0277.9536192 .S5.00+0.00
Sot.SC;.Med. Vol.34.No. 12,pp.1399-1404, 1992 Copyright
Printed in Great Britain. All nghts reserved
COMPENSATION
C= 1992 Pergamon Press Ltd
FOR HARM: THE IMPLICATIONS MEDICAL RESEARCH IAN HARVEY'~~~
FOR
RUTHCHADWICK'
‘University of Wales College of Medicine, Centre for Applied Public Health Medicine, Temple of Peace and Health, Cathays Park, Cardiff CFI 3NW and ‘Centre for Applied Ethics, University of Wales College of Cardiff, P.O. Box 94, Cardiff CFI 3XE, Wales, U.K. Abstract-The rising incidence of medical litigation in the 1980’s led to a relatively minor modification in January 1990 of the United Kingdom’s negligence-based system of medical compensation, as a result of which employing Health Authorities now accept full vicarious liability for the negligent (but nor other) actions of employed clinicians (Crown indemnity). During the same period, by contrast, bodies such as the Royal College of Physicians have strongly influenced locally established Research Ethics Committees towards favouring no-fault compensarion for harm resulting from medical research. The discordance between these two approaches to compensation has caused anxiety during the last year. There has been particular concern about the possible discouragement of non-commercially sponsored research. Hitherto however, no data has been available to inform this discussion. This paper reports the findings of a questionnaire survey of all research ethics committees in the United Kingdom (64% response rate), which indicates that 61% of committees require no-fault compensation for at least some projects and that 33% of these committees have rejected projects solely or mainly due to the lack of such provision. Rejection is significantly more common in Health Districts which contain teaching hospitals. Our critical analysis of the arguments advanced in favour of no-fault provision for research subjects suggests that they do not place adequate emphasis upon respect for the autonomy of individuals. We consider that it is in general more consistent with such respect fully to inform research subjects of the available compensation arrangements, rather than for ethics committees to make no-fault compensation a general requirement before ethical approval is given. Key wtorcis-research
I.
ethics,
no-fault
Crown
compensation,
INTRODUCTION
The provision of compensation for victims of medical accidents in the United Kingdom has been a complex and controversial subject for almost 10 years, with the advocates of no-fault compensation battling with government funded organizations reluctant to move away from negligence based arrangements. Most recently the arena for these arguments has shifted towards the realm of medical research, prompted largely by the introduction of Crown indemnity. This has had potentially worrying implications for the future of clinical research. This paper aims to describe the background to these concerns, present data about the attitudes of research ethics committees, and to present a critical analysis of some of the arguments invoked in favour of no-fault compensation. Crown indemnity
The switch to a system of Crown indemnity to compensate patients for harm resulting from medical activities within the District Health Authority controlled sector of the National Health Service occurred on I January 1990 (1,2]. Crown indemnity was a compromise arrangement designed to meet the anxieties of clinicians about the soaring cost of medical defence subscriptions [3] without undermining the basic principle that negligence must still be
indemnity
demonstrated before compensation is payable. The strong support that had been developing within the United Kingdom for the introduction of a system of no-fault compensation (4, 51 (whereby payment is made on the basis simply that harm has resulted from medical action), similar to those in New Zealand, Sweden and Finland, was deflected, at least temporarily, by this measure. The move to Crown indemnity, viewed as highly significant by the medical profession in terms of potential infringement of clinical freedom, was, from the patients’ perspective, only a minor technical change involving a shift of legal responsibility away from individual employees towards their employing body. It brought the arrangements for medical staff into line with those operating for all other NHS employees. As a consequence, from January 1990 the financial burden resulting from any act of medical negligence has fallen upon the employing health authority and membership of a medical defence organization has ceased to be a condition of employment for medical staff. Research
ethics committees
The build up to the introduction of Crown indemnity and the uncertainties that accompanied it prompted a wide-ranging examination of medicolegal
1399
1400
IAN HARVEYand RL.THCHADWICK
issues in various settings within the health service. In particular. research ethics committees were stimulated to examine more closely the arrangements in place for awarding compensation in the event of harm resulting from clinical medical research. During this process at least some ethics committees found, to their surprise, that the new Crown indemnity arrangements were at variance with the widely adopted recommendations favouring no-fault compensation that had emanated over a period of time from national bodies with an interest in ethical issues. Amongst the most influential of bodies in the field of medical ethics has been the Royal College of Physicians, which first recommended the establishment of research ethics committees in 1967 and which has produced a series of reports on ethical aspects of research over the last 8 years. The first of these [6] adopted a decidedly guarded view on the matter of compensation. It commented that relying on the demonstration of negligence was “clearly unsatisfactory” but went on to say that “the major issues of liability will have to await solution on a national basis and there is nothing that individual Ethics Committees can do about them.” Two years later the view had changed dramatically [7]. The report ‘Research on Healthy Volunteers’ recommended that universities and other institutions should make binding commitments to provide compensation on a no-fault basis. It stated: “We believe that a healthy volunteer sustaining ill health or injury as a result of participation in a research project is entitled to compensation, and also that justice requires that the parties promoting the research so organise their affairs that he (sic) receives this compensation without delay.” It went on to say that it was the responsibility of Ethics Committees to ensure that adequate arrangements for compensation existed before a project involving healthy volunteers was approved. Two reports published in 1990 have further elaborated this view [8, 91. The report entitled ‘Research Involving Patients’ recommended that: “Bodies that sponsor research, including both publicly funded bodies (such as the Research Councils, the Department of Health, and the National Health Service) and the pharmaceutical industry, should now so arrange their affairs as to implement the principle that injury due to participation in research sponsored by them or conducted by their staff with the approval of a Research Ethics Committee shall be compensated by a simple, informal and expeditious procedure.” It inserts two caveats however. First, that certain research, such as epidemiological research, might be so innocuous that it would be inappropriate to require carefully developed arrangements for compensation and, second, that until sponsoring bodies alter their policy in the recommended way, the absence of no-fault arrangements should be disclosed to the patient. The second report (Guidelines on the Prac-
