Measurement properties of the Brazilian version of the Working Alliance Inventory (patient and therapist short-forms) and Session Rating Scale for low back pain.

Abstract

In the low back pain (LBP) field, therapeutic alliance is considered a non-specific factor of interventions associated with improvements in clinical outcomes. However, there is a paucity of studies aimed to evaluate measurement properties of tools used to objectively quantify the alliance between therapist and patients, such as the Working Alliance Inventory (WAI) and Session Rating Scale (SRS). To translate and cross-culturally adapt the short-form version of WAI - therapist and SRS into Brazilian Portuguese; to investigate the measurement properties, of the WAI-Patient, WAI-Therapist and SRS in patients with LBP and their physical therapists, respectively. One hundred patients with LBP and 18 physical therapists were recruited from physical therapy clinics in Brazil. Therapeutic alliance measures were collected at the initial assessment, prior to the second session, and at 2-month follow-up. The measurement properties investigated were reproducibility, internal consistency, ceiling/floor effects and responsiveness. Although WAI-Patient, WAI-Therapist and SRS were considered to have acceptable test-retest reliability (ICC2,1 > 0.70), these questionnaires showed problems with other measurement properties. WAI-Patient showed problems with internal consistency (i.e. Cronbach's alpha < 0.70 for all subscales). Presence of ceiling effect (i.e. > 15% of participants with the maximum score) and poor internal responsiveness were found for the WAI-Patient (Effect size = 0.15; 84% CI: 0.04 to 0.29) and for the SRS (Effect size = 0.05; 84% CI: -0.22 to 0.11). The WAI-Therapist revealed slightly better measurement properties. We identified psychometric limitations with most measurement properties of the WAI questionnaires and SRS. Future studies are needed to refine these tools.