Examination of Witness (Questions 400-419)|
Professor Sir Richard Peto
14 FEBRUARY 2006
Q400 Lord Skidelsky: It is paragraph
9, page 2. You have the sentence: "For example, if investigators
were unable to achieve statistical significance at the standard
95 per cent confidence level to confirm their a priori
hypotheses, they merely lowered the confidence level to 90 per
cent." Is it the case that by lowering the confidence level
you increase the probability of accepting the wrong hypothesis
as correct? If that is so, then one may be sceptical about the
Professor Sir Richard Peto: I think this is
a caricature of what has actually gone on. Yes, certainly, as
you know, you are going to get people who are enthusiastic about
tobacco control wanting to have studies demonstrating the hazards
of passive smoking and, as a result, you will get some claims
that are not justified. It is when you look at the totality of
the evidence and try to get all of the evidence together, and
try to assess the extent to which selective publication of positive
results could cause problems, then I think you do get a consistent
picture that there is some excess risk of lung cancer among those
who are exposed to passive smoking, from other people's smoke,
on a regular long term basis. The comment by the tobacco manufacturers
is okay as a comment on some particular studies, but it is not
okay as a comment on the totality of the evidence. When you start
putting the data from several dozen studies together, the fact
that one or other of them individually claims significance or
not is just point-scoring, and that is what this is, it is just
Q401 Lord Skidelsky: Could I draw you
out on one further thing. You have been unwilling to quantify
the risks from passive smoking.
Professor Sir Richard Peto: Yes.
Q402 Lord Skidelsky: You say there is
an increase in risk but you do not know how much it is.
Professor Sir Richard Peto: Tobacco smoke is
far and away the most important cause of human cancer in the world,
and chronic exposure to it, even at low levels, is going to produce
some increase in the risk of developing cancer, not only among
lifelong non-smokers but also, more importantly, among ex-smokers,
who are well off the zero on a dose-response curve.
Q403 Lord Skidelsky: You may not want
to give your opinion on legislation, but would you say that the
risk justified the legislation which is now being proposed, or
some legislation that may be more proportionate to the risks?
Professor Sir Richard Peto: I am sorry, I am
not trying to be evasive, but when we document the extent to which
smokers kill themselves, I do not try to say what laws should
or should not be made about what smokers themselves do. I think
trying to get evidence as to what hazards are is not the same
as proposing legislation.
Q404 Lord Skidelsky: But legislation
should be based on some measure of hazards. That is the basis
Professor Sir Richard Peto: We know this is
the most serious of all human carcinogens. In terms of numbers
of deaths, just lung cancer alone is causing about one million
deaths a year worldwide and smoking kills a lot more people by
other diseases than by lung cancer. This is the most serious of
all human carcinogens. It has to be causing some risk. Whatever
risk it is causing, there is going to be uncertainty about it.
It is difficult to measure small risks reliably, particularly
on heart disease and particularly in people who are ex-smokers
and have some substantial exposure previously. But, environmental
tobacco smoke has to be causing some risk. In this country alone,
we are talking about 100,000 deaths a year from smokingactually,
it used to be more than 100,000. These are really big numbers.
Trying to minimise exposure or to limit exposure to such agents
seems attractive to many people. The fact that it is difficult
to measure these low risks is always going to be the case. Whatever
those risks are, it is going to be difficult to measure them.
Q405 Lord Vallance of Tummel: Sir Richard,
I wonder if I may change the subject a little bit because you
have already answered the question I was going to ask on smoking
en passant. Could we move on to hospital acquired infections.
Professor Sir Richard Peto: I have been fairly
useless on the previous questions and I am probably going to be
even more useless on this.
Lord Vallance of Tummel: That is fine.
Do not think that because you are being useless it is not helpful.
Chairman: You may take some comfort from
the fact that with some people who are very assertive about what
they believe we are not very convinced they are right either.
We think you are more likely to be right if you are somebody who
shows a little doubt about something.
Q406 Lord Vallance of Tummel: If you
shed doubt on an area where there is apparent clarity before,
that may be very useful indeed. Media reports have highlighted
the risks of contracting MRSA.
Professor Sir Richard Peto: Yes.
