About the Governor’s New Metrics…

by James A. Bacon

Governor Ralph Northam presented Friday his “Forward Virginia” blueprint for reopening Virginia’s economy when the COVID-19 epidemic recedes. As Virginia sees progress in five key metrics, the Governor says, he will relax his emergency restrictions in phases. The first of four phases would allow some businesses to re-open with “strict safety restrictions.”

“We will move forward, but in a way that prioritizes public health and builds public confidence,” said Northam. “Businesses know that customers will return only when they feel that it is safe to do so. Our blueprint for the path forward is data-driven and provides clear guidance, so Virginians will know what to expect and understand how we will decide to when to lift certain public health restrictions.”

Adopting a data-driven approach sounds reasonable in the abstract. The devil. of course, is in the details. The logic behind the Governor’s metrics is not self-evident, and may require more explanation. According to the Blueprint, the five metrics are:

  • Downward trend: Percentage of positive tests over 14 days
  • Downward trend: Hospitalizations over 14 days
  • Increased testing and tracing
  • Enough hospital beds & intensive care capacity
  • Increasing & sustainable supply of PPE

Here’s what the Governor’s press release offers by way of explanation:

Virginia is looking at a wide range of public health data. The Governor emphasized that key indicators will include a 14-day downward trend in confirmed cases as a percentage of overall tests and in reduced COVID-19 hospitalizations. While hospitalization rates have largely stabilized in the Commonwealth, confirmed cases continue to rise. …

To ensure the continued safety of Virginians, the Commonwealth aims to test at least 10,000 individuals per day. Karen Remley, former Commissioner of Health and current co-chair of Virginia’s Testing Work Group, outlined a four stage approach to meet this goal prior to safe reopening. The expanded testing plan includes hiring contact tracers, who will support local health departments in identifying individuals who may be exposed to COVID-19 and helping them self-isolate.

First question: How did the Governor pick the 10,000 number? Do we need 10,000 tests daily to draw statistically valid inferences about the prevalence of the disease, or was this number plucked out of thin air because it sounds good? Neither the governor’s press release nor news accounts address that question.

Now, let’s discuss these metrics one by one.

Downward trend in percentage of positive tests over 14 days.

This graph shows the percentage of positive COVID-19 tests since mid-March.

The percentage is highly volatile but has been heading in a higher direction. While the percentage of positive tests may reflect the underlying prevalence of the virus, it is also influenced by extraneous factors such as the number of tests conducted, the protocol behind the tests, and who is conducting the test.

The Virginia Department of Health will have a testing capacity of 400 to 500 per day. The other 9,500 will have to come from Virginia hospitals and private labs. Will these diverse groups are use the same or different technologies? Will their testing protocols be the same? In the past, the high percentage of positive tests has been tied to the priority given to patients showing COVID-like symptoms, health care practitioners in hospitals, and patients in prisons, nursing homes and other high-risk groups. Does the VDH propose testing the same groups with the same protocols, or does it propose shifting to a random testing of the general population?

I have argued previously that the number of positive tests is a meaningless indicator of the virus’ spread, and the same applies to the percentage of positive tests. Furthermore, it strikes me as a metric of limited value. What matters is not the percentage of the population that has contracted the virus (unless we’re trying to calculate herd immunity, which the Governor has not mentioned) but the number of people so severely afflicted that they go to the hospital…. which brings us the the next metric.

Downward trend: Hospitalizations over 14 days.

What does “hospitalizations” mean? Are we talking about the trend in patients admitted that day? Are we referring to the number of hospitalized patients confirmed by tests to have the disease? Are we talking about the total number of confirmed patients plus those being tested for the disease?

We published the graph below in a separate post earlier today. The red line shows the number of confirmed + being tested patients. Although the number has crept higher, it is far less than the number of available beds, which numbered 5,343 in data published yesterday.

What significance is the statewide average anyway? The prevalence of the disease and the burden it poses on local hospitals varies widely by region. In Northern Virginia, hospitals are far more stressed than, say, in Southwest Virginia. Emergency measures that might be appropriate for one might be inappropriate for the other.

Increased testing and tracing

The Governor says he will not relax emergency measures he he sees “increased testing and tracing.” Specifically, he proposes hiring “contact tracers” who will “support local health departments in identifying individuals who may be exposed to COVID-19 and helping them self-isolate.”

