Screening School Lesson 1
Screening School Lesson 1
This is a repost of an old blog
Screening school. Lesson 1
I’ve written a lot on
screening. Mainly for a technical audience. I often assume that’s there is
a lot more background knowledge than people really have. I was asked by someone
with little knowledge of screening to give a single key reference that
summarises the field in a few sides.
I found it quite difficult to find an article to summarise
the whole lot simply and easily.
3 key references
This article by Muir
Gray is one of the simple standard references I often give people.
Maximizing benefit and minimizing harm of screening
WHO/Europe | European Observatory on Health Systems and
Policies – Screening. When is it appropriate and how can we get it right?
(2020)
Health
Knowledge pages on screening also produce a good set of information,
if a bit more complex.
My take on some of the key concepts that are required are in
the Screening
101 lesson plan.
And of course the granddaddy of them all is the masterpiece
that is the Raffle
and Gray book. This is currently viewed as the bible. It’s quite long,
but most public health professionals are well trained in this stuff so you
could go ask them (just don’t be grumpy when they give you an answer that
screening won’t save quite as many lives as you’d thought it might.)
a good answer to the “screening saves lives” headline was
recently published.
Does
screening for disease save lives in asymptomatic adults? Systematic review
of meta-analyses
and randomized trials. Notable authors including John Ioannidis
Among currently available screening tests for diseases where
death is a common outcome, reductions in disease-specific mortality are
uncommon and reductions in all-cause mortality are very rare or non-existent
Key Messages
• We evaluated the evidence on 39 screening tests for 19
diseases where mortality is a common outcome.
• We found 48 randomized controlled trials and 9
meta-analyses that addressed either disease-specific or all-cause mortality.
• Reductions in disease-specific mortality were uncommon and
reductions in all-cause mortality were very uncommon, or even non-existent
with these screening tests.
I wrote something bespoke. It is here. It sets out some of
the background, in hopefully simple language.
Key points
1
The theory
the notion that earlier diagnosis of disease is better is
alluring, but often the evidence doesn’t support this
there are important ethical issues inherent in this – it is
a process of offering someone – in fact a whole population – who thinks they
are well a test that may determine whether or not they have a disease. The test
may lead to false positive/false negative. The diagnostic pathway &
subsequent treatment may lead to harm, and certainly may lead to harm
that outweighs benefit.
It is well accepted that all health care can lead to harm
and all health care can lead to good. This is equally true of screening is
equally valid here, but with the ethical element above leading to a different
bar.
Screening is not a process of offering a test to an
individual, but a programme offered to a population.
The key components are
·
definition of the population to be tested – is
it clear. Can we identify the population reliably.
·
Test – is it a “good” test for the condition
concerned. There are technical definitions here defining sensitivity (the test
is good at picking up true positives – ie people who “test” positive who
actually have the condition) and specificity (the test is good at ruling out
those who don’t have the condition). Often tests used in screening are not
particularly sensitive or specific.
·
the diagnostic pathway – are there risks and
benefits in the diagnostic pathway. A screening test is NOT a diagnosis.
·
is there an available treatment. Is that
treatment effective, cost effective and affordable.
·
is the whole screening process effective, cost
effective and affordable
especially given the ethical issues, what is the process for ensuring high
quality.
·
QA is critical. The QA is not just “was the scan
valid” but was the WHOLE pathway from identification of the cohort, invite, uptake,
test, diagnostic follow up and then into treatment achieving the standards set.
In the UK there is a requirement for any proposition to
screen for anything to be assessed against a set of criteria – known as Wilson
Jungner criteria.
This is overseen by the National Screening Committee. Which
makes recommendations to screen or not screen to the 4 UK governments.
2
Screening programme creep
It is fair to say many try to start screening for a whole
host of things when NSC has recommended against it
3 issues below illustrate
Prostate cancer
The test used (urine PSA) is a very poor test if used in a
general population context
Screening may pick up a condition that would never otherwise
come to clinical attention. – more people die with prostate cancer
than of it.
