Joumal ofMedical Ethics, 1996; 22: 232-237

Consent to clinical research - adequatelyvoluntary or substantially influenced?Sarah Hewlett University ofBristol

AbstractIn clinical research the giving of consent by the patientoften lies within the context of illness or thedoctor/patient relationship. On exploration of theseissues it would appear unlikely that the patient's consentis free of substantial influences, some of which may bestrong enough to be controlling.

Five categories of consent are suggested: voluntary,involuntary, coerced, enforced and partially voluntary. Itis argued that consent in clinical research is substantiallyinfluenced and thus only partially voluntary. Severalpractical strategies are proposed to ensure adequatelyvoluntary consent by reducing some circumstantialinfluences when consent to clinical research is obtained.

IntroductionClinical research is necessary to establish the safetyand efficacy of a therapy. It may include, forexample, the testing of nursing or physiotherapytechniques, as well as the testing of new drugs, thefocus of this paper.

Clinical testing of a new drug is required by theMedicines Control Agency (MCA) before a productlicence for that drug can be given. As the drug isbeing tested for safety and efficacy, patients takingpart will be put at risk of unknown side-effects andmay also be randomised to receive either theunproven drug or a placebo. When we enter apatient into such a clinical trial we are normallytesting a drug appropriate to his or her condition.However, because of randomisation we may notalways be selecting a particular drug for a particularpatient's needs. If the comparator drug is a placebothen we may not be acting in that patient's bestinterests. In some cases the patient is being used as ameans to an end: establishing the safety and efficacyof a new drug for the benefit of future patients. Thejustification for this is that the only way to establishthe most effective and safe treatments to improvehealth (desired by most members of society) isby clinical research.' By consenting, the patient

Key wordsConsent; voluntary; clinical research

knowingly agrees to this goal and makes it his own,becoming an active participant in the research sothat it cannot then be said that the patient is beingused as a means to an end. Clearly the quality of thatconsent is therefore vital.

What does clinical research mean for thepatient?The scientific gold standard for clinical research isthe randomised controlled trial (RCT) whichattempts to establish statistically the risk/benefit ratioof the drug by reducing bias and controlling for vari-ables.2 This it does by controlling the selection ofwell-characterised groups of patients, randomisingthem into treatment groups, using standardisedoutcome measurements performed by blindedassessors, and stipulating a population size powerfulenough to answer the question being asked.Although not without criticism3 the RCT is acredible test of safety and efficacy within the medicalworld and some would say it would be unethical tointroduce a new drug without RCT data.4

Patients participating in an RCT will usually needto attend hospital regularly for safety and efficacyassessments - perhaps clinical examination, bloodand urine samples, and possibly X-rays. Such visitsmay take several hours and their frequency variesfrom every few days to every month, as does theduration of the RCT (from a few days to severalyears). Neither the safety nor the efficacy of the drugis yet proven and in addition the patient may receivea placebo during the course of the study. When takentogether these are not inconsiderable inconveniencesfor the patient and although hopefully kept tominimum impact, there is sometimes an element ofrisk.

How can we protect the patient in anRCT?Patients participating in clinical research areprotected by the requirement for a Clinical TrialsExemption Certificate from the MCA; theDeclaration of Helsinki guidelines on the ethicalconduct of the research5; Research Ethics Committee

copyright. on January 25, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.22.4.232 on 1 August 1996. D

ownloaded from

Sarah Hewlett 233

approval which must be obtained beforehand6;Guidelines on Good Clinical Practice developed bythe Association of British Pharmaceuticals Industryand now adopted as an EC directive7; peer review bypublication (including details of ethics committeeapproval); public accountability through publicaccess to medical journals, media reports etc, andfinally the personal and professional codes of theresearchers. By far the strongest protection however,is consent.

The purpose of consentClinical research contains an unquantified risk ofharm for the patient, it may also involve the use of aplacebo and/or a drug which is not fully licensed,and the patient's treatment is randomly allocated:the patient is being used as a means to an end.However, the principle of respect for personsrequires that we respect the wishes of others andhave concern for their welfare.8 Consent is anautonomous authorisation by one person to permitanother person to carry out an agreed procedurewhich affects the subject and therefore by askingpatients to consent to research, we respect theirwishes, enable them to be self-governing and upholdthe principle of respect for persons. In clinicalresearch consent is vital in maintaining trust in thedoctor/patient relationship and preventing thepatient from being deceived.