tice of Ethics Committees in Medical Research involving Human Subjects) commends the Association of the British Pharmaceutical Industry Guidelines (ABPI) as a model for no-fault arrangements and adds that it is reasonable for ethics committees to expect that public bodies should make similar arrangements. In summary therefore these influential reports of the Royal College of Physicians have swung firmly behind no-fault compensation as the standard model for providing redress to those patient volunteers and healthy volunteers harmed during medical research. The arguments that have been adduced in the debate about ethical compensation arrangements in clinical research have almost exactly mirrored those put forward during the wider debate about compensation throughout routine medical practice, but the outcome has been quite different. Current problems Current uncertainties centre upon the fact that advice regarding compensation in a research setting has moved in a different direction from policy relating to compensation of National Health Service patients receiving routine medical treatment. Research policy asserts the desirability of no-fault arrangements, whilst ‘routine’ compensation rests upon the tort (or civil wrong) of negligence. The discordance that has thus developed between the demands of local ethics committees for no-fault arrangements, and the willingness of national bodies (such as the NHS) generally only to offer negligence-based compensation, undoubtedly poses a potential threat to medical research in the United Kingdom. Recent guidance from the United Kingdom Health Departments [IO] has done nothing to resolve this discordance. In that guidance it is clearly stated that “NHS bodies are not empowered to offer advance indemnity to participants in research projects.” Later it warns that “volunteers must therefore be told in advance of all known risks and be made aware that there could also be unforeseen risks and of the possible difficulties in obtaining compensation ” (our emphasis). Despite the anxiety that has arisen, particularly within universities, about the overall implications for research there has until now been no data available to describe the extent of the problem. The objectives of the second part of this paper are two-fold. First to present the results of a questionnaire survey directed towards all research ethics committees in the United Kingdom with the intention of describing the current situation. The questionnaire was designed to elicit information about research compensation requirements and to describe the impact that these requirements are currently having upon judgements about the acceptability of research submitted to ethics committees. The second objective is to provide a critical examination of the philosophical considerations that
Compensation Table I. Does the etlucs committee require. before ethical approval will be given for a project. that compensation regardless of/huh (no-full) is available lo participanls in the research. if they should suffer harm? No. of responses
Response Yes--required for Yes-required for Not required Not required-but currently being
all projects some projects
I9 71 32 27
issue is discussed
(%)
(13) (48) (21) (18)
149
underpin the strong advocacy sation in clinical research. 2.
QUESTIONNAIRE
SURVEY
of no-fault
OF ETHICS
compen-
COMMIITEES
Method A mailed questionnaire was sent to General Managers of all District Health Authorities and Health Boards in Great Britain and to the Deans of all medical schools during March and April 1991. The covering letter requested that the questionnaire be passed on to the chair or secretary of the research ethics committee. Of 234 questionnaires sent, 149 were completed and returned (64% response rate). Both the total number of research ethics committees to whom questionnaires were sent and the response rate are similar to that found in a previous national study [I I]. Results Table I shows the current requirements of ethics committees with regard to compensation arrangements. Of those committees requiring no-fault arrangements for some or all projects, 69% (45/65) of those responding to the question had introduced the no-fault requirement within the last 3 years. Thirty-eight percent (25/65) specifically mentioned the influence of the Royal College of Physicians’ Guidelines in arriving at this decision. Other influences mentioned were the Association of the British Pharmaceutical Industry’s (ABPI) Guidelines (So/,, 5/65), local thinking and the existence of Regional policies. Committees requiring no-fault provision for only some projects generally stipulated it for drug trials, invasive studies and healthy volunteer studies. Table 2 shows, of those committees with no-fault requirements, the number reporting rejection of research mainly or solel_vdue to the absence of no-fault compensation arrangements and the estimated percentage of projects rejected. There is statistically significant heterogeneity between Districts containing either an undergraduate or postgraduate teaching hospital and Districts without, with a greater likelihood of rejection in teaching Districts. Amongst the 30 ethics committees that had rejected projects owing to inadequate compensation arrangements, 23/27 respondents (85%) had rejected only drug trials-either because satisfactory agreement
for harm
1401
with the commercial sponsor had not been possible or because there was no sponsor for a new, unlicensed use of a marketed drug. Many of these committees had adopted a pragmatic approach by ignoring the absence of no-fault cover in studies where there was no possible commercial sponsor (for example surgical trials). However 4/27 committees (I 5%) reported that both questionnaire studies and surgical trials had been rejected. Three universities in the study have taken out no-fault insurance cover and two more are actively considering it. One health authority was said also to have taken out insurance and another two authorities had agreed to provide no-fault cover from their own resources for approved research Discussion It is apparent from this survey that the reports of the Royal College of Physicians have influenced many ethics committees in the direction of favouring no-fault compensation and that many committees are demanding evidence of it for at least some research projects. The results show that teaching districts are significantly more likely to reject projects on the grounds of absence of no-fault arrangements. This may be because teaching districts are more likely to be the location of non commercially sponsored studies (such as surgical trials and trials of marketed drugs for new, unlicensed indications) and/or because they adopt a different ethical approach. In a few instances questionnaire studies have been rejected owing to the lack of no-fault indemnity, presumably predicated upon the argument that questionnaires are capable of causing psychological harm [9]. Although three health authorities and three universities are offering no-fault indemnity, there remains for the most part a glaring gap between the policies, on the one hand of the ethics community, and on the other of most of the large publicly funded organizations sponsoring medical research in the United Kingdom. A battle is being fought between two notions of just compensation with clinical research as the battleground. Table 2. What percentage of projects submitted since the requirement for no-fault compensation was introduced have been rejected mainly or solely due lo the absence of no-fault arrangements? No. of responses Proportion projects rejected
of
None
Districts with under/post graduate teaching hospitals
Non-teaching Disrricrs
Total (%)
I3
47
60 (67%) lO(ll%) 8 (9%) 5 (6%) 6(7%) I (I%) 0 (0%)
Under I% 2-5% 610% I I-20% 2l-50% 51% or more
9o(lcMl%) -x2 = 13.4, with Yates correction P < 0.001
IDF
1~s HARVEYand RUTH CHADWICK
1402
3.
