Anda di halaman 1dari 7



Time for a smoke? One cigarette reduces your life by 11 minutes Stumbling into occupational
health: the influenza
EditorStudies investigating the impact on 6.5 years = 2374 days, 56 976 hours, or
mortality of socioeconomic and lifestyle fac- 3 418 560 minutes EditorIn his editorial Smith describes
tors such as smoking tend to report death 5772 cigarettes per year for 54 years = how the British government has decided on
rates, death rate ratios, odds ratios, or the 311 688 cigarettes access to treatments.1 The occupational
chances of smokers reaching different ages. 3 418 560/311 688 = 11 minutes per health service of the NHS has had a similar,
These findings may also be converted into cigarette. albeit less sophisticated, experience with
differences in life expectancy. We estimated This calculation is admittedly crudeit respect to advising on the immunisation of
how much life is lost in smoking one relies on averages, assumes that the health staff against influenza this winter. The lack of
cigarette. effects of smoking are evenly spread a recognised body to give authoritative
Our calculation is for men only and throughout a smokers lifetime, presupposes advice on occupational health has led to
based on the difference in life expectancy that the number of cigarettes smoked piecemeal arrangements by trusts.
between male smokers and non-smokers throughout a lifetime is constant, and The chief medical officer has advised
and an estimate of the total number of ciga- ignores the difficulties in classifying people that routine immunisation of healthcare
rettes a regular male smoker might consume as either lifetime smokers or non-smokers.5 workers is not recommended in the United
in his lifetime. We derived the difference in However, it shows the high cost of smoking Kingdom as part of national policy but that
life expectancy for smokers and non- in a way that everyone can understand. some health trusts may offer influenza
smokers by using mortality ratios from the The first day of the year is traditionally a vaccine to staff as part of their planning for
study of Doll et al of 34 000 male doctors time when many smokers try to stop, and on winter.2
over 40 years.1 The relative death rates of 1 January 2000 a record number might be Subsequently, NHS employers have
smokers compared with non-smokers were been advised that Ministers had concluded
expected to try to start the new millennium
threefold for men aged 45-64 and twofold that the immunisation of staff should be
more healthily. The fact that each cigarette
for those aged 65-84,1 as corroborated else- regarded as an acceptable part of winter
they smoke reduces their life by 11 minutes
where.2 Average life expectancy from birth planning arrangements.3 The purpose of
may spur them on. The table shows some
for the whole population or subgroups can this planning measure is to reduce staff
better uses for the time they save.
be derived from life tables. Applying the illness. No additional funds will be made
Mary Shaw Economic and Social Research Council available.
rates of Doll et al to the latest interim life research fellow
tables for men in England and Wales, with Richard Mitchell research fellow I surveyed trusts current arrangements
adjustment for the proportion of smokers Danny Dorling reader by contacting NHS occupational physicians
and non-smokers in each five year age School of Geographical Sciences, University of
Bristol, Bristol BS8 1SS
group,3 we found a difference in life expect-
ancy between smokers and non-smokers of Competing interests: Drs Shaw and Mitchell are
6.5 years. Advice to authors
non-smokers. Dr Dorling is a smoker (20 cigarettes
We used the proportion of smokers by a day). We prefer to receive all responses electronically,
age group, the median age of starting smok- sent either directly to our website or to the
ing, and the average number of cigarettes editorial office as email or on a disk. Processing
1 Doll R, Peto R, Wheatley K, Gray R, Sutherland I. Mortality your letter will be delayed unless it arrives in an
smoked per week in the 1996 general in relation to smoking: 40 years observations on male
household survey.4 We calculated that if a British doctors. BMJ 1994;309:901-11. electronic form.
man smokes the average number of
2 Phillips AN, Wannamethee SG, Walker M, Thomson A, We are now posting all direct submissions to
Davey Smith G. Life expectancy in men who have never
cigarettes a year (5772) from the median smoked and those who have smoked continuously: 15 year our website within 24 hours of receipt and our
starting age of 17 until his death at the age follow up of large cohort of middle aged British men. BMJ intention is to post all other electronic
1996;313:907-8. submissions there as well. All responses will be
of 71 he will consume a total of 311 688 3 Office for National Statistics. 1997 mortality statistics:
cigarettes in his lifetime. general. London: Stationery Office, 1999. (Series DH1 eligible for publication in the paper journal.
No 30.) Responses should be under 400 words and
If we then assume that each cigarette 4 Office for National Statistics. General household survey.
makes the same contribution to his death, Living in Britain. London: Stationery Office, 1996. relate to articles published in the preceding
each cigarette has cost him, on average, 11
5 Suidicani P, Hein HO, Gyntelberg F. Mortality and month. They should include <5 references, in the
morbidity of potentially misclassified smokers. Int J
minutes of life: Epidemiol 1997;26:321-7.
Vancouver style, including one to the BMJ article
to which they relate. We welcome illustrations.
Please supply each authors current
Opportunities gained in stopping smoking by amount smoked appointment and full address, and a phone or
Amount smoked Life lost Opportunity gain fax number or email address for the
One cigarette 11 minutes Telephone call to friend; read of newspaper; brisk walk; corresponding author. We ask authors to declare
or fairly frantic sexual intercourse any competing interest. Please send a stamped
Pack of 20 cigarettes 3 hours 40 minutes Long film (for example, Titanic); two football matches; addressed envelope if you would like to know
one shopping trip; Eurostar journey from London to whether your letter has been accepted or rejected.
Paris, including visit to cafe; running in London
marathon; or tantric sex
Letters will be edited and may be shortened.
Carton of 200 cigarettes 1.5 days Visit to friends or family; one very serious shopping
trip; Wagner opera; flying round the world; or romantic
night away

