Borderline personality disorder: part 2 – psychiatric management

Summary

The large volume of seemingly conflicting guidance on the management of borderline personality disorder (BPD), combined with the ongoing shortage of specialised resources, can make the task feel like an exclusive undertaking that the general psychiatrist is underprepared for. In this article, we distil current evidence to submit that sound psychiatric management principles used to treat all serious and enduring mental disorders (diagnostics, comorbidity management, rational pharmacotherapy and dynamic risk management) are readily applicable and particularly therapeutic for BPD. We offer actionable practice guidance that we hope will render the clinical management experience a more lucid and rewarding one for both practitioner and patient.

LEARNING OBJECTIVES

After reading this article you will be able to:

• • utilise current evidence on the psychiatric management of BPD to derive pragmatic, actionable strategies for clinical practice

• • identify the actual tasks of psychiatric management in BPD, and recognise that BPD management can be done effectively and therapeutically within existing mental healthcare systems

• • improve confidence in and inspire ownership of clinical decision-making when caring for this patient group.

In our first article (Garland 2020) on borderline personality disorder (BPD) we summarised the core clinical features of the disorder and discussed appropriate diagnostic practice. In this second article we turn attention to the psychiatric management of BPD.

The ‘unique’ volatilities and intensities that characterise BPD can appear to overwhelm a mental healthcare system's conventional approach, which works well for other serious and debilitating psychotic or affective disorders – an approach that prioritises adherence to medication and sees more consistent engagement, better environmental attitude towards the patient group and ready agreement that an actual ‘illness’ is at play (McKenzie 2022). Everything BPD management seems not to be (Box 1).

BOX 1 Why read this article?

You have just joined a generic community mental health team as a consultant psychiatrist as your first job after specialty training. You have had no specialised training in the management of borderline personality disorder (BPD), but patients with the diagnosis are 40% of your case-load.

Many of these patients exhibit repeated and risky ‘crisis’ presentations, and their engagement with offered services is inconsistent. Most have had referrals for psychological therapy rejected as they are deemed ‘not therapy-ready’. The mental health trust's personality disorder services are prepared to provide ‘advice’ but communicate clearly that case management is not their remit.

Two pieces of advice that you find everywhere you turn is that ‘hospital admissions are unhelpful’ and that ‘medication is not first-line management’ in BPD.

This combination of chaotic, inconsistent engagement, high-risk escalations, discouragement of and confusion regarding the role of in-patient care, and the consistently communicated ‘unsuitability’ of psychological therapy (which is supposed to be first-line management) has left you stuck. These patients show a significant morbidity burden, but answers on how to help and what to do with their clinical challenges remain elusive.

You really could do with a clear, practical strategy to manage this large group of patients that you feel genuinely confused about.

At its more intense times, the nature of the BPD management experience can force the rhetorical question of whether this is even a true mental disorder that psychiatry has a role in managing (Markham 2003). We submit instead that there are few diagnoses that offer as complete a ‘fit’ and use for all the skills of the general psychiatrist. BPD treatment requires (Fig. 1):

• • use of medical/psychiatric diagnostics

• • involvement in and direction of the management of individual cases and the collective case-load by clarifying comorbidity and identifying intervention targets

• • a genuinely disciplined and evidence-based use of pharmacological treatments

• • sound working knowledge (gained sufficiently during psychiatry training) of risk management in psychiatric disorders.

FIG 1 An overview of the components of the psychiatric management of borderline personality disorder, involving: a comprehensive and unambiguous diagnosis; assessment and management of comorbidities; simplified and standardised pharmacotherapy; and dynamic risk management. MDT, multidisciplinary team; NOK, next of kin; EUPD, emotionally unstable personality disorder.

Although psychological therapies are evidence-based treatments for the disorder, the successful management of patients with BPD requires a lot of sound and standard psychiatric management. We use the word ‘psychiatric’ here in the most ‘branch-of-medicine’ terms possible.

Cautious not to inadvertently encourage imposing a reductionist ‘medical model’ of care on BPD management, we offer the following distillation: that if one somehow removes from it the fierceness of emotional charge and the oddly permitted subjectivity, then BPD management has more in common with the management of other serious mental disorders (e.g. schizophrenia, bipolar disorder) and addictions than it has differences. We hold this position for the following reasons.

• • BPD is also a chronic condition with a considerable morbidity burden, requiring a protracted period of stabilisation during which evident symptom remission sets in and, if appropriately supported, a more holistic and very real recovery can result. As with the mental disorders listed above, the parameters of this recovery need to be collaboratively and pragmatically defined and the rehabilitation inputs required are distinct from acute treatment.

• • All these conditions have periods of stability (whose elements need reinforcing) interspersed with acute destabilisations (which require more proactive risk management). Viewing their longitudinal courses, the kind of personal and social functioning breakdowns encountered in relapses due to non-adherence or substance use in serious mental or substance misuse disorders strongly resemble those seen in emotional ‘crisis’ presentations in BPD.

