This study is a randomised controlled trial, utilising a delayed intervention design where all participants will ultimately be offered the intervention. A study schema is presented in Figure .
Setting and location
The study is situated in Port Augusta, South Australia, and includes participants from outlying communities who frequent services in Port Augusta.
To be eligible participants must be Indigenous, aged 18 years and above, and intending to reside in Port Augusta or a nearby community for the duration of the study (two years).
Participants are recruited using a variety of methods previously used successfully with this community [10
] including self-nomination, home visits, word of mouth, visits to community centres and referrals. Promotion of the study has occurred via posters in community centres and advertisements on the local Aboriginal radio station.
Funding has been provided by the National Health and Medical Research Council of Australia (NHMRC, project grant 627101).
A total of four part-time staff are employed for the initial phase of the study including a dental therapist, employed as a project manager. The remaining three staff are of Indigenous descent and are employed as project officers. In addition to recruiting participants, administering questionnaires and delivering the intervention, the Indigenous research officers play an invaluable role in providing ongoing cultural advice for investigators. The Indigenous project officers have been provided with basic oral health theory training by the project manager, enabling them to deliver the intervention. All staff have lived or worked in the local area previously.
An advisory group has been established comprising seven Indigenous community representatives. The group includes people working in health and education as well as community Elders. The advisory group has provided input into the study design, promotion of the study and the data collection instruments and techniques. The advisory group continues to advise investigators and the study team in relation to appropriate implementation of the study in the local community.
A pilot study was conducted in a neighbouring regional centre, primarily for the purpose of field testing the intervention instruments and giving the Indigenous project officers experience in delivering the sessions. Participants were invited to be involved in the pilot study through contacts at the local Indigenous health service. Written consent was obtained and participants received supermarket vouchers in acknowledgement of their time commitment. Involvement included completion of a baseline questionnaire, attendance at group intervention sessions and completion of a follow-up questionnaire. Participants were asked for immediate feedback upon completion of the questionnaire and after each intervention session as well as at the completion of the pilot study. This feedback was used to refine the questionnaire and intervention instruments for the parent study.
Consent and incentives
Participants are provided with written and verbal information about the study prior to giving consent. In acknowledgement of their time commitment, participants receive a $20 supermarket gift voucher upon completion of each questionnaire and a $10 gift voucher for each intervention session attended. At each intervention session refreshments are offered and participants provided with a variety of products to reinforce the key messages from each session, for example, water bottles, tooth brush and tooth paste, disposable dental mirrors.
The intervention consists of a series of five workshops delivered over a 12-month period.The instruments were developed collaboratively by the project manager, other study investigators and the Indigenous project officers, and are used to guide the project officers in delivering the intervention. Each workshop lasts approximately one and a half hours, including morning or afternoon tea and is conducted predominantly by the Indigenous project officers. Workshops consist of presentations, hands-on activities and interactive displays, group discussions and role plays. A key focus of the workshop series is on breaking down barriers and improving confidence of participants. Information around dental disease processes (ie dental caries, periodontal disease and dental erosion) is fundamental to workshop activities and discussion. The themes addressed in the workshops are underpinned by oral health literacy concepts and incorporate oral health-related self-efficacy, oral health-related fatalism, oral health knowledge, access to dental care and rights and entitlements as a patient.
The cluster randomisation method was selected as an appropriate approach for this study because of its acceptability to the participants and the local community (based on feedback from the advisory group and from the pilot study), as well as for the potential to increase the efficacy of the intervention through encouraging discussion amongst family groups and providing a supportive environment for participants’ development or change.
After purposive recruitment of 400 individual participants largely through local knowledge of kinship and other networks of the Indigenous project officers, 40 groups are formed based on family and social groups. Group sizes range from 8–12 people. The Indigenous project officers are responsible for assigning participants to the groups, utilising their knowledge of the local community. Family groups (clusters) are randomly assigned on a 1:1 basis to either a test-immediate or control-delayed intervention group. A computer-generated permuted block randomisation sequence is used, developed by biostatisticians at the Australian Research Centre for Population Oral Health (ARCPOH) using a random number generator. Randomly selected block sizes of 4, 6 and 8 are used, such that there is an equal number of participants in each intervention arm within the blocks. This ensures that if the study is stopped at any particular time there will be approximately an equal number of participants in each intervention arm (Figure ).
