Saturday 6 December 1997
BMJ
Profiting from closure: the private finance initiative
and the NHS
A covert, untested, and destabilising way of restructuring health care
P
rivate investment is efficient when it maximises
the returns on capital. Public investment is
efficient when it maximises returns within the
constraints of public policy goals, like meeting the
population’s healthcare needs. Given these different
aspirations, what does partnership between the private
and public sectors mean in practice? Specifically, what
happens when private finance funds hospital redevelopment, as the private finance initiative attempts to do?
In a BMA report published this week, Declan Gaffney
and Allyson Pollock of St George’s Hospital Medical
School address this question in what is the most
detailed independent study to date of the 14 private
finance initiative schemes approved in June this year.1
Reliance on private investment, the authors say, inflates
the scale of capital schemes to levels which far exceed
more prudent public proposals as bidders try to
improve their rate of return. This cost escalation puts
new demands on public revenue, which in turn leads to
the search for new economies and new subsidies. The
economies inspire bed reductions and unpiloted innovations in healthcare provision, while the subsidies
entail transfers from other health sectors and raids on
the very public funds that private finance was meant to
replace. Returns on capital come to predominate over
other policy considerations, and the health service
ends up paying more for less.
The figures are striking. From relatively modest
beginnings, the estimated costs of the 14 schemes rose
on average by 72% as investors proposed bigger
schemes involving larger loans and more equity. In
Swindon the cost rose by 229%. Hospital closures and
bed reductions of 7% to 44% helped meet the cost by
releasing land for sale and allowing economies to be
made in the new buildings.
But asset sales and conjectured “efficiency savings”
have proved inadequate to bridge the affordability
problem which cost escalation had created, and a series
of new subsidies have been introduced. Some health
authorities increased their annual commitment, taking
money from other schemes and from sectors, such as
community services, most likely to bear the burden of
cost shunting out of the hospital system. Regional
offices translated block grant capital into revenue payments, which meant subsidising privately financed
projects out of the equipment and maintenance budgets of hospitals wthout privately financed schemes. In
several cases equipment replacement was dropped
BMJ VOLUME 315
6 DECEMBER 1997
from the deals even though equipment formed part of
the estimated capital cost. And finally, the NHS Executive introduced a direct annual subsidy for the first 30
years of the private finance contract, a subsidy almost
large enough in the case of Swindon to have paid for
the original public scheme which the private finance
initiative scheme had replaced (£42m compared with
£48m).
The private finance initiative, says the report, has
been bailed out and the cost borne by other parts of
the health service. It has become not just a mechanism
for reducing hospital services but also a costly burden.
This state of affairs is unlikely to be exposed by the system of appraising the initiative, as that overlooks the
costs which the scheme shifts out of the hospital sector
on to others.
These are important findings. They suggest that the
private finance initiative results in commercial returns
unduly influencing the conduct of capital planning and
the determination of asset size. In Edinburgh, for
example, efficiency savings tied to the proposed new
hospital imply patient throughput approaching 88 finished consultant episodes per bed per year compared
with a national average for England that has levelled
out at 54.2 There is no precedent in Britain for such
levels of activity. The planning base, staffing, and
resource implications of the new model of care on
which the hospital depends are unclear and the practical arrangements remain unpiloted. This is not healthcare planning as it is traditionally understood.
The Department of Health knows this, of course, so
why is it prepared to accept the cost and the risk? The
signs are that the private finance initiative offers a vehicle for another agenda that is gaining ground among
NHS managers. Under this agenda, largescale capital
investment provides an opportunity to redesign the
hospital sector. What are the problems to which large
(and costly) capital investment is supposed to be a
solution? In Birmingham, where the health authority
started consulting last month on its own private
finance initiative plan, the problems are said to be constantly increasing referrals to relatively expensive hospitals.3 Their solution involves reducing the size of the
hospital sector by half and substituting cheaper
alternatives. This means building a new health
infrastructure—which is where the private finance
initiative comes in.
1479
Editorials
What Birmingham’s analysis omits to mention is
the role of capital charges in the pressure felt by hospital budgets. Capital charges, which force hospitals to
earn commercial returns, were introduced by the previous government to make transfer to private health
provision that much easier. This rump of a
privatisation policy continues to influence the NHS
asset base by encouraging ward and hospital closure
in much the same way as the old window tax encouraged bricking up windows (D Mayston, paper in
preparation). To pay for the privately financed project
in Birmingham, the expected rate of return on
redeveloped hospitals has been increased from the
current 6% to 13%.3 So hospital costs are artificially
inflated rather than hospitals simply being too
expensive.
