A.J. Spong, I.C.H. Clare, J. Galante, M.J. Crawford, P.B. Jones
9 min
This systematic review and meta-analysis aimed to evaluate the effectiveness of brief psychological interventions—defined as lasting six months or less—for adults with Borderline Personality Disorder (BPD). Given the limited availability of long-term therapies and high dropout rates, the authors investigated whether shorter, manualized treatments could effectively improve symptoms, mood, self-harm, and social functioning.
The researchers conducted a systematic review of 27 randomized controlled trials (RCTs) published up to April 2020. They performed random-effects meta-analyses, sub-grouping data by delivery method (individual, group, or mixed), the presence of additional support (stand-alone vs. adjunctive), and the type of control condition (manualized vs. non-manualized).
The study found that brief interventions are generally more effective when provided as an adjunct to ongoing care rather than as stand-alone treatments. While interventions showed large effect sizes when compared to non-manualized controls (like treatment as usual), these effects were significantly smaller or negligible when compared to manualized, structured clinical management. This suggests that the benefit may stem from the consistency of regular, predictable contact with a professional rather than the specific content of the brief intervention itself. Furthermore, while group-only interventions showed some promise in reducing self-harm, the high level of heterogeneity across studies makes it difficult to draw definitive conclusions about the superiority of any specific delivery mode or intervention type.
This review challenges the clinical reluctance to offer brief psychological interventions for BPD, suggesting they can be a viable component of care. However, it highlights that these interventions should likely be integrated into a broader support system rather than used in isolation. The findings underscore the need for more rigorous, standardized research to determine if symptomatic improvements from brief treatments are sustained over time and to clarify the relative value of specialist interventions versus generalist, structured support.
Alex: That covers very different programmes. What makes pooling them useful rather than just averaging unlike things?
Sam: The useful move was to split the evidence by the care arrangements. They examined the comparator, the delivery format, and whether treatment stood alone or supplemented ongoing care. Their statistical model allowed for differences between studies, though it could not make those differences disappear.
Alex: Give me a concrete example of the comparator distinction. What did “structured contact” actually mean in these trials?
Sam: Some controls offered supportive groups or therapy not specialised for borderline personality disorder. These were manualised: delivered according to a planned protocol. Other controls were waiting lists or treatment as usual that was not delivered in a planned way.
Alex: So treatment as usual was not necessarily no care. But a specialist programme could still provide more predictable contact than that comparison offered.
Sam: That is the question the authors raise. Against non-manualised controls, brief interventions showed large effects across the pooled outcomes. Against planned controls, the effects were much smaller, especially for borderline symptoms and social functioning.
Alex: Does that establish that ordinary structured care is just as effective? Or is the evidence too thin for an equivalence claim?
Sam: It does not establish equivalence. The planned-control comparison for borderline symptoms rested on four trials. Those pooled results showed little difference between groups. The authors say generic planned support may be as effective, but call for further examination.
Alex: Then the practical question is whether specialist content adds something beyond dependable contact. The review does not settle that, but it changes what a convincing comparison should include.
Sam: Now consider additional support. Some programmes were adjunctive, meaning added to ongoing treatment, usually treatment as usual. Others were stand-alone. For borderline symptoms, stand-alone programmes showed little advantage over their controls.
Alex: Were outcomes consistently better when the brief intervention came with other care, or only for that symptom category?
Sam: Effects were generally larger with additional support, except for general mental health. But the difference between these subgroups reached statistical significance only for borderline symptoms. The review therefore supports taking surrounding care seriously, not claiming improvement across every outcome.
Alex: These are comparisons across trials, though. Could support be tangled up with the programme, the comparator, or the people delivering it?
Sam: That is the central limitation. The subgroup features overlapped, and there were too few studies to control for their effects on each other. So larger effects with support cannot cleanly tell us which ingredient produced them.
Alex: Did comparing named treatments help untangle that? Researchers often want to know which programme deserves another trial.
Sam: Not enough to rank them confidently. An emotion-regulation group programme, teaching ways to manage emotions, looked stronger than dialectical behaviour therapy skills training on some outcomes. But its trials involved the treatment developer and included ongoing individual-therapist support.
Alex: Whereas some of the skills-training trials delivered treatment alone. That makes a programme comparison partly a comparison of care packages.
Sam: The authors flag precisely that problem. Replication was also limited. Some apparently large effects came with very high heterogeneity, meaning substantial variation among study results. In those cases, the pooled average was not a dependable summary.
Alex: Beyond differences between studies, what threatens confidence within the trials themselves?
Sam: Attrition and selective reporting were the main risks. Nearly half the trials reported high dropout across treatment and control groups. Only a minority had published a protocol before starting, leaving room for favourable outcomes to be selectively reported.
Alex: And symptoms at treatment completion are only part of the clinical question. What happened after the sessions stopped?
Sam: Follow-up was limited, and the studies contributing follow-up results often differed from those contributing immediate results. That makes direct before-and-after comparisons unreliable. Where variation between studies was acceptable, most follow-up treatment effects disappeared.
Alex: What about self-harm and crisis services? Those could matter greatly even if symptom scores barely changed.
Sam: Self-harm measures differed in behaviours counted, time periods, and reporting formats, restricting pooling. Service-use measures were too inconsistent to pool at all. The review could not establish an overall effect on self-harm or service use.
Alex: So there are promising signals, not a comprehensive clinical verdict. Who should read the full document, and where should they begin?
Sam: Researchers planning trials and people designing services should start with “Impact of comparator” and “Impact of additional support.” Then read the limitations and follow-up discussion. Also check the population: seven studies included only women, and many others were predominantly women. This review did not compare brief treatments directly with longer therapies.
Alex: And what should everyone else remember when they hear that a brief programme worked?
Sam: Ask what it was compared with, and what other care came with it. Brief treatment may help, but the support around it is part of the evidence.
Alex: Keep the whole care package in view.