Closing the health gap in Central Australia: reduction in Indigenous Australian inpatient self-discharge rates following routine collaboration with Aboriginal Health Workers

Indigenous Australians experience significant socioeconomic disadvantage and healthcare disparity compared to non-Indigenous Australians [1]. Historically, generations of Indigenous Australians have faced rampant systemic discrimination since British colonialisation. Between 1883 and 1969, Indigenous children were forcefully removed from their families by the Australian Government and church missionaries, with victims of this period coined the “Stolen Generation” [2, 3]. Cumulative generational trauma among Indigenous Australians has resulted in their social disenfranchisement, poor health literacy, and impaired self-advocacy [1].

Self-discharge describes events where patients leave a health service before discharge by their clinician [4]. Data from the Australian Institute of Health and Welfare showed that the age-standardised rate of self-discharge among Indigenous patients had risen from 11 to 16 per 1,000 population between 2004–05 to 2016–17 [5]. These figures are of significant concern as patients who self-discharge from hospital are more likely to be readmitted and succumb to higher rates of morbidity and mortality [6]. Self-discharge is associated with patient dissatisfaction and negative hospital experience, and hence considered a surrogate measure of patient satisfaction [7]. There is no consensus definition of self-discharge in the literature [8]. For this study, we defined self-discharge as patients who had left the service for 24 h without prior agreement with the treating team.

Alice Springs Hospital (ASH) is the major referral centre for Central Australia, with a catchment area of over 1.2 million square-kilometres, and a population of approximately 42,000 [9]. Roughly 60% of the population reside within the town of Alice Springs, and the remaining population are widely dispersed between the rural town of Tenant Creek and fifty other rural communities, some of which extends into the adjacent states of South Australia and Western Australia. As many as seventeen different languages are spoken by Central Australia’s Indigenous people [10, 11], and is often as their first and preferred language. To improve communication with this diverse population, the government inducted Aboriginal Health Workers to help with health care delivery. Though they have been an integral part of the Northern Territory workforce since the 1870s, this rich, varied, and accessible resource have not been engaged to their full capacity [12, 13].

Research by Einsiedel et. al. in Alice Springs has emphasised the dire need for better communication among clinicians, patients, and other healthcare providers. They found 73·4% of self-discharged patients from medical units had done so with little understanding of their diagnosis or their expected trajectory of disease [14]. Routine use of Aboriginal Health Workers in Australia is poorly described in literature [8]. A qualitative study by Kerrigan and colleagues showed improved communication and patient satisfaction with consistent utility of interpreters [15]. To improve communication, and provide effective health care, the Orthopaedic team at ASH incorporated Aboriginal Liaison Officers (ALO), part of the broader family of Aboriginal Health Workers, into the routine care of Indigenous patients. In this study, we aim to demonstrate the effect of regular ALO presence within the Orthopaedic multi-disciplinary team (MDT) on the rates of self-discharge among Indigenous orthopaedic patients. We also explore the stage of treatment at which patients self-discharge, risk factors for self-discharge, and the association between length of stay and self-discharge.

This retrospective cohort study was conducted at a single centre. Ethical approval was obtained from the Central Australian Human Research Ethics Committee prior to commencement of this study (Approval No.: CA-22–4376). Reporting was performed based on the recommendations by Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) [16].

Between 23^rd of May 2020 and 22^nd of November 2021, a total of 2,147 patient episodes admitted under our Orthopaedic service was identified. 1,135 Indigenous patient episodes were included in our primary analysis. The remaining 1,012 non-Indigenous patient episodes were analysed for comparison.

There were 617 patients in Group 1 and 518 patients in Group 2. The average age of patients was 31·06 ± 17·01 and 31·10 ± 18·19 years in Groups 1 and 2 respectively. This was not statistically significant. In both groups, most patients were male, never married, and unemployed. There was no significant difference between groups in terms of gender, marital status, employment status or place of residence. Majority of patients presented with an acute presentation, most commonly fracture or dislocation. Significantly more patients presented with tendon injuries in Group 1, though other diagnoses did not differ significantly between groups. The average length of stay was comparable in both groups. A detailed description of the baseline characteristics is shown in Table 1.

