This week I am talking to Dr. Zoë McLaren, Ph.D. (@ZoeMcLaren) Associate Professor in the School of Public Policy at the University of Maryland, Baltimore County. Zoë has a broad and deep set of experiences when it comes to effectively gathering data to combat infectious disease, especially in resource-limited countries, and has commented and written on multiple occasions on the current COVID19 pandemic, “No, soaring COVID-19 cases are not due to more testing – they show a surging pandemic” and quoted here.
We discuss the details of the various testing strategies and the importance of leveraging available data and resources – an important point currently as we find ourselves without widespread testing availability that means the data available is limited and comes with limitations and built-in bias. As she describes it
Every test is surveillance data
But to extract the value requires understanding the context
We explore what success looks like and the importance of setting testing goals that are driven back from the ultimate goal of eliminating COVID19 disease and any deaths related to it. In an ideal circumstance, everyone would reliably know if they were contagious and carrying the SARS-CoV-2 virus at all times and would know the status of everyone prior to interacting with them in person. But achieving that is not possible so working back from that how do we decrease the risk, what tests do we use and how do we manage the differences in test result accuracy.
Listen in to hear Zoë prescription for testing moving forward that must include rapid, widely available testing and how we communicate these complex issues effectively to everyone and increase compliance and trust in the systems to protect everyone’s safety.
Listen live at 4:00 AM, 12:00 Noon or 8:00 PM ET, Monday through Friday for the next week at HealthcareNOW Radio. After that, you can listen on demand (See podcast information below.) Join the conversation on Twitter at #TheIncrementalist.
Listen along on HealthcareNowRadio or on SoundCloud
Raw Transcript
Nick van Terheyden
And
today I’m delighted to be joined by Dr. Zoe McLaren. She is an associate
professor in the School of Public Policy at the University of Maryland,
Baltimore County. So he thanks for joining me today. It’s a pleasure to be
here, neck. So I’m especially excited to interview you, you have such an
incredible background that is so relevant to the current crisis, with deep
insights. But for the benefit of the listeners, could you share a little bit
of, you know, what got you to this point as a professor in Maryland?
Zoe McLaren
Sure. My
PhD is in economics and public policy. And so that training was really about a
statistical analysis, and then thinking about kind of government priorities and
the policy process and politics as well. And so I focused on major challenges
primarily in low income countries around health policy and economic policy. So
that’s kind of my training background. And then I spent eight years on faculty
at the School of Public Health at the University of Michigan. And so there, I
got kind of even more exposure training work within public health. And so my
area of expertise really spans economics, statistics, policy, public health.
And then I also bring to kind of my research in my professional world, deep
empathy for people that are suffering through kind of economic crises and
public health crises, as well as an interest in kind of current events and
news. And so trying to get a sense of kind of what’s actually going on in real
life so that my policy expertise is informed not just from what I’ve read in a
book, or done research on but also kind of what’s going on from day to day. So
practical policy, and I think, you know, prior to COVID-19, you were focused
on, you know, some of the diseases that are impacting our populations, HIV and
TB, you know, so in many respects, COVID-19 wasn’t new from an approach
standpoint, that true. So that’s part of why I became involved in research on
COVID-19 is because of my background had been working on HIV and tuberculosis,
primarily in Africa. But a lot of the same lessons apply in terms of thinking
about, these are both infectious diseases, and population wide. And then also the
idea about kind of low resource environments, so we don’t have all the
resources we need to fight COVID-19, you didn’t when it arrived. And so
thinking about how to make do in those types of environments is something that
I have a lot of experience with. So thinking about that maybe you can’t do
first best. So you need to figure out what second best is that you have to
make? Do, you have to leverage all of the resources you have and use them
efficiently. So I think kind of my perspective and my experience, working with
those two, those two epidemics, really helped kind of inform my approach to
COVID-19. Tuberculosis is transmitted via the air and much the same way as
COVID-19 is, and then also, the HIV epidemic has important lessons as well.
Nick van Terheyden
Right.
And you know, the best available option is always available, perfect almost
never is and finding what’s appropriate within the resources. So let’s get
right to it. I mean, we have, you know, this pandemic, sweeping across the
constantly moving state of understanding and knowledge. And one of the areas
that you’ve really sort of commented a lot on, have deep insights is around
testing. And it’s really complicated from the general perspective, because,
first of all, we’ve got lots of different tests. And then should I be guessing
your test? Tell us as simply as you can, your thoughts around testing and how
we approach this from the covid 19 pandemic standpoint.
