Can and should the research–therapy distinction be maintained? Reflections in the light of innovative last-resort treatment

Abstract

It has been debated for quite some time among bioethicists and others whether or not the distinction between therapy and research in healthcare can and should be maintained. This paper tries to clarify what the disagreement is about, and argues that the distinction can be maintained in most, if not all, situations. However, even if it can be maintained, it does not necessarily follow that it should. It is argued here that there are good reasons to maintain the distinction both for the sake of protecting research interests and the sake of patient safety. In addition, recognizing the distinction provides a barrier against unreasonable prioritization between patients. However, there may be situations, such as innovative last-resort treatment, in which regulatory requirements from both therapy and research should be considered.

There is a long tradition of conducting research at hospitals and other healthcare facilities. Research is a common and integrated part of modern healthcare. This fact seems to be a breeding ground for the view that therapy and research are somehow indistinguishable from one another, if not always then at least sometimes (Angrist and Jamal, 2015; Kass et al., 2013).

This matter would hardly be debated if it only concerned conceptual distinctions and empirical descriptions. What nourish the debate are rather the normative disagreements tied to the descriptive claims. In particular, defenders of the distinction have tended to be eager to protect research activities and resources from being treated as just another resource for patients’ health promotion, while its critics have tended to lean toward a “patient first” approach. Debates regarding the research–therapy distinction have often flared up in relation to specific issues, such as whether or not one should stop promising controlled clinical trials prematurely in order to give all participants, even those in the control group, access to the new treatment (Buchanan and Miller, 2005; Malmqvist et al., 2011; Miller and Brody, 2003) – or whether or not research participants should be considered to have the right to have access to personal data that research studies generate about them (Helgesson, 2014; Richardson and Cho, 2012). In these debates, the distinction between research and therapy has been part of the argumentation for one solution or another.

The normative interest in making claims about the (in)distinguishability of human subjects research from therapy is the idea that different sets of relationships and responsibilities are tied to the different roles of researcher and healthcare provider, and to the activities of doing research or examining and treating patients (Brody and Miller, 2013; Miller and Brody, 2007). This is also reflected in legal regulations, in which healthcare and medical research are regulated as distinct activities. ¹ Differences in roles partly relate to differences in intentions. In particular, for examinations, treatments, and care, the patient’s best interest is in focus (while gaining general knowledge is, at most, a secondary interest), whereas the main concern for research is rather to add new knowledge (while health promotion for participating patients is, at most, a secondary concern).

Franklin Miller, who defends the distinction and its moral significance, describes the different roles (Miller, 2006: 11):

Medical care has a personalized focus. It is directed to helping a particular person in need of expert medical attention. Clinical research essentially lacks this purpose of personalized help for particular individuals. … The distinctive purpose of clinical research [is] to develop generalizable knowledge.

If you agree that different responsibilities are tied to the different activities, and you establish that an activity is research and is not at the same time or in the same respects therapy, then you have established that the responsibilities tied to healthcare providers (regarding these respects) do not apply, but only those tied to doing research with human subjects. Those who deny that the distinction between research and therapy can be maintained, or that it is normatively significant, insist that clinical researchers have all the responsibilities treating clinicians do toward patients/research subjects, plus the additional ones tied to the fact that it is research (as long as they do not conflict with treatment duties), while the defenders of the distinction argue that different sets of responsibilities apply to the different sorts of activities.

This paper aims to clarify what the distinction between research and therapy is about, whether or not it can be maintained, and whether or not it should be maintained. First, some definitions of “research” and “therapy” are introduced, and then, arguments to the effect that the distinction cannot be maintained are examined. Thereafter, pros and cons of maintaining the distinction are identified and discussed, with innovative last-resort treatment as the main case.