THE MORAL ARGUMENT FOR NO-FAULT COMPENSATION
In the light of this situation, it is important to go back to first principles and ask exactly what are the arguments, if any, that support the positions being taken. Why has the requirement for no-fault compensation been introduced, in some cases, for subjects of research, at a time wheti it is not available for patients (in spite of calls for change in this area)? In the absence of a legal requirement for no-fault compensation, what is the moral argument for it? As early as 1984, the Royal College of Physicians described the legal position, where a subject has to rely on an action in negligence or an ex grafia payment, as “clearly unsatisfactory” [6] and this line is also taken in the later reports on research on healthy volunteers and on patients [7,8]. As indicated above, the report on Research on Healthy Volunteers, for example, claims that it is “unacceptable” that a healthy volunteer should have to rely on an e.y gratin payment and that a healthy volunteer sustaining ill health or injury as a result of participation is *‘entitled” to compensation. It is said that “justice” requires compensation without delay [7]. Clearly some notion of justice is being appealed to, but it is important to try to tease out exactly what it is. Justice has been said to be an ‘essentially contested’ concept, in that different concepts of justice are locked in permanent and irresolvable conflict. David Miller, however, has suggested that the conflict about justice can be reduced to a struggle between three rival contenders: desert, rights and needs [12]. In addition, two further lines of argument (based on ideas of fairness and social climate) are evident in the Royal College of Physicians’ Reports. These five perspectives will therefore be considered in turn. The desert argument The tradition which identifies justice with allocating goods according to desert interprets ‘desert’ in a particular way. It is not concerned with a person’s natural talents, or with what happens to a person, such as good or ill that befalls them through chance or the actions of others. On the contrary, it sees the individual as having free will and responsibility, and as being deserving, or not, according to the use made of that freedom. The choices the individual makes determine what he or she deserves. Whether the goods said to be deserved are to be understood as a reward for praiseworthy choices, or as compensation for choosing to undergo burdens, is not always made explicit by those who take this view of justice. In the context of the debate on research perhaps the suggestion is that the healthy volunteer and the patient, (who is also said to have the status of a volunteer in willingly undergoing research [8]) deserve compensation, in return for choosing to contribute to the
development of knowledge or for the sacrifice made in undergoing risk. This cannot be the case, however. for two reasons. First both the harmed and the unharmed volunteer undergo risk. and secondly it is explicitly stated that people should not be rewarded for undergoing risk [7, 81 partly because it is thought to be inimical to the spirit of altruism. which should be encouraged [7]. We have an apparent paradox here, on the desert model, that those \vho might be thought to deserve reward for their altruism, cannot be given it without the danger of that altruism being undermined. It might be argued that the desert argument applies to the causer of the harm-that the one who harms deserves to pay for it, as a kind of punishment. While this argument is understandable, it is clearly not the thinking underlying the Royal College reports, precisely because we are being taken away from the notion of ‘fault’ to ‘no-fault’; the emphasis throughout is on the sufferer of the harm. The rights nrgument A rights view might be suggested by the Royal College’s use of the word ‘entitled’ in relation to compensation [7]. Some take the view that individuals have certain moral rights which as a matter of justice should be protected. What these are. however, is controversial, and there is a potential for conflict between those suggested by the Royal College of Physicians. One of the reports for example, mentions the subject’s right to self-determination [7]. This arguably should include the right to take risks even if no-fault compensation is not guaranteed, provided the individual gives fully informed consent. The needs argument Here the focus is on the state to kvhich the harm reduces the person. The suffering itself provides a reason for holding that the proposed response (viz. compensation), is required as a matter ofjustice. This is an important argument, containing as it does the insight that a person’s suffering is itself a reason for helping him or her. There are two potential problems with it however. First, there is the practical problem as to what degree of harm warrants no-fault compensation-for example, does it have to bs a ‘serious’ injury, and if so what does that mean? Secondly, in taking us away from a consideration of the choices made by the volunteer and directing our attention to their suffering it fails to provide any reason why the needs of healthy and patient volunteers for research, in particular, should be dealt with in this way. A consequence of the present position being taken by many Research Ethics Committees is that higher status is being accorded to volunteers for research than to patients in the course of routine medical practice. Certainly an argument for a wider application of no-fault compensation has been made in the Royal
Compensation College of Physicians’ Report on Compensation for Adverse Consequences of Medical Intervention, in which the needs argument is highlighted [5]. In the reports on research subjects, however, the Royal College appears to be less impressed by a needs argument than by the status of research subjects as volunteers [7,8]. T/re fairness
argument
The recurring concern about the volunteer status of the research subject appears to be connected with a general point about fairness, or a notion of equilibrium between gains and losses. Gain to researchers and possibly to society is seen as being at the expense o/‘someone else, viz. the volunteer. Therefore society should pay. The report on Research Involving Patients says “there is an element of benefit to society and it seems fair that society should bear some of the risk in financial terms” [g]. This idea that some individuals’ gains are at the expense of others’ losses (which need to be made good) serves to emphasise the element of conflict between the interests of different individuals. Such an emphasis is common in an atomistic view of society. There is another view which sees the interest of individuals as being much more closely interconnected. On such a picture it is arguable that the volunteer does gain, if only indirectly, from living in a society which has the benefits of medical progress. The report on Research Involving Patients points to “the benefit of the addition to general knowledge which would be to the advantage of all patients in the long run” [8]. It must be acknowledged that this kind of argument can be used to justify too much. Clearly the interests of healthy or patient volunteers should be protected and there must be some limits to the sacrifices that can be expected of individuals for the social good. At the most extreme of course some would argue that undergoing the risk of participation in research transgresses the bounds of acceptable sacrifice unless there is the guarantee of no-fault compensation for harm. The social climate argument A final argument put forward by the Royal College is that whatever the motives of the research subjects, “In the present social climate there is no doubt that patients will expect that, should they be injured by participation in research, they will be compensated” [8]. It may be true that in contemporary society there is a tendency to expect compensation (though we might want to draw a distinction between thinking it ought to be available and thinking it is /ikeI)* to be available). It is clear, however, that the fact that people expect something is not sufficient to show that there is a moral obligation to give it to them.
