Examination of Witnesses (Question 580-599)
DR ALISON
HOLMES, PROFESSOR
MATT GRIFFITHS
AND PROFESSOR
ANEEZ ESMAIL
5 FEBRUARY 2009
Q580 Stephen Hesford: So it becomes
not exactly scientific but it has a representative nature?
Professor Esmail: Yes, and I think
the point is it has the effect of actually getting you to improve
things, and that is very important. It means that when you find
out, for example, that a patient had a blood test done and it
was not followed up, you invariably go and look at your systems
and what are our systems of follow-up. When a patient comes for
a blood test what reminders do we have, what sort of prompts do
we have to make sure that someone follows up on that result. So
it tells us to do things like that, and I think we can hugely
improve systems in primary care through those sorts of techniques.
We also have techniques like significant event audit. We have
a system whereby we ask people to make a record of something that
goes badly wrong. It is a significant event, it is not something
minor; something that they feel is really big. They will make
a record of it and we will discuss it as a team, and, as a result
of our discussion, we will say, "Gosh, this is not very good
in what happens. It highlights a flaw here. It shows we do not
do this very well", and we then actually get someone, or
a team of people, to go and make the changes and we hope that
it does not happen again, but if the problem arises again, we
have a record of what happened before, we can follow the audit
trail to see what happened. Those sorts of things can be very
important and should not be negated.
Q581 Dr Taylor: Going on with significant
event audits, do you only look at those four weeks at a time?
Professor Esmail: No, that is
an ongoing programme. We just talked about reporting. We have
a computerised base system. For those four weeks we want them
to report everything that goes wrong, from their perception, and
we will get everyone to do that—nurses, reception staff,
the people who work in the admin office—so everyone knows
that that is the four weeks we try to collect information on.
That is the purpose; it gives us an idea of how big a problem
it is. Significant event audit goes on throughout the year, because
I may find out last week something happened. If I got a very poor
discharge of someone from hospital, I did not know something that
should be happening, then I should log that so that when we come
to discuss it we know. How can we actually contact the hospital
to make sure that this does not happen again? Can we make sure
that it is followed through? Does that audit loop get closed?
Can we make sure that we have something in place that we know
it is not continuing to happen? That is what significant event
audit achieves, and that has to happen all the time.
Q582 Dr Taylor: Could you give us
some other examples of what are significant events?
Professor Esmail: I will give
you an example of a patient of mine who was discharged from hospital
with a diagnosis of cancer, which I was never told about, and
nor was she. Eventually we were able to find out. We kept insisting:
"Why have we not had a proper discharge letter?", and
we then discovered this happened. That was a terrible case really,
and what happened on that is that I discussed this. We discussed
it as a team. We wrote a letter to the hospital. We asked them
to investigate it. We expected a response back again. We hope
we have a system in place so that that should no longer be the
case, and things are improving on that. That is an example of
significant event audit.
Q583 Dr Taylor: That is really something
happening in secondary care though. What about significant events
in primary care?
Professor Esmail: A common problem
is someone has a blood test done and is found to have anaemia
and we do not follow it up. We look at it, we see that the haemoglobin
is low and it is just put down as probably due to iron deficiency
anaemia, but actually it might be something more significant.
We might have to do more investigations to see is there a cancer
underlying that, are those investigations followed up, who is
going to carry them out, and so on. It is becoming particularly
important in primary care: continuity is a real problem. You might
have one doctor seeing the patient and then someone else having
to follow it up three or four weeks later, and if the record keeping
is not good or the audit trail is not good, that might get missed.
That is a very common example, when we have a result which might
signify something more serious but is not followed up.
Q584 Dr Taylor: Has incorporating
SEAs into the QOF made any difference?
Professor Esmail: I do not think
it has yet. The reason is because the way that people do significant
event audit in general practice varies hugely from practice to
practice, and we need to have a standard way of doing it, we need
to have a standard way of collecting information, we need to have
a standard way of making sure that information is shared, and
all that the QOF does is say, "Do you do it?", yes or
no, and that is it.
Q585 Dr Taylor: So you can get the
money if you say, "Yes", even if you do not do it?