The value of contact tracing is evident in the early stages of an epidemic when it may be possible to identify rare individuals with a disease, isolate them, and alert others who may have been exposed to them. But once the virus has broken out into the community — when literally thousands or tens of thousands have it — what’s the point of tracing? The epidemic reaches a point where it is moving too fast to trace. The Governor has not explained how he expects this measure will help to control the spread of the disease at this stage of its spread.

Finally, what metric will the Governor be looking at? The number of VDH contact tracers? The number of patients contacted? The number of potential exposures traced? If we select a measure, do we know what constitutes a number that would justify relaxing the shutdown?

Enough hospital beds & intensive care capacity

Of all the metrics, this is the easiest to measure and track. Refer again to the graph above. The green and blue lines show the number of ICUs and ventilators in use. They are heading down. Hospitals have stated loudly and clearly that capacity is not an issue.

Increasing & sustainable supply of PPE

Personal protective gear is delivered to Virginia in large but infrequent batches. According to the press release, Virginia has ordered 17.4 million N95 masks, 8.3 million surgical masks, 17.1 million gloves, 1.7 million gowns, and 1 million face shields in a joint contract with Maryland and Washington, D.C. A second shipment from Northfield Medical Manufacturing, scheduled to arrive Friday, was supposed to include three million titrile exam gloves, 100,000 N95 masks, 500,000 3-ply procedure masks, and 40,000 isolation gowns.

So, which of these items will be measured — N95 masks, surgical masks, gloves, or gowns? Or will VDH measure as “basket” of equipment? What if hospitals have an ample supply of one category but a shortage in another? 

Will the VDH metric encompass state supplies only, or will it include equipment that hospitals are able to acquire on their own? And how do we define “sustainable” anyway? Must the equipment be contracted for? Or must it be delivered and stockpiled?

Alternatively, why not use the Virginia Hospital and Healthcare Association metric? The VHHA reports the number of hospitals “experiencing difficulty obtaining or replenishing PPE in the next 72 hours.” This morning, the VHHA indicated that only two hospitals were having difficulty.

In summary, the Governor’s metrics are only conceptual at this point. Many, many details need to be worked out to give them meaning as decision-making tools. Once the Governor’s people do work them out, it would be very helpful if the VDH added them to its COVID-19 dashboard.

There are currently no comments highlighted.

25 responses to “About the Governor’s New Metrics…

  1. Since we have no idea what the relationship is between hospitalized and total non-hospitalized with the virus at any given time, increased testing could be biased whatever direction the governor wants by selection criteria.

    Why not start by adding to the current test total the tested negative results which “may not be represented due to the large number of fax reports from new labs testing for COVID-19” and see where that ends up?

  2. You raise an important issue, which is regional differences. Since I don’t think COVID ill patients are being moved regionally to balance the load, then why not segment Virginia according to the areas that naturally–from a geographic perspective–feed into the various hospital systems? Measure each of those hospital-system-defined regions and free them up as they meet the metrics; then allow the local leaders to make their decisions from there.

    Just as the more rural states have an argument that they shouldn’t be under same restrictions as NY, an even more granular approach within states is needed. Otherwise, we are unnecessarily holding back regions that can go back to work.

    • The biggest reason I see against different regional rules is the number of people commuting in and out of different regions. What information do we have on the numbers not commuting now who would if restrictions are lifted?

      Mathews has 4 cases, 1 in hospital. People here say open up. But of our 8,800 population, 2,200 workers commute out to other areas, including Hampton, Henrico, James City County, and 656 “elsewhere.” 631 commute in with 200 from “elsewhere.”
      Hampton has 37,174 out-commuters including to James City, and Fairfax Counties, with 6,951 to Elsewhere. 7,163 in-commuters are from Elsewhere.

      Fairfax County has 258,412 out-commuters. 72,676 to DC, 17,460 to Montgomery MD, 14,814 to Prince William, 3,256 to Henrico, and over 100K to several NOVA areas, but 36,444 are just listed as elsewhere. 94,549 of in-commuters reside elsewhere.

      Henrico has 86,129 out-commuters, 120,793 in-commuters.

      How much community transmission already happened from in and out commuters at gas stations, convenience stores, coffee shops? How many people would even think to mention those places in contact tracing?

      https://demographics.coopercenter.org/sites/demographics/files/2018-15-03-CommutingReport_Revised_UVACooperCtr_0.pdf

      • I do a fair amount of RV camping. Right now, the RV folks are distributing the names of places that are “open” and they’re in’ the process of heading to those places en masse.

        Similarly, people that live in one place and yearn for sit-down restaurants and other less restrictive social-distancing are going to take a “vacation” to other places, motels, B and Bs, etc… If beach motels are open in SC and GA – you can bet on what is going to happen.

