Advance Statements for Borderline Personality Disorder: A Qualitative Study of Future Crisis Treatment Preferences

Borderline personality disorder is a mental disorder characterized by a pervasive pattern of instability in interpersonal relationships and self-image and marked impulsivity (1). Although some people with borderline personality disorder are able to negotiate life successfully and without major disruption, others suffer considerably and make frequent use of acute services (2,3). People with borderline personality disorder are more likely to experience life crises, and their ability to cope with such events may be impaired by poor problem-solving skills (4). Indeed, it is widely recognized that the risk of completed suicide during a crisis is elevated for someone with borderline personality disorder (5). During a crisis, people with borderline personality disorder may have a range of acute psychiatric symptoms, including impulsive aggression or short-lived psychotic symptoms, as well as intense anxiety, depression, and anger (6). However, the extent to which they are able to make considered decisions about their treatment when in crisis is often doubted by clinicians (7), and little is known about the future treatment preferences of people with this disorder (8,9).

The implementation of psychiatric advance directives (10) or advance statements (11) is one method by which service users are able to proactively express their treatment preferences for use in periods when their capacity for decision making might be compromised (12–14). Advance directives are widely available in mental health care settings, and several national organizations, including the National Resource Center on Psychiatric Advance Directives (www.nrc-pad.org), Mental Health America (15), and the National Alliance on Mental Illness (16), all provide resources, advice, and guidance on the process of creating such directives. Despite the availability of these resources, research suggests that they are not widely used (11).

A joint crisis plan (11) is a specific type of advance statement that is produced with the help of an independent facilitator working with the service user and care team; it contains a service user’s treatment preferences regarding future health care during acute crises (17). Previous research has indicated that joint crisis plans may reduce coercive treatment (18). The facilitated psychiatric advance directive (19) is a similar intervention that may improve communication between service users and clinicians, which, in turn, can enhance the therapeutic relationship (20). Joint crisis plans are a trusted collaborative tool among service users (21), and in a recent trial of joint crisis plans for people with borderline personality disorder, 70% of participants reported using the plans in a crisis (22). In this study, using established qualitative methods, we sought to undertake a detailed examination of crisis treatment preferences of a sample of community-dwelling adults with borderline personality disorder.

Methods

This qualitative study formed part of a randomized controlled trial of joint crisis plans for people with borderline personality disorder (22). Participants in the trial were randomly assigned either to receive treatment as usual or to develop a joint crisis plan plus receive treatment as usual. They were assessed six months later on a range of variables, of which the primary outcome measure was self-harming behavior. Despite finding no evidence of clinical efficacy of the intervention, the trial showed that joint crisis plans were viewed very favorably by participants, with more than 70% reporting that they had used them in a crisis and 85% stating that they would recommend them to a peer (22).

Ethics committee approval and trial registration

Data collection protocols were approved by the South London Research Ethics Committee, and the trial was registered with the International Standard Randomized Controlled Trial Registry before commencement of data collection (23).

Participants

Recruitment of participants was conducted at three community mental health teams within south London. After the study was fully described to potential participants, they provided written informed consent to participate. Participants in this qualitative study were all community-dwelling adults who created a joint crisis plan (N=41) in the aforementioned trial, all of whom met DSM-IV-TR criteria for borderline personality disorder, as measured by the borderline personality disorder subsection of the Structured Clinical Interview for DSM-IV (24,25).

Data collection

Each participant attended at his or her local community mental health team site a one-time joint crisis plan meeting, which was facilitated by a clinical psychologist (RB) independent of the participant’s treating team between January 2010 and May 2011. This meeting was also attended by the participant’s treating community mental health team clinician, and participants were encouraged to invite to the meeting a caregiver, friend, other health professional, or other advocate. At this meeting, all parties contributed to an open discussion about the information the participant wished to include in a joint crisis plan, using the crisis plan subheadings as a basis for the structured discussion. Participants were encouraged to consider the advantages and disadvantages associated with the information he or she wished to include in the plan, and opposing opinions were resolved within the group. The final information included in the joint crisis plan was entered in the participant’s own words. After the meeting, the clinical psychologist (RB) distributed copies of the plan to all parties nominated by the participant, and, with the participant’s permission, uploaded a copy onto his or her electronic medical record.

Data analysis

The qualitative methodology guiding this study was thematic analysis (26). Analysis began with two raters (RB and KT) independently conducting open coding of all joint crisis plans, with the codes rooted in the data. The two raters then compared their codes, and a preliminary coding frame was built up. This frame, and the initial categories, were scrutinized by two senior clinical researchers (RCH and PM), and the coding frame was further developed. Some units of text were assigned several codes to reflect the multifarious nature of participants’ statements. The two initial raters then actively searched for data that did not fit into the coding frame (deviant cases). Revisions to the coding frame were again cross-checked by senior clinical (RCH and PM) and service user (DR) researchers and further refined through discussions of the appropriateness of each of the codes, with any initial disagreements resolved iteratively through consensus. Microsoft Word was used for indexing material and for retrieval of text chunks pertaining to the same or similar codes. This procedure ensured reliability of the analysis in that it relied on the multiple coding and combined assessment of five raters (27). In terms of validity, this procedure is transparent, and this attribute has been argued to be an indicator of validity in qualitative analysis (28).