BMJ VOLUME 320 1 JANUARY 2000 53


throughout the United Kingdom on an conclude that a randomised controlled trial instructions. Patients values are recorded as
email network. Of 14 replies from England, of evidence based medicine versus shroud a basis for decisions on medical treatment
Scotland, and Northern Ireland, five trusts waving (as standard practice) is urgently (rather than including explicit instructions
were not offering vaccinations, three were required. on specific treatments). They identify core
adopting a low key response, three were This process of evidence based commis- values and beliefs in the context of terminal
actively targeting groups of staff such as sioning was comparatively resource inten- care that are important to the patient.3 4
accident and emergency staff, and three sive. Unless there is some perceived benefit Values histories take a goal based rather
were offering vaccinations to all their staff. for clinicians and commissioners of care, it than prescriptive approach, giving guidance
The large scale vaccination programmes will wither on the vine. We do not expect all on a policy to be implemented rather than
were being funded either by the trust or by cost effective new treatments to be funded the medical means to the end. The legal
local purchasers. automatically. However, there should be persuasiveness of them is less strong, but
What is the explanation? The case that scope for the evidence to have a greater they may be useful adjuncts when a person
routine influenza vaccinations will reduce impact on the system and for the emerging is seeking to have an advance refusal
staff absences remains to be made. Indeed, a National Centre for Clinical Excellence to respected or they may provide valuable
recent review of NHS staff absences during be able to say yes as well as no to costly guidance in their own right.5 In general,
the influenza epidemics of 1993-4 and treatments. the trend towards greater use of values in
1996-7 found that absenteeism was not June So chief pharmacist advance statements is more useful to
affected.4 Preventing nosocomial infection J Howard Scarffe professor of medical oncology patients and intensive care doctors than is
of susceptible patients merits study but has Christie Hospital NHS Trust, Manchester M20 4BX the trend towards increasingly specific
nothing to do with the millennium holiday. Elizabeth Rous consultant in public health medicine wording of treatments to be refused. Use
The main focus of risk assessment, there- Manchester Health Authority, Gateway House, of values histories should therefore be
Manchester M60 7LP
fore, seems to be the prospect of embarrass- encouraged.
Robbie Foy clinical research fellow
ment should planning go awry. Some trusts Further details on values histories may
Scottish Programme for Clinical Effectiveness in
are responding by minimising the risk of Reproductive Health, Department of Obstetrics be obtained from the Living Will and
being criticised, others are taking pro- and Gynaecology, University of Edinburgh, Values History Project, BM 718, London
fessional advice and acting accordingly. Edinburgh EH3 9AW WC1N 3XX.
Scotland and Northern Ireland are Chris Docker director
interested observers. This is a curious Voluntary Euthanasia Society of Scotland,
example of postcode prescribing but a 1 In brief. NHS to have cancer tsar. BMJ 1999;319:1152. Edinburgh EH1 3RN
(30 October.)
familiar example of postcode occupational 2 Sikora K. Cancer survival in Britain. BMJ 1999;319:461-2.
health in the NHS. (21 August.)
3 Foy R, So J, Rous E, Scarffe H. Commissioner and special- 1 Winter R, Cohen S. ABC of intensive care: withdrawal of
John Harrison senior lecturer in occupational ist perspectives of prioritising new cancer drugs: impact of treatment. BMJ 1999;319:306-8. (31 July.)
medicine the evidence threshold. BMJ 1999;318:456-9. 2 Docker C. Living wills/advance directives. In: McLean S,
Department of Environmental and Occupational ed. Contemporary issues in law, medicine and ethics. Aldershot:
Medicine, Medical School, Newcastle upon Tyne Dartmouth, 1996:179-214.
3 Gibson J. Values history focuses on life and death
NE2 4HH decisions. Med Ethics 1990;5:1-2, 17. 4 Lambert P, Gibson J, Nathanson P. The values history: an
Decisions to withdraw innovation in surrogate medical decision-making. Law Med
Health Care 1990:18:202-12.
1 Smith R. Stumbling into rationing. BMJ 1999;319:936.
(9 October.)
treatment 5 Docker C. Living wills. In: Finance and law for the elderly
client. London: Butterworths-Tolley (in press).
2 Department of Health. Influenza immunisation. CMOs
update 23. London: DoH, 1999.
3 Department of Health. Immunisation of NHS staff against Values histories are more useful than
influenza for winter 1999-2000 including the millennium
period. London: DoH, 1999. (HSC 1999/214.) advance directives Treatment can sometimes be withdrawn
4 Nguyen-Van-Tam J, Granfield R, Pearson J, Fleming D,
EditorWinter and Cohen recognise one at home
Keating N. Do influenza epidemics affect patterns of
sickness absence among British hospital staff? Infect Control of the problems with advance directives EditorIn their article on withdrawal of
Hosp Epidemiol 1999;20:691-4. when they correctly state: The advance treatment in the intensive care unit Winter
refusal of treatment is legally binding and Cohen rightly emphasise the
provided certain conditions are met. . . . A importance of ensuring the comfort of
problem still exists unless they are precisely dying patients as well as caring for the
Lack of funding will inhibit worded.1 family.1 The intensive care unit can be noisy
evidence based commissioning Traditional advance directives are and stressful for grieving relatives. It may
of cancer treatments becoming less and less useful, partly as a be difficult to provide a comfortable and
result of lack of data on when treatment private environment for families and
EditorWe welcome the appointment of becomes futile in different clinical scenarios. patients who are terminally ill.
a cancer tsar to improve the treatment of When advance directives were first intro- In appropriate circumstances it is possi-
patients with cancer in the NHS. 1 However, duced, the application of standard heroic ble to take patients to their home and with-
more resources and greater transparency measures, often without reasonable expec- draw treatment there. Since 1996 we have
in their allocation will be necessary tation of result, was far more common than taken six patients from our unit home to die.
to improve survival rates in the United it is today. In that situation, a general The table gives details of the last four
Kingdom in relation to its comparable advance directive about refusing, say, cardio- patients.
European neighbours.2 pulmonary resuscitation, was an appropri- In all cases the relatives found this
We have previously reported our efforts ate statement of common sense. The approach helpful. Relatives often state that,
in evidence based commissioning of cancer situations facing modern intensive care given the choice, the patient would prefer to
services for 1997-8.3 Since then we have units are far more complex. The tendency die at home rather than in hospital. More-
repeated the exercise in the 1998-99 and towards precise wording in advance direc- over, relatives seemed better able to cope
1999-2000 contracting rounds. However, tives to make them legally binding has made with their grief in the familiar surroundings
most available growth monies have been it difficult for them to keep up with the pace of their homes.
absorbed by large increases in activity and of medical technology.2 There are several conditions that should
pay awards. Little money has been left for An alternative approach that is finding be satisfied before a patient in intensive care
new expensive cancer drugs, despite strong increasing favour, either as an adjunct to the can be taken home to die. There has to be
evidence of their cost effectiveness and the advance directive or as a stand alone instru- medical consensus that continued intensive
increasing willingness of clinicians to set ment, is the values history. Values histories therapy is futile, and withdrawal of ventila-
priorities between cancer treatments. We relate to the declarants values rather than tion or cessation of inotropes should

