Participants

Participants were identified by regular examination of electronic medical records and consultations with ward nursing staff. The sole exclusions were individuals from other catchment areas or transfers from other in-patient facilities.

Potential participants were approached for a research interview. Those assenting were given full details of the study. Written consent was sought. The interview was stopped if there was any subsequent change in choice or resistance. Individuals were offered £5 for their time. Capacity assessments were held as close to admission as possible.

If individuals were found to lack capacity at admission they were followed up. The follow-up assessment took place at either the point of discharge from hospital or at 1 month (whichever was sooner). We chose 1 month because this was the average length of stay for people admitted to the Maudsley Hospital acute wards at the time of the study. At follow-up the capacity assessment was repeated and participants were asked about their views on surrogate treatment decision-making. The permission procedure was the same. Figure 1 shows the flow of individuals through the study.

Fig. 1 Flow of participants through the study.

Assessment of capacity

Relevant information about the participant's presenting problems, ICD–10 diagnosis ¹⁰ and treatment plan was obtained from the medical record. The interviewer (G.S.O.) clarified whether the main treatment decision was stabilisation on medication or admission to hospital for a place of safety/assessment. If it was medication, the capacity assessment centred on the decision to take the prescribed medication or not. If it was hospitalisation, it was the capacity to decide whether to be in hospital or not. Follow-up capacity assessments related to the main treatment decision at admission.

The capacity judgement was based on a clinical assessment (notes, review and clinical interview) and the administration of the MacArthur Competence Assessment Tool for Treatment (MacCAT–T). ^{Reference Grisso, Appelbaum and Hill-Fotouhi11,Reference Appelbaum12} Previous studies have demonstrated excellent interrater reliability (κ>0.8) when the MacCAT–T is used to guide a clinical judgement. ^{Reference Cairns, Maddock, Buchanan, David, Hayward and Richardson13,Reference Okai, Owen, McGuire, Singh, Churchill and Hotopf14} The MacCAT–T is a semi-structured interview that provides relevant information disclosures to the individual about their illness, the nature of treatment options and their risks and benefits. The assessor evaluates capacity in terms of four abilities relating to the disclosures: understanding, appreciation, reasoning and expressing a choice. These abilities have evolved in reflection of the case law in the USA. They also reflect the requirements of the Mental Capacity Act 2005 in English law. A key feature of the capacity test is that it requires evidence that the decision-making process is impaired rather than the decision-making outcome.

Assessment of participant's views on surrogate decision-making

All participants who were followed up were asked for their views on the surrogate decisions they received following admission and about surrogate treatment decision-making in general. Participants were asked to think back to their admission and were asked: ‘do you think the right decisions were taken on your behalf?’. Responses were recorded as yes, no or don't know. We also asked participants to respond to four statements on a five-point scale ranging from ‘strongly agree’ to ‘strongly disagree’ to allow for degrees of belief. Two questions were asked in positive and negative frames. The first addressed the need for surrogate decision-making at admission (‘I feel doctors decided when I had the ability to decide myself’ and ‘I feel doctors decided for me when I needed them to’) and the second addressed whether the surrogate decision-making was of the right kind (‘I think doctors made decisions I wish they had not made’ and ‘I think the decisions that the doctors made where the right ones for me’).

Views relating to surrogate decision-making in general were elicited by asking participants for their views on who should make treatment decisions if they were not able and whether doctors should be able to make decisions on behalf of some people with mental health problems. Answers were summarised.

Analysis

Participant characteristics, proportions of individuals with capacity and their views on surrogate decision-making are presented descriptively. Simple statistical comparisons were made.

Main findings

Doctors, particularly psychiatrists, as part of their practice, sometimes override the treatment wishes of their patients. We know that mental incapacity is common in both in-patient psychiatric and medical settings ^{Reference Owen, Richardson, David, Szmuker, Hayward and Hotopf9,Reference Raymont, Bingley, Buchanan, David, Hayward and Wessely15} and we are beginning to clarify how it relates to clinical concepts such as diagnosis and insight. ^{Reference Owen, David, Richardson, Szmukler, Hayward and Hotopf16} There are virtually no data on the overriding of treatment wishes and few on the effects of clinical improvement ^{Reference Gardner, Lidz, Hoge, Monahan, Eisenberg and Bennett6,Reference Kane, Quitkin, Rifkin, Wegner, Rosenberg and Borenstein7,Reference Schwartz, Vingiano and Perez17} despite the obvious importance of the matter in clinical practice. In this area, doctors may feel precariously balanced between liabilities for negligence on the one hand and assault on the other. Because of the high ethical premium placed on self-determination, acts of overriding treatment wishes stand out and call for justification. We studied what people think when they are on the receiving end of such acts.

