Challenge 1: Psychiatric patients eligible for DBS can be vulnerable and have high expectations

Psychiatric patients referred for neurosurgical consultation should have been deemed resistant to conventional treatment. Although there is no precise definition of treatment resistance, this is typically determined by a failure of the standard of care, such as, for example, psychotherapy and medications in depression and obsessive-compulsive disorder (OCD), and in the case of the former, failure of electroconvulsive therapy as well.4 This lack of benefit with successive treatment places the patient in a vulnerable position, one that could alter the threshold for agreeing to participate in potentially risky clinical trials. DBS for most psychiatric indications, including depression, is not currently available as an approved therapy, and as such, receiving the procedure means doing so as part of an experimental trial. The only exception is in OCD in which the US Food and Drug Administration has issued a humanitarian device exemption for the use of DBS in highly refractory and resistant cases. Research, however, into psychiatric DBS, including OCD, is still very much an area under active investigation and as such, a patient's referral to a neurosurgeon typically constitutes the crossing of an invisible border, from treatment to research. This division is rarely explicitly addressed and enhances the vulnerability of these patients by potentially disturbing the consent process, in several ways. First, failure to recognise DBS as research lends credence to DBS as therapy, leading potentially to a minimisation of the unknown risks, as well as misplaced faith in the efficacy of the ‘treatment’. Second, referral to a neurosurgeon without explicit discussion of the therapeutic misconception and the experimental and unproved nature of the intervention indirectly capitalises on patients' desperation to improve. For example, if DBS is perceived as an extension of treatment, patients may view ‘brain surgery’ as a validation and confirmation of their ‘desperate’ condition, ie, desperate times call for desperate measures. Indeed, DBS is certainly a more physically invasive intervention than other psychiatric treatments, and may therefore be perceived as more likely to get at the ‘root of the problem’. Patients could then potentially and incorrectly attach more weight to the perceived efficacy of DBS, and thus be more willing to proceed with a risky trial not because they feel it may benefit them necessarily, but as a reflection of their attitudes towards their condition and its history of treatment failure. It is important to note here that patient desperation should not be considered an exclusion criterion for DBS trials: patients who have reached the limits of conventional therapy are indeed desperate to get better and should not be denied the option to take part in clinical trials. This is not at issue here, but rather whether desperation alters the ability of patients to judge the risks and benefits of an untested ‘therapy’ adequately and if so, if researchers are indirectly taking advantage of this fact for trial enrollment. This is a fundamentally different issue from competency, which will be discussed below, which is the ability of patients to understand the goals and details of a clinical trial as a consequence of their underlying condition. Here the discussion is not whether patients understand, but whether their history of treatment resistance has altered their ability to appreciate (1) the research treatment distinction; (2) the risks involved in an untested therapy; and (3) that DBS, for example, may have little to no individual benefit, and thereby potentially violate their goals and expectations of a ‘risky’ therapy. There are currently no empirical studies that examine the ability of psychiatric patients eligible for clinical trials of surgical interventions to distinguish therapy from research or accurately gauge or recall the risks of described untested therapies. There is some literature on patient expectations in psychiatric treatment clinical trials, as these are recognised to have a significant impact on treatment and study outcomes. For example, one recent study found that in a sample of 90 patients with bipolar disorder, the ‘discrepancy between actual and hoped-for outcomes can be substantial’, and that 40% of patients receiving medications for treatment hoped to be ‘cured’ or to be ‘normal’.5 Unrealistic expectations of outcomes are particularly worrisome in clinical trials, and even more so in trials involving treatment-resistant patients, who have already failed all previous therapies. The hope that ‘this will be different’, is arguably heightened with riskier trials, such as DBS, in which risk can be viewed as a gauge of the potential benefit.

The concept of benefit in clinical trials, particularly for high-risk interventions, is complex, and akin to a cubist canvas, in which multiple perspectives, in this case the benefits to different parties, are portrayed simultaneously. The need for consistency is therefore critical. For example, the potential benefit to the individual subject (1) must be weighed and balanced, alongside the future benefit to all patients with a given diagnosis (2), as well as with the benefit to the individual researcher (3). The primary objective of a clinical trial, furthermore, as outlined in the consent form, is the disruption of equipoise in a situation in which the optimal therapeutic direction is unclear (benefit 2); it is not the individual benefit to the patient (1), although this may realistically be a motivation for enrollment for some individuals. A systematic examination of why patients enrol in high-risk clinical trials has not yet been undertaken and would help elucidate whether benefits to the individual, to the group or a combination, underlie a decision to participate.

