Group Schema Therapy Equals Group CBT for Social Anxiety with Avoidant PD — and Patient Profile Doesn't Tell You Which to Pick

For the AVPD–SAD overlap — one of the most disabling and least researched comorbidities in clinical practice — the field has had two competing offers: classic exposure-rich CBT, and Young's group schema therapy with chair work and limited reparenting. Clinicians have been picking based on intuition. This Dutch RCT (PsyQ The Hague + Leiden University + University of Amsterdam, Baljé and Greeven group) tested whether intuition tracks anything real. It doesn't.

What the data actually shows

Both arms improved, the curves were superimposable at 3 months, and they stayed superimposable at 12 months. The authors went further than a simple null result: they pre-specified moderators known to plausibly favour schema therapy (severe schemas, high emotional dysregulation, history of childhood neglect, low self-esteem) and ran causal-style subgroup analyses — including Cox regression for differential attrition. Nothing.

Then the mediator question: does GST work through schema-specific mechanisms? The cross-lagged panel design separates within-person from between-person effects. Average mode score and the avoidant-protector mode at mid-treatment predicted later social anxiety — but identically in both conditions. Self-esteem played the same role on both sides. The "mechanism" GST claims to target is improving in CBT too. Either both treatments are reaching it through different routes, or our schema-mode measurement is downstream of something more general (functional engagement, fear reduction, reduced behavioural avoidance) that both therapies hit.

This converges with the wider differential-effectiveness literature. Assmann et al. (10.1016/j.brat.2025.104899) — same month, also from this Dutch network — found a small subgroup with high baseline functioning and the "failure to achieve" schema who did better with DBT than schema therapy in BPD, but that effect washed out at follow-up. The pattern across studies: short-term moderators may exist for some specific characteristics, but personality-disorder treatments mostly converge over a year.

For your practice

Three concrete things change.

First, stop using schema severity, emotion-regulation deficits, or trauma history as the reason to push an AVPD patient towards group schema therapy specifically. The data doesn't support that decision rule. Pick based on accessibility, fit with the patient's preferred language ("learn skills" vs "understand my parts"), therapist competence, and group composition — not pseudo-precision matching.

Second, when GST works for AVPD, it is not working primarily through dismantling Early Maladaptive Schemas in some unique way. It is working through the avoidant-protector mode loosening, mode-average dropping, and self-esteem rising — outcomes you can also reach in CBT via behavioural experiments and cognitive restructuring. Track those processes across modalities rather than fetishising the model.

Third, for the Russian-speaking clinical context, this matters: ISST-certified schema therapists in Russia number in the low single digits. If a patient with SAD+AVPD asks whether they need to wait six months and pay €120/session to find a certified schema therapist abroad, the honest answer is no. A competent group CBT programme — which is far more available — will get them to the same place at 12 months. Save the schema referral for cases where CBT has already failed or where the patient needs a relational frame to engage at all.