        My bet is that even Mathews has a number of second homes and if they “open up”, a lot of those folks are going to leave the more restrictive places and go to their homes in Mathews?

        Perhaps that would be considered a “good” thing to getting their local economy going again,

        But how much risk is there in doing that?

        • Already happened. There are RVs in yards all over the county visiting family and friends. So not too much risk until the campgrounds open. And maybe not even then.

          • What I am hearing is that where the campgrounds are – there is going to be an increase in “mixing” at stores and restaurants, etc and that if more infections occur – the hospitals don’t have capacity in those smaller areas.

            Are we in favor of places like Bush Gardens to re-open?

  3. Seems like if all you are doing is counting things like hospitalizations, ICUs, ventilators – you’re just tracking the spread of the virus.

    What good is that?

    Social distancing is designed to minimize the spread of the infection and contact tracing is counting how many are becoming infected and attempting to isolate them to keep them from infecting others.

    It’s the scientists that are advocating both social distancing AND contact tracing AND it is said to be the reason South Korea and Germany have been successful in their efforts.

    re: different rules for different regions.

    Here’s a question. Roanoke.. Lynchburg “open up” and open their restaurants and barber shops, etc… and NoVa does not. Does anyone see a problem with NoVa folks making a trip down to the “open” places?

    What’s the plan for that?

  4. I may be thinking about this the wrong way, but until there is a vaccine, all that can be done is to keep the numbers of ill and gravely ill to manageable levels. We can’t bring the world to a halt until there is a vaccine. We’ll kill the patient to cure the disease otherwise.

    We have to start re-opening in measured fashion where risk is lower. And in doing so, yes, people will come and go within and from outside of ‘opened regions’ and the disease will certainly spread due to inter-regional mixing, but in my mind that’s inevitable. The key is testing and measuring. If the metrics start going the wrong way for a region, then it has to reinstitute ‘stay safe’-type measures. It’s cold-hearted to some degree that you know people will get sick and, even worse, die, but we can’t stay holed up for 12-24 months waiting on a vaccine. And non-granular approaches–at the state or national level–aren’t going to be the ultimate solution, esp. for states like VA where there is a mix of rural and urban.

    • The problem is – if the current level of infections and deaths is lower because we are socially-distancing, what happens when we start doing it less?

      You say increased risk. Are you willing to say – increased deaths?

      Are you essentially advocated more deaths now – because you think we’ll get those deaths downstream anyhow?

      When we say “holed up” for 12-24 months, what does that mean?

      Does it mean the economy will not adjust and recover in a new way?

      In terms of a vaccine – what if we never get one?

      Do we consciously agree that we ARE going to have more infected and deaths than we can handle in the hospitals?

      Is the current strategy of trying to spread out the infections so we do not overwhelm the hospitals – wrong and instead we just go forward and some folks will literally die for lack of a hospital bed?

      There are no easy answers but the question is are we making rational choices?

      • I’ll touch on your points in order, but in general there is a stark choice between staying locked down until metrics for the whole state are acceptable and being more sophisticated in our appraoch. What if NoVA has a massive increase, but rest of the state is minimal and going down? Are you going to force stay at home for the entire state? That makes no sense. It’s too blunt of a policy instrument. The re-opening can be more nuanced than where we are headed.

        -lower #’s due to social distancing (SD): yes, numbers are lower due to SD. Flattening the curve is the right thing, but some areas have essentially zero cases and have ample hospital capacity. We can’t stay in the current state of SD. We’ll start a depression if we do.

        -higher risk=increased deaths: yes, it does mean that, but I’ll keep going back to the “killing the patient analogy”. Is avoiding a depression and the agony that would bring, worth increased risk? In my mind yes.

        -what does ‘holed up’ mean: it means we are under stay at home orders and ‘holed up’ in our homes.

        -new economy: the economy will die before it adjusts to the new normal, because a large enough portion of it will cease to exist if stay at home orders persist. If unemployment stays at 15-20% for 12 months, it will be a disaster, but someone knows better than me what the tipping point actually is.

        -no vaccine shows up: even more to my point and reason to start reopening on a regional basis where possible

        -hospitals will be overwhelmed: no, I don’t agree with that statement. That’s part of the process with my suggestion. When the metrics start to turn south for a region, ‘stay safe’ is reinstated. I wholeheartedly agree with flattening the curve, but many places have no curve. I’m suggesting the plan be more targeted than just at the state level. I think that makes sense in Virginia.