Data is collected through self-report questionnaires at baseline, at 12 months and at 24 months. The questionnaires include items pertaining to the primary and secondary outcomes and covariates. Questionnaires are administered by the Indigenous project officers and completed either as an interview or self-completed, with the degree of self-completion determined by the participant. The project officers are provided with a scripted method of introducing and administering the questionnaire. A log of attendance at the intervention sessions is collected.
Primary and secondary outcomes
The primary outcome measure is oral health literacy, measured using the HeLM [18
] adapted by investigators for a dental context.
Secondary outcome measures include oral health knowledge, oral health self-care, dental service utilisation, oral health-related self-efficacy and oral health-related fatalism [19
In the initial study design and pilot study, REALD-30 was utilised as the instrument to measure oral health literacy. After feedback from pilot study participants and the advisory group, it was deemed more culturally-appropriate to utilise the HeLM, adapted by study investigators and project officers for oral health, to measure oral health literacy. Specific feedback in relation to the use of REALD-30 included: (1) participants felt they were being tested and were hence intimidated; (2) the relevance of REALD-30 to the Indigenous oral health context was not clear; (3) the use of REALD-30 may be a barrier to full participation and completion of other components of the questionnaire.
Socio-demographic covariates include age, gender, education level, employment, income source, number of people staying in the house the previous night, number of children under 18 living in the household and car ownership.
General health covariates include medical conditions, behaviours such as cigarette smoking and alcohol consumption status, and self-rated general health.
Oral health covariates include self-reported oral health, previous dental extractions and oral health-related quality of life.
Psychosocial covariates linked with oral health outcomes in the Australian population [20
], include personal control [21
] perceived stress [22
] and an adapted version of the social support measure [23
Data handling and statistical methods
There will be three main analyses. The first analyses will occur after baseline, to quantify the extent of poor oral health literacy and ascertain the relationship between oral health literacy and oral health literacy-related outcomes. The second analyses will occur after 12 months, providing a comparison between the intervention and delayed intervention (control) groups, in order to assess the impact of the intervention. The third analyses will occur after 24 months, assessing the sustainability of intervention impact 12 months post-completion in the initial intervention group.
The Generalised Linear Mixed Model (GLMM) approach will be adopted, using STATA statistical software. Initial analyses will be simple, unadjusted comparisons of individuals. If there appears to be substantial imbalances between individuals in terms of baseline covariates, adjusted analyses will also be performed. All variables that are p<0.15 in the GLMM univariate analyses will be entered into multivariate models using a stepwise approach. All effects will be estimated with 95% confidence intervals, with the threshold for statistical significance determined as a two tailed p-value less than or equal to 0.05. All participants with at least one set of follow-up oral health literacy-related outcome data will be included in multivariate modelling.
The initial study design utilised a calculation of sample size performed using PC-SIZE software (GE Dallal, 1990, Version 3). Based on the 2008 oral health literacy survey using REALD-30 [10
], it was estimated that a sample size of 310 would be necessary to detect a 7.5 percent difference in the proportion of problem-based dental attenders (pre-intervention vs post-intervention), a 25 percent difference in the proportion of those who believe teeth should be brushed none or once daily (pre-intervention vs post-intervention) and a 30 percent difference in the proportion of those who believe cordial is good for teeth, don’t own a toothbrush or own a toothbrush but didn’t brush the previous day (pre-intervention vs post-intervention) at the significance criterion of 0.05 and a power of 0.80. Allowing for an attrition rate of 25 percent after 18 months, 388 participants would be necessary at base-line, rounded up to 400 for convenience; 200 in the intervention group and 200 in the control (delayed intervention) group.
During planning stages of the study, it was evident that some changes were necessary to study design. Specifically, a cluster randomisation approach was more appropriate for this study and community, the HeLM [18
] a more appropriate and acceptable tool for assessing oral health literacy than the REALD-30, and a 12-month follow-up more practicable than the proposed 9-month follow-up. The original calculation was retained as the best indicator of sample size available in the absence of other data to perform a revised power calculation.
Ethical approval was granted by the Aboriginal Health Council of South Australia and the Human Research Ethics Committee of the University of Adelaide. The Board of Management of the Pika Wiya Health Service (PWHS), the local community controlled Indigenous health service, also gave approval for the study. Comprised of representatives from the Indigenous community, the Board of Management is the peak body which governs the delivery of PWHS services and programs.