To our knowledge this is the only study to quantify the impact of partnership between clinicians and Aboriginal Health Workers on the rate of self-discharge amongst Indigenous Australians within an Orthopaedic unit. Our findings demonstrate that routine inclusion of ALO within the Orthopaedic MDT is associated with a lower risk of self-discharge among Indigenous Australian patients.

Routine ALO input was associated with a smaller proportion of patients who self-discharged prior to definitive surgical or medical treatment, and completion of intravenous antibiotic course following surgery – which are arguably the more critical aspects of a patient’s care. We postulate that delays in allied health assessment may be a cause for patients self-discharging prior to clearance, however, this would need to be investigated further. As most patients admitted within our service would have experienced a similar treatment path (definitive surgical or medical treatment, completion of postoperative antibiotic treatment, allied health clearance, discharge medications), we did not modify our analysis. Our findings mirror that of O’Connor et. al, which showed a trend towards reduced self-discharge rates as interpreter use within the centre increased [20], as well as that of Taylor et. al., which showed reduced number of Indigenous self-discharges following creation of an Aboriginal Health Worker role [21]. Neither of these studies defined the precise role and extent of involvement of Aboriginal Health Workers.

Despite improvement in self-discharge rates following routine inclusion of ALO, they remain significantly higher than in non-Indigenous patients. On comparing risk factors for self-discharge between Group 1 and 2, patients living in Alice Springs town-camps, and those diagnosed with septic arthritis and osteomyelitis were no longer risk factors following routine inclusion of ALO. This change is encouraging, as it suggests that patients with these potentially more serious musculoskeletal conditions, were more likely to have comply with their recommended treatment in hospital. Although our statistical analysis may have suggested Perthe’s disease being a risk factor for self-discharge in Group 1, this is unlikely to be of clinical significance, as there was only a single patient diagnosed with this condition within the group. The association between length of stay and odds of self-discharge was analysed separately. Although our initial analysis of the data showed no association between length of stay and the odds of self-discharge, deeper analysis indicates a significantly lower odds of self-discharge in Group 2. This implies that patients requiring prolonged hospital treatment episodes experience a stronger beneficial relationship with ALO, generating a greater degree of trust in the system and improved patient compliance.

Despite introduction of ALO, patients residing in communities between 51 to 100 kms of Alice Springs continued to be a risk factor for self-discharge. It is possible that patients residing within this zone may feel isolated from their community, resulting in a leave event. Interestingly, patients residing in communities greater than 101 km from Alice Springs was not a risk factor for self-discharge. We speculate that patients living in these areas are still likely to experience isolation from their community but are unable to self-discharge due to a lack of access to transport back home. It is important to emphasise that these hypotheses would need to be assessed in a future qualitative analysis within our patient cohort. In addition to their residence category, patients of younger age, pensioners and those diagnosed with wound or soft tissue infections persisted as risk factors for self-discharge following the change in practice. Targeted health promotion campaigns, as well as special emphasis by clinicians on such patient demographics, may help reduce the risk of self-discharge.

Qualitative analysis on the subject of self-discharge suggests the decision to self-discharge to be a rational act of reclaiming personal autonomy and having agency over ones healthcare decisions, likely stemming from unmet needs [22]. In most patients, inability to delegate responsibilities inherent to social or cultural roles and financial constraints in combination with perceived improvement in ones’ health, often precipitate a leave event [8, 23, 24]. In our experience, ALO are often related to patients through bonds of kinship. This enables trust and confidence in the interactions between patient and the treating MDT, empowering patients to voice their concerns surrounding care, as well as openly disclose often competing priorities, for example childcare, which may otherwise be undeclared. This allows for the development of a more comprehensive management strategy and allows for early involvement of other services, such as social work, to potentially mitigate competing priorities.