Research and therapy: Conceptual and descriptive issues

There should be no doubt that research and therapy are conceptually distinct, as shown earlier in the quote from Miller. While research is defined as “creative work undertaken on a systematic basis in order to increase the stock of knowledge” (OECD, 2015: 44) or as “studious inquiry or examination; especially investigation or experimentation aimed at the discovery and interpretation of facts, revision of accepted theories or laws in the light of new facts, or practical application of such new or revised theories or laws” (Merriam-Webster Online Dictionary, 2019a), therapy is defined as “the attempted remediation of a health problem, usually following a diagnosis. In the medical field, it is usually synonymous with treatment” (Wikipedia, 2019) or “therapeutic medical treatment of impairment, injury, disease, or disorder” (Merriam-Webster Online Dictionary, 2019b).

Research and therapy may also involve different activities, so that what is an instance of research is not necessarily simultaneously an instance of therapy and vice versa. This is most obvious regarding non-clinical research: You may indeed, for instance, study the political development in Uruguay or some extraordinary events in a neighboring galaxy without thereby examining or treating a patient. However, this is equally clear regarding clinical research, whereby, for instance, a retrospective registry study, a prospective follow-up study, or an interview study does not – and certainly not necessarily – involve therapy. Considering the other direction, from therapy to research, it should be equally clear that, for instance, the treatment of kidney failure, infection, or a skin lesion does not as such necessarily involve medical or any other research.

Has modern healthcare changed the game?

One might still want to maintain that the research–therapy distinction is no longer relevant in many of the interesting cases due to the integration of research into everyday activities at hospitals and other healthcare facilities. Indeed, research and therapy activities are quite often integrated rather than segregated; a great deal of clinical research takes place at hospitals and other healthcare facilities with patients while they are there for examination, treatment, and follow-up. Many hospitals have also adjusted to these circumstances in various ways, for example by systematically requiring an additional test tube of blood for research when blood is collected for diagnostic purposes, and by collecting patient data for quality assessments of various practices, techniques, and procedures (Angrist and Jamal, 2015; Kass et al., 2013). In a learning healthcare organization, it is arguably imperative to learn for the next generation of patients while examining and treating the present one (Faden et al., 2013).

In these cases, is it impossible to maintain the research–therapy distinction? Certainly not. In most cases it is easy to identify what parts of the ongoing activities relate to research – that is, those that are not done for the purpose of standard treatment and care but rather because there is now a study protocol to follow – and what parts do not. The research activities may concern the collection of specific data, such as additional measurements, or the pooling of standard patient data from a number of patients for the purpose of statistical analysis in order to answer specific research questions. Or, it may concern a randomization of patients into an experiment group and a control group, with the research protocol specifying who gets treated in what way depending on what group they end up in. Participation in controlled randomized clinical trials, furthermore, means that the dosage is fixed, as described in the study protocol, and is not adjusted to the specific needs of the participating individuals (Brody and Miller, 2013). The reason why these things are done the way they are is because the patients are participating in research – otherwise, it would have been done another way. One might still argue, for instance, that when an additional test tube of blood is drawn as part of the routine at the hospital as an adjustment to the needs of research, it cannot really be said at the time which test tube will be used for research and which for diagnosis. Nevertheless, it is clear how research considerations change the practice.

The use of healthcare quality registries in a wide variety of clinical specialties is arguably another example of how the healthcare system has adjusted to the interests of clinical research (Mattsson, 2016). However, these registries are not exclusively used for the “external” interest of research but also for the “internal” interest of continuous quality assessment and the improvement of treatment regimes. Perhaps quality assessment and development cannot be distinguished from research – both involve the systematic collection of data and evaluation of alternative therapies and routines, without focusing on any particular individual. However, they can both easily be distinguished from therapy, so this circumstance does not provide an argument against the research–therapy distinction.

Innovative last-resort treatment: The sole exception?

