A new model of integrated primary-secondary care for complex diabetes in the community: study protocol for a randomised controlled trial

Type 2 diabetes mellitus (T2DM) is one of the most common chronic diseases and significantly impacts healthcare systems in both the developed and developing world. Diabetes currently affects about 371 million people worldwide[1] and this figure is expected to increase to 522 million by 2030[2]. In 2012, 11.7% of adult Australians were estimated to have diabetes. This prevalence is increasing steadily, being the fourth most frequently managed chronic disease in general practice, accounting for 6.8% of all chronic disease-related GP visits[3]. Internationally, waiting lists at specialist diabetes outpatient departments continue to grow, resulting in care that can be fragmented and inefficient[4‐8]. All indications are that Australia faces similar challenges with long waiting lists and difficulties related to accessing hospital-based diabetes outpatient clinics[9‐11]. It follows that diabetes is a chronic disease where the GP has a central role to play.

With appropriate support, follow-up and information technology systems, delivery of complex diabetes care in general practice can be as effective as hospital-based outpatient care[12]. The United Kingdom has introduced diabetes clinics conducted by general practitioners with special interests (GPwSIs) with demonstrated favourable improvements in HbA1c, cholesterol and blood pressure[13]. This research has demonstrated the potential of GPwSIs to take on a much greater role in diabetes screening and management to help meet the current demands on hospital outpatient services[14‐17]. In addition, an evaluation of a diabetes management program based on integrated care and management between GPs and GPwSIs undertaken in Germany showed a reduction in HbA1c from baseline[18]. Similar studies in Belgium and the USA have also demonstrated improved outcomes (HbA1c, systolic blood pressure and LDL-cholesterol) for diabetes patients using shared or integrated care between GPs and specialists[19‐21]. A multidisciplinary and coordinated approach to diabetes care including collaboration between practitioners can result in effective, efficient service delivery[17, 22].

Recently in Australia, Russell and colleagues[23] evaluated a model of complex diabetes care using a controlled pre-post study design which incorporated the principle of a GPwSI supported by an Endocrinologist and diabetes educator within a community-based general practice setting. The study showed promising results with a trend to better glycaemic control at a reduced delivery cost for the intervention group compared to the usual care group treated at a hospital diabetes outpatient clinic. Few studies have evaluated integrated diabetes services between primary care and secondary care by increasing the capacity of primary care to meet the needs of patients who would otherwise use hospital outpatient services. Even fewer have done this using randomised controlled trials. Indeed, a recent systematic review in the area concluded that there is a need to improve the design and quality of studies investigating shared care at the primary-secondary interface[24].

The specific purpose of the work reported here is to build on the promising results of Russell and colleagues’ study by using a more rigorous study design[23]. The aim of the study is to evaluate the new model of care using a range of outcomes, the primary one being glycaemic control. Secondary outcomes include: other clinical outcomes; patient-reported outcomes, such as quality of life and mental health; the acceptability of the model to patients, GPs and other healthcare providers; and the cost-effectiveness of the model.

We present a protocol for the study design and methods to evaluate a new model of diabetes care in the community through multidisciplinary collaboration and integration across the primary-secondary interface. The model of care aims to improve the quality and safety of healthcare at the interface between hospital- and community-based primary care for patients with complex diabetes. This model has been pilot-tested in a community chronic disease management service, and the results indicated benefits for diabetes patients in comparison with usual care at the hospital diabetes outpatient clinic[23, 26]. The current study builds on the promising results of Russell and colleagues’ study[23] in a larger, more robust research design and addresses the methodological issues identified in their study which utilised a non-randomised convenience sampling method. Non-inferiority randomised controlled trials are usually conducted on the premise that there is some other advantage of the treatment such as a reduction in waiting list times, reduced costs or improved access[34].

The study is important because it will provide empirical evidence about the impact on health outcomes, patient and clinician satisfaction, and economic outcomes of the improved integration of services. The results will also help us to understand the perspectives and experiences of both patients and health providers regarding the new model of care. Using a mixed methods approach will enhance the quality of the findings by enabling us to examine the model from different perspectives. It will also provide insights regarding contextual factors affecting the acceptability of the model and how it may be implemented in the future. In addition, selection bias is avoided by randomising at the level of individual patients.

However, some limitations of the study need to be considered. Participant retention in this longitudinal study will be sensitive to patient medical conditions, patient motivation as well as patient relocation out of the study area. There may be some patient issues around willingness to participate if randomly allocated to a clinic setting that is difficult to attend due to distance travel cost, and so on. Additionally, self-reported data on service use may be under-reported[35, 36]. Nevertheless, every effort has been made to present and structure the patient information sheet and survey for this present study in a simple and straightforward format. Furthermore we have conducted pre-testing and pilot studies to refine the survey in an attempt to minimise item non-response and facilitate a good overall response rate[37]. In addition, efforts will be also made through rigorous implementation of recruitment and survey strategies during the data collection process, and using a best-estimation technique for item non-response to reduce missing items. Also, we cannot rule out the possibility that change in the outcomes could in part be accounted for by an observer effect whereby participants modify their behaviour, in particular their self-care and self-management regimes, in response to being involved in the trial (for example, the intensity of the study assessment) rather than in response to being involved in the model of care. To minimise this effect, we have included both a control and intervention group in the study design that undertake the same measures. Patient surveys are incorporated into the process of patient care, and are distributed to patients by Registered Nurses at the clinical study sites for patients to self-complete. Ideally, if the model of care is shown to be effective and is subsequently embedded in usual clinical practice, routinely collected clinical data could be used to reassess patient outcomes in the longer term. Finally, a longer follow-up period would further elucidate the potential to investigate the sustainability of any observed effects. The results will need to be interpreted with consideration of the inherent weaknesses of non-inferiority trials[25].

There are some challenges around planning and setting up such a complex randomised controlled trial within a health services delivery setting, such as governance issues, the building of relationships with key research stakeholders, and adjustments to planned research delivery to accommodate healthcare reform changes in hospital and community care. Nevertheless, better integration of primary and secondary care and ‘ending the blame game’ between Commonwealth and State-funded healthcare has been a key element of Australian health reform since 2007. Better integrated care leading to improved quality outcomes for communities has been underlined repeatedly by the National Health and Hospitals Reform Commission, the Council of Australian Government’s Reform Council, the National Primary Health Care Strategy, the National E-Health Transition Authority and Health Workforce Australia[38]. To date, there have been few rigorous international publications to guide success in the area. Our randomised non-inferiority trial, based around a successful cohort pilot, will progress international learning in the area of chronic disease. If effective, the intervention could be applied to other chronic conditions and extended nationwide for integrated care delivery in the future.

Recruitment for the trial began in December 2012. To date, we have four study sites engaged and 111 patients recruited (Control: Intervention = 28:83).