What is consent?Four elements must be present for consent to bemorally acceptable: competence, information, under-standing of that information, and voluntariness.9Beauchamp and Childress suggest that a person iscompetent "if and only if that person can makereasonable decisions based on rational reasons".10Competency however, is a complex issue and is notthe focus of this paper, in which we address voluntaryconsent to clinical research by the competent patient.Information must be sufficient and unbiased, suchthat a substantially autonomous decision can bemade. Consent can rarely be fully informed but thisdoes not mean it is not adequately informed." Forexample, I might buy a car based on adequate infor-mation about its history and performance, ratherthan full information about the workings of its enginewhich I would not understand. Similarly, whilst itis not possible fully to inform patients about thepharmacological actions of the research drug it maybe possible adequately to inform them. The mannerin which information is provided will also influenceadequacy of understanding and it is the responsibilityof the health care professional to ensure the patientunderstands the proposed research.'2 This can bedone by using appropriate terminology - presented inan unbiased manner- encouraging questions, andascertaining what the patient understands. These

first three elements of consent have been widelyresearched and debated but little work has been per-formed on the fourth element voluntariness which is, Ipropose, seriously threatened in clinical research.

VoluntarinessAll decisions are made within the context andinfluence of people or circumstances. Thus it wouldbe difficult (if not impossible) to define the notion of"fully voluntary" consent. There would always bearguments that further efforts could be made toreduce influences. As with competence, informationand understanding the issue is whether voluntarinessis adequate. Beauchamp and Childress describevoluntariness as being independent of controllinginfluences exerted by others and discuss coercion(the intentional use of a credible threat), manipula-tion (of information to influence a decision) andpersuasion (convincing by presenting rationalreasons). '3 This model depends largely on theintentional actions of others. It fails to address theinfluence of circumstances, which I believe iscommon in consent to clinical research and whichposes an easily overlooked threat to voluntariness.For consent to research to be adequately voluntarytwo factors must be fulfilled: the absence of con-trolling influences and the ability to choose eitherone of at least two options.

Influences or controlling influences?Whilst circumstances and people will always influ-ence any decision we make, it is the responsibility ofthe researcher to ensure that in clinical research,these are not so strong as to be controlling. Unlikehealthy volunteers, many patients invited to partici-pate in clinical research will have an illness and theexperience of illness (which may at times includepain, disability, fear of deterioration or death,physical or emotional dependence) '" and the accom-panying psychological response (possibly depres-sion, mourning, denial, anger, anxiety, passivity andregression to an invalid role) 15 may well reduceautonomy. In addition, a request to enter a researchtrial may of necessity come at the time of, or soonafter, the shock of diagnosis. Furthermore, somepatients in hospital feel vulnerable - unaware of thenormal routine or what is expected of them, andreduced, as they are, to wearing night-clothes. Is itpossible to make a reflective decision, free of stronginfluences in such a situation?The doctor/patient relationship is centred around

patients' trust that doctors act in their best interests.Thus even though the doctor has explained that treat-ment in the research trial is randomised and accord-ing to a strict protocol, patients often still believe thatthe doctor will only act in their best interests. In arecent trial 41% of patients believed there were norisks in a phase II trial of a new anti-inflammatory

copyright. on January 25, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.22.4.232 on 1 August 1996. D

ownloaded from

234 Consent to clinical research - adequately voluntary or substantially influenced?

agent, despite being told it was unlicensed and beingtested for safety.'6 So strong is this trust that patientsmay agree to anything the doctor suggests and eventhe invitation to participate may be viewed as a rec-ommendation rather than a request. Patients may feelflattered by the request and under an obligation tohelp because of past care received.

For many patients the relationship is an unequalone, with the doctor being perceived as a powerfulfigure on whom they depend, making it difficult forthem to take an unnaturally dominant role andrefuse the doctor's request. Patients may fear that ifthey do so, the doctor will be displeased and theirfuture care will be jeopardised. Patients may viewclinical research as a means of access to care (whichmay be true for patients in countries where manypeople do not have health care insurance) and mayalso equate the frequent safety visits with improvedcare.