Are there any areas in healthcare where research and therapy activities cannot be distinguished from each other? Perhaps the most promising candidate is research carried out in relation to so-called innovative or experimental last-resort treatment; that is, treatment in situations in which the treating physician no longer has any established therapies to offer but still makes an additional life-saving effort – or an urgent effort of potentially profound effect on the health and wellbeing of the individual. ² When research is carried out in relation to innovative last-resort treatment, what is done at the clinic is arguably done for the purposes of research and treatment at the same time and in the same regard. This, I imagine, would be research of a special kind: with no fixed protocol beforehand and involving no specifically research-related interventions, meaning that there is no research interest in acting in any other way than what is perceived as best for the patient in the situation. This means, among other things, that research in such situations has the flexibility to allow adjustments to the needs of the individual, in contrast to, for instance, the situation in randomized clinical trials. Treatment is the only intervention consideration, while research consists of systematic follow-up, with the publication of intervention-related data in mind. The research perspective might still have some practical influence – perhaps the monitoring is more thorough than what might otherwise have been the case. But it might just be that it helps identify what ought to be monitored anyway, considering the kind of treatment and lack of knowledge in relation to it.

Also, in situations like this it seems possible to identify what is done for research (and development) that is not done for the sake of therapy. This concerns the “add-ons” that would not be done if it were only about treating this patient on this occasion, plus generating the relevant data for future follow-up: Mainly, the systematic monitoring of predefined measures and potentially taking additional tests (Brody and Miller, 2013). The most obvious add-on, apart from the considerations just mentioned, is the research-related work done with the collected data, with the practical aim of scientific publication.

In conclusion, in principle, there seems to be no case in which the research–therapy distinction cannot be maintained. The most likely case in which it cannot be maintained in practice is when an involved party intentionally muddles the borders. If we want to keep the research and therapy aspects distinct, we can. ³

The value of maintaining distinctions for both research and patients

That the distinction between research and therapy can be maintained does not necessarily mean that it should, or that it should in all circumstances. Hence, there is no immediate step from a successful defense of the distinction to policy recommendations. In what follows, I will argue that there are good reasons to maintain the distinction for both the sake of protecting research interests and the sake of patient safety. In addition, recognizing the distinction provides a barrier against unreasonable prioritization between patients. However, there may be situations in which policies should not rest on the distinction – or, rather, in which both the therapy and the research aspects should be considered when deciding what conditions to raise in relation to the activities concerned.

Maintaining the distinction for the sake of research

For the sake of research, including the attainment of new knowledge relevant for the improvement of healthcare, it is important to maintain the distinction in order to ensure that research studies are not spoiled for the sake of present treatment interests, for instance, because of the attitude that it is legitimate to treat all clinical research as “just another treatment option or healthcare resource” that can always be turned into additional resources for present patients. One example, mentioned above, is the early cessation of randomized clinical trials with promising intermediate results in order to distribute treatment to all participants, at the cost of abandoning the research protocol and not being able to answer important research questions (most obviously questions regarding long-term effects) (Bryant and Wolmark, 2003; Guyatt et al., 2012). Another example, also mentioned above, is the requirement that individual study information be returned to research participants – a requirement that, apart from not necessarily doing more good than harm (Helgesson et al., 2007), might put enough strain on research resources to make the study impracticable to carry out with the resources at hand.

It might be worth clarifying that I am not suggesting that these conflicts in the past should be characterized as conflicts between those who were able to recognize the distinction and those who were not. Rather, the conflicts have been normative throughout, between those who are of the opinion that the distinction makes a moral difference to how research resources may be handled (basically reasoning that if it is research, and not therapy, then the resources dedicated to it should be used primarily for research and not for therapy) and those who are of the opinion that other concerns, such as the idea that patient interests should always be given priority, trump research considerations even if the resources were initially dedicated to research.

That studies must not be spoiled can be convincingly defended in consequentialist terms: It is generally ill-spent research money to first invest in a study, then interrupt it in a way that makes it impossible to draw important conclusions from it. If a study is stopped early based on preliminary results in order to make the treatment available to all study participants, then potential but uncertain benefits for the few are prioritized over verified benefits for the many (present and future patients). Such prioritization does not make consequentialist sense. Leaving the consequentialist corner, such distribution also is deeply unfair since it gives unjustified priority to the present patients participating in the study. It can also be argued that it is unacceptable behavior toward research funders to use funding allocated to research for completely different purposes.

