BACKGROUND: Working alliance is the interaction between the patient and therapist. It is a crucial part of the physiotherapeutic process. One instrument to measure working alliance is available in Dutch/Flemish language and validated in psychotherapy setting. OBJECTIVE: This study aims to validate the Working Alliance Inventory Short-Form in a Dutch physiotherapy setting. DESIGN: A prospective cohort study in primary-care physiotherapy. METHOD: To validate the Dutch/Flemish version of the working alliance inventory short-form (WAV-12) a RASCH analysis was used. RESULTS: Sixty-six physiotherapists enrolled in total 389 patients with an average age of 50 years and a mean duration of shoulder pain of 33 weeks. A total of 274 patients filled in one or more items of the WAV-12. The WAV-12 showes good discriminative abilities and all items contributed to a one-dimensional construct. Due to the selective nature of the missing items, we believed rewording was necessary to make it more suitable to the physiotherapy setting. We performed a Delphi study and revised the WAV-12 into the PAS (Physio Alliance Scale). The validity of the revised version is unknown and is therefore not sufficiently strong to be implemented as a measurement tool. LIMITATIONS: The response rate for three items especially was low and we found ceiling effects in ten items. CONCLUSION: Although the measurement instrument shows good internal consistency and reliability, we made adjustments to the WAV-12 for Dutch physiotherapy setting.