        • re:
          “there is a stark choice between staying locked down until metrics for the whole state are acceptable and being more sophisticated in our appraoch. What if NoVA has a massive increase, but rest of the state is minimal and going down? Are you going to force stay at home for the entire state?”

          Is the state as a whole – an either / or proposition?

          “That makes no sense. It’s too blunt of a policy instrument. The re-opening can be more nuanced than where we are headed.”

          is the virus a blunt instrument? Are we to do this irregardless of the reality of the virus?

          “-lower #’s due to social distancing (SD): yes, numbers are lower due to SD. Flattening the curve is the right thing, but some areas have essentially zero cases and have ample hospital capacity. We can’t stay in the current state of SD. We’ll start a depression if we do.”

          I agree on that – right now – what is the plan if they “follow” the other areas and start to ramp up also?

          “-higher risk=increased deaths: yes, it does mean that, but I’ll keep going back to the “killing the patient analogy”. Is avoiding a depression and the agony that would bring, worth increased risk? In my mind yes.”

          Do you know how much death you are voting for?

          “-what does ‘holed up’ mean: it means we are under stay at home orders and ‘holed up’ in our homes.”

          except that’s not true. I see LOTS and LOTS of people in the grocery stores, the home improvement stores, walking in the parks, etc… most of us are not “holed” up.. but we are less free than we used to be but the bigger point is – if “holed” up means social distancing and social distancing means less infections and less deaths – what exactly are you advocating as a remedy?

          “-new economy: the economy will die before it adjusts to the new normal, because a large enough portion of it will cease to exist if stay at home orders persist. If unemployment stays at 15-20% for 12 months, it will be a disaster, but someone knows better than me what the tipping point actually is.”

          I do not know and I do not diminish it at all but I do think the economy will adjust – it already has been. this is just accelerating it.

          “-no vaccine shows up: even more to my point and reason to start reopening on a regional basis where possible”

          if “opening up” means a huge increase infections and death…. what then?

          “-hospitals will be overwhelmed: no, I don’t agree with that statement. That’s part of the process with my suggestion. When the metrics start to turn south for a region, ‘stay safe’ is reinstated. I wholeheartedly agree with flattening the curve, but many places have no curve. I’m suggesting the plan be more targeted than just at the state level. I think that makes sense in Virginia.”

          so if infections and death skyrocket – you will then shut down again?

          what are the metrics for that? Are the states that are opening up – do they have such “shut down again” metrics?

        • I think the idea that we could “get back the economy” by starting to “open up” the areas which don’t have a lot of cases – is probably missing an important aspect and that is most of these areas of low infection are also low population and low numbers of jobs.

          Most of the economy of Virginia is in the Metro Areas – probably 6 million or more of the 8.5 million and probably more than that in terms of economic activity.

          Opening up the less-urban areas is not going to save the economy because the economic activity in those places is minimal to start with.

          This thing is on the virus terms – not ours.

          • “Opening up the less-urban areas is not going to save the economy because the economic activity in those places is minimal to start with.”

            Maybe opening up the less-urban areas would help save the economies of… the less urban areas. Ever think of that?

          • That’s a legitimate point but most of these places don’t really have a viable economy to start with beyond some sit-down restaurants and a regional Walmart.

            They still have their local govt, education, post-office, trades, auto repair, etc… what else is not open?

            Walmart has pretty much decimated most of the mom/pop already.

            I just don’t think it’s going to do much but perhaps if you made it local option – which I’m pretty sure they would try to keep people who do not live there, away.

            agree?

          • I think the crux of the issue between our views is this:
            “I do think the economy will adjust – it already has been. this is just accelerating it.”

            I see it accelerating into a deep, deep hole and I’m not sure you share that view. Whether right or wrong, I think that’s why I’m willing to be more aggressive.

            None of this is easy, but i should also say that I’m really glad to see Northam starting down a path towards reopening. I’m just advocating a more granular approach that attempts to recognize ‘one size doesn’t fit all”

          • re: ” I think the crux of the issue between our views is this:
            “I do think the economy will adjust – it already has been. this is just accelerating it.”

            I see it accelerating into a deep, deep hole and I’m not sure you share that view. Whether right or wrong, I think that’s why I’m willing to be more aggressive.”

            I truly don’t know. I only know the various opinions I hear from various folks who have economic credentials and a significant difference is that this downturn was NOT from an economic cause but more like a war or other external force.

            The predictions of how much more damage will happen are indeed sobering and one would be a fool to heed them. We have not seen the worst yet.