• • In-patient care has a specific (and similar) role to play in all these conditions – i.e. to provide a setting of safer risk management during which other treatments (medical, psychological) may be adapted but should continue. The thresholds for in-patient care and its duration will differ, but the approach to and role of in-patient care would appear similar.

• • In BPD treatment, medication use addresses only a selection of specific target symptoms (hostility, impulsiveness, aggression), as others (chronic emptiness, perceived abandonment) simply do not respond to treatment with any medication. A similar selectiveness is seen in how antipsychotics used to treat schizophrenia can control hallucinations and delusions, but have little effect on neurocognitive and negative symptoms.

In this article we do not suggest that psychiatric management will ‘replace’ psychological therapies. If anything, wider use of the ‘common factors’ of psychotherapy (Mulder 2017) would enhance all (non-psychotherapy) patient interventions by any member of the care team. Instead we hope to:

• • convey that the framework of psychiatric management for BPD has more (rather than less) in common with other psychiatric disorders, thereby generating an ownership of the management of this disorder

• • illustrate that the psychiatrist's core skill set is (without any major change or additional training) adequate to care successfully for these patients

• • ‘normalise’ and equate the management of BPD with that of other serious mental disorders.

Comprehensive and unambiguous diagnostics

Establish a clear and current diagnosis of BPD

Even ‘inherited’ cases of patients with existing BPD diagnoses invite a longitudinal review to establish the presence and context of diagnostic criteria, and should generate discussions with colleagues (including those outside the immediate multidisciplinary team, such as in-patient teams, personality disorder teams) to confirm or discount the diagnosis for oneself. This process serves to answer two central questions:

Does my patient actually have BPD?

Patients with BPD typically present with symptoms of other disorders and are frequently misdiagnosed with other primary conditions (Biskin 2012). Burgeoning case-loads and competing demands on clinical time have made the clinical task of distinguishing BPD from bipolar disorder, depression and post-traumatic stress disorder (PTSD) all the more challenging (Zimmerman 2010), even though clinicians are probably well aware of the validity of distinction between these conditions and BPD.

This exercise is especially challenging (and necessary) in the subset of patients who reject or disengage from the BPD diagnosis and prefer relating to another diagnosis instead, for example bipolar disorder, attention-deficit hyperactivity disorder (ADHD) or autism. Their doing so appears to be associated with a lack of understanding about BPD or their own and other people's lack of interest in talking about it (Ng 2019).

Does my patient still have BPD?

When the multi-axial classification system was introduced in DSM-III, personality disorders were placed on Axis II, with the stated understanding that they ‘begin in childhood or adolescence and persist in stable form (without periods of remission or exacerbation) into adult life’ (American Psychiatric Association 1987).

This idea continued through DSM-IV, which referred to ‘an enduring pattern of inner experiences and behaviour [ … ] manifested in cognition, affectivity, interpersonal functioning, or impulse control [ … ] which is [ … ] inflexible and pervasive [and] stable and of long duration’ (American Psychiatric Association 2000).

The perception these descriptions established was that personality disorders were stable, lifetime disorders to be distinguished from the more episodic disorders diagnosed on Axis I. However, this notion is universally undermined by findings on diagnostic stability – even accommodating their numerous methodological shortcomings, a host of studies from the DSM-III era found that fewer than 50% of people diagnosed with personality disorders retained these diagnoses over time (Paris 2003). In more rigorous modern studies, this low diagnostic stability was even more certain – in the McLean Study of Adult Development (MSAD) (Zanarini 2006) and Collaborative Longitudinal Personality Disorders (CLPS) study (Gunderson 2011), 88 and 85% of people with BPD experienced remission at 10 years.

Over time, two consistent findings emerge about BPD symptom stability:

• • personality psychopathology improves over time at unexpectedly significant rates – a clear departure from the impression of these being life-long conditions

• • although personality psychopathology improves (enough to revoke a diagnosis), residual effects often persist in the form of sustained functional impairment, continuing behavioural problems, reduced quality of life and ongoing Axis I psychopathology (Zanarini 2004a).

A clear and collaboratively reached diagnosis:

• • forms a basis for setting accurate, realistic and actionable treatment goals

• • instils the managing team with a confidence and ownership (that counters attempts by the disorder to introduce ‘splitting'.

Explain the diagnosis in terms of a medical disorder

BPD is a debilitating disorder to live with. Repeated cycles of acute behaviour dysregulation, chronic feelings of emptiness and recurrent interference with meaningful life goals leave this patient group with very poor quality-of-life outcomes.