An important principle of psychiatric care is that mental illness does not affect patients alone. For patients who have a support network of family or caregivers, a mental illness such as depression or OCD can have a profound and devastating impact on those around them. It is important, therefore, for clinicians to recognise the broader, social impact of mental illness not only because patients with support have better outcomes, but because the impetus to get treatment, and possibly to enrol in clinical trials, may be external to the patient. This can represent an additional, largely unrecognised, source of vulnerability in psychiatric patients, namely that the influence of those around them can potentially translate into taking on additional ‘involuntary’ risks, including clinical trial enrollment.

Treatment resistance, outcome expectations, the therapeutic misconception, and external pressure to participate, can all affect the patient's threshold for agreeing to enrol in a DBS clinical trial, putting at risk their ability to decide autonomously and voluntarily on participation. Furthermore, this risk is independent of any possible effect that the underlying condition can have on their ability to consent. These patient factors need to be appropriately isolated and addressed if consent acquisition is judged to be ethically sound.

Challenge 2: Psychiatric diseases may affect patients' competence to consent

Diagnostic criteria for all psychiatric diseases include, by definition, disturbances of mental function, to the point of impediment in everyday life. The aim of psychiatric treatment is a restoration of that function to achieve individual reintegration into life, work and family, whereas the objective of psychiatric research is to gather data on new and emerging therapies, their safety and efficacy, with the ultimate aim of improving the lives of all patients with the disease in question. In discussions of research consent and mental illness, it is important first to draw a distinction between disorders of impaired reality testing and those in which reality testing is intact. The former, which includes psychotic disorders such as schizophrenia, are beyond the scope of this paper, and will not be discussed in detail. Suffice it to say that such disorders present unique challenges to informed consent and trial enrollment, and will by necessity involve substitute decision makers and proxy consent. Here, the discussion will be limited to non-psychotic disorders, and specifically those disorders that are currently the subject of ongoing surgical clinical trials, namely depression and OCD. More broadly, these conditions belong to a group of psychiatric disorders that are classified as ego-dystonic, or conditions that generate anxiety in patients as a result of an incongruity between the way the patient wants to think and behave, and the way they do behave. Treating patients with ego-syntonic conditions can be ethically problematical, but is sometimes nevertheless necessary, as in individuals with troublesome paraphilias or conditions that may lead to self or other harm.

The question then becomes whether non-psychotic, ego-dystonic disorders that are treatment resistant affect the ability of patients to consent to research. Work done by Grisso et al6 has examined competency for treatment decisions in psychiatric patients, as it relates to four dimensions: understanding, reasoning about risks/benefits, appreciation of their situation and the expression of a choice. The resulting scale, the MacArthur competency assessment tool–treatment (MacCat–T) has since been modified for use exclusively in research populations, the MacCat–clinical research (CR).7 The initial validation of the MacCat–CR in 26 depressed women eligible for research studies found ‘few impairments’ in the ability to decide on trial enrollment.7 A valuable review by Dunn et al8 outlined 23 decisional capacity instruments, 10 of which are dedicated to research consent. As the authors point out, the heterogeneity in the field, as evidenced by the variable expertise required and diverse administration times of the 10 research instruments, prevents any broad conclusions to be made. However, the MacCat–CR stands out with its history of empirical validation, validity, reliability and assessment of all dimensions of informed consent.

