Routine Outcome Monitoring as an Alliance-Building Tool: Structured Feedback Discussions Produce Hedges' g = 1.33

Routine Outcome Monitoring (ROM) systems — regular psychometric assessment of patient progress and systematic feedback to therapists — were originally designed as early warning systems. The evidence base for ROM as a tool to prevent deterioration and catch non-responders is well established. What this study adds is different: structured feedback discussion as a specific intervention to actively improve therapeutic alliance, not just detect problems.

What the Study Did

Thirty-four outpatients in routine psychotherapy were enrolled in a pre-registered single-case experimental design with multiple baselines. Each patient had a baseline phase of varying length (3 multiple baselines stagger the intervention start), followed by the intervention phase: structured, therapist-patient discussion of the psychometric feedback data — not just handing the scores to the therapist, but explicitly discussing them together as part of the therapeutic work.

Alliance was measured across the first 20 sessions using a validated self-report measure. The single-case effects were aggregated using random-effects meta-analysis to estimate the group-level effect and its variability.

The effect size is striking: Hedges' g=1.33. In psychotherapy research, g above 1.0 is uncommon outside of highly controlled efficacy trials. Achieving this in routine care, with a brief structured discussion intervention, suggests that ROM feedback is not merely an administrative monitoring tool — it is a clinically active alliance-building procedure when used explicitly.

Why This Is a Causal Finding

Most ROM research is correlational: practices with ROM tend to have better outcomes. The mechanisms are inferred. This study establishes causality through design. The multiple baseline structure means each patient served as their own control, with the baseline phase ruling out natural improvement over time. Pre-registration rules out selective outcome reporting. The absence of significant moderator effects from baseline impairment or personality functioning means the effect is not driven by easy-to-treat patients — it generalizes across case complexity.

The implication: the structured discussion is the mechanism, not the measurement. Simply administering questionnaires and filing the scores does not produce g=1.33. Deliberately engaging the patient in a collaborative review of the psychometric data — "here is what the scales show about your experience of our relationship, let's discuss it" — appears to function as an alliance intervention in its own right.

How to Implement This in Practice

The intervention is low-tech and replicable. Components from the literature on effective ROM feedback discussion:

1. Select a brief alliance measure — OQ-45, WAI-Short Form, CORE-10, or the Session Rating Scale (SRS) — administered before or at the start of sessions
2. Review scores together — not just flagging deterioration signals to the therapist, but explicitly showing the patient their scores and discussing them
3. Ask about discrepancies — especially where patient perception differs from therapist expectation, or where scores change between sessions
4. Use scores to open metacommunication — "I notice the alliance score dropped after last session — what was happening for you?"
5. Treat the data as relational material — not administrative feedback, but clinical content

The Session Rating Scale (SRS; Miller, Duncan & Johnson) is a 4-item visual analogue scale taking under 60 seconds to complete. It is freely available for clinical use and is the measure most specifically designed for integration into routine sessions.