Extemal pressures may be brought to bear on thepatient: the patient may be asked to decide immedi-ately, without time to reflect; family and friends maysuggest the patient "ought" to participate; the factswe give the patient will be laden with our beliefs; wemay use closed questions or statistics which areloaded in a persuasive fashion (for example 75% ofpatients do well rather than 25% do badly); doctorsmay be under pressure to reach target numbers,with academic and financial kudos resulting from acompleted trial. Thus the influences on the patientare multiple and complex.

Theoretical options or realistic options?Is it possible that the option of consenting or notconsenting to clinical research may be theoreticalrather than realistic? For example, there is currentlyno effective treatment for AIDS and so an invitationto test a potential therapeutic agent will be the onlymeans of access to the only possible alternative todeath. Whilst the purist might suggest that thispatient does still have a choice, in the real world Ibelieve few patients would find death a realisticoption. Patients invited to participate in clinicalresearch may be limited by practical difficulties. In adraft research proposal, general practices were to berandomised to particular treatments for osteo-arthritis (OA) as specific arms of a research trial.As standardised treatment regimens in differentpractices were being compared, all OA patients fromeach GP would be treated in this new manner. Apatient has the choice of declining the research, butif this is the only treatment his GP can now offer,does he have any real option? This initial proposalhas now been extensively revised after these issueswere raised.

These two illustrations offer different examples ofhow options can be limited. For the patient withAIDS this is by circumstance but for the patient withOA, the initial trial design itself limited his options.

Most choices in life are limited in some way butwhere those limits are imposed by others then thequality of the consent, being deliberately con-strained, may be morally unacceptable.We have looked at influences on the patient and

limitations of options, but in practice are thesestrong enough to make it difficult for the patient torefuse consent? Refusal rates are low, with none atall in recent years in a neighbouring unit (personalcommunication) and in our own unit, in one studyonly four out of 28 patients refused.'6 We comparedthese four patients who declined to enter a phase IIItrial of a specific anti-rheumatoid drug ("refusers")with 17 patients who consented to take part in asimilar Phase II trial ("consenters"). The consentersrated how difficult it would have been for them tosay "no" if they had wanted to refuse and therefusers rated how difficult it actually was (10 cmvisual analogue scale [VAS], very easy to very diffi-cult). There was a significant difference between theconsenters' median score (0 9, range 0-5 4) and therefusers' (7-15, range 5 2-9-6) (p=0-0036, MannWhitney U test).'6 Therefore, assuming this to be areasonable representation of patients invited to takepart in drug trials, we must appreciate that patientsexperience considerable difficulties when it comes torefusing the doctor's request, such that their consentmay not be adequately voluntary.

What is consent if it isn't voluntary?I propose that there are a number of situations inwhich consent is not adequately voluntary and mayor may not be acceptable.'7 As a simple illustrationlet us assume that I wish to purchase a jacket andhave sufficient funds to buy either red jacket A orgreen jacket B. If I decide to buy jacket A because Ilike the way I look in it then this would be a reflectivedecision, free of constraints or coercive pressures - anadequately voluntary choice. However, if I had acompulsion to wear green I might automaticallychoose jacket B. This reflex action which I could notcontrol would be an involuntary choice. Third,suppose my employer had insisted that all staffshould wear red jackets and that non-compliant staffwould be sacked then my subsequent decision to buyred jacket A would have been coerced by the action ofothers. However, if I had lost C10 on the way to theshop and had to buy jacket B because it was all that Icould afford, then my choice would have beenenforced by circumstances. There is however, still aform of consent which fits none of these categories.Suppose a friend who accompanies me declares thatjacket A suits me best and is of a better quality. Thisopinion, coming from someone I respect, leads me todecide to buy jacket A. Such a choice is not voluntarybut neither is it involuntary, coerced by use of acredible threat; nor is it enforced by circumstances.But it is not free from pressure (however well-intentioned) and has been substantially influenced. I

copyright. on January 25, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.22.4.232 on 1 August 1996. D

ownloaded from

Sarah Hewlett 235

would suggest that such a decision is partially volun-tary.

