An evaluation of the telehealth facilitation of diabetes and cardiovascular care in remote Australian Indigenous communities: - protocol for the telehealth eye and associated medical services network [TEAMSnet] project, a pre-post study design

Background Despite substantial investment in detection, early intervention and evidence-based treatments, current management strategies for diabetes-associated retinopathy and cardiovascular disease are largely based on real-time and face-to-face approaches. There are limited data re telehealth facilitation in type 2 diabetes management. Therefore, we aim to investigate efficacy of telehealth facilitation of diabetes and cardiovascular disease care in high-risk vulnerable Aboriginal and Torres Strait Islanders in remote/very remote Australia. Methods Using a pre-post intervention design, 600 Indigenous Australians with type 2 diabetes will be recruited from three primary-care health-services in the Northern Territory. Diabetes status will be based on clinical records. There will be four technological interventions: 1. Baseline retinal imaging [as a real-time patient education/engagement tool and telehealth screening strategy]. 2. A lifestyle survey tool administered at ≈ 6-months. 3. At ≈ 6— and 18-months, an electronic cardiovascular disease and diabetes decision-support tool based on current guidelines in the Standard Treatment Manual of the Central Australian Rural Practitioner’s Association to generate clinical recommendations. 4. Mobile tablet technology developed to enhance participant engagement in self-management. Data will include: Pre-intervention clinical and encounter-history data, baseline retinopathy status, decision-support and survey data/opportunistic mobile tablet encounter data. The primary outcome is increased participant adherence to clinical appointments, a marker of engagement and self-management. A cost-benefit analysis will be performed. Discussion Remoteness is a major barrier to provision and uptake of best-practice chronic disease management. Telehealth, beyond videoconferencing of consultations, could facilitate evidence-based management of diabetes and cardiovascular disease in Indigenous Australians and serve as a model for other conditions. Trial registration Australia and New Zealand Clinical Trials Register (ANZCTR): ACTRN 12616000370404 was retrospectively registered on 22/03/2016


Background
Diabetes and cardiovascular disease (CVD) are leading causes of personal and healthcare system disease burden, globally [1], and in Australia [2]. It is anticipated that this burden will disproportionately affect vulnerable populations, such as those in low-income countries and Indigenous populations [3], particularly in remote settings.
A wide range of interventions aimed at improving chronic disease outcomes and reducing care costs has been tried. These interventions mostly fall into one of two categories. One aspect is quality-improvement that focuses on service delivery in primary care and support services for patient self-management. The other aspect is patient-based disease-management. Evidence suggests the interventions that not only target disease management, but also aim to improve care-delivery design may be more effective than those that target disease management alone [4,5]. Targeted e-health applications that facilitate both service delivery and disease management have potential to meet such an objective. To date, clinical e-health applications mainly relate to the use of electronic health record systems (often alongside older paper-based systems), telecommunication of medical reports and videoconferencing consultations between a clinician and patient.
Telehealth is, in the context of the present study, that area of e-health that centres on the provision of healthrelated services and information using telecommunicationbased technologies. Telehealth in the clinical setting is in its infancy in Australia [6], where the use of this term has been largely limited to one Federal government funded healthrelated servicethat of a videoconferencing consultation between a specialist clinician and patient, subject to specific criteria [7]. By contrast, global uptake of a broad spectrum of e-health technologies, particularly mobile-health, is high [8].
Telehealth facilitation of service delivery and disease management has the potential to improve health outcomes for people with diabetes and to improve the quality of service delivery in the primary care setting. Currently, there are inadequate data to support the widespread clinical uptake of telehealth technologies. Therefore, the 'Telehealth Eye and Associated Medical Services Network' [TEAMSnet] study aims to address the knowledge gaps in three areastelehealth facilitation of 1. diabetic retinopathy screening 2. evidence-based diabetes and CVD management (via decision support and survey tools) and 3. patient self-management strategies.
This paper presents the protocol for the TEAMSnet study. This technology-based pre-post intervention study aims to describe the baseline characteristics of the participants with diabetes and clinical workflows in partner health service and to investigate the impact of telehealth facilitation in the management of CVD, diabetes and diabetic retinopathy. We hypothesise that telehealth facilitation of diabetes management will lead to improvements in patient engagement [attendance rates and survey participation] and clinical outcomes, such as blood pressure and metabolic control.

Study design
This is a pre-post study of technology-facilitated management of diabetes, CVD, and diabetic retinopathy. We plan to enrol 600 adults (age ≥18 years) with clinically-diagnosed diabetes from three remote Northern Territory Aboriginal-controlled health services -Central Australian Aboriginal Congress [CAAC] in Alice Springs, Miwatj Health Aboriginal Corporation [Miwatj] that incorporates the four clinics at Nhulunbuy, Galiwin'ku/ Elcho Island, Yirrkala, Gunyangara/Ski Beach in East Arnhem Land and Wurli-Wurlinjang Health Service [Wurli] in Katherine. Recruitment and intervention are anticipated to occur over a 3-year period. Participants will receive retinal imaging, diabetic retinopathy screening, including diabetes education, at baseline, followed by lifestyle assessment six months later, including self-management support. An electronic CVD and diabetes decision support tool that includes an Indigenous-specific CVD risk calculator, based on bestpractice as generally defined by guidelines in the current Standard Treatment Manual of the Central Australian Rural Practitioner's Association [CARPA], will be developed in conjunction with a clinical working group and used to generate a clinical recommendations report to assist clinicians with evidence-based management of CVD, diabetes and diabetic retinopathy approximately 6-18 months after baseline. Mobile tablet technology based on clinical input and feedback will be developed to support clinical staff and participants with engagement in diabetes management and selfmanagement, respectively.