The patient's subjective experience is equally unfortunate – internally, BPD is characterised by feelings of estrangement, hopelessness, inadequacy, shame and self-stigma (Perseius 2005). Patients come to hold personal life-narrative themes that the diagnosis is a ‘dustbin’ or ‘wastebasket’ label, essentially communicating that nothing can be done for them and that they should forego hope (Horn 2007). The invalidating attitudes of involved professionals when caring for people with BPD are well documented (McKenzie 2022), and at its weaker moments the mental healthcare network can even communicate that the very diagnosis is invalid (the individual is ‘bad’ rather than ‘mad’) (Markham 2003).

Furthermore, if practitioners working with patients with BPD show a reluctance to convey the diagnosis or fail to give them enough clear information about the condition (Stalker 2005), this can elicit powerful negative emotions in patients and reinforce their perception of being stigmatised and rejected (particularly by professionals and services).

The experience of hopelessness this combination of internal and external invalidations can generate has one very clear remedy: a well-delivered diagnosis.

Having an open, information-based dialogue during the early contacts about the condition brings objective clarity to an area otherwise fertile for confusion and frustration. Towards this aim, we suggest considering the model of BPD as a medical condition and covering discussion points as shown in the patient information leaflet provided as Supplementary material 1, available at https://doi.org/10.1192/bja.2024.20. Reiterating the ‘medical reality’ of BPD validates the experienced distress and also allows a large volume of sought information and answers to be shared.

The importance of how the diagnosis is delivered cannot be emphasised enough. It appears that adverse emotions initially elicited in the patient by a poorly delivered diagnosis (hopelessness, shame and inadequacy) are maintained through subsequent clinical encounters when living with the now-named condition (Rüsch 2006). However inaccurate their doing so may be, patients do appear to define themselves through the diagnosis, and a tangible perception of the diagnosis as an identity is generated (Lester 2020).

However, the opposite outcome is just as possible – when time and care are taken to explain the diagnosis in the context of the person's life, it is not only received more positively, but can actually come as a relief. It can provide patients with a clearer understanding of themselves, helping them make sense of their difficulties, connecting them with other people and becoming a part of their collective owned identity (Fallon 2003; Horn 2007). After receiving a well delivered diagnosis, one patient said:

‘I had something that I could firmly grasp, and, you know, I could find out about and try and resolve it [ … ] The more I found out about it, the more I fitted, and then I want[ed] to tell my mum how I'm feeling’ (Horn 2007).

Systematic case-management: managing comorbidity and sustaining cohesion

Case management is a process of integration: coordinating the involvement and collaboration of diverse services and interested parties to give a patient the best chance of optimised health and life outcomes. This task is just not possible in BPD without psychiatric grounding. Two major reasons are that (a) the actual focus of ‘management’ in BPD is obscured by the considerable and labile comorbidity burden and (b) the disorder tends to generate acute help-seeking in distinct parts of the care team, which can then cause team splitting and a departure from the overarching treatment goal (e.g. physical health symptoms may seize priority).

Sound psychiatric (medical) management reveals and sustains focus on the problem at hand and spells out the task that needs doing:

• • If the patient has a personality disorder, is it being managed consistently?

• • If the patient has another mental disorder, has that been identified and managed?

• • If there is no active psychiatric disorder, is there some alternative (non-mental health) resource that can be collaborated with to support stabilisation and improve the patient's quality of life?

Expect and systematically manage comorbidities

Psychiatric comorbidity is near universal in BPD and the relationship is skewed – Axis II personality disturbances worsen the course and prognosis of a comorbidity (Goddard 2015). However, research indicates that personality disorders diagnosed in the presence of another mental disorder, specifically major depressive disorder (MDD), have a clinical course and outcomes very similar to those of personality disorders diagnosed in the absence of MDD (Beatson 2013). Given the imbalanced interaction, identifying the appropriate focus of treatment (Table 1) early on greatly increases chances of treatment being successful. In the most pragmatic terms (Gunderson 2014):

• • the comorbidity becomes the primary target if it interferes with active social and cognitive learning or engagement (e.g. substance use disorder, mania, PTSD) or is characterised by low motivation as a cardinal feature (e.g. antisocial personality disorder, anorexia nervosa)

• • BPD becomes the primary treatment target when the comorbidity is either likely to relapse or unlikely to remit unless BPD is brought to remission (e.g. MDD, panic disorder, remitted bipolar disorder, bulimia nervosa).

TABLE 1 Determining the appropriate focus of treatment in borderline personality disorder (BPD) comorbid with other psychiatric disorders

PTSD, post-traumatic stress disorder.

Adapted from Gunderson (2014).

Maintain care-team cohesion and minimise unnecessary referrals

A standard design can be applied to understanding the referrals generated by a fragmented team when caring for someone with BPD: that a decompensating system will ‘transmit’ onwards pressure that it is unable to ‘contain’. This will take the form of symptom-driven referrals, investigations and interventions.