for harm
1103
4. COMPENSATION ARRANGEMENTS AND THE ETHICS OF RESEARCH
In our view the strongest of the philosophical arguments for no-fault compensation is that based on need, whereby the suffering of those who have been harmed is itself a reason for helping them. Even if it is accepted however that there are sound reasons for making no-fault compensation available there is a different question as to whether it is ever appropriate to withhold approval of a research project sole& on the grounds that there is no provision for it. as the questionnaire results indicate is currently happening. As in any other area of life, the needs of some people are liable to clash with the needs of others. and the value of responding to those needs comes into conflict with other values. In order to make any headway, as a first step the utility of any policy has to be considered. Suppose a blanket ban were instituted on any research not guaranteeing no-fault compensation. One consequence of this would be that some research that would otherwise have been carried out would not be carried out. It might be argued that this is a price worth paying to protect the interests of research subjects and we have already acknowledged that there must be some limits to the sacrifice of individuals for the general good. The ban could, however, in fact lead to less protection for the interests of patients. Researchers who could not afford to guarantee no-fault compensation could be tempted to define their work as not requiring scrutiny by a Research Ethics Committee. Given the contentious nature of the boundary between research and, in particular, innovative therapy [8, 131 this possibility cannot be ignored. Anecdotal evidence suggests that such re-definition is already happening. Secondly, a ban would in practice operate to the systematic disadvantage of certain institutions, particularly those whose funds are limited and who cannot afford to guarantee no-fault compensation. Universities, for example, are not in command of the kind of resources which drug companies have at their disposal. This could have the effect of making university based research further dependent on external funding from drug and other companies, thus increasingly laying researchers open to charges of comindependence. The promising their academic protection of research subjects has, on a utilitarian model, to be weighed against other considerations such as these. An attempt to maximise utility does not require the subordination of protection of research subjects to other interests, but does seek ways of protecting them that are at least nor incompatible with concern for those other interests. The Royal College itself recognises the importance of other values, saying that “it is particularly important to respect autonomy, right of self-determination and safety” (71. There is clearly potential for conflict here between the paternalistic desire to protect and the aim of respecting autonomy. In so far as
1404
IAS HARVEY and RUTH CHADV. ICK
the paternalistic road is taken, it is arguable that sufficient protection can be achieved by strengthening the requirements of minimal risk and informed consent, rather than by a ban on research which does not guarantee no-fault compensation. The moral arguments, therefore, support no more than a selective ban. Although the number of committees imposing a blanket ban on all research lacking no-fault provision is small, it is worth noting that such a policy might not itself meet the requirements of justice in that it does not allow for relevant differences to be taken into account in considering applications. As has been pointed out above, there is a vast difference in the degree of risk imposed by various research programmes, and some involve virtually no risk at all. This suggests that a case-by-case approach is necessary. Some research may involve risks so serious that the possibility of uncompensated harm outweighs the prospect of potential benefit. In such cases provision for no-fault compensation may properly be required. In general, however, we would suggest that potential subjects have a capacity for autonomy which, provided the conditions for informed consent are met (such as the availability of relevant information and the absence of coercion or undue influence) enables them to decide whether or not to undergo the risks of research. It is consistent, moreover, with respect for autonomy that the presence or absence of no-fault compensation should be one of the key factors they themselves are asked to take into consideration in deciding whether or not to give informed consent. Such an approach would protect the interest of subjects, researchers, and society as a whole.
Acknowledgements--We gratefully acknowledge the help of all those who completed and returned the questionnaire, and the invaluable assistance of Mrs Gaynor Thorne in circulating the questionnaire and preparing the manuscript.
REFERENCES
I Havard 2 3
4.
5.
6.
7. 8. 9.
10. II.
12. 13.
J. D. J. Doctors and medical negligence. or. Med. J. 300, 3433344. 1990. Welsh Office. Claims of medical neghgence against NHS hospital and community doctors and dentists. (WHC (89) 70). Welsh Office, Cardiff, 1989. Ham C., Dingwall R., Fenn P. and Harris D. Medical Negligence: Compensarion and Accouniabilil_s. King’s Fund Institute, London. 1988. British Medical Association. Report of the Working Party on No-Fault Compensation for Medical Injury. British Medical Association, London. 1989. Royal College of Physicians. Compensarionfor Adverse Consequences of Medical Imerrenrion. Royal College of Physicians, London, 1990. Royal College of Physicians. Guidelines on rhe Practice of Ethics Committees in Medical Research. Royal College of Physicians, London, 1984. Royal College of Physicians. Research on Healthy Volunfeers. Royal College of Physicians, London, 1986. Royal College of Physicians. Research Incolcing Patients. Royal College of Physicians, London, 1990. Royal College of Physicians. Guide-lines on the Practice of Ethics Commitlees in Medical Research Incohing Human Subjects. Royal College of Physicians, London, 1990. Department of Health. Local Research Erhics Committees. Department of Health, London, 1991. Neuberger J. Research ethics committees. In Healih Care UK 1990(Edited by Harrison A.). Kings Fund Institute, London, 1990. Miller D. Social Jusrice. Clarendon Press, Oxford. 1976. Caplan A. Is in vitro fertilization research or therapy? In Beyond Baby M: Ethical Issues in Vew Reproducrice Techniques (Edited by Bartels D. M.. Priester R., Vawter D. E., Caplan A. L.), pp. 1499170. Humana Press, Clifton, NJ, 1989.