Maintaining the distinction for the sake of patient protection

There is at least one important reason to maintain the research–therapy distinction in practice from the perspective of protecting patient interests. This is in order to guarantee voluntariness among patients in relation to research participation. More specifically, it is important that therapy and research be kept sufficiently clearly apart in order to prevent researching healthcare providers from making research participation a requirement in order for a patient to get standard treatment. For instance, examinations and tests that are related entirely to research and irrelevant to medical diagnostics should be described as such to the patient, and not as necessary steps for identifying the proper therapy. In other words, what is done in the name of examinations and treatment must be done in the interest, and according to the will, of the patient, not for a hidden research agenda. If patients are to be included in clinical studies, procedures for informed consent should be used. In order for these to be adequate, the distinction between research and therapy needs to be maintained and kept clear.

Policy reasons to use double standards

As hinted above, there may be situations in which policies should not rest on the distinction; that is, when it may be less helpful than harmful to stress it. This would concern cases in which it is more appropriate to apply regulations from both the healthcare and research areas, rather than from just one of them. One general example would be cases in which double purposes are present – that is, when the measure is taken for both the purpose of helping a patient and the purpose of research (Dheensa et al., 2018). Then, both sets of regulations apply. Another example concerns incidental findings in research of great clinical relevance – here we may prefer a system whereby some treatment-related action is required of the researchers, such as informing the concerned participant and guiding him/her to a relevant physician, to a system whereby researchers are permitted to ignore such findings even when they could save lives or avoid great suffering by passing on the information.

Poor understanding among patients is no reason to transfer research to therapy

Patients and relatives sometimes do not recognize the difference between therapy and research (as early recognized by Appelbaum et al., 1987). Does this justify a practice in which the distinction is not maintained? For instance, if additional information efforts do not suffice to clarify the situation, or if it can be convincingly predicted that they are not likely to, is it then justified to apply both sets of requirements and act according to the one that best protects the interests of patients/research participants?

I think not. It is clear that this idea is not valid when we consider the specific directions the mistake on behalf of the patients or their relatives may take. First, say they believe that what is done for research, such as steadfastly keeping to the protocol in a randomized clinical trial, is also the primary choice for therapy when in fact it is not. What reasons for action are there in this situation? First, to provide better information so that study participants understand the conditions under which they are participating and how this relates to their treatment interests – and let them withdraw from the study if they so choose. Providing proper information includes avoiding persuasive and misleading recruitment efforts, like claiming that participation in a clinical drug trial is equivalent to being offered a miracle cure. Transparency vis-à-vis patients and relatives regarding one’s roles is also important, particularly in clinical research environments where healthcare providers also do research.

Should we also start treating the research project as the primary therapy choice? Not if this means changing the content of the intervention to adjust to the treatment interests of the participating individuals, as this would spoil the study (as discussed above). Spoiling the study by conflating research and therapy cannot be the appropriate reaction to the insight that patients or relatives have not realized the difference.

And when the mistake goes in the other direction – patients are treated, but think they are (exclusively or also) participating in a research study – the reasonable step here is also to better communicate what is in fact going on. People believing they are participating in research when they are not is perhaps not harmful, but it should be corrected by giving them adequate information if possible, out of respect for their autonomy.

What about clinical trials of potentially life-saving drugs, such as cancer therapies, in their early phases (which concern establishing a treatment effect and side-effects)? This seems to be a case where what is happening is research, but that is also often perceived as a last resort for terminally ill patients. If patients are appropriately informed and understand the conditions, and still want to participate because they want to try an intervention rather than doing nothing, then this would be acceptable. But in such situations it is even more important that concerned patients understand the difference between research and therapy, as well as what participants can and cannot expect from participating in such studies.