54 BMJ VOLUME 320 1 JANUARY 2000


Courts can help resolve disagreement in

Details of four patients who had intensive treatment withdrawn at home
difficult cases
Age (years) and sex Diagnosis APACHE II score Length of stay (hours) Ventilated
EditorIn the same issue as Winter and
51, male Cellulitis, septic shock 33 79 No
Cohens article on withdrawing treatment in
54, female Pneumonia 32 19 Yes intensive care,1 Clare Dyer reported on a
61, male Cardiac arrest outside 24 52 Yes recent Court of Appeal case involving the
treatment of a 12 year old child.2 3 One
58, male Cerebellar haemorrhage 24 436 Yes
important aspect of the case was the view
Written consent was obtained from relatives for publication of patients details. expressed by the court and by Michael
APACHE=acute physiological and chronic health evaluation.
Wilks, chairman of the BMAs ethics
committee, that if there was fundamental
preferably lead quickly to death. Organ on the basis that death was imminent or disagreement between the parents and doc-
donation should not be an option, and the that prolonging treatment was futile. A tors an application to the family division of
patients death should not require referral combination of worsening severity of the High Court could be an appropriate way
to the coroner. Ideally the patient should illness, minimal response to aggressive forward. This is an important point to bear
live locally. An intensive care nurse and treatment, and poor future quality of life in mind when faced with the problems iden-
doctor need to be available to accompany were often the stated reasons for with- tified by Winter and Cohen.1
the patient and to manage the process of drawal. However, our objective analysis of The idea of futility is a quagmire.4
withdrawal of treatment once at the these deaths using severity of illness Despite this I would agree that in most cases
patients home. Transport needs to be measurements with the paediatric risk of careful and sensitive discussion does lead to
booked in advance. mortality (PRISM II) score, showed that 22 a resolution of any disagreement. As the
A clear drug palliation plan should be (50%) of these children had mortality risks recent case2 shows, however, the disagree-
formulated that uses opiates and benzodi- > 50% and only 10 (23%) had mortality ment sometimes cannot be resolved without
azepines to ensure that the patient is sedated risks > 80%. Twenty (45%) had conditions the benefit of the views of the court. Doctors
and absolutely comfortable. It is essential to that were associated with extremely poor should be sensitive to the existence of such
discuss the plan with the patients general long term prognosis. Therefore it seems situations and be aware that a short notice
practitioner so that ongoing care for the that most decisions concerning futility of application to the court is a potential course
family can be instigated. Under certain of action in these cases.
treatment were based on subjective
circumstances the general practitioner may
appraisal of the childs deteriorating condi- Raj Mohindra specialist registrar, general medicine
issue a death certificate. Heatherwood Hospital, Ascot SL5 8AA
tion. A study in a paediatric intensive care
Once all the conditions are satisfied the
unit in Malaysia has shown that personal
possibility of taking the patient home can be
bias of the intensive care doctor and the
offered to the family. The process should be 1 Winter B, Cohen S. ABC of intensive care: withdrawal of
patients sociocultural background may
discussed in detail and they should be told of treatment. BMJ 1999;319:306-8. (31 July.)
the risk of death during transfer. It is essen- influence these decisions.2 The concept of 2 R v Portsmouth Hospitals NHS Trust ex parte Carol Glass.
futility may differ quite substantially among Times 26 July 1999.
tial that all family members agree before the 3 Dyer C. Mother fails to win right to control treatment for
patient is transferred. intensive care doctors and also among son. BMJ 1999;219:278. (31 July.)
patients. 4 Zucker MB, Zucker HD, eds. Medical futility and the evalua-
We believe that when intensive care is tion of life sustaining interventions. New York: Cambridge
deemed futile, consideration should be We disagree with Winter and Cohen that University Press, 1997.
given to withdrawing treatment in the it is easier to withhold a treatment than to
patients home. withdraw it once it has been instituted. We
often find ourselves in situations where the
Paul Frost consultant in intensive care
inappropriateness of intensive care and
So much time for so little:
on behalf of medical and nursing staff from the
department of intensive care medicine, limitations of treatment have not been Italys pharmaceutical industry
Middlemore Hospital, Auckland, New Zealand discussed with the family before referral to and doctors information needs
the intensive care team. In the current audit
1 Winter B, Cohen S. ABC of intensive care: withdrawal of EditorTwo local experiences in northern
treatment. BMJ 1999;319:306-8. (31 July.)
all the patients had received mechanical
Italy may be relevant to the discussion follow-
ventilation and a variable period of aggres-
ing the recent editorial by Griffith on doctors
sive intensive care before treatment was
reasons for not seeing drug representatives.1
Most decisions are based on subjective withdrawn.
In 1992, eight general practitioners from
appraisal With advances in critical care these ethi-
Guastallae evaluated the amount of time
EditorWinter and Cohens review of cal dilemmas are expected to intensify. Until spent with pharmaceutical representatives.
withdrawal of treatment in intensive care such a time that objective criteria for futility On average, each doctor had 435 visits per
touched on several issues that continue to become sufficiently accurate for individual year from 102 different companies and
pose vexing problems to intensive care doc- patient prognostication, we agree with the spent a total of 58 hours every year talking
tors.1 These ethical problems apply to both authors that the principles that guide inten- to pharmaceutical representatives. To
adult and paediatric intensive care. We agree sivists on end of life decision making should reduce this time, four doctors allowed up to
with the authors that death in intensive care be based on beneficence and non- two visits per day, and four allocated a single
often follows a process of withholding or maleficence to our patients.1 3 weekly session, allowing up to eight visits,
withdrawal of treatment and that these Adrian T Goh visiting fellow paediatric intensive care but this approach did not lead to the desired
issues are becoming increasingly common results (table).
as well as complex. Quen Mok consultant paediatrician in intensive care In 1994, four of these general practition-
In a prospective audit of deaths over Paediatric Intensive Care Unit, Great Ormond ers contacted 102 drug companies, asking to
Street Hospital for Children, London WC1N 3JH
nine months in a paediatric intensive care receive information based on systematic
unit in the United Kingdom, we found that evaluation of available evidence rather than
44 (84%) of the 52 deaths resulted from a glossy booklets and favourable trials only.
process of withholding or withdrawal of 1 Winter B, Cohen S. ABC of intensive care: withdrawal of Nothing changed except that staff specialists
treatment. BMJ 1999;319:306-8. (31 July.)
treatment. Withdrawal of active treatment, 2 Goh AYT, Lum LCS, Chan PWK, Bakar F, Chong BO.
joined the visit. In 1996, drug companies
such as inotropes or renal replacement Withdrawal and limitation of life support in paediatric were asked to send their representatives only
therapy or extubation from mechanical intensive care. Arch Dis Child 1999;80:424-8. when invited or when relevant information
3 Royal College of Paediatrics and Child Health. Withholding
ventilation, was by far the commonest or withdrawing life saving treatment in children. A framework for was available in advance. Most representa-
process. These decisions were often made practice. London: RCPCH, 1997. tives stopped visiting doctors surgeries, and