Zoe McLaren
And so I
come to this thinking about decision making. So the policy decisions that are
made by both the government and also the the population constituents
determining what they want to lobby, the government for, it is all about
decision making. And there’s a lot of insert uncertainty involved. And so what
I do is think about how we can leverage the data that we have to produce
evidence. So the data doesn’t always tell us exactly what’s going on, we need
to think about how to process that data in a way that tells us what’s actually
going on. And then once we know what’s going on how to translate that into
useful guidance for actual policy. So thinking what actually what the
government can can do and also what individuals can do. So that’s what I think
about a lot and testing is an very, very important source of information. And
you’ve heard the phrase knowledge is power. That is testing is incredibly
powerful because it produces knowledge and that knowledge can be used by
another have different people. So the person who gets the test gets knowledge
that helps them guide their own behavior and communicate with those around
them. When a test is positive, that information is generally reported back to
public health authorities that can make decisions about contact tracing about
allocation of resources about the need for potential hospital beds in the
future. And then information is also available for kind of more surveillance
prophecies where we want to let the public know about the the, the dynamics of
the epidemic. So everybody has looked at the case counts, the hospitalization
counts, the mortality counts, that information comes from our testing programs.
And that information is what people use all the time to determine whether they
should cancel Thanksgiving with their family, whether they should travel for
Christmas, how safe they feel, going out and interacting with other people how
important it is to wear a mask. And so testing provides information for all of
these components. So one single test feeds into all of those pathways. And
that’s one reason why testing is so important.
Nick van Terheyden
So when
you think about testing, and the current levels that we have, we hear sort of
different versions of positivity. Can you explain a little bit behind the
various positivity, right? So there’s, you know, the general population as a percentage,
what are those numbers mean? And how do we interpret them both on a policy
level, but also as an individual level? If I hear that the positivity rate is x
in my community? What are those things mean to you? And what how should we
interpret?
Zoe McLaren
So the
idea with all the data that we have, and so I think it’s important to think
about, well, we want to understand what the data mean, we need to think about
how that data are collected. And so right now, our data about COVID-19, comes
through testing. And we’re not doing very many kind of population surveys, not
everybody’s getting a test, people are individually deciding that they want a
test or need a test, and are going and seeking out a test. And that’s how
people end up in the data that we see. Or they’re realizing they need to go to
the hospital. And then they’re ending up in our data. So if we look at our
case, data, it really is about what proportion of the people who decide they
need a test or who are advised they should get a test are actually testing
positive. And that gives us some information about the overall prevalence in
the population. But it’s a biased estimate of the overall prevalence. So when
we see a test positivity rate of 12%, doesn’t mean that 12% of the population
is infected with COVID-19. It means actually much less than that. But we don’t
know exactly how much less. And so part of it is saying, well, when we have
this data that’s imperfect, it’s a biased estimate, is it totally useless? Or
is there still some information in this biased estimate that we can still use.
And so for example, if test positivity, we know that even though we can’t back
out directly from a particular number of pot, test positivity, what the actual
prevalence is that we know that when test positivity is high, it’s going to
suggest two things, one, that there are a lot of cases in the population. And
to that there are a lot of cases in the population that aren’t getting a test
and haven’t been tested and may not know, and maybe spreading kind of COVID-19,
unawares. And so that’s one kind of that we know a lot, a lot of cases and a
lot of cases spreading.
Nick van Terheyden
So you
obviously look internationally, you know, what we’re doing, you know, we have
our own process here, we’re not doing surveillance testing in this country,
other other countries that you can point to where they’ve actually managed to
implement that kind of policy that has helped inform some of the decision
making and to success.