The problem for Birmingham and Edinburgh, and
the other areas which are using capital spending to
drive out labour from the hospital sector, is that no one
is yet clear what sort of health service will result or
whether it will save money. In the past the impact of
capital charges was felt by chronically sick and elderly
patients. The impact of the private finance initiative will
be wider as early discharge and prevented admission
have their effects across the board. How will these
patients react when they are directed to cheaper alternatives and what sort of care can they expect to find
there?
David Price Research fellow
Social Welfare Research Unit, University of Northumbria, Newcastle
upon Tyne NE7 7XA
1
2
3
Gaffney D, Pollock AM. Can the NHS afford PFI? London: BMA, 1997.
Dunnigan MG. Lothian Health Board’s Integrated Healthcare Plan
1996-2003: A high risk strategy. Glasgow: University of Glasgow, 1997.
Birmingham Health Authority. Birmingham’s health care future? The
consultation document. Supplement: new models for health care. Birmingham:
Birmingham Health Authority, 1997.
Preventing RhD haemolytic disease of the newborn
Give anti-rhesus(D) immunoglobulin to all pregnant women who are Rh negative
I
n 1994 we saw the 25th anniversary year of the
worldwide introduction of anti-rhesus (D) immunoglobulin prophylaxis, one of the most successful
prophylactic programmes in medical history. In 1977,
110 cases of stillbirth or neonatal death due to RhD
haemolytic disease were registered in Britain; by
1992, this figure seemed to have dropped to nine
cases.1 In this issue of the BMJ, however, Whitfield and
colleagues show that this is a serious underestimation
(p 1504).2 Extrapolation of their Scottish figures
suggests that there are 50 deaths a year due to RhD
haemolytic disease in Britain. They conclude that the
discrepancy is due to underreporting because deaths
due to early abortion (before 24 weeks) and late
neonatal death (second to fourth week of life) are not
included in official figures.
The policy introduced in 1969 was to give anti-D
immunoglobulin to RhD negative women only after
the birth of a RhD positive infant or after a sensitising
event in pregnancy, such as antepartum haemorrhage.
The question now is whether the policy should be
extended so that all pregnant women who are RhD
negative receive an antenatal dose of anti-D immunoglobulin. More than half the deaths in Whitfield’s
study were due to sensitisation of the mother between
the 28th and 40th week of her first pregnancy. Routine
antenatal prophylaxis with anti-D immunoglobulin
would largely have prevented sensitisation in the first
pregnancy, and neonatal death after a subsequent
pregnancy. Another paper in this issue reports that
introducing a programme of antenatal prophylaxis in
Derbyshire reduced the sensitisation rate from 1.12%
of women at risk to 0.28%, a finding that is consistent
with previous work.3-5 This paper also shows that the
shift from hospital to community based antenatal
care did not reduce the programme’s success. The
lower incidence of sensitisation might also have been
1480
due to a heightened awareness among general
practitioners of the need to give anti-D immunoglobulin to women with antepartum haemorrhage; the
number of such events doubled during the period
under study.
There are two reasons why RhD immunisation and
sensitisation of pregnant women still occurs. Firstly,
because not all women receive anti-D immunoglobulin
when they should, and, secondly, because small, undetected leaks of fetal blood into the maternal circulation
can occur during the third trimester of pregnancy.
Both issues were extensively discussed in April during
a consensus conference organised by the Royal
College of Physicians of Edinburgh and the Royal College of Obstetricians and Gynaecologists. Contributors
were worried that current guidelines for prophylaxis
with anti-D immunoglobulin were not being followed.6 7 Potentially sensitising events during pregnancy such as blunt abdominal trauma and antepartum haemorrhage require that the mother is protected
by an injection of anti-D immunoglobulin, preferably
after a Kleihauer test or equivalent to asses the size of
the fetomaternal bleed. Kleihauer tests may be difficult
to perform and standardise,8 but health professionals
should at least recognise sensitising events and
consider giving anti-D immunoglobulin “blind” if testing is not available. Pregnant women who are RhD
negative should also be educated about the condition,
so that they recognise the need for treatment should a
sensitising event occur.