The majority of individuals who regained decisional capacity (80–90%) indicated that they thought they needed surrogate decision-making by doctors during their recent admission and that it was of the right kind. Between 5 and 20 per cent indicated that they thought it was not needed or not of the right kind. Most people regaining decisional capacity stated that doctors or family should make decisions if they were unable. Almost all participants said that doctors should be able to make decisions on behalf of some people who have mental health problems.

There were no clear differences in the views expressed about surrogate decision-making between those who were treated involuntarily under the Mental Health Act on admission compared with those who were treated informally. In Kane et al's 1983 study in the USA, ^{Reference Kane, Quitkin, Rifkin, Wegner, Rosenberg and Borenstein7} 69% of patients who were judged by their doctor to be 80% improved said they thought it fortunate they were detained when asked about it 2 months later. Our study suggests that when regaining capacity rather than clinical improvement is used as the outcome variable, the levels of endorsement of surrogate decision may be higher.

The views of people who did not regain capacity were more mixed. Fifty-nine per cent did not give retrospective endorsement of the decisions that were made on their behalf. Many of them thought professionals or family members should be able to make surrogate decisions but there were also views that only God should be able to make surrogate decisions or that no one should make surrogate decisions. Although these views on surrogate treatment decisions have to be interpreted in the context of present incapacity, clinicians need to have regard to them when planning future treatment. These individuals' relative lack of endorsement of the treatment might have had some justification – after all, the treatment had not led to their regaining capacity over approximately 1 month – but firm conclusions cannot be drawn here. Information after longer follow-up would be valuable particularly in light of the results from Priebe et al ^{Reference Priebe, Katsakou, Amos, Leese, Morriss and Rose8} mentioned above.

Limitations

This study has limitations. It is a series drawn from consecutive admissions to psychiatric in-patient units within a single UK urban district with results that may not generalise to other in-patient settings. Further, consideration needs to be given to the face validity of the hindsight views we elicited. A range of contextual factors may influence the expressions of hindsight approval. Thus, participants who expressed approval might have felt that ‘saying the right thing’ would shorten their stay in hospital. In contrast, others who did not give hindsight approval might have found it difficult to reverse strongly held and expressed differences of opinions with the treating team during the acute phase of an admission. Further, the manner in which treatment decisions were overridden, which may be reflected in the strength of ongoing therapeutic alliance, may be important in influencing the views expressed by participants. We tried to minimise these contextual factors by emphasising the researcher's independence from the clinical team and giving assurances that the interview would not affect the treating team's decisions. Capacity assessments were made prior to eliciting views on surrogate decision-making in order to reduce observer bias in the capacity assessment. We used both positively and negatively framed questions to elicit retrospective views.

Implications

Most participants who regained capacity agreed with the surrogate decision-making that had taken place during their admission to hospital. They may not have much liked the process of admission or the services received – there may be no ‘thank you’ – but they indicated approval nonetheless. This was the case even in people whose treatment wishes were overridden by doctors using the powers of the Mental Health Act. This implies that most individuals when regaining capacity (where consents or refusals carry legal respect – whether wise or unwise) retrospectively judge their incapable states of mind to have required surrogate decision-making. If this can be taken to indicate that most people with mental illness who regain capacity would use their past experience to make anticipatory decisions about treatment that would favourably acknowledge the potential role for surrogate decision-making then it might assuage some of the concerns surrounding the use of psychiatric advance directives, at least when they are drafted following the experience of acute illness. There is some evidence that psychiatric crisis planning resembling advance directives reduces compulsive treatment. ^{Reference Henderson, Flood, Leese, Thornicroft, Sutherby and Szmukler18}

Psychiatry is under an obligation to reflect upon practices that cause the public concern. Overriding the treatment wishes of individuals is one such practice. This study contributes some empirical data to help this reflection. We think it contributes a moderate degree of reassurance both concerning surrogate decision-making by psychiatrists and advance decision-making by people with mental illness.