Sot.SC;.Med. Vol.34.No. 12,pp.1399-1404, 1992 Copyright
Printed in Great Britain. All nghts reserved
COMPENSATION
C= 1992 Pergamon Press Ltd
FOR HARM: THE IMPLICATIONS MEDICAL RESEARCH IAN HARVEY'~~~
FOR
RUTHCHADWICK'
‘University of Wales College of Medicine, Centre for Applied Public Health Medicine, Temple of Peace and Health, Cathays Park, Cardiff CFI 3NW and ‘Centre for Applied Ethics, University of Wales College of Cardiff, P.O. Box 94, Cardiff CFI 3XE, Wales, U.K. Abstract-The rising incidence of medical litigation in the 1980’s led to a relatively minor modification in January 1990 of the United Kingdom’s negligence-based system of medical compensation, as a result of which employing Health Authorities now accept full vicarious liability for the negligent (but nor other) actions of employed clinicians (Crown indemnity). During the same period, by contrast, bodies such as the Royal College of Physicians have strongly influenced locally established Research Ethics Committees towards favouring no-fault compensarion for harm resulting from medical research. The discordance between these two approaches to compensation has caused anxiety during the last year. There has been particular concern about the possible discouragement of non-commercially sponsored research. Hitherto however, no data has been available to inform this discussion. This paper reports the findings of a questionnaire survey of all research ethics committees in the United Kingdom (64% response rate), which indicates that 61% of committees require no-fault compensation for at least some projects and that 33% of these committees have rejected projects solely or mainly due to the lack of such provision. Rejection is significantly more common in Health Districts which contain teaching hospitals. Our critical analysis of the arguments advanced in favour of no-fault provision for research subjects suggests that they do not place adequate emphasis upon respect for the autonomy of individuals. We consider that it is in general more consistent with such respect fully to inform research subjects of the available compensation arrangements, rather than for ethics committees to make no-fault compensation a general requirement before ethical approval is given. Key wtorcis-research
I.
ethics,
no-fault
Crown
compensation,
INTRODUCTION
The provision of compensation for victims of medical accidents in the United Kingdom has been a complex and controversial subject for almost 10 years, with the advocates of no-fault compensation battling with government funded organizations reluctant to move away from negligence based arrangements. Most recently the arena for these arguments has shifted towards the realm of medical research, prompted largely by the introduction of Crown indemnity. This has had potentially worrying implications for the future of clinical research. This paper aims to describe the background to these concerns, present data about the attitudes of research ethics committees, and to present a critical analysis of some of the arguments invoked in favour of no-fault compensation. Crown indemnity
The switch to a system of Crown indemnity to compensate patients for harm resulting from medical activities within the District Health Authority controlled sector of the National Health Service occurred on I January 1990 (1,2]. Crown indemnity was a compromise arrangement designed to meet the anxieties of clinicians about the soaring cost of medical defence subscriptions [3] without undermining the basic principle that negligence must still be
indemnity
demonstrated before compensation is payable. The strong support that had been developing within the United Kingdom for the introduction of a system of no-fault compensation (4, 51 (whereby payment is made on the basis simply that harm has resulted from medical action), similar to those in New Zealand, Sweden and Finland, was deflected, at least temporarily, by this measure. The move to Crown indemnity, viewed as highly significant by the medical profession in terms of potential infringement of clinical freedom, was, from the patients’ perspective, only a minor technical change involving a shift of legal responsibility away from individual employees towards their employing body. It brought the arrangements for medical staff into line with those operating for all other NHS employees. As a consequence, from January 1990 the financial burden resulting from any act of medical negligence has fallen upon the employing health authority and membership of a medical defence organization has ceased to be a condition of employment for medical staff. Research
ethics committees
The build up to the introduction of Crown indemnity and the uncertainties that accompanied it prompted a wide-ranging examination of medicolegal
1399
1400
IAN HARVEYand RL.THCHADWICK
issues in various settings within the health service. In particular. research ethics committees were stimulated to examine more closely the arrangements in place for awarding compensation in the event of harm resulting from clinical medical research. During this process at least some ethics committees found, to their surprise, that the new Crown indemnity arrangements were at variance with the widely adopted recommendations favouring no-fault compensation that had emanated over a period of time from national bodies with an interest in ethical issues. Amongst the most influential of bodies in the field of medical ethics has been the Royal College of Physicians, which first recommended the establishment of research ethics committees in 1967 and which has produced a series of reports on ethical aspects of research over the last 8 years. The first of these [6] adopted a decidedly guarded view on the matter of compensation. It commented that relying on the demonstration of negligence was “clearly unsatisfactory” but went on to say that “the major issues of liability will have to await solution on a national basis and there is nothing that individual Ethics Committees can do about them.” Two years later the view had changed dramatically [7]. The report ‘Research on Healthy Volunteers’ recommended that universities and other institutions should make binding commitments to provide compensation on a no-fault basis. It stated: “We believe that a healthy volunteer sustaining ill health or injury as a result of participation in a research project is entitled to compensation, and also that justice requires that the parties promoting the research so organise their affairs that he (sic) receives this compensation without delay.” It went on to say that it was the responsibility of Ethics Committees to ensure that adequate arrangements for compensation existed before a project involving healthy volunteers was approved. Two reports published in 1990 have further elaborated this view [8, 91. The report entitled ‘Research Involving Patients’ recommended that: “Bodies that sponsor research, including both publicly funded bodies (such as the Research Councils, the Department of Health, and the National Health Service) and the pharmaceutical industry, should now so arrange their affairs as to implement the principle that injury due to participation in research sponsored by them or conducted by their staff with the approval of a Research Ethics Committee shall be compensated by a simple, informal and expeditious procedure.” It inserts two caveats however. First, that certain research, such as epidemiological research, might be so innocuous that it would be inappropriate to require carefully developed arrangements for compensation and, second, that until sponsoring bodies alter their policy in the recommended way, the absence of no-fault arrangements should be disclosed to the patient. The second report (Guidelines on the Prac-