The conference agreed that the second cause of
RhD immunisation—fetomaternal bleeding during the
last trimester—would be largely eliminated by giving
anti-D immunoglobulin antenatally in one of two possible dose schedules: two doses of 500 IU, one at 28
weeks and the other at 34 weeks, or one dose of 1000
IU given between 28 and 30 weeks. Both options are
BMJ VOLUME 315
See p 1504
6 DECEMBER 1997
Editorials
effective.3 9 Antenatal programmes are cost effective if
reserved for women in their first pregnancy, as they
were in the Derbyshire study.10 11
Although the conference consensus panel considered that this restriction could not be justified on ethical or economic grounds, it seems that the tremendous
global shortage of anti-D immunoglobulin may
prevent extension of antenatal prophylaxis to all pregnant women at risk. The shortage of anti-D
immunoglobulin from voluntary RhD negative donors
is a major concern worldwide. The risk of transmitting
viruses with the immunising RhD positive red cells
cannot be completely eliminated and has caused the
withdrawal of many voluntary donors.12 Most countries
now accept immunoglobulin preparations from
commercial sources, prepared from paid donors’
blood, and these seem to be safe. Once human monoclonal anti-D immunoglobulin is available, shortage
will be a thing of the past; but, while it is being
developed and tested, the need for immunised donors
will continue well into the next century, and the debate
about how best to use this scarce resource will
continue.
1
6
2
3
4
5
Clarke C, Hussey RM. Decline in deaths from rhesus haemolytic disease
of the newborn. J R Coll Physicians London 1994;28:310-1.
Whitfield CR, Raafat A, Urbaniak SJ. Underreporting of mortality from
RhD haemolytic disease in Scotland and its implications: retrospective
review. BMJ 1997;315:1504-5.
Tovey LAD, Townley A, Stevenson BJ, Taverner J. The Yorkshire antenatal
anti-D immunoglobulin trial in primigravidae. Lancet 1983;2:244-6.
Huchet J, Dallemagne S, Huchet C, Brossard Y, Larsen M,
Parnet-Mathieu F. Application anté-partum du traitement préventif
d’immunisation chez les femmes Rhésus negatif. J Gynecol Obstet Biol
Reprod 1987;16:101-11.
Thornton JG, Page C, Foote G, Arthur GR,Tovey LA, Scott JS. Efficacy
and long term effects of antenatal prophylaxis with anti-D immunoglobulin. BMJ 1989;289:1671-3.
Bob van Dijk Consultant in blood transfusion
SeroConsult, 9752 EV 19 Haren, Netherlands
Ghosh S, Murphy WG, Implementation of the rhesus prevention
programme: a prospective study. Scott Med J 1994;39:147-9
7 Huggon AM, WatsonDP, use of anti-D in an accident and emergency
dept. Arch Emerg Med 1993;10:306-9
8 Duguid JKM. Antenatal serological testing and prevention of haemolytic
disease of the newborn. J Clin Pathol 1997;50:193-6
9 Bowman JM. The prevention of Rh immunization. Transfus Med Rev
1988;2:129-50
10 Torrance GW, Zipursky A. Cost-effectiveness of antepartum prevention of
Rh immunization. Clin Perinatol 1984;11:267-81
11 Vick S, Cairns J, Urbaniak S, Whitfield C, Raafat A. Cost-effectiveness of
antenatal anti-D prophylaxis. Health Econ 1996;5:319-28
12 Crespigny L de, Davison G. Anti-D administration in early pregnancyTime for a new protocol. Aust NZ J Obstet Gynaecol 1995;35;385-7
Helicobacter pylori and its interaction with risk
factors for chronic disease
We are not quite ready to recommend green tea and saki to the exclusion of coffee
See p 1489
BMJ VOLUME 315
I
n this issue Brenner and colleagues document the
effects of lifestyle on Helicobacter pylori infection in
447 patients in a German rural practice (p 1489).1
They found that the H pylori infection rate appeared to
be reduced by alcohol consumption (>100 g per week),
increased by coffee drinking (>3 cups per day), and
non-significantly increased by smoking. In the past,
when idiopathic gastric hyperacidity was considered to
be the chief cause of dyspeptic symptoms, smoking, coffee, and alcohol were often implicated as exacerbating
the condition and advice given to eliminate these habits.