Should consent to clinical research also bethought of as partially voluntary? It is rarely involun-tary, coerced or enforced but as I have shown thereare many potential influences on the patient. Whilstsome pressures may be subtle and not strong enoughto enforce or coerce consent, it is quite possible thata combination of the factors described may sub-stantially influence consent so as to make it onlypartially voluntary. We may have to accept thatmuch consent in clinical research is partially (andpossibly adequately) voluntary, just as it is partiallybut adequately informed.The point at which influences cease to be sub-

stantial but become sufficiently controlling as tomake consent morally unacceptable may vary fromperson to person but there are ways in which we can

and should try to reduce those influences or makethe patient aware of them.

Reducing influences on consent to clinicalresearchWhilst it could be argued that the adequatelyautonomous patient is the best person to decide if hisor her decision is adequately voluntary, as health care

professionals we have a professional responsibility totry and promote adequately voluntary consent by themanner in which we obtain it. This is similar to our

responsibilities in trying to ensure adequatelyinformed consent by the manner in which we informpatients. Areas where influences could be reducedinclude the doctor/patient relationship, selection ofpatients, education of the researcher and informationgiven during recruitment.

The doctor/patient relationshipInterventions here might allow the patient to stepback from the relationship and consider theproposal in a more detached manner. This could bedone by using a researcher/doctor as well as a

carer/doctor, by using a patient advocate or byrecruiting patients in groups. Separating the doctorsinto researcher/doctors and carer/doctors'8 may

allow patients to feel less anxious about future care

if they refuse. However, the doctors are likely to be,or to become, close colleagues; the carer/doctor willhave to be responsible for recruitment, whichinvolves him in the research; the researcher/doctorwill need to know about the patient's care; and bothdoctors will need to deal with clinical problems as

they arise, thus causing practical problems to ariseas a result of creating this split. In therapeuticresearch, care and research are so closely interlinkedthat it is impractical and impossible not to mix thetwo.A patient advocate might ask questions on the

patients' behalf, act as an impartial sounding board,

and deter the doctor from pressurising or hurryingthe patient. A lay advocate (friend or relative) mightnot be able to interpret technical details any betterthan the patient, might also be in awe of the per-ceived power of the doctor, and might influence thepatient by the strength of their own relationship. Anon-doctor health care professional might be wellsuited to advocacy because of his or her familiaritywith the health care system, rapport with patientsand ability to interpret technical data. However, sucha professional may be seen as part of the doctor'steam and may well be employed by the doctor so thathe or she has a vested interest in encouraging recruit-ment. However, a trained advocate, perhaps fundedby the health authority or trust, could be specificallytrained and supervised by the research ethics com-mittee. Such posts, separate from any research orcare team, have the potential to develop into anindependent advocacy system.

Talking to patients about clinical research ingroups rather than individually may reduce pressureon patients - more outspoken patients may askquestions which others are reluctant to voice. How-ever, patients may also find it difficult to go againstthe group decision if they do not agree with it.

Selection of patientsIt has been suggested that we should first approachpatients who are best able to understand theresearch, and who are most highly motivated andleast captive, such as health care professionals whohave the particular disease in question."9 Thenumbers in this group are likely to be so small as tomake research impractical and where there is a linkbetween educational level and disease, selectinghighly educated subjects might bias the study.

Education of researchersEducation of health care professionals in researchethics and obtaining consent should ease pressure onpatients by increasing the researchers' awareness.This is gradually happening as nursing and medicalschools include medical ethics in their curricula.

Information to patientsSpecific trial information sheets are vital andpatients must have time to read and discuss them.However, as they are written by the researchers whowish the patient to consent there is the real possibil-ity that they will be biased. Consumers for Ethics inResearch (CERES) produce a standard leaflet onmedical research which is not written in relation tospecific projects.20 Having seen this leaflet andfollowing our own research into consent, we havedeveloped a Patient's Guide to Medical Research(figure 1) which is given to all rheumatology patientsas they are invited to participate in clinical research,

copyright. on January 25, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.22.4.232 on 1 August 1996. D

ownloaded from

236 Consent to clinical research - adequately voluntary or substantially influenced?

alongside the specific trial information sheets.2' It isindependent of any single researcher or researchtrial, is designed to cover many of the issues raised inthis article and has an "easy" readability level.22We have made other changes to the consent

process, including giving the responsibility for infor-mation-giving and initial interview and consent tothe research nurses, who are perhaps seen as lesspowerful figures than the doctor. We give patients atleast 48 hours to reflect on the proposal and encour-age them to discuss it with family, friends and GP,before telephoning them to discuss the trial, answerquestions and take their decision, which is passed onto the doctor on their behalf. We believe that thesestrategies reduce the difficulty patients experience indeclining to take part in research and we will beassessing this in future trials.