While a patient's mental health symptoms are relatively well contained by mental health teams and relatively well identified by physical health teams, the same does not appear to hold true for the physical symptoms which are often a part of the BPD symptom cluster (Box 2). The psychiatrist is uniquely placed to assess, contextualise and interpret new or inconsistent (especially medical) symptoms and buffer their impact on team cohesion. By communicating proactively with other teams involved in the patient's care (Box 2) the psychiatrist can also bring a similar cohesion to the overall care and recovery process.

BOX 2 Somatisation and its management in borderline personality disorder (BPD)^a

Somatisation and physical healthcare utilisation

• • Compared with people without borderline personality symptomatology, those with such symptoms consistently show higher rates of healthcare utilisation in the primary care setting. These include a greater number of surgery visits and documented prescriptions,^1,2 more contacts with the treatment facility (e.g. telephone calls)¹ and more frequent referrals to specialists.¹

• • A prevalence of somatisation disorder as high as 36% has been detected in a sample of people with BPD³ and somatisation has even been conceptualised as a dimension of personality disorders.⁴

• • Moderate-to-high correlations have consistently emerged between somatic preoccupation and BPD symptomatology,^5,6 with no individual symptom or pattern emerging as predominant in people with BPD features.⁷

• • People with BPD in remission are shown to have far fewer visits to the emergency room and use significantly fewer analgesics and hypnotics than people with BPD not in remission.⁸

Supporting physical health teams in the management of people with BPD

With the patient's consent, establishing consulting communications with the involved physical health teams is an invaluable case-management input. Contrary to it feeling an encroachment, general practitioners (GPs) already perceive an ongoing lack of support from specialist mental health services when navigating complexities of the patient–doctor relationship associated with this patient group.⁹ The patient's profile of medical service utilisation can benefit greatly from positive shaping using consulting advice informed by psychiatric expertise:¹⁰

• • stress to the treating team the importance of communicating that their intervention will be one of ‘managing’ rather than ‘treating’ a condition and that, although symptom resolution is unlikely, symptom management will be a realistic and strived for treatment goal

• • advise that the team maintain an emotionally neutral treatment environment (self-monitoring one's responses to the patient, avoiding direct expression of anger) while remaining supportive of the patient

• • advise scheduling multiple brief appointments to address the needs of patients who struggle with strong attachment dynamics

• • encourage preventing the patient from unnecessarily entering high-risk medical situations, including unnecessary prescriptions, investigations and referrals.

Such a shaping of the treatment relationship, along with as much consistency as possible in and on the part of the treatment provider (e.g. the patient sees the same physician on serial reviews) establishes the clinic as the centre of the treatment process, and should discourage the use of emergency services except in genuine emergencies.

a. The numbered sources cited in this box are listed in the Supplementary material.

Simplify and standardise management of psychiatric medication

A central message communicated through all BPD treatment guidelines is to prioritise psychological treatments and accord medications (if any) a targeted supporting role (American Psychiatric Association 2001). Certain guidelines actively advise against the use of medical treatments for BPD (National Collaborating Centre for Mental Health 2009; National Health and Medical Research Council 2012). This position has likely been reached using the clear benefits that have emerged from studies that utilised psychological therapies as treatment interventions (Storebø 2020) and the lack of benefit shown by those that utilised medication. However, as identified in many studies assessed in a Cochrane Collaboration review on psychotherapies for BPD (Storebø 2020), even if referred for psychological therapy, many patients continued to receive medication. This suggests that evidence of the efficacy of psychological therapies is not necessarily derived in isolation from pharmacological treatment. Both patients and practitioners also endorse real-world symptom improvement and stabilisation with medication use (Vita 2011).

Given this dichotomy, decisions on the use of medication are best made collaboratively with each individual patient, incorporating their unique clinical indications. The following general principles of medication use can generate a more rational and less harmful utilisation of medication.

Establish and agree to a stance on medication with the patient

We support a departure from the rigid positions that concrete interpretations of guidelines can generate – i.e. asserting that medication will/will not be helpful/harmful. The project of collaboratively working towards recovery would be much better supported by holding an attitude of measured hope, founded on uncertainty, caution and participation, and encouraging thought about the role and impact of medication:

‘We could try this medication, but should recognise that whether it will help is not certain. I'll need you to help me assess its effectiveness. It would help if you could read as much as you can on this medication, and then check if it's helping you with the symptoms it's intended to affect. Will you do this?’ (adapted from Gunderson 2014).

In keeping with patients’ expressed need for information (Bonnington 2014), provide the patient with up-to-date knowledge on the status of medication treatment in BPD (an example we use is provided as Supplementary material 2).

Emphasise the value of patient collaboration in medication management and review

The parameters to be discussed could include:

• • identifying treatment goals for medication – for example a reduction in the intensity of meltdowns or intrusiveness of voices when paranoid ideation escalates

• • preparing the patient for potential side-effects and involving them in the evaluation of side-effects

• • raising patient awareness about the effect certain psychiatric medications (and combinations) can have on their health and metabolism (and, by extension, on their life expectancy)

• • reinforcing the need to hold realistic expectations of medication – for example particularly that the chronic emptiness and abandonment sensitivity of BPD will not respond solely to medical treatment.