Innovative treatment in healthcare

In the aftermath of the so-called Macchiarini case – concerning trachea surgery with artificial implants as novel treatment at Karolinska University Hospital in Stockholm, Sweden, and a series of publications in medical journals relating to these interventions (Nature Editorial, 2016) – there has been a discussion regarding innovative treatment, how it should be regulated, and how it relates, and should relate, to research. Two commissions in Sweden have dealt with this issue and have reached the shared conclusion that a single patient case of innovative last-resort (or urgent) therapy without a proper research protocol and ethical approval is acceptable under certain circumstances, including severe urgency of treatment, but not for repeated cases: If it is expected that there will be more than a few cases, the exploration of the novel treatment should be carried out in a research project, after approval from the research ethical review board (The Royal Swedish Academy of Sciences and The Swedish Society of Medicine, 2016; The Swedish National Council on Medical Ethics, 2016).

What is proposed in the commissions’ reports is that the standard and ideal way to evaluate new therapies is through controlled clinical trials, the implicit message being that this is to be much preferred to piling up individual cases and conducting retrospective studies on what was not officially considered part of a research study at the time treatment was taking place. Both scientifically and ethically, there are good reasons to follow this advice. Scientifically, if planned as a research project, it is more likely that the relevant monitoring will be in place and that well considered endpoints are followed up and, eventually, reported, compared to a focus purely on the health condition of the present patient. With the “scientific relay” mentally switched on, aspects that need to be noted to complete the future scientific reports will more likely be recognized. In this way, researching clinicians will not need to regret afterwards that they overlooked controlling for or recording some of these aspects.

One great ethical benefit of following up innovative treatment in clinical studies is that there will be an external evaluation of the ethical feasibility. Unacceptable features of the intervention overlooked by those directly involved in planning it may be recognized when reflected on by ethical review board members, as it were, from the outside. But it is also an ethical imperative to systematically follow up innovative medical treatment – how else would we be able to tell whether it is likely to offer a feasible solution for similar cases in the future? And if it does, we should have it documented as well as possible. Thirdly, if this requirement is implemented and enforced, it will serve a gatekeeping function against attempts to do clinical research unannounced, for instance, by building up sets of cases of innovative treatment without advertising beforehand what is going on. For the sake of clarity at the hospital, and for the sake of trust in healthcare and medical research among the general public, it is most likely helpful to apply the motto “Whenever in doubt, don’t be” by making it the natural step in cases of doubt to write an application to the ethical review board instead of lingering over the question of whether or not it actually is research.

The procedure presented and defended in the commission reports mentioned above is that innovative treatment should be handled in a research study from the first patient, if possible, or otherwise after the first patient. However, it could be debated whether such studies should allow greater flexibility than standard clinical tests – for instance, whether it should be permitted to adjust treatment to the needs of the individual patient. One might argue that it would be absurd if innovative last-resort therapy, initiated due to an urgent need from a particular patient group, were not allowed to be used in the best interest of the patient with reference to a strict research protocol. That is to say, perhaps the best interest of the patient should prevail here, despite the fact that the treatment is provided in the context of a research study. However, it remains of great importance that the healthcare system be able to learn for the future from such studies.

Demarcation problems

A complication when it comes to basing a policy on “innovative treatment” is that there will be demarcation problems. How small or great a step away from standard applications of established practice can you take without ending up with innovative treatment? For instance, if the treatment is established for adults, can it then be applied to young adults and older children without being considered innovative? What about somewhat younger children? Young children? Infants? Newborns?

One factor that might make a difference in some of these cases is whether or not therapy concerns metabolism, since the metabolism of newborns, infants, and young children is different from that of adults. Another potentially relevant boundary may be if the step from one patient group to another involves a shift in technology. But what if the technology shift is mainly a matter of size, or if we already have knowledge of the technologies in question but only one specific kind of technology has been used in the present context? If the criteria to be used allow only small steps away from standard applications of established practice, then a large proportion of shifts made in treatment procedures or applications might have to be considered innovative treatment. If, on the other hand, the criteria are generous when it comes to departing from the standard applications, then much more – perhaps too much – will pass without specific considerations.

