Hypotheses
We hypothesised that the AIMhi Stay Strong App would be superior to both a contact control (another app - Hep B Story) [21] and usual care, in reducing psychological distress and depressive symptoms, and improving quality of life, and dialysis adherence at 3 months. We also expected that delivering the Stay Strong app to control groups after the 3-month assessment would result in those groups showing improvements in these outcomes between 3 and 6 months.
Study design and participants
The study methods have been detailed previously [17] and are summarised below.
This was a three-arm, waitlist, single-blind randomised controlled trial with 2:2:1 allocation ratio testing the efficacy of the Stay Strong App intervention in improving wellbeing among Indigenous clients undergoing haemodialysis for ESKD in Alice Springs and Darwin. Assessments occurred at Baseline, 3 and 6 months (see Fig. 1 for CONSORT diagram). The three treatment conditions were: 1) Immediate treatment with the Stay Strong App (“ISS”), 2) Contact control/Delayed Stay Strong treatment (i.e. patients are engaged with the researcher for a similar time using the Hep B Story app then received the Stay Strong App at 3 months; “HepB/DSS”), and 3) Treatment as usual/Delayed Stay Strong treatment after 3 months (“TAU/DSS”; see Fig. 1).
Inclusion criteria were Aboriginal and Torres Strait Islander and aged ≥18 years, receiving maintenance haemodialysis in Alice Springs or Darwin for more than 6 months. Exclusion Criteria were aged < 18 years, guardianship order in place, or inability to provide informed consent (e.g. because of cognitive or visual impairment). No major changes to the study protocol occurred after trial commencement.
Consent, ethics and culturally appropriate approach
This study was performed in accordance with the Declaration of Helsinki. Approvals were granted by the Central Australian Human Research Ethics Committee (CAHREC No: HREC-16-406) and the Human Research Ethics Committee (HREC) for the NT Department of Health and Menzies School of Health Research (HREC-16-2599), including an Aboriginal subcommittee. Fully informed oral consent was obtained from all participants using pictorial information sheets and flipcharts in plain English with Aboriginal language versions available. Demographic information and outcome measures were collected using a tablet device including pictorial prompts and Aboriginal language recordings (choice of 11 NT languages). Interpreters were utilised where necessary.
Interventions
In addition to their allocated treatment, all participants received usual care from their renal service. Usual care was carried out according to the norms prevailing in the renal service, informed by the needs of the client. Interventions are described in detail elsewhere and outlined below [17].
Immediate AIMhi stay strong app treatment (ISS)
Baseline
Participants randomised to ISS completed a ~ 20-min interview using the AIMhi Stay Strong App at baseline, with a second ~ 20-min session using the App within 2–4 weeks. Session 1 explored family, strengths, worries and goal setting. Session 2 reviewed information entered previously, refined the goals and addressed any barriers to goal attainment, setting new goals as appropriate. Participants received a text message or phone call 1 week following the initial treatment reminding them of their goals and steps for making changes.
Three Months
Participants received a further two sessions using the AIMhi Stay Strong app following the 3-month follow-up assessment. The two 20-min sessions occurred 2–4 weeks apart following a similar format to the earlier sessions. A text message or phone call was sent 1 week following the initial treatment to remind participants of their goals and steps for making change and the time for the next session.
Hep B story contact control/delayed stay strong treatment (HepB/DSS)
Baseline
Participants randomised to HepB/DSS received 20 min of contact with the researcher using a culturally appropriate health app (i.e. The Hep B Story) at baseline, with a further 20-min session using the same app after 2–4 weeks. This ensured that each group received the same contact time and controlled for use of an app to structure the control session, as well as assisting with participant blinding. In Session 1 the participant interacted with the Hep B Story app with support from the researcher, focusing on app navigation and content. A ‘goal’ to talk to someone else in their family about the app content before the next session was set. Participants received a pictorial summary (utilising similar colours and images to the intervention summary). Session 2 reviewed the information discussed in Session 1. Participants received a text message or phone call reminding them of their goal (to talk with someone about the information) in the intervening weeks.
Three months
Participants received a 20-min interview using the AIMhi Stay Strong App (on a tablet device) following their 3-month assessment, with one further 20-min session using the App within 2–4 weeks, following the format of the sessions received by the ISS group at baseline.