BMJ VOLUME 320 1 JANUARY 2000 55


Effect of temazepam on Although we did not measure respira-

Interactions between drug representatives and tion directly, our data indicate that a low
eight general practitioners in Guastalla, Italy, in ventilatory response at dose of a short term benzodiazepine can
1992-3 moderate altitude impair respiration at moderate altitude.
No of Total These findings seem to contradict Dubow-
visits Mean GP time
EditorDubowitzs study of the effect of
itzs conclusion. Treatment with temazepam
per duration (hours/ temazepam on oxygen saturation at high
at stable conditions after altitude acclimati-
Mode of access year (min) year) altitude found that benzodiazepines do not
sation may not impair respiration, but initial
Unrestricted (1992) 435 7 58 have a depressant effect.1 He explains the
stages of acute respiratory adaptation to
Maximum two visits/day 400 8 55 discrepancy between his findings and those
(1993) hypoxia at altitude are inhibited. Similar
of previous studies by the fact that other
Group appointment (6-8 340 10 50 results were found after 50 g alcohol at
studies have investigated the effect of
visits per week) (1993) moderate altitude.5 Caution in the use of
long acting benzodiazepines.2 Dubowitzs
benzodiazepines to treat sleep disorders at
only three out of 102 agreed to answer spe- probands were investigated after altitude altitude is therefore necessary, especially in
cific questions. acclimatisation while walking to Everest base the initial stages of altitude acclimatisation.
In 1992, four general practitioners from camp, whereas climbers in Europe mainly
Georg Rggla head of department
Imola, after the publication of a drug formu- engage in short periods of mountaineering.
lary,2 asked 105 drug companies to submit We therefore evaluated the effect of 10 mg
Berthold Moser student
only clinically relevant information and temazepam on respiration in non- Department of Internal Medicine, Municipal
answer doctors specific questions. This put a acclimatised Alpine climbers at moderate Hospital of Neunkirchen, Neunkirchen, Austria
complete stopwhich still persiststo visits altitude. Martin Rggla lecturer
from drug representatives. We performed a randomised, double Department of Emergency Medicine, University of
We analysed some typical information blind, placebo controlled, crossover trial in Vienna, Austria
packages and found that drug oriented seven healthy men aged 21 to 27. Partici-
information is often flawed, biased, or pants at 171 m altitude were randomised to 1 Dubowitz G. Effect of temazepam on oxygen saturation
misleading3 4: protocols of ongoing studies receive either 10 mg temazepam or placebo. and sleep quality at high altitude: randomised placebo
are used as evidence of clinical benefits; controlled crossover trial. BMJ 1998;316:587-9.
Three days later the men were given the 2 Rggla G, Rggla M, Wagner A, Seidler D, Podolsky A.
unpublished data on file are quoted as same medication and taken by cable car to Effect of low dose sedation with diazepam on ventilatory
reliable references; pharmacological or response at moderate altitude. Wien Klin Wochenschr
3000 m. The procedure was repeated after 1994;106:649-51.
molecular effects are overemphasised; and one week, with the men crossed to the other 3 Hills M, Armitage P. The two-period cross-over clinical
all this is almost regularly accompanied by arm of the study. Arterial blood samples trial. Br J Clin Pharmacol 1979;8:7-20.
4 Spiro SG, Dowdeswell IR. Arterialised ear lobe blood sam-
an invitation to gather your own experiences were obtained from the ear lobe before and pling for blood gas tensions. Br J Dis Chest 1976;70:263-8.
with this drug and then judge. This is not one hour after temazepam or placebo was 5 Rggla G, Rggla H, Rggla M, Binder M, Laggner AN.
patient oriented information drawn from a Effect of alcohol on acute ventilatory adaptation to mild