Professor Esmail: That is lying,
and that should not be the case. The point is you can do it, but
are we learning from it. That is the critical question. Also important
is: are we sharing the information? I might have a problem in
my practice. I work in a primary care trust where there are 50
other GPs, and what I have found is a problem may also be very
relevant to others. Are there mechanisms whereby we can share
that sort of learning? That does not happen consistently—even
the mechanism of recording them, for example. The Royal College
of General Practitioners has just launched a significant event
audit tool kit which suggests a format for collecting information,
for reporting on it and so on, and that is only recent, but that
has been the problem. So the QOF merely asks you: "Do you
do it?", yes or no, but really the question is: "How
do you do it, what do we learn from it and how can we make sure
it does not happen again?"
Q586 Dr Taylor: Even though it has
been going for some few years, you have only just got a tool kit
of how to do it?
Professor Esmail: Yes. No, a tool
kit for standardising it.
Q587 Dr Taylor: Which is what it
has to be?
Professor Esmail: What it has
to be, yes. Actually significant event audit pre-dated QOF. GPs
were doing it in some areas of the country ten years back, because
they saw it as an important way of driving improvement. So when
someone dies unexpectedly many good practices were reviewing that
information and saying, "Why did they die unexpectedly? What
was going on here?", and they look at the period of care.
So that was very, very important. We do it with every cancer death,
and we do it, not because of some terminal cancer stuff, but we
actually look at the quality of care. Was there a good follow-up?
Were people informed correctly what should happen when they died?
All this information is very important. We do that routinely,
for example, for every cancer death. We do a review of it.
Q588 Dr Taylor: Does the tool kit
standardise what is significant?
Professor Esmail: Yes, it will
have. The point about significant event, we tend to leave it to
the practitioners and the reporters to determine it. You cannot
say that is irrelevant, because from a patient's point of view
it might be very important, so we tend to leave that open: either
the patient raises it and we say, "Gosh, that should not
have happened", or the practitioner raises it. You cannot
predefine a significant event. I think it has to be determined
a lot by the circumstances and the context it occurs in.
Q589 Sandra Gidley: I must admit,
I am somewhat horrified that the evidence is that GPs will not
report the lower level stuff if they have to do it for more than
three or four weeks, because one would have hoped that there would
have been sufficient interest in improving things for patients
that this would become endemic, embedded good practice.
Professor Esmail: It is not that
they will not. The National Learning and Reporting system, as
you probably know, if you look at the reports from primary care
it is point 5% per cent, and most of those from general practice,
so it does not even include community pharmacy, it does not include
district nursing and areas like that. Is it because they do not
want to or is it because they have been asked the wrong question
and asked to report the wrong thing? I think that if you look
at a practice like ours, when probably about 500 or 600 patients
are seen every day—we have 18,000 registered patients—the
volume of work is immense. The reality is that you are so caught
up in the day-to-day running of things that to have a system without
doing a lot more work and without the culture changing, where
every incident is reported. We even have to ask ourselves, "If
you report everything what is going to happen to it?", and
we do not have the capacity to even analyse everything that comes
in.
Q590 Sandra Gidley: I am sure there
was that attitude in hospitals once, but they seem to have managed
it in hospitals. I do not buy into this, that primary care so
is rushed that they cannot do things that are good practice elsewhere?
Professor Esmail: All right. I
understand what you are saying and I realise that it comes across
that we do not want to or we do not care. I am just saying that
this is based on a research project of ours, and that is what
we found and we have had to react to it. We found this was not
just the case in Britain, we found it was the case in Australia,
in the Netherlands, in Canada, in America. This was a universal
problem. I think we have to ask ourselves: how can we sustain
that level of reporting? One of problems is this. A lot of things
that happen in primary care, whilst they should not happen, have
no consequence, and that is another problem, which is why things
like significant event audit have got a catechism in general practice
which has been very powerful: because when something terrible
goes wrong it really concentrates the mind.
Q591 Sandra Gidley: That is fine,
but the lower level events and the near-misses can often educate
about a potential problem as much as a significant event?