Increasingly since 1984, however, when Warren and
Marshall hypothesised that campylobacter-like organisms were the cause of peptic ulcer disease,2 there has
been a need to re-evaluate the role of these traditional
risk factors. This need has been addressed by Brenner et
al with some seemingly unexpected results.
There are reasons why alcohol, coffee, and smoking
might have little effect on, or even increase, the hostility
of the gastric environment to H pylori. The acid gastric
pH prevents most organisms from thriving or even surviving in the stomach. H pylori, however, has an electropositive internal milieu; twice the number of basic amino
acids, argenine and lysine, as Haemophilus influenzae and
Escherichia coli; and powerful urease activity, with the
ability to produce both ammonia and factors that inhibit
parietal cell acid production.3 All these attributes make
survival of H pylori in the stomach less influenced by the
reduction in pH which may accompany coffee, smoking,
and alcohol consumption.4-6
6 DECEMBER 1997
Alcohol and smoking tend to increase the rate of
gastric emptying.4 7 However, H pylori has at least five
different adhesins to attach itself to gastric epithelial
cells. These, together with the fact that the surface
lipopolysaccharide of H pylori is several orders of magnitude less immunoganic than other enteric bacteria,
may allow it to adhere to the gastric mucosal surface
and create a quasi “stagnant loop syndrome” despite
adequate luminal flow. These functions have been
highlighted by the recent complete sequencing of the
H pylori gene.3 One of the remaining questions is why
alcohol is effective in the face of this powerful bacterial
genetic machinery. Is it the time honoured antiseptic
effect of alcohol, to which H pylori has developed no
defence?
H pylori infection is the most common human
chronic bacterial infection, affecting half the population and associated with duodenal and gastric
ulceration, chronic active and atrophic gastritis, gastric
carcinoma, and mucosa associated lymphoid tissue
lymphoma.8 As well as the risk of cancer, Brenner et al
draw attention to the role of H pylori in cardiovascular
disease in a fashion similar to Chlamydia pneumoniae.9
They speculate that the beneficial effect of low alcohol
consumption on coronary heart disease may relate to
its ability to reduce H pylori infection. A corollary might
be that smoking and coffee consumption contribute to
the risk of coronary heart disease by increasing
susceptibility to H pylori infection. Furthermore, since
coffee raises serum cholesterol concentrations unless it
1481
Editorials
is filtered to remove the offending diterpenes,10 will filtered coffee lacking deterpenes still promote H pylori
infection? Should we screen patients at high risk of
coronary heart disease for H pylori? If so, can we use
the more generally available IgG test requiring simply
a serum sample or should a 13C urea breath test be the
standard? Finally, will alcohol consumption be advised
to help eliminate H pylori?
Brenner et al outline the possible positive effect of
alcohol and the negative effect of coffee on the
elimination of H pylori. H pylori is important because of
its association with a broad range of diseases. However,
in view of its high prevalence, other factors, including
genes and environment, are likely to be pivotal in the
pathogenesis of disease. The high prevalence of this
bacteria also raises the question of routine screening
and, if so, which test to use. In addition, should changes
in lifestyle to reduce H pylori infection be encouraged—
1
2
3
4
5
6
7
8
Brenner H, Rothenbacher D, Bode G, Adler G. Relation of smoking and
alcohol and coffee consumption to active Helicobacter pylori infection:
cross sectional study. BMJ 1997;315:1489-2.
Warren JR, Marshall B. Unidentified curved bacilli in the stomach of
patients with gastritis and peptic ulceration. Lancet 1984;i:1311-5.
Tomb J-F, White O, Kerlavage AR, Clayton RA, Sutton GG, Fleischmann
RD, et al. The complete genome sequence of the gastric pathogen Helicobacter pylori. Nature 1997;388:539-47.
Endoh K, Leung FW. Effects of smoking and nicotine on the gastric
mucosa: A review of clinical and experimental evidence. Gastroenterol
1994;107:864-78.
Singer MV, Leffmann C, Eysselein VE, Calden H, Goebell H. Action of
ethanol and some alcoholic beverages on gastric acid secretion and
release of gastrin in humans. Gastroenterol 1987;93:1247-54.
van Deventer G, Kamemoto E, Kuznicki JT, Heckert DC, Schulte MC.
