Adaption of the “Working Alliance Inventory” to general practice to measure therapeutic alliance – the WAI-GP

Problem

A tool to measure the therapeutic relationship between a general practitioner (GP) and their patient would be useful as the therapeutic relationship is recognised as an important variable in patient outcomes. In psychological research and practice, the “Working Alliance Inventory” (WAI) is used to quantitatively measure the relationship between psychologists and clients with a high score associated with better patient outcomes. The WAI is based on the theoretical framework of Bordin who describes a three-part model of the therapeutic alliance: the bond between the two parties, collaborative goal setting, and agreement on the required tasks to reach the goals.

Approach

Our aim was to adapt the WAI for use in general practice and determine concurrent validity with a measure of shared decision making (DYDADIC) and depth of relationship (PDDR).Two rounds of online survey feedback from 55 GPs and 47patients were used to adapt the wording of the WAI for the general practice context – the WAI-GP. The tool was then completed by 142 patients in waiting rooms after seeing their GP, and by 18 GPs at the end of their session. Concurrent validity was determined using Spearman Rho correlations as WAI-GP scores were highly skewed. Patients were also asked to complete two social desirability surveys.

Findings

The suggestions for re-wording of the original tool were detailed with participants highlighting words and phrases that were perceived as judgmental or derogatory. In the clinical setting, 89 (62.7%) of the patients who completed the WAI-GP were female, most were over the age of 45 years (52.9%) and 72 (50.7%) self-reported having a chronic illness. 107 patients (75.4%) said they were seeing their preferred GP that day. Overall the patients had a good experience in their consultation (based on their report of being listened to and respected, and enough time spent) and this was reflected in the positively skewed WAI-GP data. The patient measure of WAI-GP was strongly correlated with DYDADIC (0.705, p 0.0001); and PDDR (0.591, p 0.0001). WAI-GP was not correlated with the measures of social desirability.

Consequences

The WAI has been adapted for use in general practice and shows validity with measures of shared decision making and depth of relationship. The WAI-GP has been developed to be used in research, clinical practice, and teaching to measure the therapeutic alliance between a GP and their patient. Further research to validate its application in other healthcare settings and different populations groups would be welcome.