The treatment may cause harm. Treatment can have a side
effect of making men impotent and incontinent, this is not infrequent. Prostate
cancer screening (compared to no screening) does not change
in all-cause mortality, but there is a change in cause
specific mortality. This changes the cause of death not the time of death than
screening often doesn’t save lives.
Thus we don’t screen for prostate cancer
Lung Cancer screening (compared to not screening) (NB at
time of the repost of this blog, NSC is consulting on this)
There is little doubt that lung cancer screening, in those
“found” likely leads to a significant positive shift in stage at diagnosis (ie
cancers found at an earlier stage, with more getting surgery with curative
intent, as opposed to chemo or radiotherapy).
Whether that surgery leads to cure is unknown. The
evidence to date says lung cancer screening leads to a change in mortality from
lung cancer but NOT all cause mortality. Again, doesn’t change the time of
death, but does change the cause
Also, there is harm from radiation – CT scans. Harm from
over diagnosed nodules. Lung cancer screening os not cost effective and
arguably not affordable
Thus the NSC does not recommend lung cancer screening. It
is, however, being introduced now in a scenario where the National Screening
Committee is likely to NOT recommend screening.
Long story.
Dementia
This one is even trickier.
One of the UKs foremost academic in dementia studies at
Cambridge – Prof Carol Brayne – wrote in 2012 that there is no evidence base
for proposed dementia screening – https://www.bmj.com/content/345/bmj.e8588/rapid-responses
And this “Is Dementia Screening of Apparently Healthy
Individuals Justified?” – https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5591898/
The NSC recommendation is to not screen
However we went ahead and set up a programme
There are lots of other similar scenarios where the
policy recommendation is against screening for something but people ignore
them. Often this is NHS England(!). Atrial Fibrilation is one such example at
the moment.
Often this is pursued on the premise that ‘it’s case
finding’ and this is different to ‘screening’ thus no NSC assessment is needed.
There is no difference between these concepts. This
blog sets out why.
Avoiding the scrutiny of the NSC is bad for cost
effectiveness, and bad for equity & bad for the NHS as a whole.
3
On screening programmes we SHOULD be undertaking
Then for things we DO screen for, there is a major issue
around ensuring good quality commissioning and QA processes. Read Raffle and
Gray for a comprehensive overview.
NHSE are the commissioner for screening, it is a public
health service undertaken under section 7a of the 2012 Health and Social Care
Act. The National
Audit Office published a report last week setting out some of the
national failures around the management of breast and cervical cancer
screening.
It makes for rather difficult reading.
4 pay (a
lot of) attention to over diagnosis
this piece by Critle on the
good, the bad and the ugly is excellent explaining some of the core
concepts.
“The birds have already escaped the barnyard: they are the
fastest growing and most aggressive cancers, those that have already spread by
the time they are detectable. Screening cannot help with the birds; the
question is whether treatment can. The rabbits are more slowly progressive
cancers, and they can be caught early. They are the cancers with which
screening can potentially help. Then there are the turtles: there’s no need
catch them because they’re not going anywhere anyway”.
It very much
articulars poorly recognized (or hidden) harms and sets out the epidemiologic signatures
of overdiagnosis see this chart on the incidence and mortality or early stage
incidence vs late stage incidence to give an indicator of overdiagnosis.
(that said recognising this at individual level is very much
harder indeed,
This piece on lung cancer
screening in the States and association with stage shift, then survival,
over a decade also sets out the sometimes spurious link between stage shift
being equated to “lives saved”. In hindsight it is a shame that no effort to
measure actual mortality over the 10yr. The analysis of survival can give
biased impression akin to the PFS / OS discussion - See here for good example
of that pertinent to lung cancer https://t.co/8OSzorUn2Y).
Comments
Post a Comment