Study aims and objectives
The overarching aim of the TEAMSnet project is to improve health care, and therefore health outcomes, for Indigenous Australians in under-resourced communities through the integration of existing electronic health record systems, specifically Communicare™ (Healthconnex, Perth, Australia), with the customisable, web-based health information technology called Chronic Disease Management Platform [CDMP], an existing, licence-free platform that thereby facilitates sustainability. A secondary aim is to evaluate the costefficacy of the intervention. There are three main objectives under the broad aim. Apart from the development and implementation of the four technologies, the primary outcome includes improved adherence to clinical care, such as eyecare visits. Secondary outcomes include reduction in hospitalizations and cost-efficacy of the suite of technologies, both individually and collectively.

Site recruitment
An invitation was circulated to Aboriginal-Controlled Community Health Organizations [ACCHO] in the Northern Territory to attend an information workshop at the Alice Springs Hospital. Following this, an expression of interest was sought from attendees. A discussion paper was circulated for internal discussions within the health service to the three health services that formally expressed an interest in the project. Once health service board approval had been obtained, a formal letter of support from the health service for the TEAMSnet study was sought and Memoranda of Understanding and service agreements were executed between partner sites and the administering institutions, the Universities of Melbourne and Sydney. Ethics applications were submitted by investigators to the Central Australian and Menzies School of Health Research Human Research Ethics Committees. Recruitment commenced once written ethics approval had been obtained.

Participant eligibility Inclusion criteria
Participants must be clients of the partner health services who are aged 18 or over; can provide informed consent; can successfully undergo retinal imaging; are willing to undertake surveys and provide relevant clinical information; and are willing to consult their doctor on at least two occasions over the following 18 months.

Exclusion criteria
Clients of the partner health services are ineligible if they been clients for less than six months prior to enrollment or have had diabetic retinopathy treatment.

Sample recruitment
Both community-based and direct recruitment strategies are being used to recruit study participants. These include; flyers in medical waiting and consulting rooms, and contact with potential referral sources, for example, general practitioners, chronic care team memberspodiatrist, dietitian, cardiac nurse, renal nurse, diabetes educator, text and letter invitations.
Prospective participants express their interest to the study co-ordinator at a face-to face meeting during which the individual is screened. If the individual meets the inclusion criteria, they are consented to the study (consent form signed after patient information sheet discussed) and diabetic retinopathy screening is performed (baseline visit) in order to reduce the time burden on participants and improve compliance with the study protocol. (3) The referral protocol (Fig. 2) is based on the 2008 NHMRC Guidelines for the Management of Diabetic Retinopathy. (4) The reporting protocol (Fig. 3) has been developed with a view to highlighting the essential management recommendation in the TEAMSnet retinopathy screening report that is sent to Communicare at the participating site via Argus messaging (a third party vendor providing secure communications between health services) for the primary care health professional to Fig. 3 Diabetic Retinopathy Screening reporting protocol evaluate and action in the context of the participant's prior ocular history in the EHR [as is the case with pathology reports]. (5) Subsequently, the disc-centred images in a specific format will be used by graders for retinal vessel calibre measurements, which have been shown to predict CVD and retinopathy onset and progression (Fig. 4)  Trained study staff will administer the suite of surveys during the second study visit some 6-months after the imaging visit. The SNAPE report (Fig. 5) can be viewed either in the EHR systems (within 24 h) or immediately via login to CDMP or the mobile tablet. The primary care clinician upon review of the report will generate a referral as indicated. High depression scores are brought to the immediate attention of the participant's treating clinician. Participants' comments around goals and barriers/motivators to change, such as food security (access to healthy food), will be documented by the survey administrator.

Study procedure
3. EDS development, use and reporting protocol Once partner sites were identified, we identified the CARPA standard treatment manual as the most appropriate guideline for CVD and diabetes EDS development. As Communicare is the EHR system used by all partner sites, ideally the TEAMSnet EDS application would need to be interfaced with Communicare for automated syncing of as much of the clinical data required by EDS as possible so as to avoid duplication and improve clinical efficiency. The developers of the CARPA standard treatment manual and Communicare were advised of the TEAMSnet project and agreed to support the project informally as advisors and collaborators, particularly in relation to timing of their product updates and releases. In return, TEAMSnet would provide both organisations with potentially useful user feedback during the project.
TEAMSnet EDS will provide clinicians with three CVD and diabetes management tools.
(1) The CVD risk calculator and slide rule that enables baseline measurement and regular monitoring of a participant's overall CVD risk (Fig. 6). (2) The detailed CVD and diabetes risk-analysis tool that provides a clinician with specific recommendations and optional learning points for best-practice management of a participant's CVD risk (Fig. 7). (3) Active CVD and diabetes decision-support tool a proactive clinical process whereby each recommendation is reviewed by the clinician and either accepted or rejected, in which case a textbox appears for the explanation. An EDS report will be generated for each participant as soon as possible after the s econd clinical visit [at which surveys are administered]. A second and final EDS report will be generated for participants prior to the end of the study (Fig. 8).