taken.4 Arterial oxygen partial pressure and hypoxia at moderate altitude. Ann Intern Med
systematic evaluation of available evidence. 1995;122:925-927.
carbon dioxide partial pressure were ana-
Transforming advertising into reliable
lysed on an IL Synthesis 25 blood gas
information is probably impossible. Pharma-
analyser (Instrumentation Laboratory,
ceutical companies are well equipped to pro-
duce and disseminate information about
Milan, Italy). Differences in blood gas Guidelines are needed if drug
concentrations before and after temazepam testing of those arrested by the
their drugs, and they are one of the main
or placebo at each altitude were analysed by
sources of information for doctors. Health
paired t tests. police becomes compulsory
authorities should arrange a similar system,
The table shows the results of blood EditorThe prime ministers announce-
using dedicated professionals to provide doc-
gas analysis before and after temazepam. ment at the last Labour Party conference that
tors with valid and unbiased information.
Interventions of this sort should be included At 171 m blood gas concentrations did the government proposes to introduce
in the current effort by Italys health service to not change significantly after temazepam. compulsory drug and DNA testing for
produce and implement practice guidelines. At 3000 m the arterial oxygen pressure people arrested for indictable offences before
Local centres could be used to make this decreased and carbon dioxide pressure they have been convicted will raise some
information easily available to doctors and to increased significantly after temazepam. important ethical issues for healthcare pro-
determine and evaluate the best strategies for The mean decrease in arterial oxygen fessionals if it does eventually become law.1 As
dissemination and implementation.5 concentration between altitudes was 0.77 pointed out in the Economist,2 the upshot of
(95% confidence interval 8.02 to 3.69) compulsory DNA testing might be that every
Emilio Maestri general practitioner
Gilberto Furlani general practitioner kPa (P < 0.01) and the mean increase in alleged shoplifter could be held down forcibly
Guastalla, Reggio Emilia, Italy arterial carbon dioxide concentration was while a mouth swab is taken.2 Collecting sweat
Fabio Suzzi general practitioner 0.3 (0.46 to 4.11) kPa (P < 0.05). Placebo did for a drug test by wiping the forehead of a
Imola, Bologna, Italy not affect blood gas concentrations at either restrained and resisting subject with a swab
Annalisa Campomori pharmacist altitude. would be no more dignified.
Giulio Formoso epidemiologist
Nicola Magrini head
Unit of Drug Evaluation and Evidence-Based Arterial oxygen (PaO2) and carbon dioxide (PaCO2) concentrations (kPa) of seven men before and one
Primary Care, Centro per la Valutazione della
hour after 10 mg temazepam at 171 and 3000 m
Efficacia della Assistenza Sanitaria (CeVEAS), 41100
Modena, Italy PaO2 PaCO2
171 m 3000 m 171 m 3000 m
1 Griffith D. Reasons for not seeing drug representatives.
BMJ 1999;319:69-70. (10 July.) Case No Before After Before After Before After Before After
2 Magrini N, Vaccheri A, Suzzi F, Montanaro N. 150 farmaci 1 12.2 12.2 9.3 8.6 4.5 4.3 4.3 4.4
per il medico di famiglia. Rome: Il Pensiero Scientifico
Editore, 1993. 2 11.2 11.6 8.6 8.4 4.7 4.9 4.4 4.7
3 Ziegler Mg, Lew P, Singer BC. The accuracy of drug infor- 3 12.1 12 8.9 8.2 4.4 4.5 4.4 4.7
mation from pharmaceutical sales representatives. JAMA
4 11 11.4 9.1 4.9 4.9 4.8 4.0 4.3
4 Herxheimer A. Basic information that prescribers are not 5 10.9 11 8.5 8.1 4.7 4.4 4.1 4.4
getting about drugs. Lancet 1987;8523:31-3.
6 12.5 12.2 9.1 8.1 4.5 4.8 4.4 4.7
5 Thorsen T, Mkel M, eds. Changing professional practice.
Theory and practice of clinical guidelines implementation. 7 12 12.2 9.4 8.4 4.4 4.5 4.0 4.8
Copenhagen: Danish Institute for Health Services Research, Mean (SD) 11.7 (0.63) 11.8 (0.48) 9 (0.29) 8.3 (0.2) 33.4 (1.4) 34.6 (1.8) 4.2 (0.19) 4.5 (0.2)