Disorders with impaired reality testing or those conditions with organic brain pathology (eg, brain trauma, delirium, etc.) render decision making unreliable, but the situation becomes less obvious, and the need for capacity measurement more important, in conditions of borderline competence, as well as in cases in which all previous treatments have failed. Elliot9 has argued that patients with severe depression may have a disregard for their own wellbeing and this may affect their ability to weigh the risks and benefits of trial enrollment. He further argues, in a paper published before the development of the MacCat–T or MacCat–CR, that testing reasoning and memory alone cannot provide an adequate picture of competency. This is an important and valid point; however, it is as important to avoid overstating the importance of emotions in decision making as it is to downplay their importance, which is Elliot's concern. Clinical studies have established that patients with severe depression can have impairments in reasoning, information processing and memory.5 10 Neuropsychological and psychiatric testing before consent acquisition and as part of the trial eligibility process will address these issues and flag them as exclusion criteria. The emotional features of depression, however, are components of the disease process itself, similar to overemphasising negative valuation of stimuli, or perceptual errors of emotional stimuli. Indeed, it has been the experience of one of the senior authors (AML) that some patients eligible for depression DBS do not necessarily view death as a negative outcome. Minimisation of risk to self is a feature of the disease process that could represent one of the endpoints and outcome measures of a clinical trial. Patients with severe depression who are eligible for clinical trials, in other words, should not be handicapped by the essential features of their condition, but should instead, as Elliot suggests, have these better incorporated into assessments of capacity.

There is a hidden vulnerability, or a hidden source of competence-at-risk, in the severely depressed patient that can manifest in a skewed perception of trial risks that are seen and judged through the lens of years of previous therapy. An argument could be made, for example, that ongoing treatment in the face of persistent therapeutic nill effect blurs the line between treatment and research, especially in the absence of empirical evidence guiding therapeutic decision making on the part of the treating physician. For the patient with a strongly ego-dystonic condition, such as depression or OCD, treatment failure for several years may constitute an effective plateau whereby participation in a clinical trial may for the patient be no different (ie, no riskier) than taking the latest poly-pharmacy formulation their psychiatrist prescribes. This hidden face of the therapeutic misconception is thus enhanced by the severity, chronicity and resistance of the illness to treatment.

Challenge 3: DBS is inherently risky neurosurgery with unknown efficacy and an unclear mechanism of action

DBS has been used as a therapeutic tool for the past 20 years, primarily in the management of movement disorders. The procedure involves the image-guided selection of a deep brain target, which is then stereotactically implanted with electrodes, in an effort to disrupt the activity of the target structure and its dysfunctional circuit.11 12

Although effective in controlling symptoms and disrupting physiological activity, the mechanisms of action of DBS are unknown. As such, it is impossible to tell, a priori, precisely what the risks of a DBS procedure are for a specific psychiatric indication. Heterogeneity in the literature with respect to target selection for the same indication, as well as idiosyncrasies with regard to surgical techniques at individual centres, compound the ambiguity and will affect discussions of risk with patients preoperatively.

As far as risk of harm to the physical brain or body, there is a robust literature to guide these discussions with patients. For example, it is known that DBS procedures, which are technically identical for movement disorders and psychiatric indications except for the target, can be associated with a 2–3% risk of intracerebral haemorrhage, a 5–8% risk of infection and a 0.4% risk of death related to surgery.13 The ambiguity surrounding the precise mechanism of action of DBS is outweighed, as in chemotherapy and antipsychotic clinical trials, by the history of safe and effective use in other populations, as well as the emphasis on observed outcome with close follow-up. The key, as in other novel therapy trials, is full risk disclosure to the patient, and a re-orientation towards the still unanswered questions of DBS therapy.

More difficult to address, and murkier still in terms of ambiguity, are the potential risks of psychiatric DBS in the postoperative period. These can range from unexpected psychiatric side effects (eg, mania, impulsivity) to more global changes (eg, changes in personality). Some authors have suggested risks such as changes to personal identity and self, given the unique substrate of DBS in these cases, namely the patient's general mood and features of their personality that have presumably become firmly entrenched in their self-concept.14

The challenges that DBS technology offer to the trial consent process are different from other therapies being tried in new indications, and different even from DBS trials in non-psychiatric indications. This stems from at least two directions, namely that DBS clinical trials test a ‘double innovation’, as well as that DBS confers risk on patients, above and beyond standard psychiatric treatments, in which electroconvulsive therapy is considered the most physically invasive therapy. ‘Double innovation’ suggests that DBS trials not only test efficacy in a new indication (eg, depression), but that the same trials test the safety and utility of a novel technology in general. Both types of innovation research should be described to patients enrolling in surgical psychiatry clinical trials, also emphasising that neither type of research explicitly aims to benefit the individual patient.