tice of Ethics Committees in Medical Research involving Human Subjects) commends the Association of the British Pharmaceutical Industry Guidelines (ABPI) as a model for no-fault arrangements and adds that it is reasonable for ethics committees to expect that public bodies should make similar arrangements. In summary therefore these influential reports of the Royal College of Physicians have swung firmly behind no-fault compensation as the standard model for providing redress to those patient volunteers and healthy volunteers harmed during medical research. The arguments that have been adduced in the debate about ethical compensation arrangements in clinical research have almost exactly mirrored those put forward during the wider debate about compensation throughout routine medical practice, but the outcome has been quite different. Current problems Current uncertainties centre upon the fact that advice regarding compensation in a research setting has moved in a different direction from policy relating to compensation of National Health Service patients receiving routine medical treatment. Research policy asserts the desirability of no-fault arrangements, whilst ‘routine’ compensation rests upon the tort (or civil wrong) of negligence. The discordance that has thus developed between the demands of local ethics committees for no-fault arrangements, and the willingness of national bodies (such as the NHS) generally only to offer negligence-based compensation, undoubtedly poses a potential threat to medical research in the United Kingdom. Recent guidance from the United Kingdom Health Departments [IO] has done nothing to resolve this discordance. In that guidance it is clearly stated that “NHS bodies are not empowered to offer advance indemnity to participants in research projects.” Later it warns that “volunteers must therefore be told in advance of all known risks and be made aware that there could also be unforeseen risks and of the possible difficulties in obtaining compensation ” (our emphasis). Despite the anxiety that has arisen, particularly within universities, about the overall implications for research there has until now been no data available to describe the extent of the problem. The objectives of the second part of this paper are two-fold. First to present the results of a questionnaire survey directed towards all research ethics committees in the United Kingdom with the intention of describing the current situation. The questionnaire was designed to elicit information about research compensation requirements and to describe the impact that these requirements are currently having upon judgements about the acceptability of research submitted to ethics committees. The second objective is to provide a critical examination of the philosophical considerations that
Compensation Table I. Does the etlucs committee require. before ethical approval will be given for a project. that compensation regardless of/huh (no-full) is available lo participanls in the research. if they should suffer harm? No. of responses
Response Yes--required for Yes-required for Not required Not required-but currently being
all projects some projects
I9 71 32 27
issue is discussed
(%)
(13) (48) (21) (18)
149
underpin the strong advocacy sation in clinical research. 2.
QUESTIONNAIRE
SURVEY
of no-fault
OF ETHICS
compen-
COMMIITEES
Method A mailed questionnaire was sent to General Managers of all District Health Authorities and Health Boards in Great Britain and to the Deans of all medical schools during March and April 1991. The covering letter requested that the questionnaire be passed on to the chair or secretary of the research ethics committee. Of 234 questionnaires sent, 149 were completed and returned (64% response rate). Both the total number of research ethics committees to whom questionnaires were sent and the response rate are similar to that found in a previous national study [I I]. Results Table I shows the current requirements of ethics committees with regard to compensation arrangements. Of those committees requiring no-fault arrangements for some or all projects, 69% (45/65) of those responding to the question had introduced the no-fault requirement within the last 3 years. Thirty-eight percent (25/65) specifically mentioned the influence of the Royal College of Physicians’ Guidelines in arriving at this decision. Other influences mentioned were the Association of the British Pharmaceutical Industry’s (ABPI) Guidelines (So/,, 5/65), local thinking and the existence of Regional policies. Committees requiring no-fault provision for only some projects generally stipulated it for drug trials, invasive studies and healthy volunteer studies. Table 2 shows, of those committees with no-fault requirements, the number reporting rejection of research mainly or solel_vdue to the absence of no-fault compensation arrangements and the estimated percentage of projects rejected. There is statistically significant heterogeneity between Districts containing either an undergraduate or postgraduate teaching hospital and Districts without, with a greater likelihood of rejection in teaching Districts. Amongst the 30 ethics committees that had rejected projects owing to inadequate compensation arrangements, 23/27 respondents (85%) had rejected only drug trials-either because satisfactory agreement
for harm
1401
with the commercial sponsor had not been possible or because there was no sponsor for a new, unlicensed use of a marketed drug. Many of these committees had adopted a pragmatic approach by ignoring the absence of no-fault cover in studies where there was no possible commercial sponsor (for example surgical trials). However 4/27 committees (I 5%) reported that both questionnaire studies and surgical trials had been rejected. Three universities in the study have taken out no-fault insurance cover and two more are actively considering it. One health authority was said also to have taken out insurance and another two authorities had agreed to provide no-fault cover from their own resources for approved research Discussion It is apparent from this survey that the reports of the Royal College of Physicians have influenced many ethics committees in the direction of favouring no-fault compensation and that many committees are demanding evidence of it for at least some research projects. The results show that teaching districts are significantly more likely to reject projects on the grounds of absence of no-fault arrangements. This may be because teaching districts are more likely to be the location of non commercially sponsored studies (such as surgical trials and trials of marketed drugs for new, unlicensed indications) and/or because they adopt a different ethical approach. In a few instances questionnaire studies have been rejected owing to the lack of no-fault indemnity, presumably predicated upon the argument that questionnaires are capable of causing psychological harm [9]. Although three health authorities and three universities are offering no-fault indemnity, there remains for the most part a glaring gap between the policies, on the one hand of the ethics community, and on the other of most of the large publicly funded organizations sponsoring medical research in the United Kingdom. A battle is being fought between two notions of just compensation with clinical research as the battleground. Table 2. What percentage of projects submitted since the requirement for no-fault compensation was introduced have been rejected mainly or solely due lo the absence of no-fault arrangements? No. of responses Proportion projects rejected
of
None
Districts with under/post graduate teaching hospitals
Non-teaching Disrricrs
Total (%)
I3
47
60 (67%) lO(ll%) 8 (9%) 5 (6%) 6(7%) I (I%) 0 (0%)
Under I% 2-5% 610% I I-20% 2l-50% 51% or more
9o(lcMl%) -x2 = 13.4, with Yates correction P < 0.001
IDF
1~s HARVEYand RUTH CHADWICK
1402
3.