Zoe McLaren
So I
think there’s a little bit of a misunderstanding, but what the word
surveillance testing means, I think you’re thinking about doing these kind of
random population samples to kind of get an idea of prevalence. But actually,
any testing data is technically some sort of surveillance data. And we have
different types of surveillance data we use. So sometimes you might use like an
early warning surveillance data, where if there’s very low prevalence of
something that if there’s ever a single case that needs to be reported, and
then we get kind of a warning, there’s cases in a particular area, we may do
these population wide surveys. That’s a way of doing surveillance. But actually
just the routine data that’s collected through testing, we also consider
surveillance data. But it’s the type of search for surveillance data that we
need to analyze more carefully, in order to determine what it tells us, even
though we can potentially follow trends of test positivity. So for example,
when test positivity is rising, that’s generally a concern when test positivity
is rising, and then all of a sudden it starts to fall. Well, that’s important
information we need to think about, well, what changed, why is it all of a
sudden kind of turning around and starting to fall? Is it something about
changes in testing? Is it about changes in prevalence is it about changes in
the demand for testing or who’s just determining that they need a test. And so
it does give us important surveillance information, but not this population
wide random sample or even universal testing where everyone is tested. There’s
a number of different We can do surveillance. And that’s kind of my academic
area of expertise to saying when we don’t have the type of surveillance data
that we want, that makes this stuff easy. What can we do with the kind of
biased and imperfect surveillance data we actually have? And so I’ve done work
on tuberculosis in South Africa saying, well, we just have the same thing, just
people who decided they needed TB test, got a TB test they’re in we don’t have
a random Population Survey. We aren’t doing any kind of particular Sentinel
testing. And so how do we take that information and learn about prevalence and
changes in prevalence from that?
Nick van Terheyden
So
that’s off, no restrictions, resources, capabilities compliance. In a perfect
world, what would perfect surveillance testing be for a population is that 100%
of the time every day?
Zoe McLaren
So for
surveillance, the purpose is to use surveillance testing, to figure out kind of
to learn what’s actually going on with the dynamics of the epidemic. So that’s
kind of one purpose of testing. But that’s surveillance testing is actually not
when I think right now, but what the goal is for testing, it’s not actually my
number one goal is actually not surveillance. My number one goal is making
testing widely accessible to everybody. Because knowing where the epidemic is,
does not save lives, as well as making sure that every individual knows whether
they’re contagious or not, and is able to change their behavior or get support
to be able to slow the transmission. So the ultimate purpose of testing in an
epidemic is to slow transmission is to save lives is to prevent people from
being hospitalized. And so we can use surveillance data to do that. But in some
ways, surveillance data is an indirect route to saving lives. And basically
getting information to people who need it to prevent the spread and, and
prevent cases is actually one of the best ways that we can employ testing.
Nick van Terheyden
So for those
of you just joining, I’m Dr. Nick the incrementalist and today I’m talking to
Dr. Zoe McLaren. She is an associate professor in the School of Public Policy
at the University of Maryland, Baltimore County, we were just talking about
testing, surveillance and the various forms of surveillance. And I think if I
heard you correctly, it doesn’t matter where this data is coming from, even
with the bias that comes with it through that pre selection process, there are
still means of extracting useful information that can inform public policy, and
even clearly individual decision making. So as you think about the current
testing program, and you know, obviously, the desire to increase that as much
as possible, first of all, how do we increase it? And then how do we measure
success of the testing program? Not in terms? I mean, success is, you know,
disappearance and COVID-19. I think that, you know, steps back from that, what
does success look like for you?
Zoe McLaren
So maybe
I’ll start by answering the question about success, and then, and then move on.
So that the ultimate goal of testing is to end the pandemic, it’s to eliminate
death, it’s to prevent hospitalizations, it’s to prevent people getting
seriously ill, even if they’re not hospitalized, it’s prevent it’s preventing
infection. And it’s also allowing us to reopen the economy and engage in all of
our usual activities without the fear of contracting COVID-19. So that is
really the ultimate goal of testing. And we think Well, what does testing need
to do in order for us to get there. So testing needs to provide information
that we can then use for public health programs and for changing individual
behavior. And also, for general policy, economic policy should be informed by
test results for COVID-19 as well. And so we need to have that information to
be able to make good decisions that are then going to lead to all of these
positive outcomes that we’re aiming for. And in an ideal situation, everybody
would know at all times whether they were contagious with COVID-19. Because if
they knew they were contagious, then they would be able to individually take
precautions to avoid spreading it. It also would be helpful if people other
than the person themselves knew about infections in certain cases, for example,
by saying you’re not actually allowed to come into this restaurant, unless you
have a negative COVID-19 test. And so that’s one way that we can make things
safer. So the whole point is to slow transmission. And part of that is through
individuals being aware of this and changing their behavior accordingly, as
well as implementing some safeguards, because people may be more willing to
take risks, be less careful about infecting other people. So we need the
safeguards within society, this idea about public health to protect the health
of all.
