Now let's change tracks a bit, let's talk
about a logical construct, called necessariness
and sufficiency. This is a topic in classical
logic, but it is relevant for medical science
as well, because when we consider causal factors,
we need to know whether or not the causal
factor is both necessary and sufficient to
cause the outcome we're interested in.
So, there are four possibilities here when talking
about a causal factor and an outcome. Either
the causal factor is necessary and sufficient
to cause the disease I'm interested in or
it is necessary, but not sufficient or it's
sufficient, but not necessary, or it's neither
sufficient nor necessary. Let's go through
each of these cases with an example. The first,
necessary and sufficient. Think about the
disease called SARS, which we know was caused
by a variation of the coronavirus. So the
causal factor here is a coronavirus, that
causes SARS. It's necessary to have the coronavirus
before you can have SARS, there is no such
thing as SARS without infection of the coronavirus.
Is it sufficient? Absolutely, it's all you
need, you don't need anything else to cause
SARS, just the coronavirus. So that virus
is both necessary and sufficient to cause
SARS. Next, necessary but not sufficient.
Alright, consider an environmental trigger
that produces a certain disease. An environmental
trigger must trigger a gene that becomes activated,
so without the environmental trigger, the
gene is not activated. Without the gene, there
is nothing to be activated and this causes
the disease that we're interested in. So each
one of those factors is necessary, but individually
they are not sufficient, they must be present
together to cause the disease. So neither
the gene nor the trigger are sufficient, but
they're both necessary. Next, sufficient but
not necessary. Alright consider radiation
and benzene poisoning, each one of those exposures
can cause leukemia, alone. They are both sufficient,
but neither one of them is necessary, because
in absence of one, the other can do the job,
so either radiation alone or benzene alone
is sufficient to cause leukemia. Next, neither
sufficient nor necessary. This one is complicated.
Consider prostate cancer. Prostate cancer
is caused by a whole bunch of complex interactions
and variables. You've got hormones. You've
got the things that you eat. You've got inflammation.
You've got the kind of person demographically
that you are. Individually one of those risk
factors is not sufficient to cause prostate
cancer, but combined with one or two of the
other ones, it might be. Which one or two
of the other ones? Well it's a soup, you can
pick and choose which variables, any two or
three of those could probably cause prostate
cancer. It can be argued that this is the
scenario that is most likely for most diseases
that we perceive, but this is a case where
the causal factors are neither sufficient
nor necessary to cause the disease in question.
Now let's talk about validity and reliability.
These are important concepts in scientific
proof, reliability is reproducibility. Can
the results that you have collected in your
study for example, can they be reproduced
if you do the study again?
That's actually the hallmark of science, reproducibility.
That's why we publish our methodologies in
papers, so that other people can read our
methods, reproduce our studies and hopefully
get the same results. Validity on the other
hand is, do my results represent the real
world? Are they valid when I extrapolate to
the rest of the world? So validity and reliability
as I mentioned are philosophical cornerstones
to science. They are what we consider to be
the bases of scientific proof.