THE MORAL ARGUMENT FOR NO-FAULT COMPENSATION
In the light of this situation, it is important to go back to first principles and ask exactly what are the arguments, if any, that support the positions being taken. Why has the requirement for no-fault compensation been introduced, in some cases, for subjects of research, at a time wheti it is not available for patients (in spite of calls for change in this area)? In the absence of a legal requirement for no-fault compensation, what is the moral argument for it? As early as 1984, the Royal College of Physicians described the legal position, where a subject has to rely on an action in negligence or an ex grafia payment, as “clearly unsatisfactory” [6] and this line is also taken in the later reports on research on healthy volunteers and on patients [7,8]. As indicated above, the report on Research on Healthy Volunteers, for example, claims that it is “unacceptable” that a healthy volunteer should have to rely on an e.y gratin payment and that a healthy volunteer sustaining ill health or injury as a result of participation is *‘entitled” to compensation. It is said that “justice” requires compensation without delay [7]. Clearly some notion of justice is being appealed to, but it is important to try to tease out exactly what it is. Justice has been said to be an ‘essentially contested’ concept, in that different concepts of justice are locked in permanent and irresolvable conflict. David Miller, however, has suggested that the conflict about justice can be reduced to a struggle between three rival contenders: desert, rights and needs [12]. In addition, two further lines of argument (based on ideas of fairness and social climate) are evident in the Royal College of Physicians’ Reports. These five perspectives will therefore be considered in turn. The desert argument The tradition which identifies justice with allocating goods according to desert interprets ‘desert’ in a particular way. It is not concerned with a person’s natural talents, or with what happens to a person, such as good or ill that befalls them through chance or the actions of others. On the contrary, it sees the individual as having free will and responsibility, and as being deserving, or not, according to the use made of that freedom. The choices the individual makes determine what he or she deserves. Whether the goods said to be deserved are to be understood as a reward for praiseworthy choices, or as compensation for choosing to undergo burdens, is not always made explicit by those who take this view of justice. In the context of the debate on research perhaps the suggestion is that the healthy volunteer and the patient, (who is also said to have the status of a volunteer in willingly undergoing research [8]) deserve compensation, in return for choosing to contribute to the
development of knowledge or for the sacrifice made in undergoing risk. This cannot be the case, however. for two reasons. First both the harmed and the unharmed volunteer undergo risk. and secondly it is explicitly stated that people should not be rewarded for undergoing risk [7, 81 partly because it is thought to be inimical to the spirit of altruism. which should be encouraged [7]. We have an apparent paradox here, on the desert model, that those \vho might be thought to deserve reward for their altruism, cannot be given it without the danger of that altruism being undermined. It might be argued that the desert argument applies to the causer of the harm-that the one who harms deserves to pay for it, as a kind of punishment. While this argument is understandable, it is clearly not the thinking underlying the Royal College reports, precisely because we are being taken away from the notion of ‘fault’ to ‘no-fault’; the emphasis throughout is on the sufferer of the harm. The rights nrgument A rights view might be suggested by the Royal College’s use of the word ‘entitled’ in relation to compensation [7]. Some take the view that individuals have certain moral rights which as a matter of justice should be protected. What these are. however, is controversial, and there is a potential for conflict between those suggested by the Royal College of Physicians. One of the reports for example, mentions the subject’s right to self-determination [7]. This arguably should include the right to take risks even if no-fault compensation is not guaranteed, provided the individual gives fully informed consent. The needs argument Here the focus is on the state to kvhich the harm reduces the person. The suffering itself provides a reason for holding that the proposed response (viz. compensation), is required as a matter ofjustice. This is an important argument, containing as it does the insight that a person’s suffering is itself a reason for helping him or her. There are two potential problems with it however. First, there is the practical problem as to what degree of harm warrants no-fault compensation-for example, does it have to bs a ‘serious’ injury, and if so what does that mean? Secondly, in taking us away from a consideration of the choices made by the volunteer and directing our attention to their suffering it fails to provide any reason why the needs of healthy and patient volunteers for research, in particular, should be dealt with in this way. A consequence of the present position being taken by many Research Ethics Committees is that higher status is being accorded to volunteers for research than to patients in the course of routine medical practice. Certainly an argument for a wider application of no-fault compensation has been made in the Royal
Compensation College of Physicians’ Report on Compensation for Adverse Consequences of Medical Intervention, in which the needs argument is highlighted [5]. In the reports on research subjects, however, the Royal College appears to be less impressed by a needs argument than by the status of research subjects as volunteers [7,8]. T/re fairness
argument
The recurring concern about the volunteer status of the research subject appears to be connected with a general point about fairness, or a notion of equilibrium between gains and losses. Gain to researchers and possibly to society is seen as being at the expense o/‘someone else, viz. the volunteer. Therefore society should pay. The report on Research Involving Patients says “there is an element of benefit to society and it seems fair that society should bear some of the risk in financial terms” [g]. This idea that some individuals’ gains are at the expense of others’ losses (which need to be made good) serves to emphasise the element of conflict between the interests of different individuals. Such an emphasis is common in an atomistic view of society. There is another view which sees the interest of individuals as being much more closely interconnected. On such a picture it is arguable that the volunteer does gain, if only indirectly, from living in a society which has the benefits of medical progress. The report on Research Involving Patients points to “the benefit of the addition to general knowledge which would be to the advantage of all patients in the long run” [8]. It must be acknowledged that this kind of argument can be used to justify too much. Clearly the interests of healthy or patient volunteers should be protected and there must be some limits to the sacrifices that can be expected of individuals for the social good. At the most extreme of course some would argue that undergoing the risk of participation in research transgresses the bounds of acceptable sacrifice unless there is the guarantee of no-fault compensation for harm. The social climate argument A final argument put forward by the Royal College is that whatever the motives of the research subjects, “In the present social climate there is no doubt that patients will expect that, should they be injured by participation in research, they will be compensated” [8]. It may be true that in contemporary society there is a tendency to expect compensation (though we might want to draw a distinction between thinking it ought to be available and thinking it is /ikeI)* to be available). It is clear, however, that the fact that people expect something is not sufficient to show that there is a moral obligation to give it to them.