ConclusionVoluntary consent to clinical research may be com-promised because ofthe circumstances under which itmust be obtained and the idea that consent must beeither voluntary or not voluntary is too naive aconcept in this situation. I have argued that consent,which requires two realistic options and freedom fromcontrolling influences, may be voluntary, involuntary,coerced by others, enforced by circumstances or sub-stantially influenced (partially voluntary). Consent bypatients in clinical research must often be onlypartially voluntary, because it lies within the contextof illness or the doctor/patient relationship. The dutyof health care professionals is to ensure that partiallyvoluntary consent is adequately voluntary, just as wedo for information in consent. I have proposed severalroutes to reduce influences on voluntariness, the mostpromising of which appear to be the use ofhealth careprofessionals initially, the use of an advocate (whoshould be trained, independent and employedthrough the research ethics committee), well-educated researchers, time for the patient to reflectand the use ofan independent guide to research. Suchmoves are very necessary if we are to ensureadequately voluntary consent in clinical research. AsJonas argues, the consequences of ruthless researchmay be far-reaching:

"Let us remember that a slower progress in theconquest of diseases would not threaten society,but that society would indeed be threatened by theerosion of those moral values whose loss, possiblycaused by too ruthless a pursuit of scientific progress,would make its most dazzling triumphs not worthhaving".23

AcknowledgementsThe author would like to acknowledge the supportof Miss Louise de Raeve and Dr Don Evans (Centre

for Philosophy and Health Care, University Collegeof Swansea) and Professor Paul Dieppe and DrJohn Kirwan (University of Bristol RheumatologyUnit).

Sarah Hewlett, MA, RGN, RM, is Clinical ResearchManager in the Rheumatology Unit at Bristol RoyalInfirmary.

References1 Jonas H. Philosophical reflections on experimenting

with human subjects. Daedalus 1969; 98, 2: 230.2 Levine RJ. Ethics and regulation of clinical research

[2nd ed]. New Haven: Yale University Press, USA,1988: 185.

3 Hellman S, Hellman DS. Of mice but not man;problems of the randomised clinical trial. New EnglandJ7ournal ofMedicine 1991; 324, 22: 1585-9.

4 Chalmers TC, Block JB, Lee S. Controlled studies inclinical cancer research. New England J7ournal ofMedicine 1972; 287: 75-8.

5 British Medical Association. Medical ethics today: itspractice and philosophy. London: BMJ PublishingGroup, 1993: 330-3.

6 Royal College of Physicians. Guidelines on the practice ofethics committees: medical research involving humansubjects [2nd ed]. London: RCOP, 1990.

7 Association of British Pharmaceuticals Industry.Principles of good clinical (research) practice. London:ABPI, 1988.

8 Harris J. The value of life. London: Routledge, 1985:192-200.

9 Gillon R. Philosophical medical ethics. Chichester: Wileyand Sons, 1986: 113.

10 Beauchamp TL, Childress JF. Principles of biomedicalethics [3rd ed]. Oxford: Oxford University Press, 1989:83.

11 Veatch RM. Case studies in medical ethics.London:Harvard University Press, 1977: 290-1.

12 See reference 5: 29.13 See reference 10: 106-11.14 Pellegrino ED, Thomasma DC. For the patient's

good. New York: Oxford University Press, 1988:16-8.

15 Bennet G. Patients and their doctors. London: BailliereTindall, 1979: 45-52.

16 Hewlett SE. Consent to research in rheumatoid arthri-tis - a difficult decision? British Journal of Rheumatology1994; 33: 104.