Agree on and attempt to contract a policy for medication use

• • Regular prescribed medication – try to reach agreement that if the patient fails to respond to a medication, you will taper it off and only then begin another medication (unless the patient is severely distressed, in which case you should cross-taper). Doing so reduces opportunities for medication in general and polypharmacy in particular to receive undeserved credit for improvement (given how crisis presentations are in themselves short-lived). As part of a treatment policy, resist in most instances augmentation of one medication with another. Try to adhere consistently to your policy to take away a medication before changing to another (Gunderson 2014).

• • Crisis medication – agree on a short course of medications that will be resorted to in crisis, identifying what the thresholds for such medication use will be. The case can be made to refrain from quick medication changes or additions in times of crisis, because most often crises pass quite quickly and the medication may again falsely be given credit for improvement. This will depend on the stage of involvement, as early on in a therapeutic relationship, medication serves to ‘hold’ the anxiety and distress of a crisis until a skill set of non-medical management is devised. Begin work on setting up these non-medical skill strategies early and collaboratively.

Be systematic in choosing of medications: guidance does exist

The ongoing dearth of first-line psychological resources for BPD is unlikely to resolve in the near future. In such a landscape the need for rational medication-based treatment roadmaps is evident. The first practice guideline to provide one for BPD was developed in 2001 by the American Psychiatric Association (American Psychiatric Association 2001). It was based on only seven trials and as it was developed prior to most of the now available body of research – the document today does feel dated. However, a pivotal feature was its introduction of symptom domains and advice to consider differing approaches to the management of each of these domains. The identified domains were:

• • affective dysregulation symptoms – mood lability, rejection sensitivity, inappropriate intense anger, depressive ‘mood crashes’, outbursts of temper

• • impulsive behavioural/dyscontrol symptoms – impulsive aggression, self-mutilation, self-damaging binge behaviour (e.g. promiscuous sex, reckless spending)

• • cognitive perceptual symptoms – suspiciousness, referential thinking, paranoid ideation, illusions, derealisation, depersonalisation, hallucination-like symptoms.

The symptom domains concept has held good (Box 3) and has since been further developed. For example, Ingenhoven et al (2010) further dissected the symptom constellation and described specific medication classes and their respective effect sizes. The analyses included placebo-controlled trials involving people with BPD and/or schizotypal personality disorder. A summary is provided in Table 2.

BOX 3 Symptom domains as the focus of treatment in borderline personality disorder (BPD)

• • A Cochrane review¹¹ on the pharmacotherapy of BPD concluded that, although overall illness severity did not appear an actionable treatment focus for medication, sufficient effect had been demonstrated to consider the targeted treatment of specific symptoms.

• • Differences between drugs were not statistically significant and effect sizes remained at best modest, but all studied symptom domains of BPD showed improvement with psychopharmacological treatment.

• • The symptom domains and beneficial medications operationalised in the review were:

  • • affective dysregulation: haloperidol, aripiprazole, olanzapine, lamotrigine, divalproex and topiramate

  • • behavioural dyscontrol/impulsivity: aripiprazole, topiramate and lamotrigine

  • • cognitive/perceptual symptoms: aripiprazole.

• • Selective serotonin reuptake inhibitors (SSRIs) did not improve any symptom domain.

• • Core BPD symptoms such as avoidance of abandonment, chronic feelings of emptiness, identity disturbance and dissociation were unaffected by any medication.

• • Indicating a widening acceptance of the construct, symptom domains and calibrated pharmacotherapy have been incorporated in treatment recommendations issued by:

  • • the Swiss Association for Psychiatry and Psychotherapy¹²

  • • Finnish Current Care Guidelines for BPD¹³ developed by the Finnish Medical Society Duodecim, in cooperation with the Finnish Psychiatric Association

  • • practice guidelines issued by the Dutch Psychiatric Multidisciplinary Guideline Committee.¹⁴

^a. The numbered sources cited in this box are listed in the Supplementary material.

TABLE 2 Borderline personality disorder (BPD) symptom domains, medication effect sizes and dosages

Data collated from Ingenhoven et al (2010).

Adhere to rational prescribing

Avoid (or rationalise) polypharmacy

However one defines the practice, polypharmacy is widespread in BPD management. There are numerous explanations, including a pressure to focus on medical treatment (given the shortage of psychological resources), an increasing number of psychiatric medication options, and the tangible push for remission (as opposed to response) as the acceptable treatment outcome (Hoffman 2011).

The building evidence base indicates the following.

• • Combining antidepressants in any treatment context has not been shown to increase mood symptom remission (Rush 2011). Even the well-accepted olanzapine–fluoxetine combination has shown little additional benefit when used in BPD (Zanarini 2004b). To date, an augmentation of valproate with omega-3 fatty acids is the only pharmacotherapy combination that has demonstrated some advantage in terms of decreasing BPD symptom severity and self-harm, and for the secondary outcomes of affective instability and impulsivity (Bellino 2014).