Regarding the risk of bodily harm, the challenge for the patient is to balance the risk of trial enrollment, specifically the known risks to body, the less-known risks to mind and no explicit a priori known individual benefit, with the risk of their untreated disease and/or the risk of additional ongoing ‘treatment’. In DBS trials, full disclosure of unknown risks becomes as important as disclosure of known risks, and a concerted effort needs to be made by the clinician researcher not to understate the potential impact on both mind and body.

Navigating the challenges: discussion and implications for trial design

Informed consent issues are heightened when dealing with surgical neurmodulation trials in which risks are arguably greater and when patients are most certainly sicker (ie, treatment resistant). DBS for psychiatry is the only field that balances the sickest and most vulnerable patients with the risk of brain surgery; in no other field of research are these extremes brought so close together.

In the case of DBS, there is a growing body of evidence surrounding its efficacy in refractory mood and anxiety disorders that can lead to a recommendation for late phase clinical trials, and in some cases and in specific anatomical targets for its standard use in the treatment of those conditions. This is not yet the case for other psychiatric conditions, such as the psychotic disorders, in which there remains no consensus regarding the neurobiological underpinnings of these conditions and thus no consensus on surgical targeting, and it is unclear if, and how, ego-syntonic conditions should, or could, be subject to surgical intervention.

For indications in which there is a strong research base with an established DBS history, such as Parkinson's disease, redundancy measures such as third party consent are not necessary, and traditional conceptions of competence should be used in consent procedures. For more novel trials utilising DBS, the first conditions to be trialled should be strongly ego-dystonic with intact reality testing (eg, mood and anxiety disorders). The reasons for this are both practical and scientific. Patients with ego-dystonic conditions, such as depression and OCD, are more likely independently to seek as well as to comply with treatment. Furthermore, these conditions' pathophysiological underpinnings are better understood than the psychotic disorders, for example, therein allowing a stronger rationale for a somatic, targeted, surgical intervention. Even so, as DBS in psychiatry remains a double innovation, and as some authors have voiced concern regarding the competency of these patients (ie, depressed) to consent, a third party should participate in the consent process, and the patient evaluation protocol should be evaluated and validated by a non-study researcher, as is the practice at some academic institutions.3 Should empirical evidence accumulate over time, as it is now beginning to, that competency is indeed unaffected in these conditions (ie, major depression) third party consent may one day prove unnecessary. Similarly, psychotic disorders could one day become substrates for surgery, should enough neurobiological evidence accumulate supporting a somatic intervention, and enough experience with DBS psychiatry be coupled with selecting patients compliant but non-responsive to standard therapy.

The challenges to consent outlined in this paper, with respect to patient, disease and technology factors, are not straightforward and do not have simple answers. What one can do is outline the questions that each raises, and make suitable suggestions, such that the consent process asymptomatically approaches an ethical ideal (table 1). With regard to patient factors, the key questions stem from the internal and external pressures on patients to participate in a clinical trial. These can include ambiguity surrounding the treatment–research distinction, as well as unrealistic expectations of trial outcomes. Both can tip the balance of enrollment in favour of trial participation, independent of the patient's true beliefs and knowledge of the trial and its objectives. With disease factors, redundancy measures must be put in place to avoid enrolling ‘hidden’ incompetent patients, or those whose disease process prevents an adequate estimation of trial risk. Third party consent, neuropsychological testing and equal attention to both reasoning and emotional deficits will help here. The establishment of clinical equipoise and the disclosure of both known and unknown risks helps to mitigate the challenge posed by DBS as a novel technology in a new indication.

A final word can be said about the duty of the clinician investigator. A fourth challenge to the consent process can be identified, namely that the presence of the clinician researcher, whose dual role implies an implicit conflict of interest, tips the balance of consent towards participation and further enhances the therapeutic misconception. This ‘researcher factor’ can then influence each of the patient, disease and technology aspects of consent, outlined above, if the clinician researcher does not maintain an adequate distance from the enrolling, and enrolled, individual.15 Here too, mitigating factors can be employed to reduce, but not eliminate, the influence of the clinician researcher, whose presence in academic medicine is ubiquitous and necessary for both scientific and academic development. Mitigating factors can include third-party evaluation of eligible patients, including colleagues in the assessment process, and delegating the discussion of study details, including risks, to other study personnel.