A complex intervention to improve intercultural communication and Aboriginal patient outcomes: which components worked?
Abstract Background Aboriginal language interpreters are under-utilised in healthcare in northern Australia. Self-discharge from hospital is an adverse outcome occurring at high rates among Aboriginal people, with poor communication thought to be a contributor. We previously reported increased Aboriginal interpreter uptake and decreased rates of self-discharge during implementation of a 12-month hospital-based intervention. Interrupted time-series analysis showed sudden increase and up-trending improvement in interpreter use, and a corresponding decrease in self-discharge rates, during a 12-month intervention period (April 2018 - March 2019) compared with a 24-month baseline period (April 2016 – March 2018). This paper aims to explore reasons for these outcomes, and further explore the likelihood of a causal association between study activities and outcomes. Methods We used the ‘Template for Intervention Description and Replication’ (TIDieR) as a framework to describe intervention components and evaluate their relative importance. Information on intercurrent activities that may have contaminated study findings was reviewed. The relationship between proportion of hospital separations among Aboriginal people ending in self-discharge and numbers of Aboriginal interpreter bookings made during April 2016-March 2019 was tested using linear regression. Results One full-time Aboriginal Interpreter Coordinator was employed for the intervention period who identified language needs, promoted interpreter use and mentored interpreters. The intervention period start date corresponded with commencement of this role. Three ‘Working with Interpreter’ training sessions were held during the intervention period reaching 83 clinicians, and three medical officers volunteered as champions of interpreter use in hospital practice. Employment of the Aboriginal Interpreter Coordinator was deemed the most important component of the intervention, based on reach compared to the other components and timing of the changes in outcomes. There was an inverse association between interpreter bookings and self-discharge rate among Aboriginal inpatients (p = 0.02). This association, the timing of changes and assessment of intercurrent activities at the hospital indicated that the study intervention was likely to be casually related to the measured outcomes. Conclusions Communication in healthcare can be improved through targeted strategies, with associated improvements in patient outcomes. Health services with high interpreter needs would benefit from employing an interpreter coordinator.