• • Large and adequately powered studies demonstrate that while polypharmacy has questionable benefits in any mental health condition, it generates a clear worsening of side-effect burden (Rush 2011).

• • Patient health status is adversely affected by polypharmacy – a study of 659 patients with depression found that their rates of cardiovascular problems increased from 8.8 to 30.7% after only 6 weeks of polypharmacy (Lötscher 2010).

These findings would make the practice of polypharmacy a complete mismatch for the BPD patient population, in which:

• • chronic feelings of emptiness and abandonment are core symptoms that remain unresponsive to conventional pharmacological treatment (Stoffers 2010)

• • exaggerated sensitivity to side-effects is well recognised

• • all-cause mortality is already four times higher than in a comparative general population, and the increased risk of mortality (from natural and disease causes) results in 17.7 and 18.7 years of life lost in men and women respectively (Fok 2012)

• • symptom remission does not directly translate into improved social functioning (Gunderson 2011).

Given this mismatch, it is disconcerting to see the rampant prevalence of what has been called ‘irrational polypharmacy’ (haphazard combinations of medications with no distinguishable clinical/pharmacological rationale) (Hoffman 2011) in BPD treatment. In one in-patient sample of people with BPD, 56% were taking ≥3 and 30% were taking ≥4 psychiatric drugs (Pascual 2010); in another sample 6.9% reported taking ≥5 psychotropic medications (Zanarini 2015).

Use medication washouts as a pharmacological instrument

When evaluating for the first time a patient with BPD receiving polypharmacy, even after a detailed history and diagnostic review, it can be a challenge to distinguish between (a) incomplete medication effects and adverse effects and (b) the combination of active Axis I symptoms and more pervasive Axis II symptoms unique to BPD.

Good clinical care further invites ascertaining whether a given patient even needs medication anymore after a period of prolonged use. As evidence builds on the stabilising role of the ageing process in BPD (Oldham 2013), it seems appropriate to evaluate whether medications remain necessary if such an adequate psychological and neurophysiological maturity has set in.

Medication washout is an underutilised tool for such scenarios. It can give a clinician unfamiliar with the patient a way to see the person's baseline clinical state and to assess whether it has already shifted. Although not suitable for every patient, if appropriate and done correctly (refer to Supplementary material 3), washout can contribute to a mutual confidence in both patient and prescriber, helping both gain trust in the relationship, and provide the patient with a more realistic understanding of the role of their medications (Hoffman 2011). Washout is accomplished differently by different practitioners, but tapering medication and allowing 5–7 half-lives for clearance is generally accepted.

Managing suicidality and non-suicidal self-harm

Suicidality is a cardinal symptom of BPD and suicide is a tragically real possibility. The clinician needs to accept that some (a very few) patients die by suicide.

The risk profile of BPD (Fig. 2) is qualitatively and temporally distinct from that of other psychiatric disorders in that:

• • the baseline risk level in BPD remains chronically and consistently elevated (above that of the general population)

• • the baseline risk level elevates acutely (but transiently) at times of crisis: there is no doubting the catastrophic potential of these time-limited escalations, as their emotional intensity can overwhelm the patient's weighing-up of pros and cons in decisions regarding their own safety

• • usually, apart from a very brief increase of risk-containment actions (e.g. considering crisis team input or a brief crisis admission), very little change of regular treatment inputs is required to bring risk levels back to their (chronically elevated) baseline. Each crisis does, of course, become an opportunity for learning.

FIG 2 The risk profile of people with borderline personality disorder (BPD) compared with profiles for people with serious mental illness and the general population. In BPD there are very intense, but usually short-lived and self-correcting, risk escalations. Often, little input save safety-netting is required to manage the risk state. This is distinct from risk in other serious psychotic/affective disorders, where often a complete revision of treatment is required.

Principles of risk management

When assessing risk, clinically assess dangerousness

When treating any psychiatric disorder, clinicians should respond to (even indirect) communications of self-endangering behaviours with concern and curiosity. When working with patients with BPD, this response invites an additional assessment of the specific dangerousness of the situation presented. The central question that needs answering is whether this is a distinct acute-on-chronic risk presentation. Some markers suggesting such a distinct escalation of dangerousness include:

• • any recent negative life events that may have introduced a destabilisation in routine, lifestyle or environment

• • any objective or perceivable abandonments or rejections that have transpired or are now anticipated

• • any acute changes in mental state (onset of depression, exacerbation of anxiety) or change in substance use status (relapse or cold-turkey abstinence).

Their impact is gauged against the patient's current self-regulation skills and the support resources that can be called on (Fig. 3).