Nick van Terheyden
So let’s
talk a little bit about the various tests that are available. So we I think,
you know, everybody’s vocabulary has expanded to include PCR. I think, you
know, there’s some discussion or At least knowledge around antibodies and then
antigens. I’m going to put antibodies to one side because I think, you know,
research, I’m not sure that it contributes to the conversation, certainly here.
And but when you think about the PCR and the antigen test, we don’t have a lot
of antigen testing currently. But I think we’re starting to see more of that.
There’s discussions around accuracy, you know, that’s a non granular term. Help
Help us understand the underlying important elements of this, you know, do we
do we have to have 100% accurate test? Or can we do with something that’s less
and how do we cope with that? And how do you respond to people that say,
because I hear this all the time, the antigen test is inaccurate, it’s no good.
help people understand and sort of tease out the reality of that.
Zoe McLaren
Sure. So
I want to I agree with you about let’s put antibody testing aside. So antibody
testing is a test of either current infection or past exposure. And let’s focus
just on PCR testing, and the antigen screening tests that are both tests for
current infection, more thinking about slowing the transmission of COVID-19, we
really want to focus on the test that are tests of current infection, because
that’s what informs whether we need to quarantine or self isolate, and tipard
prevent the spread. And so there are different types of testing, PCR rapid
screening test, even within the rapid antigen test, there are different
manufacturers that have different accuracy levels and different methods of
doing the testing. And also there’s different methods within within PCR testing
as well. And so this idea about does the test need to be 100% accurate? The PCR
test is not even actually 100%. accurate, very, very highly accurate. But the main
thing we need to think about is not does this test, for example, with the
antigen test, does it match up exactly what the PCR and more can we figure out
what this test is telling us, and how we can accurately interpret that, so I
focus not on the test result being accurate, but on the interpretation of the
test result being accurate. And I can give one example about the weather
report. And accurate weather report might tell you there’s a 50% chance of rain
tomorrow. That’s an accurate weather report, it tells you that there actually
is a 50% chance of rain. And we all know how to interpret that. So with a an
instant test, it may only be able to tell you, it may only have an accuracy
rate of 98% or 90%. That’s still high enough, if we can think okay, well, we
know it’s not 100%. So how do we adapt how we think about this test. And we’ve
seen that the antigen tests are very, are highly accurate for detecting when
somebody is contagious. So there may be situations where we know there’s
situations where people have virus in their body, but they’re no longer able to
transmit it. And that might be a situation where the PCR test would come back
positive, but the rapid, the rapid antigen test would come back negative. So
the idea about these rapid screening tests is they’re really very good test
for, for whether somebody is contagious. The challenge is that they don’t catch
all of the contagious cases, but they catch the majority of them. So given that
information, we can make really smart decisions about how to use these rapid
antigen tests. So the idea that if you get a negative rapid antigen test, it
does not mean that you’re not infectious, and it does not mean that you’re not
going to become infectious the very next day. So what do we do we take some
relief in the fact that right now, we’re not contagious. We’re unlikely to be
contagious, but we know there’s still a risk of transmission. And so we want to
adjust our behavior accordingly. So the rapid screening tests are very, very
good for reducing the prevalence of COVID-19 in the population. They’re not
necessarily very good for giving somebody at 100% certainty that they’re
negative with with COVID-19. So they reduce the population prevalence. So what
does that mean? That means that the kind of things we might want to use with
schools, where we’re testing everybody in schools, and we know we’re not going
to catch all of the cases, so we need to have other ways, kind of backup ways
we should be masking, we should be social distancing, we should be cautious
about symptoms, and employing quarantine when necessary. But adding another
layer of doing the rapid screening tests means that we’re going to be able to
reduce the prevalence within the school population, even if we can’t catch
every case. And so as long as we understand what these tests can and can’t do,
there’s no reason that we shouldn’t be using them. The potential drawbacks to
them are very, very, almost certainly unlikely to outweigh the positive
benefits of introducing this test. And I know people are concerned about these
potential negative consequences. But I think about them as you know, there’s
going to there’s going to be a lot of benefits and some drawbacks, but if the
drawbacks don’t outweigh the benefits, we should do it anyway and figure out
ways to address these drawbacks. So there’s some concerns about false
negatives. Isn’t false positives. So you want to think about how we can
actually address those. But I will say that there is the example of Slovakia,
which a few weeks ago did kind of a math testing program, they use these rapid
screening tests, it was entirely voluntary. And the isolation following a
positive test was also entirely voluntary. And you can see that they were able
to flash the population prevalence of COVID-19, really, within just a week. And
it was seemed very, very effective to do that. And the important thing to note
is that as far as I know, and I’ve looked into this, I don’t think they did
anything at all to address potential false negatives or false positives. And
yet, they were still able to make a massive dent in their population
prevalence. And so I think that’s a really good sign that it is possible for
these tests to be implemented in a way that is really powerful and effective.