Results

Forty-one of the 46 participants (89%) randomly assigned to the intervention arm created joint crisis plans, and their demographic data are shown in Table 1. A majority of participants were female, white, single, and in their thirties.

Experience of crises

When reflecting on previous crises, all participants described a state of acute distress associated with a disruption in their daily functioning. When they were in this state, many participants reported failing to meet basic self-care requirements, such as eating, drinking, and bathing. Many reported withdrawing from the outside world and avoiding contact with friends and family members for the duration of the crisis. In addition, it was common for such isolation to result in further complications, such as not paying essential bills.

Connecting with or disconnecting from others in a crisis

Many participants emphasized the importance of having the opportunity to connect with people in their personal networks during times of crisis. Conversely, some participants preferred to disconnect completely from other people during crises. Table 2 provides examples of situations and actions perceived by participants as being helpful or unhelpful during crises.

Exacerbating factors

Many participants described situations and behaviors that exacerbated the distress experienced during previous crises and elected to include these in their crisis plan as reminders of what to avoid during future crises. These fell into two categories: interpersonal interactions (such as seeing friends or family members whom they did not wish to see or, conversely, being isolated when they did not wish to be alone) and self-destructive behaviors (including self-harming, misusing substances, engaging in risky sexual behavior, or spending vast amounts of money).

Interactions with mental health professionals during crises

Many participants recalled unhelpful interactions with mental health professionals during previous crises and expressed a desire to avoid similar interactions in the future. Terms used to describe clinicians who had interacted in an unhelpful way included “judgmental,” “dishonest,” “dismissive,” “condescending,” “disrespectful,” “sarcastic,” “misleading,” “impatient,” “patronizing,” “not taking [the participant] seriously” and “treating [the participant] like a child.” Specific actions performed previously by clinicians that were deemed unhelpful during a crisis were described as “not following through on promises” (including not returning phone calls), “not being discreet in front of others,” “comparing [the participant] to other clients,” and “giving [the participant] religious advice.” Table 3 presents statements from participants’ joint crisis plans describing how they wanted clinicians to treat them during future crises.

Specific refusals regarding treatment

Thirty-seven participants (90%) included at least one specific refusal regarding treatment during future crises. Preferences regarding medication and involuntary treatment were the two most common refusals. The 23 medication refusals were most often based on a preference to avoid being treated with specific medications (56%, N=13), undesirable side effects of specific medications (17%, N=4), or known allergies to certain medications (12%, N=3). Eleven participants (27%) expressed a preference not to be involuntarily hospitalized for treatment when in a crisis.

Dissemination of joint crisis plans

All participants requested a copy of their joint crisis plan, and, as requested, copies were distributed to a range of health and statutory representatives involved in their care. The two most frequently requested recipients were community mental health team care coordinators (95%, N=39) and primary care physicians (90%, N=37). Twenty-three participants (56%) requested a copy of their plan to be uploaded to their electronic medical records. Participants named additional recipients as follows: other health professionals (36%, N=15), parents and children (20%, N=8), friend (17%, N=7), partner (12%, N=5), the emergency department at the local hospital (5%, N=2), probation officer (2%, N=1), and other family member (2%, N=1).

Discussion

To our knowledge, this is the first study to rigorously examine the advance statements of a sample of community-dwelling adults with borderline personality disorder. In their plans, participants made clear and considered choices to include useful information both for themselves and for health care professionals with whom they might interact during future crises. Participants reported that during previous crises they had failed to meet their basic self-care requirements, including eating, drinking and bathing. Subsequently, many participants chose to include in their advance statement a reminder to address these basic needs during future crises, suggesting that thoughtfully constructed written advance statements can provide helpful, immediate reminders during an acute crisis. Key themes included the desire to be treated with dignity and respect and to receive support from clinicians. Other themes related to preferences for specific treatment refusals during crises, including particular types of psychotropic medication and involuntary treatment.

Participants were divided on the issue of connecting with others while in a crisis. For many, having the opportunity to connect with others was seen as both life affirming and vital, whereas many others described it as unhelpful, instead preferring to be left alone during crises. It is also possible that the same service user may wish to be alone during one crisis, yet with others during a different crisis. This divergence highlights the need for clinicians to actively discuss such issues with patients when reviewing crisis plans. The National Institute for Clinical Excellence (NICE) guideline on the management of borderline personality disorder (6) recommends that regular reviews of care plans be conducted with the service user and his or her family members or caregivers if possible and that service user autonomy be promoted at all times. For persons seeking to be left alone during a crisis, being “forced” into being with others is also potentially traumatic and can increase underlying feelings of powerlessness that often play out in interactions with clinicians (29).