FIG 3 Factors that affect suicide risk levels in borderline personality disorder (BPD). In people with BPD, the acute-on-chronic level of suicide risk (two-headed arrows) can change more quickly than in the general population and will be modified by several factors that can cause (upwards arrow) and several that might reduce (downwards arrow) an acute exacerbation of risk.

To make the treatment experience more consistent and therapeutic, the clinician and patient should at some stage collaborate on a comprehensive ‘risk profile’ of the patient. This is a more detailed and bespoke work, beyond the ritualised and often overpowered ‘risk assessment’ (self/others’/own health and mild/moderate/severe). Such a profile should additionally explore:

• • the distinction between non-lethal self-harm behaviour (verbalised or carried out) and ‘true’ suicidal intention

• • what a patient's personal set of usual self-harm acts are, and the need to investigate whether any changes to this profile emerge (Fig. 4)

• • the threshold risk level beyond which changes to intervention are required.

FIG 4 A decision matrix to estimate the level of risk and guide choosing the management setting for an individual with borderline personality disorder who is self-harming. SHB, self-harming behaviour. Adapted from Rao S et al (2017).

Working on this profile may be a difficult conversation to have, and is best done when the patient is in a more stable mental state. But collecting this information can reduce the detrimental outcomes that routine risk assessments often generate. The better-risk-informed team can assess the need for hospital admissions with more confidence, limit the patient's secondary gains from self-harm and then proactively shape the person's crisis communication. In time, suicidality can cease to be the only ‘call for help’ used to seek support.

Invite the patient to participate in devising a risk management strategy

The invitation:

• • helps underscore the patient's need to be active on their own behalf (reinforcing the sense of agency and the need to think) and

• • discourages the use of suicidal or self-harm threats as a way to escape, avoid or externalise responsibility.

National Institute for Health and Care Excellence (NICE) guidance endorses that a crisis plan should, as a minimum, identify potential triggers that could lead to a crisis, specify self-management strategies likely to be effective and establish how the individual can access services when self-management strategies alone are not enough (National Collaborating Centre for Mental Health 2009). To make a comprehensive crisis plan, involve the patient in as detailed an exploration and review of past crises as possible. We provide an example as Supplementary material 4, drawn from the work of Borschmann et al (2013).

It must be said that the best crisis plan is quite useless if an out-of-hours team does not even know it exists or cannot find it. Placing a statement like ‘The most recent crisis plan dated 22/12/22 is located in the Correspondence section of EPR [electronic patient record]’ on the profile launching page can be invaluable.

It is very possible that the patient may not participate in crisis planning, but the clinician can then recognise (and would do well to record) that the ability to help a patient is compromised by their non-participation (Gunderson 2014).

Involve colleagues in collective risk management

Given the high rate of suicidality and completed suicides in this patient population, for a more authentic evaluation it is necessary to hold in mind the impact of suicide on the following.

• • The people involved in the life of the patient – suicide can have severe emotional impacts on family members and friends, and mental health team members.

• • The professional careers of practitioners – any suicide will invite some form of inquiry, potentially attendance at a coroner's court and, if liability is attributed, culminate in sanction or even termination of registration.

• • The functioning and nature of practice within the mental health team – the suicide of a patient changes the way a team works. Risk assessment and management thresholds shift, the event generates overcautious decision-making, a reluctance to work with suicidal patients, compulsive urges to help/save the patient, an increased volume of clinical notes and increased frequencies of contacts, lower thresholds to admit to hospital, more conservative patient selection, an increased distance between practitioners and patients and an increased desire to change jobs (Sandford 2021).

• • Clinicians who work with such patients must ensure that they recognise and pre-emptively respond to the very real liability potential a suicide carries. We stress doing this as leaving the issue unaddressed does direct damage to the clinical confidence and case ownership within a team, but if effectively done can further insulate the team from ‘splitting’. Such pre-emptive actions can involve:

  • ◦ helping the patient's significant others understand that suicide is a very possible outcome when trying to manage the disorder (an evidence-based information leaflet for discussion has been provided as Supplementary material 5)

  • ◦ consulting with another psychiatrist to confirm that other treatment options would not be safer

  • ◦ ensuring that discussions are had and recorded within the wider multidisciplinary team to share the risk burden and liability potential.

• • It must be added though that although the risk of liability in BPD is higher than for other patient groups, it still remains low (<1%) and becomes negligible among experienced clinicians. The source of this liability is mostly countertransference enactments – excessive availability, personal overinvolvement, holding a punitive or hostile attitude and working with impressions of omnipotence or omniscience (Gunderson 2014).

Remain pragmatic and recognise your own limits in risk management

During the course of longitudinal case management

Remind yourself and the patient that you are not an omniscient or clairvoyant being – you are a clinician trying to encourage the extinction of well-established patterns of risky behaviour. It is also unrealistic for a patient's safety to depend on your availability – it is always advisable to prepare patients for any absences but to do so as an ongoing process of building their skills and stabilisation resources.