And also that if we do that, we need to make sure that we put in the safeguards
to make sure that if we do it here in the US, it would be it would be as as
effective. But that’s the idea is that these can work if they’re implemented
properly. And so we need to figure out as soon as possible, how to implement
them properly. And if they seem like they’re going to be effective, we need to
massively scaled. And
Nick van Terheyden
so I
think, you know, what strikes me about this is the level of complexity. And you
know, as you rightly point out, it’s all about the interpretation, we have
information, but we need to interpret it in the context of where it’s coming
from the context of the surrounding, sort of experiences. And my sense is that
we’ve done a pretty poor job in communicating this in a way that people can
understand. Because this takes a lot of, essentially, deeper understanding of
all of the details. In the remaining minutes that we’ve got, can you share a
little bit of where you think this will go? And how we can get better? Not just
by expanding the testing? I think that’s clear. But how do we get better at
communicating it? So that was better compliance, people understand why they’re
doing it, we don’t have the resistance. We’ve just done a poor job, that seems
to be part of the sort of challenge.
Zoe McLaren
Yes, I
think, I mean, we have a pandemic, that is basically making it very difficult
for people to, to function in a lot of ways people are overloaded, people don’t
have access to childcare, people are managing a whole lot of work, people have
lost a lot of their social support. And so we need to be really mindful of the
fact that people are not functioning at 100%. And what can we do from a public
health and policy standpoint, to support people. And so with the rolling out of
the rapid screening tests, I just disagree with some people who are focused
very much on having it be an at home test, my ideal situation would be to have
these rapid screening tests rolled out in what I call a food truck model, or
the Starbucks model, the idea that you can go and swing by somewhere very
easily get a very quick test, you get the results sent to you. But then you’d
also be connected to the public health infrastructure. And part of that is
about we’ve just talked about in terms of interpretation. So a rapid screening
test result is actually interpreted slightly differently. Depending on whether
you have symptoms or not, it’s actually interpreted slightly differently,
depending on what the population prevalence is in your community. And so giving
people just the test, without enough information to really understand how to
accurately interpret them is not is not ideal. And so this idea about what we
want people to have really easy access to testings, they can do it frequently
and get the results quickly, is there a way to do that while also providing the
support in terms of communicating the results, so can guiding people’s
behavior. And the other component of it is that if we have it’s through this
food truck or Starbucks model, the idea is that then that test result actually
is part of surveillance. And so that actually gives us even more information to
to evaluate the testing program to figure out what’s going on in terms of
pandemic, just testing at home with volunteer reporting is a real big challenge
when it first surveillance, because people report for some reason, but not for
others. Whereas going to seek a test, we do understand a fair amount about how
that selection works. Because for example, I’ve done a lot of research, many
people have done a lot, a lot of research into that. So that process we
understand better, we know how to back out what’s going on with the epidemic
based on kind of sent somewhat centralized testing or the Starbucks food truck
model. We know less about what would happen with the COVID-19 epidemic in terms
of surveillance, if we rely purely on people volunteer reporting or at home
tests. So that’s one concern I have. But I think this idea about the food truck
model is really very powerful. It can have just as big an impact as the home
testing and perhaps anybody In bigger impact, because we also get the
surveillance data that feeds into so many other components to improve the
public health response.
Nick van Terheyden
Fantastic,
as usual, and unfortunately, especially given the relevance of the topic we’ve
run out of time. I think fantastic insights, obviously more testing is coming,
perhaps is going to be a blend of some of those various models. And, you know,
the importance of where that data is coming from and how it informs our
decisions is clearly going to be an imperative. Just remains for me to thank
you for joining me on the show. And so it’s been a great pleasure. Thanks for
joining me. Thanks, Nick.