17 Hewlett, SE. Voluntary consent to research? [MAdissertation]. Swansea, Wales: University CollegeSwansea, 1994: 71-3.

18 Spiro HM. Constraint and consent on being a patientand a subject. New England Jrournal of Medicine 1975;293: 1134-5.

19 See reference 1: 235.20 Consumers for Ethics in Research (CERES). Medical

research andyou. London: CERES, 1991.21 See reference 17: 110-3.22 Flesch A. A new readability yardstick. Journal ofApplied

Psychology 1948: 32: 221.23 See reference 1: 245.

copyright. on January 25, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.22.4.232 on 1 August 1996. D

ownloaded from

Sarah Hewlett 237

Figure 1

PATIENT'S GUIDE TO MEDICALRESEARCH

Your doctor has invited you to take part in medicalresearch. This leaflet may help you make yourdecision.

What is medical research?Medical research is a way of finding out aboutillnesses, such as what causes them. It can also be a

way of testing different or new treatments, such as

new tablets. Your doctor will explain the aim of thisstudy and what it involves.Who looks after my interests?The hospital has a Research Ethics Committee(made up of doctors, nurses and members of thepublic) which must approve all medical research. Itlooks at every research study to make sure that it isfair (or ethical) in the way it treats patients. Onlywhen it has given permission can the study start.As well as this, if a doctor wants to test a new tabletor medicine they must get permission from a

Government Committee. This Committee makessure that the trial tablet has had reasonable safetychecks, before giving the doctor permission to use it.During the study your doctor will look after yourinterests and continue to give you the best possiblecare. If the doctor is not happy with your conditionduring the research you will be asked to stop thestudy.Do I have to say yes?No. Taking part in medical research is voluntary.Your doctor will not think badly of you if you sayno. You do not have to give a reason if you say no.

Your doctor will continue to care for you whateveryou decide. If you say yes at first, you can still leavethe study later on if you change your mind.What do patients think about taking part?Some patients take part because they feel they havea duty to help medical progress. Others believe itwill help their own illness and some see it as a wayof helping other people. On the other hand, somepatients decide not to take part because they believeresearch should not involve patients or because theyfeel it is risky or inconvenient. All these feelings are

understandable and reasonable.However, some patients feel obliged to take partbecause of the care they have received or becausethey feel the doctor will not look after them in thefuture if they say no. Others are simply tooembarrassed to say no. Feelings like this must notaffect your decision.This leaflet contains questions which you shoulddiscuss with the doctor who invited you to takepart. The doctor will not mind answering yourquestions at all. You must make the decision which isright foryou. You cannot do that unless you find outas much as you can about what is involved.

Before you make a decision about taking part,we suggest you ask the doctor the followingquestions

1 What is the aim ofthe research?Is it to test a new treatment or to follow the courseof my illness? Why does it need to be done?2 What will I be asked to do?Will I be asked to take tablets? Will I have to stopmy usual tablets? Will I have to come up for bloodtests or X-rays? How often? How long will the studylast?Ask as much as you can and find out what takingpart means. Your doctor should give you aninformation sheet to take home.3 What are the benefits or risks for me if I takepart?Ask how the research might help you. Often theaim is not to benefit you, but to help futurepatients. Do you mind this?Ask your doctor if there are any risks in taking part.Don't sit and worry about possible side-effects -ask!

Extra questions to ask ifyou are invited to testa medicine

4 How will my treatment be decided?Several treatments may be tested and to make surethe study is not biased, one of the trial treatmentsmay be given to you by chance, not by yourdoctor's decision. Would you mind?5 Will I get a dummy treatment?Some studies compare one treatment with adummy treatment, called a placebo. Ask yourdoctor if there is a chance you might be given aplacebo. How would you feel about that?6 If this is a new medicine am I insured ifthings go wrong?You should check with your doctor. TheResearch Ethics Committee has also asked thisquestion.7 Could I have standard treatment instead?Ask about the standard treatments that are availableto you if you decide not to take part in the researchstudy.If there is time, you may like to discuss the studywith your family or your General Practitioner.

When you have all the facts you can makeyour decision without feeling under anypressure

Thank you for thinking about our medical research.We hope this leaflet has helped you reach thedecision that is right for you.Produced by Sarah Hewlett, Research Sister,University of Bristol and United Bristol HealthcareTrust. 1993

copyright. on January 25, 2022 by guest. P

rotected byhttp://jm

e.bmj.com

/J M

ed Ethics: first published as 10.1136/jm

e.22.4.232 on 1 August 1996. D

ownloaded from