            Still – large parts of the economy are still functioning and look like they will continue to.

            “None of this is easy, but i should also say that I’m really glad to see Northam starting down a path towards reopening. I’m just advocating a more granular approach that attempts to recognize ‘one size doesn’t fit all” ”

            I just think the bigger places with more population and more jobs is where the focus has to be.

            Not really opposed to different rules for places that don’t have significant virus – with the stated concern that – just like prisons in rural areas – small towns can end up the same way – one or two people who visit from somewhere else can start it and then once that happens – some of the people who live there will leave and go elsewhere to escape the contagion.

            And these small areas are not going to get the larger economy going – it’s the thousands and thousands that are unemployed in the big METRO areas that is the problem with the economy.

            So the regional /one size thing is a bit symbolic and if you listen to folks who want that – they actually want MORE than that – they want the rules loosened across the board – so we can get the economy back.

            And that’s just not going to happen even if Govt would just take all the restrictions away – the virus is going to run wild…

            it’s just an awful reality – that we don’t want to accept.

            there just are no good answers.

            We’re frustrated and mad and want to blame someone.

  5. With regard to rural “opening up”. If most of the states population is in the dense areas like NoVa, “opening up” the rural is not really going to help the economy, right?

    Would it be more effective to loosen restrictions on the places with the most population and jobs?

    If I show you a map of how many were actually unemployed on a county by county basis in Virginia what would it show in terms of where the largest number of unemployed were?

    Not rural. Right?

  6. Some people can never be satisfied. Finally, the Governor has come up with a rational plan to begin opening the economy and he is not getting any credit for that. Yes, the outline leaves rooms for questions, but you have to remember the context for this roll-out: a press conference with Power Point slides. That does not lend itself to a lot of detail (unless you 50 slides and an hour discussion). I would like to think that the Governor’s team is working out the details.

    To be meaningful, the percentage of positives should come from a random sample, not just those showing symptoms. It does not have to be 10,000 tests. He did not say that the state had to be administering 10,000 tests per day in order for the criteria to be satisfied. The 10,000 is a goal. The number tested to determine the number of positive cases could be smaller. This is not a tightly controlled experiment, so the technologies used by the various testers, state, university, private labs, do not have to be the same. The only requirement is that the groups tested should be quasi-random. You would need to represent multiple age groups and ethnicities. Ideally, it would be regional, but any regional boundaries would be arbitrary and it would be harder to coordinate. Also, there will be regional cross-over as other commenters have noted.

    Hospitalizations–measure new hospitalizations. That is the trend we should be looking at.

    Hopefully, the mechanics for working out the testing metric will be worked out soon and the VDH website will begin showing the results.

  7. Here’s a question about testing.

    Are we testing solely to figure out how many infected/hospitalized there are

    or are we ALSO testing to find out as many infected as we can before they infect other and get them isolated?

    If we’re only going to test to find out how many are infected – does that imply there really is not real plan for “opening up” – i.e. just open up and see what happens?

    • They are going to disinfect them.

    • I think there are two reasons for it First, to get an indication if community spread is increasing or decreasing. Second would be to find those infected by the virus and isolate them and, to the extent possible, determine who they had had contact with.

  8. that was my understanding also but seems like the idea of contact-tracing is not taken seriously by more than a few who apparently feel that “testing” is just something we do to keep track of the level of infection and basically we’re just going to ride it out – and the metrics are only to tell us where we are on that journey.

  9. Four EUA, 107 antibody test devices didn’t even bother.

    https://www.npr.org/sections/coronavirus-live-updates/2020/04/26/845164212/antibody-tests-go-to-market-largely-unregulated-warns-house-subcommittee-chair

    Yeah, I know, the free market will sort it out… by 2045.

  10. “CBS News’ Weijia Jiang was the only reporter who was able to get a question in to anyone. She asked Hahn if it was true that the FDA has no review of the antibody tests on the market, and there is no way to check their accuracy.

    Hahn said under the policy, manufacturers have to validate their tests, tell the FDA the tests have been validated and let people know that those tests were not authorized by the FDA.

    Additionally, Hahn said the FDA has authorized four tests and there are “more in the pipeline.” Hahn said they are working to approve the tests that have been “self-validated.”

  11. The next time somebody tells you that the unconstitutional quarantine of the healthy should continue … ask them how they are, or are planning to, fund their retirement. A shocking number will reply that they have government pensions. In other words, they expect the government (using threat of force) to confiscate enough private property to keep them comfortable no matter what happens to the economy.

Leave a Reply