Unlike most non-Indigenous patients, Indigenous patients have everyday lived experiences of personal and institutional bias and racism [22], and unsurprisingly many may anticipate and perceive discrimination regardless of its presence [25]. Healthcare facilities in Australia often embody policies shaped by Western perception of health, which are often focussed on the recognition and treatment of disease [26]. In contrast, Indigenous people have a more holistic view of health, which encompasses the physical wellbeing of the individual in combination with the social, emotional, and cultural well-being of one’s whole community [27]. In this regard, ALOs help provide cultural brokerage, allow for shared decision-making process, and provide a culturally safe environment for patients, encouraging patients’ requests for services outside the scope of Western medicine. This may include the service of Ngangkaris (Traditional Healers), who perform healing rituals. Recent evidence have proposed self-discharge as a marker of surgical cultural competency and cultural safety [28]. In the context of our study, routine ALO collaboration is a positive step towards cultural competency. However, we acknowledge that there are further measures that could be taken, such as increasing employment of Indigenous health care providers and cultural safety training for staff members, which may continue to reduce the rates of self-discharge.

The use of ALO also helps improve communication between the patient and their healthcare providers. It has been well documented in literature that Indigenous patients often feel frustrated and misunderstood, and even intimidated by healthcare professionals [15, 29]. In an interview with an Indigenous patient for a qualitative study by Kerrigan et. al, it was remarked that he could comprehend about half of what was explained to him [15]. Although ALO are different from trained interpreters, most of the ALO in our hospital are proficient multilinguists. On rare occasions, where the ALO felt that they could not communicate in the patient’s language, this was highlighted, and alternative arrangements were made. It is postulated that a patients’ comprehensive understanding of their disease, treatment, and progress, would better incline patients to engage fully in their treatment and reduce risk of their self-discharge. Previous studies have shown improved patient engagement in rehabilitation following close partnership between specialist nursing staff and ALO [30].

Our study has a few limitations. Firstly, by virtue of its retrospective design the accuracy of our data was dependent upon the quality of data collection and entry during each treatment episode. Secondly, although we used self-discharge as a surrogate tool for improvement in service provision, there is potential that some patients may have self-discharged for reasons other than their dissatisfaction with our service. Analysis of non-Indigenous patients suggest that there were no significant changes in rates of self-discharge between the two groups. This implies that the change in self-discharge is likely due to implementation of routine ALO. Thirdly, there may be other changes within the hospital structure, apart from our ALO inclusion across the study period, which may have influenced self-discharge rates. Fourthly, we were unable to quantify the harm or increase in healthcare burden inflicted by self-discharge. It is plausible to assume that at least some self-discharged patients may have developed early complications or disease recurrence and presented to local health clinics, whilst others could present with late complications, either to our service or elsewhere.

Another limitation of our study was that there was also no definition of an Indigenous patient. Given the nature of the study design, patients were classified as Indigenous or non-Indigenous based upon self-identification. We were not able to further classify our patients into Aboriginal, Torres Strait Islander, or Aboriginal and Torres Strait Islander people. Furthermore, we did not have data on the role of caregivers or families in informing our patient’s self-discharge. In addition to this, some patients in Group 2 would not have exposure to an ALO, as this service would have only been provided on weekdays (excluding public holidays). Of the 42 patients who self-discharged in Group 2, ten patients have done so over the weekend, and hence had no exposure to ALO. Lastly, the regression model of risk factors only considered independent variables that would be available from routinely collected patient data and may not represent the full spectrum of probable causes.

In conclusion, routine incorporation of ALO within the Orthopaedic MDT was associated with a significantly lower risk of self-discharge among Indigenous Australian patients. This entails no significant added costs, simply a shift in mind-set within the Orthopaedic MDT to fully engage and empower the existing workforce of ALO within the hospital. Patients were less likely to leave prior to completion of the most critical aspects of their care. These results are likely generalisable to other centres within Australia. The potential for future out of business hours, weekend and public holiday ALO service provision in ASH is perceived by the authors as a likely positive driver for further improvement in patient care. Targeted health promotion campaigns may aid in further lowering rates of self-discharge. Further research into the risk factors for self-discharge, as well as qualitative research on patient perception of our proposed healthcare model should be considered.