00:00
Hello and welcome to epidemiology. Are you
a medical student or a doctor or some other
kind of clinician or maybe a health scientist
of some kind? Either way you probably know
that by the time a patient comes to see you
and is showing symptoms, they've already have
a disease state that has gone beyond a certain
progression and it would have been nice probably
to have been able to detect them before there
is symptomatic. But how do you identify a
potential patient before they're showing symptoms?
That's why we have screening tests, you probably
know about screening tests. Screening tests
are things like taking people's blood pressures
in the mall, or even digital rectal exams
or Pap smears in the office. These are kinds
of tests we do to identify people for a later
investigation. So in this lecture we are going
to learn some things about how to assess the
qualities of screening tests, epidemiologically.
00:50
You're going to lo learn how to set up a contingency
table, that's how we put our numbers together
to compute the appropriate indicators for
the quality of a screening test. You're going
to be able to compute and interpret certain
functions of a screening test, in particular,
the sensitivity and specificity. You've probably
already heard of those terms, probably from
reading the descriptions of tests, like a
pregnancy test on the side of the box, that
has numbers relating to its sensitivity and
specificity. You'll be able to compute and
interpret the positive predictive value and
the negative predictive value of a test. And
also you will be able to compute and interpret
this thing we called the likelihood ratio.
01:29
So screening tests are meant to identify individuals
who are probably in the early stages of disease
and that allows us to have an earlier intervention
that we assume gives them a better chance
of being cured or to improve their prognosis
in some sense. Things that affect the ability
of a screening test to be useful include,
its potential to overdiagnose an issue, if
you overdiagnose people, you're finding people
who don't actually have the disease but you
think they do and that causes some stress,
that causes financial burdens on the healthcare
system and so forth. Sometimes screening test
can misdiagnose a situation and that's confusing
for a lot of different reasons, sometimes a test itself
could be painful or damaging, we don't like
that, we try to avoid those situations. And
sometimes the screen test doesn't offer any
ability to help a patient, by which I mean,
some diseases, there is no advantage to finding
out if someone has it earlier, it doesn't
change the prognosis any. Maybe you could
argue that psychologically it's useful for
a patient to know that they have a disease
earlier rather than later, but functionally
and practically, and from the perspective
of public health officials, it's not that
advantageous on a systematic level to do so.
02:42
An important point is that screening tests
are not diagnostic tools, they are meant to
identify individuals who are likely deserving
of a deeper investigation. What kind of deeper
investigation? A more invasive, probably expensive
test. So that means that to compute the diagnostics
of a screening test, we need some objective
measurement of what constitutes truth or real
disease status. A screening test give us a
suspicion of disease status, we need another
test, a more invasive, expensive and accurate
test to confirm that suspicion. That's for
the purposes of computing our diagnostics.
On an everyday basis we don't do this, but
keep in mind, if someone tests positive on
a screening test, then they go for deeper
investigation with a specialist probably.
If they test negative, they go along their
happy way never knowing anything more about
that disease state until an opportunity arises
again for them to be tested. This means that
everything we know about screening tests are
in study communities, research programs. Only
in research programs or on studies do we know
what happens to those who test negative, but
in every day, those who test positive are
followed up on. So let's think about how we
follow up on certain positive tests. So a
Pap smear is usually followed up with a biopsy.
If you test positive on a Pap smear, there'll
be a deeper investigation to determine if
that positive test was real. If you test positive
on a home pregnancy test, you think you're
pregnant, you'll probably get an ultrasound
to confirm that suspicion. If you test negative,
you probably won't. If you had a digital rectal
exam, you probably get a biopsy as well if
you test positive on the DRE.