Decisions on whether a change is large enough to introduce an innovative treatment need to be made on a case-to-case basis (The Swedish National Council on Medical Ethics, 2016). What should steer the decision is whether or not the introduced changes or novelties are such that what is done is no longer sufficiently evidence-based. This idea is admittedly vague, but points in the direction of what matters: do we really know that this works or could work, and do we really know enough about the potential risks and advantages? The greater the risks involved, the more careful one has reason to be.

Ethical review of innovative last-resort treatment

In this context, it is also worth considering how a decision made by a research ethical review board should influence what happens at the clinic. More specifically, how should it relate to the evaluation that must be made by the formally responsible person at the hospital, such as the head of the clinic, before initiating the innovative therapy? An easy case is one in which the hospital wants to initiate innovative treatment for a specific patient, has applied to the research ethical review board, and gets the application approved. But what if there is disagreement? Here it needs to be noted that research ethical review boards have no authority over decisions about therapy for individual patients – this is the responsibility of the healthcare provider (within the constraints of the legal system). This means that if the review board approves the application, this does not – and must not – stop the hospital from deciding not to treat the specific patient with the innovative therapy, in the interest of the patient. It also means that if the review board says no to the clinical study, the healthcare provider still has the option to offer the patient the innovative therapy, if justified on the legal and ethical premises of healthcare. However, it cannot be offered within a clinical study.

Nevertheless, a rejection of the application gives healthcare providers and clinical researchers a reason to reconsider – is there potential for improvement, so that a revised application could be approved? The reasonable policy is that the goal is to conduct ethically approved studies on innovative last-resort (or otherwise urgent) therapy, even if the healthcare providers in an individual case find it necessary to proceed with the innovative treatment without research ethical approval.

In order to avoid novel last-resort treatments being arbitrarily introduced in healthcare, and to get an internal review of innovative treatment proposals, it might be an adequate measure for the healthcare organization to appoint a committee dedicated to such considerations, consisting of, for example, a legal expert, an ethical expert, and one or a few medical experts. Such a committee was introduced at the Karolinska University Hospital in 2017 as a response to the Macchiarini case (Karolinska University Hospital, 2017). The idea is that those proposing an innovative treatment should provide an argument to the committee, supported by the available evidence, that the balance of the benefits and risks of introducing the novel therapy is likely to be positive. In a report from 2016, also written in response to the Macchiarini case, The Royal Swedish Academy of Sciences and The Swedish Society of Medicine suggest that such evaluations be handled by a national advisory committee, but leave the final decision to the person formally in charge of the patient’s treatment (The Royal Swedish Academy of Sciences and The Swedish Society of Medicine, 2016).

If policy requires that innovative treatment be evaluated in prospective clinical studies, should the patient still have the right to get the last-resort innovative treatment even if he or she refuses research participation? It seems to me that the answer to this should be no. Just as you cannot insist on getting the experimental treatment while refusing to participate in the trial in any other clinical studies, there seems to be no good reason why you should be able to do it here. Innovative treatment means that it is not standardly offered, and there is no legal or ethical requirement in healthcare to offer any patient such treatment. However, research participation in cases like this mainly amounts to being exposed to premeditated rigorous monitoring in addition to receiving the therapy, in order for clinical researchers, and in the longer run healthcare, to learn from the individual case. Patients should not be allowed to make it impermissible to obtain such knowledge, which is needed to determine whether the innovative treatment should be offered as standard in the future.

Conclusions

Medical research and therapy can and often should be distinguished from each other, which is important to note since there are different responsibilities tied to the different kinds of activities. It is important to maintain the distinction in order to protect both clinical research interests and the voluntariness of research participation, and to avoid the unfair distribution of treatment benefits. This does not exclude that there are cases, such as innovative last-resort treatment, in which it might be more appropriate to apply regulations from both the healthcare and research areas, rather than from only one of them.