Lower esophageal sphincter pressure, acid secretion, and blood gastrin
after coffee consumption. Dig Dis Sci 1992;37:558-69.
Jian R, Cortot A, Ducrot F, Jobin G, Chayvialle JA, Modigliani R. Effect of
ethanol ingestion on postprandial gastric emptying and secretion,
bilipancreatic secretions and duodenal absorption in man. Dig Dis Sci
1986;31:604-14.
Cover TL, Blaser MJ. Helicobacter pylori infection, a paradigm for
for example, green tea and saki rather than coffee in
Japan? Certainly, these data are the closest we have had
to support St Paul’s injunction to “take a little wine for
thy belly’s sake.” The emphasis, however, may be on
“little”—that is, 1-2 drinks daily. Above this level, despite
the continued lower risk of coronary heart disease, the
risk of cancer, stroke, and all cause mortality rises.11 12
Finally, recent publications from the health professionals study of 47 806 men found no increased risk of
duodenal ulcer associated with coffee, smoking, or
alcohol13 but a negative association with dietary fibre
and vitamin A.14 We require more studies before we are
confident in giving advice on this important and complex topic.
David J A Jenkins Professor of nutritional sciences
and medicine
Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada
M5S 3E2
9
10
11
12
13
14
chronic mucosal inflammation: Pathogenesis and implications for eradication and prevention. Adv Int Med 1996;41:85-117.
Patel P, Mendall MA, Carrington D, Strachan DP, Leatham E, Molineaux
N, et al. Association of Helicobacter pylori and Chlamydia pneumoniae
infections with coronary heart disease and cardiovascular risk factors.
BMJ 1995;311:711-4.
Urgert R, Meyboom S, Kuilman M, Rexwinkel H, Vissers NM, Klerk M, et
al. Comparison of the effect of cafetiere and filtered coffee on serum concentration of liver amino transferases and lipids: six month randomized
controlled trial. BMJ 1996;313:1362-6.
Boffetta P, Garfinker L. Alcohol drinking and mortality among men
enrolled in an American Cancer Society prospective study. Epidemiology
1990;1:342-8.
Marrot M, Brunner E. Alcohol and cardiovascular disease: the status of
the U shaped curve. BMJ 1991;303:565-8.
Aldoori WH, Giovannucci EL, Stampfer MJ, Rimm EB, Wing AL, Willett
WC. A prospective study of alcohol, smoking, caffeine, and the risk of
duodenal ulcer in men. Epidemiology 1997;8:420-4.
Aldoori WH, Giovannucci EL, Stampfer MJ, Rimm EB, Wing AL, Willett
WC. A prospective study of diet and the risk of duodenal ulcer in men.
Am J Epidemiol 1997;145:42-50.
Pressure to prescribe
Involves a complex interplay of factors
T
wo thirds of consultations with general
practitioners end with the issuing of a prescription.1 The decision to prescribe is influenced by many factors, to do with the doctor, the
patient, the doctor-patient interaction, and the wider
social context, including the effects of advertising and
the financial incentives and disincentives for all
parties.2-6 Hardline advocates of rational drug use do
not look kindly on variations in prescribing patterns
that cannot be explained by purely clinical factors.1
The prescriber who allows the “Friday night penicillin”
phenomenon to sway his or her clinical judgment
tends to do so surreptitiously and with a guilty
conscience.