Q407 Lord Vallance of Tummel: Do you
have an assessment of the scale of this problem in the UK versus,
say, other European countries? Do you think the media reports
are accurately conveying the scale of the problem or are they
Professor Sir Richard Peto: I am sorry, I can
only make an uninformed comment. I suspect that there is a considerable
exaggeration. And, very often, when things are actually going
quite well in this country in terms of medical benefits, it gets
represented otherwise. We have the best decrease in the world
in lung cancer mortality, and we have the best decrease in the
world in breast cancer mortality, but these do not get emphasised.
You are always told how bad things are. In some respects we do
very well but this does not come across. There is a tendency,
it seems to me, to seek fault in the hospital system. The hazards
of infections acquired in hospital causing the death of patients
who otherwise would have had some chance of a reasonable life,
it seems to me, have probably been exaggerated. My impression
on reading about them is that they probably have been, but that
really is not a serious scientific comment.
Q408 Lord Vallance of Tummel: Are you
aware of any international comparisons between the UK and other
Professor Sir Richard Peto: Not serious ones,
but that does not mean that there have not been any. I am sorry,
I just cannot help you.
Q409 Lord Vallance of Tummel: This could
be helpful. You are really saying
Professor Sir Richard Peto: I am talking about
my ignorance. I am not saying it is not known. I am saying that
I do not know it.
Q410 Lord Roper: The same problem may
arise on this question, as this is one, more generally, on methodology.
A report by the National Audit Office, Improving Patient
Care by Reducing the Risk of Hospital Acquired Infection published
in 2002, suggested "there has been limited progress in improving
information on the extent and costs of hospital acquired infections"
and that "progress in preventing and reducing the number
of infections acquired while in hospital . . . continues to be
constrained by lack of data". Looking at this more generally,
could you give us your views about the quality of data collection
and statistical analysis provided within the National Health Service?
Does current statistical practice within the NHS adequately support
health care risk assessments and cost-benefit analysis? If not,
do you have any suggestions as to what could be done to improve
Professor Sir Richard Peto: Are all the statistics
gathered that could possibly be useful? I am speaking now as a
professor of medical statistics, and if you send everybody who
is supposed to be treating patients and running around wards chasing
statisticsmore and more and more statisticsyou may
not be doing patients a favour. There are things which are worth
recording and there are some things which are not. There is one
thing that can be done to reduce hospital acquired infections
however, and that is at the time of surgery. There are cases where
surgeons definitely choose not to use prophylactic antibiotics
(just antibiotics in case the patient gets an infection) because
they are afraid of encouraging the emergence of resistant strains.
But I think that one could reduce peri-operative infection rates
by the use of prophylactic antibiotics. Experts do disagree as
to whether this is an appropriate strategy. My view, however,
is that at least we know that prophylactic antibiotics can help
protect the individual undergoing surgery, and the future hazards
are somewhat theoretical, although there are some clear examples
of antibiotic resistance emerging.
Q411 Lord Roper: Would it be worthwhile
for us to invest more in statistical analysis and data collection?
Professor Sir Richard Peto: I think you have
to be very particular as to what it is you are going to do. If
you lay down blanket rules about needing more and more and more
information about this and that, and people have to run around
completing tables and filling in forms, then this may not be helpful
to effective patient care. There are times when things go wrong
and they need to be recognised, and there are times when the collection
of routine statistics helps with this. But you have to be quite
careful when you try to insist on the collection of additional
statistics. I would like to make one point about statistics in
this country, if I may. Over the last half century, until the
last few years, we have had a tradition of medical statistics
being fairly freely available to bona fide medical researchers
to help them do studies and to understand the causes of diseases.
Concerns, mostly within the last ten years, over personal privacy
have now produced a situation where really serious studies cannot
be done; where they are so impeded by data protection regulation
which really is not of benefit to anybody, that serious damage
is caused to medical research that could save lives. I think it
would be very useful if there were to be a serious reduction in
the extent of the control of the use of personal data for medical
research by bona fide medical researchers. We have reached
a situation where great damage is done to serious research. It
is, of course, very difficult to point to specific examples of
where research would have been done but was not, because you do
not know what would have been discovered, but we can look at some
things that were discovered in the past and realise that they
would not be able to be discovered now. For example, the hazards
of German measles in pregnant women were discovered, among other
things, by an inquiry where they took the records of women who
had applied for sick leave for measles or whooping cough or German
measles and who then applied for maternity support later on, and
then went back and tried to interview them to find out which of
the children had any disabilities and to relate that to the illness
that the mother had suffered. This inquiry confirmed that it was
specifically German measles, not ordinary measles and not chicken
pox, which was producing blindness and brain damage, and, as a
result, pregnant women were protected. You could not do that now:
it would be using records for purposes for which they were not
obtained. And, the question is just: why not?
