Borderline personality disorder (BPD) is frequently misdiagnosed. For instance, mood symptoms in BPD may be misattributed to a primary mood illness such as major depression or bipolar disorder when they actually result from the core borderline pathology. ^{Reference Gunderson, Morey and Stout1} Similarly, anxiety symptoms in the context of interpersonal relationships may be attributed to generalised anxiety disorder when they are actually secondary to BPD. ^{Reference Paris2} The misdiagnosis of BPD is clinically consequential insofar as it may lead to the use of psychotropics in a group of patients who generally do not respond to them, while also contributing to inaccurate or unhelpful psychotherapeutic interventions. Most experienced clinicians are familiar with patients who have undergone numerous ineffective medication trials before receiving an accurate diagnosis of BPD.

This paper addresses a particular error in diagnostic reasoning that I believe has been neglected in the existing literature on BPD and probably contributes to the misdiagnosis of these patients. Specifically, it involves the undervaluation of borderline dynamics in the absence of an intense treatment relationship. In other words, the therapist does not perceive the patient as presenting like other borderline patients in psychotherapy because the characteristic dynamics of dependency, neediness, interpersonal control and fears of abandonment are not strongly enacted within the transference. The patient, for instance, does not threaten termination following minor frustrations, contact the therapist frequently in between appointments or react with overt panic or rage to ordinary disruptions in the therapeutic frame. As a result, the patient fails to conform to the therapist’s prototype of the ‘difficult borderline patient’.

The most common explanation for this, as described originally by Gunderson, ^{Reference Gunderson3} is that the therapist does not occupy the position of ‘primary object’ in the patient’s life. A primary object is the person upon whom the patient is most emotionally dependent and whose availability, responsiveness or perceived rejection most strongly mediates the patient’s affective stability and behaviour. Because borderline symptoms tend to emerge most intensely in relation to the primary object, therapists may encounter these patients in a comparatively organised state when the central attachment relationship exists outside the treatment, such as within a romantic or family relationship. Under such circumstances, the absence of intense transference may lead the clinician to underestimate the severity, or even the presence, of the underlying borderline disorder.

This aspect of BPD is incompletely captured by some theoretical models of the condition. For instance, dialectical behavioural therapy conceptualises many borderline symptoms primarily in terms of deficits in emotion regulation and behavioural control, potentially under-emphasising the extent to which these symptoms are mediated by the patient’s relationship to a primary attachment figure. Even some psychodynamic models, such as Kernberg’s transference-focused psychotherapy, underestimate the degree to which symptoms such as anger are reactive to the realistic aspects of interpersonal events. ^{Reference Gunderson3} By contrast, Gunderson’s interpersonal hypersensitivity model ^{Reference Gunderson3,Reference Gunderson and Lyons-Ruth4} places the patient’s heightened sensitivity to attachment disruption and perceived abandonment at the centre of borderline psychopathology.

According to Gunderson’s object relations–ego psychology framework, ‘the borderline person’s manifest psychopathology can be understood in terms of relationships with major objects’. ^{Reference JG.5} Indeed, it is ‘these characteristic disturbances in interpersonal relations [that] uniformly provide the most distinguishing aspect of the borderline syndrome vis-à-vis a variety of other diagnoses’. ^{Reference JG.6} When the primary object is present and supportive, the depressive, lonely and masochistic features of BPD predominate. When this other person is experienced as frustrating, the angry, devaluative and manipulative features become prominent. And, when the major object is absent, the most severe symptoms appear: panic, impulsivity and transient psychosis. These ideas later became the basis for Gunderson’s general psychiatric management approach to treating BPD.

I largely agree with this model and would add only that borderline symptoms are marked also by their paradoxical quality: when the major object is experienced as being too close, the patient acts in ways to create distance; when the object is experienced as being too distant, the patient attempts to draw him or her near. ^{Reference Ruffalo and Kreisman7} This leads to a self-perpetuating cycle in which the patient’s efforts to preserve psychological equilibrium simultaneously undermine the very relationship upon which that equilibrium depends, creating the characteristic instability, contradiction and ‘push–pull’ interpersonal pattern so central to BPD. Adler ^{Reference Adler8} has linked this to the borderline patient’s basic need–fear dilemma: a simultaneous need for, but fear of, closeness with other persons.

One implication of this formulation is that BPD cannot always be inferred reliably from the patient’s presentation in a single relationship, including psychotherapy. Some patients appear relatively organised, reflective and sufficiently integrated providing that the primary object remains someone other than the therapist. In these cases, the clinician may conclude prematurely that the patient’s difficulties are better explained by other psychiatric pathology, or that the patient is organised at a neurotic or healthy level of personality functioning. Nevertheless, a more comprehensive assessment often reveals a characteristic pattern of instability centred on emotionally significant relationships: repeated crises involving abandonment fears, oscillations between idealisation and devaluation, frantic efforts to prevent object loss, paradoxical approaches to intimacy and severe emotional dysregulation following perceived rejection or separation. The absence of intense borderline enactments within the treatment relationship should not therefore be taken as evidence against the diagnosis. Rather, the clinician must ask where the primary object relationship is located and how the patient functions within that relationship.

It is thus incorrect to conclude, as some earlier psychoanalytic writers have suggested, that BPD can be reliably diagnosed on the basis of the transference and countertransference, or that these patients uniformly evoke countertransference difficulties in therapists. Although the therapeutic relationship may eventually become the principal arena in which borderline dynamics are enacted, this is not invariably the case, particularly early in treatment or when another person occupies the position of major object. Clinicians who rely too heavily on the immediate emotional intensity of the treatment relationship may therefore fail to recognise borderline pathology in patients whose most severe disturbances emerge primarily outside of psychotherapy. There remains a chance that, even in prolonged treatments, the patient’s outside relationships are so affectively dominant that the therapist never fully assumes the role of primary object within the patient’s internal world. In such cases, the most characteristic borderline dynamics may continue to emerge principally in external relationships rather than within the transference itself.

It is the experience of many senior clinicians that supervisees not infrequently present patients for consultation only to discover, often belatedly, that the patient with whom they are struggling suffers from BPD. Therapists may exclaim, ‘But the patient didn’t seem borderline!’, only to realise later that the absence of intense borderline phenomena within the treatment relationship reflected not the absence of BPD itself, but rather the fact that the patient’s central conflicts were being enacted elsewhere. Such oversights are unfortunately facilitated by theoretical models of BPD that insufficiently appreciate the extent to which borderline psychopathology is organised around interpersonal relationships and mediated by the patient’s experience of the major object.

In this respect, Gunderson’s interpersonal model remains indispensable because it locates the core of BPD in the patient’s disturbed relationship to the major object, whether inside or outside of psychotherapy.