56 BMJ VOLUME 320 1 JANUARY 2000


The BMA has issued helpful guidelines remaining patients with this diagnosis to was 9.1 for second generation oral contra-
for police surgeons who have been asked to have idiopathic deep vein thrombosis after ceptives.3 In the Leiden study we found four-
examine people in police custody when review of clinical histories.3 Thus we fold to sevenfold risk increases for third
consent for such examination is not concluded that many young women with a versus second generation oral contracep-
forthcoming.3 I hope that similar guidelines computer diagnosis of thrombophlebitis do tives among younger women (most new
will be produced for medical practitioners not have documented idiopathic deep vein users would be young); this was over and
who may be asked to participate in thrombosis. above an overall threefold to fourfold risk
collecting body fluids for drug testing from The results from the General Practice increase for second generation oral contra-
arrested people without their consent and in Research Database provided in Lawrenson ceptives.4
analysing, interpreting, and using the results et als letter are uninterpretable because, Jick et al found 9.2-fold and 5.6-fold
in samples obtained in this way. firstly, it is unclear whether the general prac- increases in risk between third and second
A R W Forrest professor of forensic toxicology titioners reported incidence according to generation oral contraceptives during the
Department of Forensic Pathology, University of number of users rather than person time at first six months of use.5 Multiplying these
Sheffield, Sheffield S3 7ES risk, and, secondly, presumably the diagnosis cumulative relative risks with an overall
of venous thromboembolism was not vali- baseline incidence of, say, 5 per 100 000
dated from clinical records. By contrast, woman years yields incidences of the order
1 Another fine mess [editorial]. New Law Journal 1999;149:
1429. cases of venous thromboembolism that we of magnitude found by Herings et al,2
2 Tony Blair and liberty [editorial]. Economist 1999; reported were validated as idiopathic from thereby confirming the suspicion that the
3 Pownall M. Doctors should obtain informed consent for
clinical records.3 incidences are higher than generally
intimate body searches. BMJ 1999;318:310. An interpretation of estimates of the believed, especially in young and first time
incidence of venous thromboembolism users.
related to current use of oral contraceptives Jan P Vandenbroucke professor, department of
requires that the method for identifying rel- clinical epidemiology
Incidence of venous evant cases as well as the method for
thromboembolism in users of estimating person time at risk be clearly Kitty W M Bloemenkamp registrar, department of
combined oral contraceptives detailed,4 as in our paper.3 In the absence of obstetrics, gynaecology and reproductive medicine
Frits R Rosendaal professor, thrombosis and
such information, comparison of rates of ill- haemostasis research centre
Methods for identifying cases and ness between different reports is unjustified Frans M Helmerhorst lecturer, department of
estimating person time at risk must be and surely misleading. obstetrics, gynaecology and reproductive medicine
Leiden University Medical Centre, PO Box 9600,
detailed Hershel Jick associate professor of medicine 2300 RC Leiden, Netherlands
Boston Collaborative Drug Surveillance Program,
EditorLawrenson et al reported a crude Boston University School of Medicine, 11 Muzzey Competing interests: Professors Vandenbroucke
incidence of venous thromboembolism of Street, Lexington, MA 02421, USA and Rosendaal have no competing interests. Dr
38 per 100 000 women who used combined Bloemenkamp has been involved in, and Dr
Competing interests: None declared. Helmerhorst has supervised, studies sponsored or
oral contraceptives, based on information assigned by various pharmaceutical companies that
derived from the General Practice Research 1 Lawrenson RA, Whalley A, Simpson E, Farmer RDT. DoH manufacture oral contraceptives, but none of these
Database, without describing how they seems to have underestimated incidence of venous throm- companies has funded their studies on the
boembolism in users of combined oral contraceptives. BMJ comparative merits of second and third generation
derived their estimate.1 This estimate is 1999;319:387. (7 August.) oral contraceptives.
closely similar to the incidence reported in a 2 Farmer RD, Lawrenson RA, Thompson CR, Kennedy JG,
prior publication of theirs, which was based Hambleton IR. Population-based study of risk of venous
thromboembolism association with various oral contra- 1 Lawrenson RA, Whalley A, Simpson E, Farmer RDT. DoH
on a different automated medical database.2 ceptives. Lancet 1997;349:83-8. seems to have underestimated incidence of venous throm-
In their letter they provide no details on the 3 Jick H, Jick SS, Gurewich V, Myers MW, Vasilakis C. Risk of boembolism in users of combined oral contraceptives. BMJ
idiopathic cardiovascular death and nonfatal venous 1999;319:387. (7 August.)
validity and specificity of the diagnosis of thromboembolism in women using oral contraceptives 2 Herings RMC, Urquhart J, Leufkens HGM. Venous
venous thromboembolism or the presence with differing progestogen components. Lancet 1995;346: thromboembolism among new users of different oral con-
1589-93. traceptives. Lancet 1999;354:127-8.
of medical risk factors and no information 4 Jick H, Garca Rodrguez LA, Prez Gutthann S. Principles 3 Poulter NR, Farley TMM, Chang CL, Marmot MG, Meirik
on the person time at risk for current users of epidemiologic research on adverse and beneficial drug O. Safety of combined oral contraceptive pills. Lancet
of oral contraceptives. effects. Lancet 1998;352:1767-70. 1996;347:547.
4 Bloemenkamp KWM, Rosendaal FR, Helmerhorst FM,
An incidence must be based on person Bller HR, Vandenbroucke JP. Enhancement by factor V
time at risk. Despite apparent deficiencies, Risk is particularly high with first use of Leiden mutation of risk of deep-vein thrombosis
the authors provide an incidence that is oral contraceptives associated with oral contraceptives containing third-
generation progestagen. Lancet 1995;346:1593-6.
roughly twice as high as the incidence that EditorLawrenson et al show that the risk 5 Jick H, Jick SS, Gurewich V, Myers MW, Vasilakis C. Risk of
colleagues and I reported in a paper derived of venous thromboembolism during use of idiopathic cardiovascular death and non-fatal venous
thromboembolism in women using oral contraceptives
from the General Practice Research Data- oral contraceptives might be higher than with differing progestagen components. Lancet 1995;
base; we included only idiopathic cases of estimated by the Medicines Commission of 346:1589-93.
venous thromboembolism (those possibly the United Kingdoms Department of
related directly to use of oral contracep- Health.1 That is especially likely for first use. Authors reply
tives).3 Lawrenson et al conclude that our In a paper that was published at about EditorOur letter in response to the
reported estimate was substantially lower the same time Herings et al found an Department of Healths statement calcu-
than the true incidence. incidence of 90 per 100 000 woman years in lated an incidence of idiopathic venous
On the basis of their previous publi- the first episode of use of third generation thromboembolism of 36.5 per 100 000
cation,2 Lawrenson et al seem to have oral contraceptives; for second generation exposed woman years in users of combined
included both outpatients and inpatients contraceptives the figure was 24 per oral contraceptives. This was based on a
with any one of five computer recorded 100 000; for the youngest ages and during structured review of published papers and
diagnoses as patients with venous thrombo- the first year of use the incidences became compared well with the findings from our
embolism, without documentation from higher than 100 per 100 000.2 These data recent study, for which we used the General
clinical records. The five diagnoses include a are in line with those found in earlier Practice Research Database. That study
computer recorded diagnosis of thrombo- case-control studies. included only cases of diagnosed venous
phlebitis. Colleagues and I found that, A stratified analysis of the World Health thromboembolism with supporting evi-
on the General Practice Research Database, Organisations study found a 21.6-fold dence of anticoagulation treatment. Further
over 90% of subjects with this computer increase in the risk of venous thrombosis information from general practitioners
recorded diagnosis alone were treated among first ever users of third generation showed that in 83% of cases there was
as outpatients, so clinical histories were oral contraceptives during the first year of supportive evidence (venograms, Doppler
not available; we did not find any of the use (relative to never use); this relative risk ultrasound scans, or lung scans).