Professor Esmail: Yes, and that
is why we have said if we can do it for a month at a time and
vary that over a period of years, we will build up a very, very
important profile, and we can sustain that and have good quality
data and we can then work on that, and that is very important
as well. I think that sort of model might work better in primary
care than in hospital, where you have better structures and organisations,
and so on, to do that.
Q592 Sandra Gidley: I am not convinced,
but I think we are going to go round and round in circles. Let
us get to some specifics. Is there a problem with prescribing
and medicines management in primary care?
Professor Esmail: The research
data shows that there is a lot of interaction. This has come about
from all sorts of work where patients are admitted to hospital,
for examples. Many admissions are due to drug interaction, so
again we understand there is a problem there. I think at one level
we have quite good safeguards in primary care. We have been computerised
for a long, long time and our computer systems are constantly
improving. For example, we get warned now if you prescribe something
and there is interaction with another drug. That is quite sophisticated
and has been around for a long, long time. We know that there
are problems with the system still, but they are constantly being
improved and I think people are aware of that, so at that level
we are moving towards electronic prescribing. Virtually no-one
writes handwritten prescriptions because there are transcription
errors, for example.
Q593 Sandra Gidley: They are not
allowed to; that is why they do not.
Professor Esmail: Yes, but the
point is that that is an error that rarely exists now. Sometimes
when you go on home visits you have to write handwritten prescriptions,
so that still happens, but that has become minimal now, so those
areas have gone. We, for example, recognise the problem and we
employ a pharmacist. In fact the research shows that the solutions
are not necessarily in the technology but in the people, and because
so much prescribing goes on in general practice, it is so important,
we have actually employed a pharmacist who works alongside the
doctors and helps us towards identifying problems where things
are not going right, develops systems so that things happen properly,
and in our case I would say that has made the most impact. We
have good computerised systems and I think that that has helped
tremendously, and where the problems arise probably are when people
go into hospital, because there is not good exchange of information
that goes across that, and, secondly, when people are discharged
from hospital. I think when general practice handles the prescribing,
on the whole, it works pretty well and I think that pharmacists
play an increasingly important role in helping us reduce error
in that area.
Q594 Sandra Gidley: Is there a problem
with variable quality of diagnosis in primary care?
Professor Esmail: We have always
known about. I would not say variable quality. The research was
based really on our small studies which we did, where we got people
to report things and find out what went wrong, but also by looking
at the litigation databases, for example, and what had people
complained about, and what was surprising, what stood out very
clearly in primary care, was that 50% of the cases in the litigation
databases we looked at were because of delayed diagnosis. What
happens is that someone has a condition and it is picked up quite
late. Let us take a simple example of a bowel cancer. The problem
is, of course, that they present sometimes with very non-specific
symptoms, it may take three or four months before the diagnosis
becomes known and they get referred at that point. So we know
that delayed diagnosis is a big problem, it stands out, it really
hits you in the face, and of course the reason for this, when
we do work around diagnosis in general practice, is that things
that come to general practice in primary care are very undifferentiated;
people do not come with a glaring symptom. If someone came to
me saying, "I am bleeding rectally", that hits me in
the face and says, "Yes, we need to examine and understand
why that is happening", particularly if you are in a particular
age group, but what happens is they come in and say, "I have
not been feeling right for a while and I am not quite right here
and I have got a bit of pain in my stomach." You investigate,
you look at this, and it is a rolling process and when you have
completed the process, you say, "Gosh, this is a cancer",
and yet you can see they presented three or four months down the
line, and from the patient's perspective they say, "I have
had this for three or four months", but, of course, if I
just referred everyone to hospital with those non-specific symptoms,
we would bankrupt the system; it just could not sustain itself.
So there is always that balance of trying to say when is it serious,
when are the problems arising and how can we investigate as much
in primary care before we refer people on? There are many examples
like that: headaches which people talk about. We know that headaches
are a sign of brain tumours, yet if we scanned everyone with a
headache we just could not sustain that; our system cannot allow
that.
Q595 Sandra Gidley: Is there not
some evidence that there is a huge difference in the way individual
GPs would react to those sets of symptoms, which is why the referral
centres were introduced?