But such behaviour is the rule rather than the
exception. Several studies have shown that the
prescribing behaviour of doctors is heavily influenced
by their perceptions of the social background, beliefs,
attitudes, and expectations of the patient,2 as well as the
uncertainty of the diagnosis.5 7
1482
Bradley identified several patient factors associated
with doctors’ discomfort when prescribing (or refusing
to prescribe) drugs: extremes of age and of social class,
non-white ethnic group (because of perceived differences in expectations), patients with a medical or paramedical background, and patients whose history the
doctor either did not know or knew only too well (frequent attenders, “heartsink,” and “fat file” patients).5
Although the powerful placebo effects of drugs
prescribed in such situations are well documented,8 so
are the maladaptive behaviours that follow. Patients
who receive a prescription for a self limiting condition
are more likely to expect (and receive) one if the same
symptoms recur.9 The profession is regularly called on
to acknowledge its vulnerability to the allegedly intense
pressure to issue a prescription when none is needed,5
yet hard data on the extent of this pressure remain
sparse. Patients originating from the Asian subcontinent in particular have been accused of attending their
general practitioners expecting prescriptions for
BMJ VOLUME 315
See p 1506
6 DECEMBER 1997
Editorials
“trivial complaints” and “ill defined conditions,”10 and
patients of Pakistani or Indian origin are more likely to
be given a prescription than those from white and West
Indian ethnic groups.7
Two recent studies in the BMJ,11 12 one of them published this week, address the gap between the prescriptions patients actually expect and what their doctors
assume that they expect. An Australian study confirmed previous findings that about half of all patients
have clear expectations for a prescription.11 After controlling for presenting condition and patients’ actual
expectations, doctors’ perceptions of these expectations also independently influenced the decision to
prescribe. A British study by Britten and Ukoumunne
(p 000) found that doctors’ perceptions of patients’
expectations for a prescription were significantly
related to patients’ hopes and educational level, and to
the broad category of diagnosis, as well as to characteristics of the individual doctor.12 The decision to
prescribe was strongly related to the doctor’s
perception of the patient’s expectations, and, overall,
doctors classified 21% of their own prescriptions as
“not strictly necessary.”
Cockburn and Pitt speculate that failure to
ascertain patients’ expectations is a major reason why
practitioners prescribe more drugs in total than
patients expect.11 But although these two studies
suggest that a doctor’s assessment of a patient’s expectation is wrong in about a quarter11 and a sixth12 of
cases, neither study provides direct evidence that this
misconstruction leads to substantial overprescribing.
Of the 255 patients in Cockburn and Pitt’s study who
expressed their expectations as anything other than
“Don’t know,” 71% were correctly classified by their
doctor as either expecting or not expecting a prescription.11 A further 21% were incorrectly classified as not
expecting a prescription; despite this, half of them
received one. Only 8% (20) of the 255 patients did not
expect a prescription when their doctor thought they
did, and 16 of these received one.
In Britten and Ukoumunne’s study, doctors admitted feeling pressure to prescribe in 66 out of 540
encounters (12%), and were more likely to prescribe
for this group of patients.12 But the number of excess
prescriptions given to this group (that is, the number
beyond what would have been expected if there had
been no perceived pressure to prescribe) amounted to
16 in 540 encounters and 5% of all prescriptions
issued.
These two most recent studies concur with
previous findings that patients who expect a prescription are many times more likely to receive one than
those who do not. This evidence is compatible with the
stereotype of demanding and manipulative patients
repeatedly forcing the hands of their reluctant doctors.
But it is also compatible with the fact that patients may
have more insight into their medical needs than their
doctors give them credit for, and that both doctor and
patient may legitimately take account of non-medical
factors in deciding whether a particular drug is necessary at a particular point in time.
The act of issuing a prescription is the culmination
of a complex chain of decisions.2 It is open to biomedical, historical, psychosocial, and commercial influences,
no aspect of which can be singled out as the “cause” of
non-rational prescribing. The search should continue
for methods to measure the interplay of these
disparate factors on the decision to prescribe.
Trisha Greenhalgh Senior lecturer
Paramjit Gill Senior lecturer
Department of Primary Care and Population Sciences,
UCLMS/RFHSM, Whittington Hospital, London N19 5NF
(email p.greenhalgh@ucl.ac.uk)
1
Audit Commission. A prescription for improvement: towards more rational
prescribing in general practice. London, HMSO, 1994.
2 Bradley CP. Decision making and prescribing patterns—a literature
review. Fam Pract 1991;8:276-87.
3 Inman W, Pearce G. Prescriber profile and post-marketing surveillance.
Lancet 1993;342:658-61.
4 Chen MM, Landefield CS. Physicians’ behavior and their interactions
with drug companies. JAMA 1994;271:648-9.
5 Bradley CP. Factors which influence the decision whether or not to prescribe: the dilemma facing general practitioners. Br J Gen Pract
1992;42:454-8.
6 Armstrong D, Reyburn H, Jones R. A study of general practitioners’ reasons for changing their prescribing behaviour. BMJ 1996;312:949-53.
7 Gill P, Scrivener G, Lloyd D, Dowell T. The effect of patient ethnicity on
prescribing rates. Health Trends 1995;4:111-4.