Treatment as usual/delayed stay strong treatment (TAU/DSS)
Baseline
Participants who are randomised to TAU/DSS only received the questionnaires and no other researcher intervention at baseline.
Three months
After the 3-month assessment, participants received the AIMhi Stay Strong app intervention, using the same procedures as the HepB/DSS group.
Fidelity of the intervention
The interventions were delivered by trained researchers with reference to the AIMhi Stay Strong Planning Brief Treatment Manual [22]. Reviews of App data and ongoing booster sessions were used by the research team to provide regular feedback to redirect and adjust their mode of delivery as needed [17]. We defined an acceptable level of fidelity as two sessions of treatment delivered, one of which was at least 15 min. This applied to both the contact control (Hep B app) and intervention (Stay Strong app).
Outcomes
All participants were assessed at the beginning of the study (T0), after 3 months (T3) and again after 6 months (T6) with the following instruments: (1) the Kessler Distress Scale (K10, 2) the adapted Patient Health Questionnaire (PHQ-9); and the EuroQol (EQ-5D) 5 level. The K10 is a 10-item measure of psychological distress that is sensitive to symptoms of both anxiety and depression [23]. Responses are on a 5-point Likert scale. The PHQ-9 is a 9-item measure of depressive symptoms, scored on a 4-point Likert scale and has been adapted and tested for Australian Aboriginal people [24]. EQ-5D is a self-report measure of quality of life in 5 domains (mobility, self-care, usual activities, pain and discomfort, anxiety and depression). Respondents rate their health today on 5 levels of severity. It also includes a visual analogue scale which is used as a quantitative measure of overall health status. These assessments and their justification are described in detail elsewhere [17].
Sample size and power
The sample size calculation aimed to detect a minimum difference of 5 points on the K10 (with 90% power and an alpha of .05) leading to a sample size of 62, 62 and 32 for the ISS, HepB/DSS and TAU/DSS groups respectively. Based on previous data we considered a difference between the group mean scores of 5 to be clinically significant. This calculation allowed for 10% attrition.
Randomisation
Participants were randomised using a block sequential random number sequence (block sizes of 2 & 3) and an envelope system of randomisation, stratified by site, level of psychological distress (high or low i.e. < or > 25) and access to respite dialysis in home community as a person’s ability to return home to family (for visits) may influence wellbeing. An independent statistician created the allocation schedule with a computerized random number generator and investigators were blind to this schedule. Following baseline assessment, research assistants selected the next envelope in the sequence based on the participants site, level of distress and access to dialysis at home. Participants were allocated to ISS, HepB/DSS or TAU/DSS at a ratio of 2:2:1. The allocation was concealed to participants.
Blinding
Participants and outcome assessors were blinded to treatment condition. Research assistants who delivered the intervention were restricted from conducting follow up assessments with those participants to maintain blinding. The contact control involved an intervention of similar length of time and utilised an app presented on a tablet device to minimise client awareness of their treatment group assignment.
Statistical analyses
The primary analyses were conducted according to the intention-to-treat principle.
Descriptive statistics of baseline characteristics are provided by randomisation group. Continuous variables are summarised with mean and standard deviation (SD) when appropriate, or median and interquartile range (IQR) when not normally distributed, categorical variables were summarized with frequency and percentage.
Linear mixed models were used to estimate, for each of the K10, the EQ-5D and PHQ-9 score, the change between: (i) baseline assessment and 3-month assessment (ii) baseline and 6-month assessment (iii) 3-month and 6-month assessment. The changes described above were also compared between groups (ISS vs HepB/DSS; ISS vs TAU/DSS; HepB/DSS vs TAU/DSS) with mixed models that included a categorical variable for the three intervention arms, for each time point and an interaction term between arm and time. The use of mixed models has the advantage of allowing the modelling of the correlation between repeats of outcome recorded within the same participant and within site. Sensitivity analyses including baseline scores were included in the models when descriptive analyses showed important between intervention arm differences for baseline scores. The above described analyses were carried out for subgroups defined by baseline score severity (below or above 25 for K10 and below and above 10 for the PHQ-9), by region (Top End and Central Australia), access to home community dialysis or not.
Zero inflated Poisson models were used to calculate and compare between allocation arms, mean and 95% confidence intervals of numbers of missing dialysis sessions during the baseline to 3-month period, and the 4–6-month period.