Professor Esmail: That is right,
and they have made a huge difference, I think, in making early
diagnosis available, but we have not actually done the research
to see whether it has reduced that.
Q596 Sandra Gidley: Some would say
that it has gone the other way and when GPs have wanted to refer,
if they are seen to be referring more than the people down the
road, they are asked to look again. Who is to say who is right?
Professor Esmail: We do not have
the research evidence on that—that is what I am trying to
say. We have put in these innovations. It is too early for us
to even understand whether it has had an impact on things like
delayed diagnosis, but what you say is absolutely right: we know
there is huge variability. Of course the thing that struck me
when I was looking at the research in this area is sometimes you
have what I call `barn door cases' and you just cannot understand
why it is that people miss them. You look at the case and you
read the medical legal summary and you think, as an outsider,
that is an obvious case of a heart attack, and yet it was missed
by the doctor, and then you ask yourself: "Is it because
they did not know?", and you think that may be the case,
but it is highly unlikely, it is something that you learn really
early on, so what else was going on there? I think that when we
begin to understand this we begin to realise the complexities
around the situation. For example, the patient had had a long
history of chest pain which has never amounted to much and has
been diagnosed as having indigestion, but on this occasion it
was because of myocardial infarction, for example, and it can
be very difficult to distinguish. These sorts of things become
much more complex. So we are beginning to realise that when we
look at diagnosis it is as dependent on the patient as it is on
the clinician as it is on the context of where you are and when
you were, and so on, and it becomes very, very difficult.
Q597 Sandra Gidley: You are describing
something that is more of an art than a science by the sound of
it. Is there a role for automated decision support systems that
would mean you have better outcomes?
Professor Esmail: I think they
can help, definitely, in some areas. The problem is that because
things are so dependent upon the context, we do not have good
enough systems that can tell us about it. For example, if I get
a prompt with someone who comes in with rectal bleeding and the
prompt simply says, "Think about cancer", if it is a
20 year old, that is not helpful, but if it is a 50 year old,
it is. Somehow you need to have a system which can not only look
at it by age, for example, which is a simple example, but, more
importantly, might able to look at it by things like ethnicity,
might be able to look at it by other factors that they have, other
co-morbid conditions, diseases and so on, and can give that information.
That is the holy grail we all look for, and I think we are a long
way away from that because it is actually very, very complicated.
I think the NHS is littered with examples where someone has come
along and said, "This will do it" and people have invested
huge amounts of money into it and it has not delivered it. I think,
yes, we should try and achieve that, but it is far more important
probably to try and get doctors to work together, to understand
these things, to feed back relevant information, for example,
to look at cancer referrals, see what is happening about that.
We can do a lot around training and improvement before trying
to say: let us put all the money into trying to find a technological
solution which I am not entirely certain will work.
Q598 Dr Naysmith: Professor Esmail,
I want to ask you a question in a minute or two about the role
of commissioners and also about the role of registering of GPs
and so on. First of all, I would like to ask you: why is it there
is still such a large amount of over prescribing of antibiotics
by GPs? Dr Holmes went on at some length about the importance
of not allowing the emergence of resistant bacteria, but GPs still
do it, and we know they still do it, so why?
Professor Esmail: I would question
whether. I think, for example, it is very common now; it is part
of the commissioning process. We get feedback on the antibiotics
prescribed for conditions and we have targets, in effect, and
we use those internally. If we see that we are over prescribing,
for example, if we look at our diagnosis of upper respiratory
infections and we say, "Why is the prescription writing very
high?", we look at that and say "Gosh, what is going
on?", and we even look at it by doctor then and, as part
of our review, we take it up with each other and say, "What
is happening here?" So I would question that. I think there
are many improvements that have happened. It is always given as
example that GPs over prescribe. Looking at it from my part, I
get people home, discharged from hospital and I think, "Gosh,
they have been given a whole paraphernalia of drugs which I do
even go near."
Dr Naysmith: No-one would dare over prescribe
in Dr Holmes' hospital; they would not dream of doing it.
Q599 Dr Stoate: It is our job to
stop it all, is it not!
Professor Esmail: That is right,
yes
|