BMJ VOLUME 320 1 JANUARY 2000 57


We have argued that general practition- Some patients are happy for accident and three had ischaemic heart
ers in the United Kingdom do not treat disease. Thirty five (40%) patients were
patients with anticoagulants without the
doctors to make decisions ex-smokers (mean age 57.0 (13.4)), with a
support of diagnosis by a hospital consult- EditorI agree with Bastian that the mean smoking burden of 26.4 (18.8) pack
ant.1 The general practitioners responses relationship between doctors and their years. In this group one had had a
indicated that all our cases had been patient is fundamentally changing and that cerebrovascular accident and six had ischae-
referred to hospital, although for a fifth this many patients should be more closely mic heart disease. Only 17 (20%) patients
had not been recorded on the General Prac- involved in the decision making regarding (mean age 53.8 (9.1)) had never smoked, and
tice Research Database. We believe that the their health needs.1 I do, however, not accept none of these had vascular disease.
strategy used by Jick et al of requiring a Bastians statement that doctors do not have These results suggest that smoking is a
computer record of hospital admission led a sophisticated understanding of the consul- risk factor for Q fever and show that in our
to a 20% underestimate of cases.2 In all our tation process and wish to raise a point that patients it has been the current or
studies of combined oral contraceptives we is often overlooked by many people in con- ex-smokers who have developed cardiovas-
have used person time at risk for calculating sumer organisations. cular disease. It is therefore essential that
incidences, based on a 28 day pack. Jick et al An appreciable body of patients in the current and past cigarette smoking are
used 30 days, which would have led to a fur- community is more than happy to offload added to a reanalysis by the Geneva group
ther 7% error. the decision making about their treatment before C burnetii can be taken as the
Vandenbroucke et al cite a further study to the doctor. Many of these patients are explanation of the excess cardiovascular
by Herings et al that supports our assertion intelligent, motivated people who approach morbidity and mortality observed.
of a higher incidence of venous throm- the medical profession for advice based on Martin Wildman research fellow in respiratory
boembolism than that reported by the knowledge and experience. As an anaesthet- medicine
ist I have spent much time discussing with
Department of Health. Their letter cites
examples of higher incidences of venous patients the pros and cons of one form of Jon G Ayres professor of respiratory medicine
Heartlands Research Institute, Birmingham
thromboembolism among users of third perioperative analgesia versus another, and, Heartlands Hospital, Birmingham B9 5SS
generation compared with second genera- almost invariably, when I ask the patient
tion combined oral contraceptives. We agree about their preference the response is, 1 Lovey P, Morabia A, Bleed D, Peter O, Dupuis G, Petite J.
that new users of combined oral contracep- Whatever you think best, doctor or Which Long term vascular complications of Coxiella burnetii
one would you have? infection in Switzerland: cohort study. BMJ 1999;319:
tives seem to be at an increased risk of idio- 284-6. (31 July.)
pathic venous thromboembolism; this is Every patient we see is unique and 2 Smith DL, Ayres JG, Blair I, Burge PS, Carpenter MJ, Caul
should be treated as such. Some patients EO, et al. A large Q fever outbreak in the West Midlands:
important when considering the risk clinical aspects. Respir Med 1993;87:509-16.
ascribed to third generation combined oral clearly gain reassurance from the medical 3 Hawker JI, Ayres JG, Blair I, Evans MR, Smith DL, Smith
contraceptives. profession adopting the politically incorrect EG, et al. A large outbreak of Q fever in the West Midlands:
windborne spread into a metropolitan area? Commun Dis
In our studies we have compared users paternalistic approach, and these people Public Health 1998;1:180-7.
of levonorgestrel 150 g plus ethinyloestra- must not be forgotten. 4 Ayres JG, Flint N, Smith EG, Tunnicliffe WS, Fletcher TJ,
Hammond K, et al. Post-infection fatigue syndrome
diol 30 g with users of formulations of Ian Taylor specialist registrar in anaesthesia following Q fever. Q J Med 1998;91:105-23.
combined oral contraceptives containing Princess Margaret Hospital, Swindon SN1 4JU
desogestrel and gestodene and found no
difference in risk of venous thromboembo- 1 Richards T. Australias consumer champion. BMJ 1999; Authors reply
319:730. (18 September.)
lism between the formulations. Interestingly, EditorThe data reported by Wildman
a further analysis has shown that, despite a et al cannot be interpreted as evidence that
large change in prescribing practice since smoking is a risk factor for Q fever. The
November 1995 from third to second Long term vascular prevalence of current smoking in these 110
generation formulations, the overall risk of complications of Coxiella patients was unusually high (55%), even
venous thromboembolism in users of burnetii infection compared with the general local population
combined oral contraceptives has not of East Birmingham (35%).1 2 These patients
changed.3 had mostly been admitted to hospital and
The key point of our letter was that the Cardiovascular risk factors cannot be been identified on the basis of a recent diag-
true rate of venous thromboembolism in ignored nosis of pneumonia or fever of unknown
users of combined oral contraceptives is EditorIn their paper on the potential for origin. Smokers may therefore have been
probably nearer 37 per 100 000 exposed infection by Coxiella burnetii to be a risk factor overselected simply because they were more
woman years and that the figures quoted for cardiovascular disease Lovey et al suggest severely affected by respiratory disease. This
by the Department of Health are under- that the established mode of transmission of overrepresentation of smokers among cases
estimates. C burnetii is unlikely to be associated with risk may in turn explain the positive association
R A Lawrenson senior lecturer in public health factors for cardiovascular disease. They also of smoking and Q fever observed in the
A Whalley research officer say that the unavailability of baseline data on study by Ayres et al, in which 71 patients
E Simpson research officer such risk is unlikely to influence their with Q fever were compared with patients in
R D T Farmer professor of pharmacoepidemiology and
findings.1 However, in an outbreak of Q fever general practice who were free of febrile ill-
public health
European Institute of Health and Medical Sciences, pneumonia affecting 147 patients in the ness.3 By comparison, in the Swiss outbreak
University of Surrey, Guildford GU2 5RF United Kingdom in 1989 (not referred to by of 1983, 191 patients had fever, but 224
Lovey et al) we found that of 110 patients in others were identified by serum analysis of
Competing interests: The department in which the whom smoking history was available for the the general population (n = 3036).4 Q fever
authors work has received funding from Organon
and Schering for investigating venous thrombo-
time of the infection, 60 were current pneumonia affected only 68 patients (36%
embolism in women using oral contraceptives. smokers, 28 were ex-smokers, and only 22 of the patients with symptoms and 16% of
had never smoked.2 3 the whole group). Only 8 of the 415 patients
A subsequent case-control study in this (2%) were admitted to hospital.
1 Farmer RDT, Lawrenson RA. Third generation oral
contraceptives and venous thrombosis. Lancet
cohort confirmed smoking to be a risk factor Wildman et al also argue that cardio-
1997;349:732-73. for Q fever.4 Follow up of 87 (59%) patients in vascular disease does not occur in patients
2 Jick H, Jick SS, Gurewich V, Myers MW, Vasilaskis C. Risk of clinic nine years after the original outbreak infected with Q fever who never smoked.
idiopathic cardiovascular death and nonfatal venous
thromboembolism in women using oral contraceptives identified 31 (35%) as current smokers (mean Three cases of ischaemic heart disease in 31
with differing progestagen components. Lancet 1995;346: age 51.2 (SD 10.2)), with a mean smoking smokers may, however, not be different from
3 Farmer RDT. The impact of the 1995 pill scare in the UK. burden of 33.0 (15.8) pack years. In this group no cases in 17 people who have never
Pharmacoepidemiol Drug Safety 1999;8(suppl 2):S187. one patient had had a cerebrovascular smoked. Their numbers (31 + 35 + 17) do