Much of the content of the advance statements was related to previous unsatisfactory interactions that patients had experienced with health care professionals and an understandable wish to avoid such negative interactions in the future. These concerns indicate staff training issues, with research showing that negative clinician attitudes toward this population can be improved as a result of targeted clinical education about borderline personality disorder (30–32). A common theme in participants’ joint crisis plan was that they wanted clinicians to listen to them, to treat them with dignity and respect, and to actively respond to their needs. As such, we think that staff training initiatives should highlight the fact that people with borderline personality disorder are able to make considered choices around planning their future treatment and should be treated as equal partners in planning meetings.

Treatment preferences regarding medication and involuntary hospital treatment were the two most common treatment refusals. The NICE guideline (6) states that although medication is commonly started when a person with borderline personality disorder presents in crisis, there is no evidence for the use of any specific drug or combination of drugs in crisis management. In addition, there are currently no medications licensed in the United Kingdom for the treatment of borderline personality disorder. The choices expressed by participants regarding medication type therefore appear to be ones that clinicians should not have difficulty following. In this regard, our findings are similar to research into the content of psychiatric advance directives produced in the United States by people with severe mental illnesses (33,34).

At the request of participants, the advance statements were distributed to a range of health and statutory representatives involved in their care. More than 90% requested that their primary care physician be provided with a copy of their crisis plan. People with borderline personality disorder are high consumers of primary health care services (35), with primary care physicians often being the first contact in a crisis. As such, advance statements that can be shared with primary care physicians may be particularly valuable. Goals of crisis interventions for people with borderline personality disorder typically include returning people to their precrisis level of functioning and mobilizing both internal and external resources (36). These were, for the most part, reflected by our participants and mirror the goals of the recovery orientation, which is central to mental health policy in the 21st century throughout the Western world (37,38). Participants included clear statements in their joint crisis plans relating to a desire to recover from the crisis and to establish a more optimal level of social functioning. This finding challenges misconceptions held by many clinicians that people with borderline personality disorder either consistently make self-destructive life choices (39) or lack the requisite capacity to make sensible life choices (7,40,41). We think this highlights the fallacy whereby clinicians may make unfavorable assumptions about the level of functioning of people with borderline personality disorder during their daily lives based on their presentation during a crisis. Our findings also demonstrate that the same issues of dignity, respect, and autonomy identified in global surveys of discrimination among people with severe mental illness (42,43) are important to people with borderline personality disorder.

Our findings have important implications for clinicians involved in the daily care of people with borderline personality disorder, at a time in the evolution of health care services when shared decision making (44) is becoming the norm rather than the exception. Clinicians may wish to use joint crisis plans in routine practice as a method of discussing, in considerable detail, the differing risk factors, protective factors, and treatment preferences of the individual service user. One of the strengths of joint crisis plans is that they are written in the service user’s own words and are therefore largely free of the medical jargon seen in patient records and earlier forms of crisis plan.

Our findings need to be considered in view of some limitations. First, most participants in our sample were female, and what impact, if any, this may have had on the data collected remains unclear. Nevertheless, the profile of our patients was typical of that seen in clinical practice. Second, we were unable to verify our coding framework with participants (that is, to conduct a respondent validation exercise); however, our research team included a senior service user researcher who was closely involved in the analytic process. Third, all joint crisis plans were created in the presence of participants’ care coordinators, and this may have affected the statements made by participants. Against these points, however, the content of the joint crisis plans suggests that participants felt comfortable to discuss previous experiences when clinicians had been neither helpful nor respectful. Our study also had strengths. We used a randomized sample of community-dwelling people with borderline personality disorder that was clearly defined with DSM-IV criteria. These sample attributes increased the external validity of our findings. In addition, two experienced researchers (who achieved good interrater reliability) were involved in coding the qualitative data, which were then independently checked by other members of the research team.

Conclusions

Our study highlights several important issues. First, the advance statements made by the participants mostly expressed a clear desire to recover from the crisis. Second, practitioner interactions that preserve patients’ dignity, respect, and autonomy are as critically important to people with borderline personality disorder as they are to people with a severe mental illness, and to patients in general. Third, the variation observed in the treatment preferences of participants in our study underscores the importance of developing individually tailored advance statements and involving service users with borderline personality disorder in collaborative therapeutic relationships. Participants in our study wanted to be listened to, to be treated respectfully, and to be actively involved in decision making concerning their care. Our findings highlight several clear domains in which people with borderline personality disorder can be more involved in planning their future care during crises and important lessons for practitioners and policy makers regarding future care planning for this population.