8 Chaput de Saintonge DM, Herxheimer A. Harnessing placebo effects in
health care. Lancet 1994;344:995-8.
9 Webb S, Lloyd M. Prescribing and referral in general practice: a study of
patients’ expectations and doctors’ actions. Br J Gen Pract 1994;44:165-9.
10 Ahmad WIU, Baker MR, Kernohan Elizabeth EM. General practitioners’
perception of Asian and non-Asian patients. Fam Pract 1991;8:52-6.
11 Cockburn J, Pitt S. Prescribing behaviour in clinical practice: patients’
expectations and doctors’ perceptions of patients’ expectations. BMJ
1997;315:520-3.
12 Britten N, Ukoumunne O. The influence of patients’ expectations of prescriptions on doctors’ perceptions and the decision to prescribe. BMJ
1997;315:1506-10.
Evidence based practice in mental health
New journal acknowledges an approach whose time has come
W
hy has it proved so difficult to narrow the
gap between research and practice in
psychiatry and mental health? The provision of mental health services is determined by many
factors, including government policy, public demand,
the behaviour of general practitioners and mental
health professionals, and the financial pressures under
which purchasers and providers of services work.
These groups often have widely disparate views about
the nature of mental disorder and the most appropriBMJ VOLUME 315
6 DECEMBER 1997
ate services, and many forces exist to keep their views
apart. Now is the time for a different approach based
on the optimum application of the available evidence—
heralded by the publication early next year of a new
journal, Evidence- Based Mental Health. This approach
will not provide easy answers and there will still be
room for discussion about interpretation of even the
very best evidence. Nevertheless, an approach that,
firstly, acknowledges that mental health services should
be fundamentally evidence based and, secondly, helps
1483
Editorials
define what constitutes the best available evidence
should clarify decision making.
The task is formidable. The public view depression
as being mainly caused by life events, may be reluctant
to consult their general practitioner, and believe that
counselling is more effective than antidepressant
drugs, which they consider to be addictive.1 Differences
also exist between the views of all the disciplines working with people with mental health problems. General
practitioners, psychiatrists, clinical psychologists, and
mental health nurses are educated and trained in
unidisciplinary structures and organisations. These
professional organisations often sustain interdisciplinary rivalries. Even within the single discipline of psychiatry considerable differences exist between clinicians who subscribe to different (and often competing)
schools of thought.
How much evidence is available on which to base
mental health services? In fact, there is much evidence,
which, although often difficult to find, is gradually
being systematically reviewed by organisations such as
the Cochrane Collaboration.2 3 Psychiatry was one of
the first medical specialties to use extensively the
randomised controlled trial, and one of the founding
principles of the profession of clinical psychology in
the 1950s was that practice should be based on the
results of experimental comparisons of treatment
methods. Multicentre randomised controlled trials
showed the effectiveness of antidepressant and
antipsychotic drugs in the 1960s.4 5 The recognition of
international variations in diagnostic practice led to the
development of explicit diagnostic criteria such as the
Diagnostic and Statistical Manual, third edition, of the
American Psychiatric Association.6 7 Methodological
innovations such as meta-analysis were first used in
health care by psychologists in psychotherapy.8
Despite these undoubted advances, however, a considerable gap remains between research and practice.
For example, important variations exist in the
treatment of depression, in the use of electroconvulsive
therapy, and in the use of stimulant medication for
attention deficit hyperactivity disorder.9-11 In mental
health nursing the recent increase in the amount of
published research has rarely been reflected by
changes in practice.12 13 In clinical psychology it has
been asserted that “in clinical practice empirically supported methods are routinely ignored in favour of
intuition and clinical experience.”14 Moreover, the public perception of mental health services has not kept up
with advances in research and practice.1 9 Others have
argued that mental health policy has usually been
influenced more by political values than evidence.15
As elsewhere, one essential ingredient required to
make mental health services clinically effective is to
ensure that clinicians know how to use evidence. There
are many workshops aimed at helping mental health
clinicians of all disciplines acquire the skills required
for evidence based practice. Clinicians also need easy
access to high quality evidence.16 In addition to the
problems of keeping up to date common to all
clinicians,17 mental health practitioners are often
geographically isolated from information resources.