58 BMJ VOLUME 320 1 JANUARY 2000


not add up to 87, nor the proportions would accept such new technology if they psychosocial care are few, and people rely
(35 + 40 + 20) to 100%. were shown how to use it. heavily on the charity sector for palliative
Overall, the direct evidence for a Jadad mentioned email, which is care. The Calman-Hine network is only par-
possible confounding effect of smoking on another facility I offer to my patients and tially and patchily implemented, with little
the relation between Q fever and coronary local practitioners for rapid response to any new funding. There is still gross inequity in
heart disease and stroke is weak. Moreover, queries or anxieties they may have and to the quality of cancer care in the United
smoking does not seem to have been a avoid delays caused by post and telephones.5 Kingdom. And there is no National Cancer
strong confounder in other studies that It is vital to ensure equal access to Centre to set the gold standard.
investigated the infectious origin of cardio- technology and information. Even compara- Ominously for the NHS are the clear
vascular disease.5 We therefore think it tively poor countries may be able to afford signs that there will be major improvements
unlikely that a confounding effect of telecommunications facilities as the costs are in cancer treatment over the next decade.
smoking, if it exists, could be strong enough coming down, but they paradoxically remain Next year will see the licensing of herceptin
to fully explain the observed association unable to afford clinical expertise because of for breast cancer by the European Medi-
between acute infection with Coxiella burnetii a lack of trained specialists in such countries. cines Evaluation Agency. This will be the
and the incidence of coronary heart disease Cyberclinics of the type I have developed forerunner of sophisticated targeted treat-
and stroke. may enable patients from these countries to ments that will require integrated molecular
Pierre-Yves Lovey senior resident, department of obtain medical advice in a more cost and therapeutic solutions. The human
internal medicine effective way than in many other formats of genome project, scheduled for completion telemedicine currently available, which can in 2003, will lead to new ways of predicting
Alfredo Morabia director, division of clinical be constrained by cost and technical failure. which people are at high risk of developing
epidemiology cancer. In the United Kingdom the basic
University Hospital of Geneva, Rue Micheli du Badal Pal consultant rheumatologist
Crest 24, 1211 Geneva 14, Switzerland Department of Rheumatology, Wythenshawe cancer services need urgently to be put in
Hospital, Manchester M23 9LT order to meet these new challenges, allowing
1 Smith DL, Ayres JG, Blair I, Burge PS, Carpenter MJ, Caul
British patients to reap the benefits of global
EO, et al. A large Q fever outbreak in the West Midlands: progress. If the NHS cannot do it the private
clinical aspects. Respir Med 1993;87:509-16. 1 Jadad AR. Promoting partnerships: challenges for the
2 Hawker JI, Ayres JG, Blair I, Evans MR, Smith DL, Smith internet age. BMJ 1999;319:761-3. (18 September.) sector willbut at a price.
EG et al. A large outbreak of Q fever in the West Midlands: 2 Shepperd S, Charnock D, Gann B. Helping patients access The National Cancer Forum is a deriva-
windborne spread into a metropolitan area? Commun Dis high quality health information. BMJ 1999;319:764-6.
(18 September.)
tive of the Calman-Hine group and is
Public Health 1998;1:180-7.
3 Ayres JG, Flint N, Smith EG, Tunnicliffe WS, Fletcher TJ, 3 Pal B, Laing H, Estrach C. A cyberclinic in rheumatology. convened irregularly. At its most recent
Hammond K, et al. Post-infection fatigue syndrome J R Coll Phys London 1999;33:161-2. meeting this topic was raised, but the forum
following Q fever. Q J Med 1998;91:105-23. 4 Pal B. Cyber/virtual clinic in osteoporosis? A question-
4 Dupuis G, Petite J, Pter O, Vouilloz M. An important naire based feasibility study towards a rapid consultation was told that ministers have gone too far
outbreak of human Q fever in a Swiss Alpine valley. Int J and advisory service [abstract]. Ann Rheum Dis 1999: down the road on this one. Why? It has
Epidemiol 1987;16:282-7. (suppl) No 1010.
5 Nieto FJ, Adam E, Sorlie P, Farzadegan H, Melnick JL, 5 Pal B. Email contact between doctor and patient. BMJ never been on the agenda of the forum.
Comstock GW, et al. Cohort study of cytomegalovirus 1999;318:1428. During my two years with the World Health
infection as a risk factor for carotid intimal-medial thick- Organisation I have visited many countries
ening, a measure of subclinical atherosclerosis. Circulation
1996;94:922-7. in different economic environments.
New guidelines for urgent Nowhere does a state health department
have a two week policy or send out such gra-
referral of patients with cancer tuitous advice to doctorsyet many have
Internet helps communication are waste of energy better outcomes for cancer care.
between doctors and patients Although streamlining the diagnostic
EditorGuidelines have recently been sent
process and improving treatment resources
EditorI found the articles by Jadad and out by the Department of Health for wide
makes good sense, these new targets are a
Shepperd et al on using the internet to pro- consultation in the United Kingdom. 1 The
waste of energy. Along with NHS Direct and
vide patients with information helpful.1 2 aim is try to reduce mortality from cancer in
its latest variantits interactive websitethey
The internet is beginning to transform the United Kingdom by guaranteeing that
are simply window dressing. A significant
health care, but there is fear that clinicians all patients who present to their general
effort to improve the quality of cancer care is
may remain ill prepared unless action is practitioner with symptoms possibly due to
essential if we are really going to make an
taken now. Jadad was unable to identify my cancer are seen in a hospital clinic within
impact and save lives.
article on this subject.3 I have developed the two weeks. But there is no evidence whatso-
concept of cyberclinics in rheumatology and ever that delay in diagnosis is a large Karol Sikora global vice president
Clinical Research (Oncology), Pharmacia and
osteoporosis with active participation of problem in the United Kingdom. Efforts to Upjohn, via Robert Koch, Milan 20152, Italy
patients and am now able to offer consulta- implement the so called two week policy are
tion and advice on the net.3 4 misplaced. The guidelines themselves repre-
The internet can provide vast amounts sent a reasonable distillation from basic 1 Department of Health. Referral guidelines for suspected
of information, but the material is variable medical textbooks, but they are patronising cancerconsultation document. London: DoH, 1999. (HSC
1999/241.) (
in quality. I have created one of the first to doctors, who have the longest university
departmental websites in rheumatology training of any professional group. Most
and osteoporosis ( copies are likely to end up in the bin.
rheumat/), which features information that I Those in the United Kingdom know that Breast feeding and obesity
have written personally or selected (after it continues to lag behind its European Two biometricians should have been
studying it online) and consider reliable so neighbours in cancer survival. Politicians included with Johannes Hebebrands as
that patients can access this information must be educated about this rather than be authors of the second letter in this cluster.
with confidence. A further point raised by allowed to continue with their obsession They are Frank Geller and Andreas Ziegler,
Jadadstrategies to increase health lit- about waiting list targets. The whole exercise both from the Institute of Medical Biometry
eracy in generalis also important. I have is reminiscent of the man who loses his keys and Epidemiology at Philipps-University of
recently held a road show to demonstrate on the dark side of the road and looks for Marburg at Marburg in Germany.
to patients the basic aspects of computers them under the light across the street. In the
and the internet, and how the internet can United Kingdom the shortage of cancer
be used to access information relating to specialists is more than 500; 1.2bn is
health care. Patients were able to participate needed to bring radiotherapy equipment up
and gain hands-on experience. My overall to date; and an extra 170m a year is needed
impression was that even elderly people for chemotherapy. Facilities for good

BMJ VOLUME 320 1 JANUARY 2000 59