Q412 Chairman: This strikes me as being
Professor Sir Richard Peto: This is something
which would be a really important area for concern.
Chairman: The law about unintended consequences
of increased privacy produces this problem. Has there been anything
published on that? Is there any documentation or article we could
read? It is really quite interesting.
Q413 Lord Roper: It has come up in recent
legislation, I think. We did debate it in the House.
Professor Sir Richard Peto: Yes.
Q414 Lord Roper: If Professor Peto has
some other sources, that would be very helpful.
Professor Sir Richard Peto: Perhaps I could
answer that in writing subsequently or I could refer to the article
written by the late Professor Sir Richard Doll in 2001 in the
British Medical Journal on the subject. It was a brief
article and he argued very strongly that, for moral reasons, the
amount of restriction on the collection of medical statistics
has now reached a situation which he described as immoral. He
was talking as somebody who had discovered quite a lot of things
from the collection and appropriate use of routine statistical
Chairman: All right. We have one or two
more questions, and we should get on.
Q415 Lord Sheldon: In talking about distortion
by the media of medical risks, there is the example, of course,
of the MMR vaccine.
Professor Sir Richard Peto: Yes.
Q416 Lord Sheldon: How are we going to
get a balanced picture of the risks involved here?
Professor Sir Richard Peto: So far there seems
to be no evidence of any risks involved, and yet that has not
stopped it being a major scare. That is an extraordinary example
because there is no good evidence of any such risks, yet it just
runs and runs.
Q417 Lord Sheldon: Should the health
care professionals not get involved with this with the media?
Professor Sir Richard Peto: Yes, but then the
media will try to get a "balanced" view by finding somebody
who says it might be dangerous. This is one of the prices of having
a free press, that they print whatever rubbish they like in circumstances
like this. In this case, to be fair to the media, there was an
eminent professor making the claim, so they are bound to report
it, but it is an absurd story. It is an absurd episode, taken
as a whole. I do not really blame the media for it, it is just
one of these things that happens, and more such things will happen
again and again.
Q418 Lord Macdonald of Tradeston: As
a supplementary to that, if we take the role played by politicianswhich
can sometimes seem political or innumerate or cynicalthey
perhaps have a greater duty of responsibility than the press.
Is there a way in which medical professionals could try to add
more weight and evidence to the kind of debates that are held
in Parliament and to influence the politicians perhaps to be better
informed and more responsible?
Professor Sir Richard Peto: I do not know of
any obvious steps, I am sorry. My main concern, still, is with
the extent to which people are getting killed by tobacco in this
country. I am sorry to come back to it, but it is such an absurd
situation, where you have 100,000 deaths a year in this country
alone. However, this country has done a lot better than many other
countries. Indeed, we have a very nice control group just across
the Channel in France, where the French delayed about 20 or 30
years longer than we did before trying to take tobacco seriously.
The result was a really marked contrast in lung cancer trends
between France and Britain.
Q419 Chairman: I wonder if I could press
you on a bit with another side to our interest: How serious is
the threat to antibiotic resistance?
Professor Sir Richard Peto: Certainly there
are bacteria that are difficult to treat and in some cases this
has been produced by widespread use of antibiotics, but, overall,
there is the most extraordinary decrease in mortality from infections.
My colleague Dr Gary Whitlock has recently produced a graph describing
the patterns of mortality in Britian, running back to 1838, when
statistics first began to be collected. Because of public health
measures between 1850 and 1950 the death rates kept on dropping
and dropping until the middle of the 20th century, at which point
effective drugs came in, since when the residual risks of death
from infection have kept on dropping. For example, the probability
of a five year old dying before age 15 is now only about one-third
what it was 150 years ago when some of the portraits in this committee
room were painted. In looking at what is wrong, it is really worth
remembering what we are getting right. We have had the most extraordinary
reductions in the probabilities of death in infancy, death in
childhood, death in early adult life and death in middle age in
this country. There have been extraordinary changes over the last
100 years, and largely favourable. So when we do talk about the
use of antibiotics and the encouragement of antibiotic resistance,
remember that many peopleI, for onehave had their
lives saved by antibiotics. That is the main thing. They do work,
but they do not work if they are not used. The damage that they
do is a small fraction of the good that they do.