Mobile tablet application
Mobile tablet development and use will enable display and review of much of the data derived from GRI, SNAPE surveys and EDS applications. Surveys can be administered and an educational resource library can be populated and accessed via the tablet, and this can take place in an out-of-clinic setting, thereby facilitating patient engagement.

Data collection and outcome measures
Demographic data (including age, gender, postcode, contact details) and self-report measures are obtained from each participant at baseline and/or at the 6-month survey administration visit. Pre-and post-study clinical data such as blood and urine biochemistry, blood pressure and anthropometric measures including height and weight are also collected. Pre-and post-study health care resource utilization is captured through EHR records and includes information on hospitalizations and eye-care consultations.

Pre-study data
Site partners will collaborate to audit the Communicare EHR for each participant in order to obtain the required pre-study data i.e. the data that correspond with post-study data relating to CVD, diabetes or diabetic retinopathy risk and outcomes and each participant's use of health services, such as eye-care consultations and hospitalizations.

Data management
The study coordinators at respective sites will be responsible for electronic scanning and storage of hard copies of participant consent and report forms for entry into the password protected central database. Hard copies will be kept in a secure cabinet at the respective sites. If requested, participants can access their individual results at study completion.

Study integrity
Approval to conduct the study was received from Central Australian and Menzies School of Health Research Human Research Ethics Committees in the Northern Territory. Written informed consent will be obtained from all participants. A local translator will assist in consenting patients for whom English is not their first language.

Sample Size
For a conservative differential in adherence to subsequent care of 15%, and with 80% power, the minimum sample required is 490. Estimates are based on a 25% attrition rate. We will therefore need 600 participants across the three partner sites.

Data analysis
Investigators and their supporting statistics departments will conduct data analysis. Site differences will be analysed. Analyses will provide important descriptive data, such as prevalence rates and lifestyle data, proportion treated to target for CVD, diabetes and diabetic retinopathy risk factors and predictors of diabetic retinopathy and CVD outcomes. Analysis of covariance (ANCOVA) will be used to compare differences between groups. To assess confounding, covariates will include age, gender, and established and emerging risk factors for CVD, diabetes and diabetic retinopathy, as appropriate. Non-parametric statistics will be used when assumptions for parametric methods are violated. All tests will be conducted using an alpha level of 0.05 and 95% confidence intervals will be reported.
Complete case analysis and, where reasonable, multiple imputation will be employed for the management of missing data. Missing data may occur due to drop-out, missed assessments or item non-response. With knowledge of the missing data mechanism, all available information can be used. One approach will be to use the multilevel model, a full-likelihood method that allows for a response to be missing at random, where missingness depends only on observed data and not on unobserved data. Another strategy involves imputing missing values by multiple imputation procedures.
The proposed economic evaluation aims to include pre-and post-study resource utilization costs for both the costs and consequences and a cost consequence analysis (CCA). Standard unit costs i.e. Australian Medicare Benefits Schedule will be applied to the resource-use data. Costs will be presented in total and disaggregated Of the recruitment target of 600 Indigenous Australians with diabetes, to date 576 participants have been recruited and undergone retinal imaging for diabetic retinopathy screening.

Discussion
It is well-known that a significant number of individuals with diabetes fail to adhere to screening recommendations and/or adequately self-manage diabetes-related risk factors and consequences. Furthermore, primary care service delivery for diabetes and its complications is challenging, even in well-resourced environments with a large proportion of patients at low risk of diabetes complications and related conditions, such as depression. Importantly, the disease burden of diabetes in Australia is greatest amongst Indigenous Australians, particularly in remote settings.
Telehealth facilitation of service delivery and diabetes management has potential benefits. To date, there are insufficient data to answer the question 'Can telehealth applications facilitate best-practice in diabetes and CVD service delivery and disease management in remote Indigenous Australian communities?' The TEAMSnet study will provide evidence regarding the efficacy and cost-efficacy of telehealth facilitation in the management of diabetes, its complications and related conditions in remote and very remote Indigenous Australian communities. We anticipate benefits in the area of new and modified diabetes management strategies alongside improvements in patient self-management.
This multi-pronged telehealth approach may facilitate the management of diabetes, an increasingly prevalent and complex chronic condition and public health problem. The benefits of telehealth facilitation of retinal screening, lifestyle modification and electronic decision support might then be extended to other complications of diabetes and common co-morbidities.