither the Governor nor the Legislature has any idea what to do.
It used to be at least somewhat different. Governors used to have a fully staffed planning office to develop solutions to difficult problems, and very serious people worked there…that all gone now. As for the Legislature, it is amateur hour. The members are part-time, and they have no staff. The leadership of the House and Senate each have an assistant or two; and there is a small Joint Fiscal Office group that mainly serves the interest of the most powerful legislator. There can be very strong players in the various departments, but at the upper reaches of the executive there is no credible policy making expertise, especially on issues that cross department lines. Something to ponder.
Variation on a theme:
As we have seen in the past (I won’t bother with any more links), failures by the press can be extremely troublesome. In my last post, I cited the claim by Joe Woodin, the CEO of Copley Hospital, that the antibody test he was running on his staff was a “breakthrough, a game changer.” This claim was amplified by worshipful coverage in VTDigger. Others may be tempted to make the same claim. They may want to consider the warnings published over the last week or so.
From the Washington Post, April 19:
“Testing experts warn the risks of inaccurate testing are high. A wrong result could, for example, indicate individuals have immunity against the virus when they don’t, potentially resulting in behavior that would endanger themselves and their families…Even for high quality tests, scientists say, they don’t know the level of antibodies to make someone immune from the virus or how long the protection might last.”
From The New York Times:
In a story about the dangers and misuse of the antibody tests, the Times story said the little federal guidance on the use of antibody testing “is so confusing that health care providers are administering certain tests unaware they may not be authorized to do so. Some are misusing antibody tests results to diagnose the disease, not realizing they can miss the early stages of infection.
“People don’t understand how dangerous this test is,” said Michael T. Osterholm, an infectious disease expert at the University of Minnesota. “We sacrificed quality for speed, and in the end, when it’s peoples lives that are hanging in the balance, safety has to take precedence over speed.”
So, what do we still need to know? Probably a lot of things, but there are a few things that should be easy. They include the record of hospitalization of Covid patients by hospital, not just county; that information will be critical to health care reform going forward. And Mike Pieciak should disclose the inside of the black box of the Northeastern epi model. And given the very strong likelihood that both Dartmouth and UVM are running their own models (they couldn’t train infectious disease doctors or epidemiologists or virology specialists without them) the Scotties ought to disclose those also.