Although it is uplifting (and, arguably, demonstrates an ability to generate trust) too much is often made of ‘uncovering’ or being ‘the first one to be told’ of a risk. Be clear with patients that you cannot be expected to know about their suicidality or self-harm unless they tell you. Furthermore, even when you know about it, you may not be able to help the patient prevent it, and when such behaviours have already occurred, you may not be able to repair the damaging effects.

Such openness may generate the question ‘So then what use are you?’, at which point you can only apologise for your limits, while assuring the patient you will do everything realistically possible to support them in making their life more stable and their riskier situations safer (Gunderson 2014). Then do so. Deflating the ‘aura’ of there needing to be any unique or special ‘bond’ or ‘secret’ to the management allows discussions to take place on a more down-to-earth and clinical basis.

During management of an acute crisis presentation

Clinicians making decisions in acute presentations should be vigilant of how their decision-making can be distorted by loudly voiced opinions that in-patient treatment is detrimental for BPD, or by attending staff's frustration from caring for these patients. It can feel like one is showing leadership by completely ruling out admission as an option for BPD instead of conducting a more objective and thought-through decision. Such actions may feel stirring and even invite immediate praise from the same frustrated colleagues. Giving in to such hubris, though, can have disastrous consequences.

Each encounter must be risk assessed on its own merit, and if a clinician feels the profile has become unsafe, using short-term containment/stabilisation admissions for risk management is completely sensible. It is now largely accepted that extended in-patient admissions have little therapeutic value. They are usually problematic, as suicide precautions used in hospital settings reinforce the very behaviour being treated. However, it is equally clear that enforcing an absolute (personal or systems) embargo on brief admissions for risk management is just as inappropriate (Box 4).

BOX 4 Case vignette: the patient asking to be admitted for ‘suicidality’

It is 03.00 h. You have been called to the hospital emergency department by the psychiatric nurses on duty to see a patient. She is a 23-year-old woman with documented diagnoses of borderline personality disorder (BPD), attention-deficit hyperactivity disorder and bipolar disorder. She is on the ‘frequent attenders’ list at the hospital, and has been discharged by her community mental health team for ‘non-engagement’. There have been numerous in-patient admissions in the past for ‘suicidality’, her in-patient admissions have been protracted and chaotic, and her notes suggest to ‘avoid admission wherever possible’. There is no crisis plan on record. The nursing team say that although she appears relatively calm, she refuses to offer any assurances of safety and is demanding to be admitted as she is ‘suicidal’.

On your review, although the patient is irritable and brittle, she does not look emotionally overwhelmed, only repeatedly saying ‘I will kill myself if I leave here’. Any attempt to further explore is turned down.

How will you manage the situation?

This fictional account speaks to an often encountered out-of-hours scenario and reinforces the value of pragmatism in managing people with BPD. In this case, given the patient's potential for escalation and refusal to discuss alternatives, plus the genuine risks of liability, it would feel your hands are tied.

Similar to an approach offered by Gunderson (2014), we would advise a clear line of communication with the patient:

‘I am willing to admit you despite my concerns that it will be unhelpful. I will do this because I fear that you will display even more suicidal behaviour if I don't. Am I right about this? It would help us all if we could understand your situation and find a better alternative.’

Such a submission serves several purposes:

• • it takes the magic out of the ‘silver bullet’ of an admission and demonstrates the lack of importance you personally place on it

• • it calls on your patient to discuss why they want an admission, giving you more material with which to understand and solve the problem at hand

• • it disarms anyone who would otherwise be critical of your lack of decisiveness.

These kinds of patient contact are good opportunities to communicate (if the patient can tolerate the discussion) defining facts about BPD and the approach to its management:

• • that a patient in crisis tends to feel better in hospital because it offers structure and perceived safety

• • that life cannot be lived in a hospital

• • that the aim of treatment and recovery is to help build a skill set and a life to provide the same semblance of control and structure

• • that this process of recovery and skills building can only genuinely happen in the community

• • that hospitals certainly have a role in BPD management, but that role is to serve as a safe place to contain acute and short-lived risk escalations: there is ample evidence to indicate that such admissions are best kept short

• • that if, after a patient is discharged, a crisis re-emerges and reaches a similar degree of dangerousness, there is nothing stopping a reconsideration for admission.

The ‘loop’ of intervention that such a contact generates can be completed by communication to the patient's care team about the discussion that took place. Any knowledge shifts or skill gains can then be further incorporated into the patient's crisis plan, and the progress towards stabilisation is added to.

Conclusion

Managing BPD within the existing mental health framework can be confusing (and hence challenging), given unclear and often conflicting guidance and widespread subjective practice. As the knowledge base on the disorder builds, management roadmaps will become clearer. At present, utilisation of standard psychiatric management principles (sound diagnostics, identification and treatment of comorbidity, rational pharmacotherapy and risk management) still not only apply well, but can contribute to observable stabilisation and improvements.