Each discipline has its own journals which, even if read,
may not contain the most important research. To
improve access to the best evidence as it is published
the BMJ is starting a new journal, Evidence-Based Mental
1484
Health, in collaboration with the Royal College of Psychiatrists, the British Psychological Society, and the
Royal College of Nursing. Evidence-Based Mental Health
will be a sister journal to Evidence-Based Medicine and
Evidence-Based Nursing, and ACP Journal Club, using the
same methods and the same editorial office in the
health information research unit at McMaster University. The aim of the journal will be to provide all mental
health clinicians with the very best information about
mental health care in the form of “value added”
abstracts. The first issue will be in published in
February 1998 and a launch conference will be held in
London on 16 February 1998.
Evidence based practice offers a way of making
sure that clinical practice is based on the best available
evidence. But we also need a culture change with better
integration of patient values into the implementation
of research and a need to go beyond professional rivalries and other barriers to provide the best available
care for patients.
John Geddes Senior clinical research fellow
Centre for Evidence Based Mental Health, Department of Psychiatry,
University of Oxford, Oxford OX3 7JX
(john.geddes@psychiatry.ox.ac.uk)
Shirley Reynolds Academic Course Organiser
Clinical Psychology Doctorate Programme, School of Health Policy
and Practice, University of East Anglia, Norwich NR4 7TJ
David Streiner Professor
Departments of Clinical Epidemiology and Biostatistics and
Psychiatry, McMaster University, Hamilton, Ontario, Canada
Peter Szatmari Professor
Department of Psychiatry, McMaster University and Center for the
Study of Children at Risk, Hamilton, Ontario, Canada
The authors are the editors of Evidence-Based Mental Health
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Priest RG, Vize C, Roberts A, Roberts M, Tylee A. Lay people’s attitudes
to treatment of depression: results of opinion poll for Defeat Depression
Campaign just before its launch. BMJ 1996;313:858-9.
Geddes JR, Game D, Jenkins NE, Petersen LA, Pottinger GR, Sackett DL.
What proportion of primary psychiatric interventions are based on randomised evidence? Quality in Health Care 1996;5:215-7.
The Cochrane Library [database on disk and CD ROM]. Cochrane
Collaboration. Oxford: Update Software; 1997. Updated quarterly.
National Institute of Mental Health Psychopharmacology Service Center
Collaborative Study Group. Phenothiazine treatment in acute schizophrenia. Arch Gen Psychiatry 1964;10:246-61.
Clinical Medical Research Council. Clinical trial of the treatment of
depressive illness. BMJ 1965;i:881-6.
Cooper JE, Kendell RE, Gurland BJ, Sharpe L, Copeland JRM, Simon R,
et al. Psychiatric diagnoses in New York and London. London: Oxford
University Press, 1972 Maudsley Monograph No 20).
American Psychiatric Association. Diagnostic and Statistical Manual of
Mental Disorders 3rd ed. Washington, DC: APA, 1980.
Smith ML, Glass GV. Meta-analysis of psychotherapy outcome studies.
American Psychologist 1977;32:752-60.
Hirschfeld RM, Keller MB, Panico S, Arons BS, Barlow D, Davidoff F, et al.
The National Depressive and Manic-Depressive Association consensus
statement on the undertreatment of depression. JAMA 1997;277:333-40.
Pippard J. Audit of electroconvulsive treatment in two national health
service regions. Br J Psychiatry 1992;160:621-37.
Valentine J, Zubrick S, Sly P. National trends in the use of stimulant
medication for attention deficit hyperactivity disorder. J Paediatr Child
Health 1996;32:223-7.
Yonge O, Austin W, Zhou Qiuping P, Wacko M, Wilson S, Zaleski J, et al. A
systematic review of the psychiatric/mental health nursing research
literature 1982-1992. Journal of Psychiatric and Mental Health Nursing
1997;4:171-7.
McKenna HP. Dissemination and application of mental health nursing
research. British Journal of Nursing 1995;4:1257-63.
Wilson GT. Treatment manuals in clinical practice. Behav Res Ther
1997;35:205-10.
Ham C, Hunter DJ, Robinson R. Evidence based policymaking. BMJ
1995;310:71-2.
Smith, R. What clinical information do doctors need? BMJ
1996;313:1062-8.
Mulrow CD. Rationale for systematic reviews. BMJ 1994;309:597-9.
BMJ VOLUME 315
6 DECEMBER 1997