for harm
1103
4. COMPENSATION ARRANGEMENTS AND THE ETHICS OF RESEARCH
In our view the strongest of the philosophical arguments for no-fault compensation is that based on need, whereby the suffering of those who have been harmed is itself a reason for helping them. Even if it is accepted however that there are sound reasons for making no-fault compensation available there is a different question as to whether it is ever appropriate to withhold approval of a research project sole& on the grounds that there is no provision for it. as the questionnaire results indicate is currently happening. As in any other area of life, the needs of some people are liable to clash with the needs of others. and the value of responding to those needs comes into conflict with other values. In order to make any headway, as a first step the utility of any policy has to be considered. Suppose a blanket ban were instituted on any research not guaranteeing no-fault compensation. One consequence of this would be that some research that would otherwise have been carried out would not be carried out. It might be argued that this is a price worth paying to protect the interests of research subjects and we have already acknowledged that there must be some limits to the sacrifice of individuals for the general good. The ban could, however, in fact lead to less protection for the interests of patients. Researchers who could not afford to guarantee no-fault compensation could be tempted to define their work as not requiring scrutiny by a Research Ethics Committee. Given the contentious nature of the boundary between research and, in particular, innovative therapy [8, 131 this possibility cannot be ignored. Anecdotal evidence suggests that such re-definition is already happening. Secondly, a ban would in practice operate to the systematic disadvantage of certain institutions, particularly those whose funds are limited and who cannot afford to guarantee no-fault compensation. Universities, for example, are not in command of the kind of resources which drug companies have at their disposal. This could have the effect of making university based research further dependent on external funding from drug and other companies, thus increasingly laying researchers open to charges of comindependence. The promising their academic protection of research subjects has, on a utilitarian model, to be weighed against other considerations such as these. An attempt to maximise utility does not require the subordination of protection of research subjects to other interests, but does seek ways of protecting them that are at least nor incompatible with concern for those other interests. The Royal College itself recognises the importance of other values, saying that “it is particularly important to respect autonomy, right of self-determination and safety” (71. There is clearly potential for conflict here between the paternalistic desire to protect and the aim of respecting autonomy. In so far as
1404
IAS HARVEY and RUTH CHADV. ICK
the paternalistic road is taken, it is arguable that sufficient protection can be achieved by strengthening the requirements of minimal risk and informed consent, rather than by a ban on research which does not guarantee no-fault compensation. The moral arguments, therefore, support no more than a selective ban. Although the number of committees imposing a blanket ban on all research lacking no-fault provision is small, it is worth noting that such a policy might not itself meet the requirements of justice in that it does not allow for relevant differences to be taken into account in considering applications. As has been pointed out above, there is a vast difference in the degree of risk imposed by various research programmes, and some involve virtually no risk at all. This suggests that a case-by-case approach is necessary. Some research may involve risks so serious that the possibility of uncompensated harm outweighs the prospect of potential benefit. In such cases provision for no-fault compensation may properly be required. In general, however, we would suggest that potential subjects have a capacity for autonomy which, provided the conditions for informed consent are met (such as the availability of relevant information and the absence of coercion or undue influence) enables them to decide whether or not to undergo the risks of research. It is consistent, moreover, with respect for autonomy that the presence or absence of no-fault compensation should be one of the key factors they themselves are asked to take into consideration in deciding whether or not to give informed consent. Such an approach would protect the interest of subjects, researchers, and society as a whole.
Acknowledgements--We gratefully acknowledge the help of all those who completed and returned the questionnaire, and the invaluable assistance of Mrs Gaynor Thorne in circulating the questionnaire and preparing the manuscript.
REFERENCES
I Havard 2 3
4.
5.
6.
7. 8. 9.
10. II.
12. 13.
J. D. J. Doctors and medical negligence. or. Med. J. 300, 3433344. 1990. Welsh Office. Claims of medical neghgence against NHS hospital and community doctors and dentists. (WHC (89) 70). Welsh Office, Cardiff, 1989. Ham C., Dingwall R., Fenn P. and Harris D. Medical Negligence: Compensarion and Accouniabilil_s. King’s Fund Institute, London. 1988. British Medical Association. Report of the Working Party on No-Fault Compensation for Medical Injury. British Medical Association, London. 1989. Royal College of Physicians. Compensarionfor Adverse Consequences of Medical Imerrenrion. Royal College of Physicians, London, 1990. Royal College of Physicians. Guidelines on rhe Practice of Ethics Committees in Medical Research. Royal College of Physicians, London, 1984. Royal College of Physicians. Research on Healthy Volunfeers. Royal College of Physicians, London, 1986. Royal College of Physicians. Research Incolcing Patients. Royal College of Physicians, London, 1990. Royal College of Physicians. Guide-lines on the Practice of Ethics Commitlees in Medical Research Incohing Human Subjects. Royal College of Physicians, London, 1990. Department of Health. Local Research Erhics Committees. Department of Health, London, 1991. Neuberger J. Research ethics committees. In Healih Care UK 1990(Edited by Harrison A.). Kings Fund Institute, London, 1990. Miller D. Social Jusrice. Clarendon Press, Oxford. 1976. Caplan A. Is in vitro fertilization research or therapy? In Beyond Baby M: Ethical Issues in Vew Reproducrice Techniques (Edited by Bartels D. M.. Priester R., Vawter D. E., Caplan A. L.), pp. 1499170. Humana Press, Clifton, NJ, 1989.
