Personalised care for people with type 2 diabetes: Developing a shared decision-making tool embedded in the electronic medical record in primary care.

Talk Code: 
John Furler
Brea Kunstler, Jo-Anne Manski-Nankervis, Hamish McLachlan, Dougie Boyle, Sean Lo, Elizabeth Holmes Truscott, Jane Speight, Gary Kilov, Mark Kennedy, Ralph Audehm, Ken Clarke, Sophia Zoungas, David O'Neal
Author institutions: 
University of Melbourne, Deakin University, Monash University


In the 2018 Declaration on Primary Health Care, the WHO envisioned high-quality primary health care that empowers and engages individuals to maintain and enhance their health and well-being. This is a particular challenge in managing long-term conditions such as type 2 diabetes (T2D). T2D is progressive, requiring stepwise, evidence-based treatment intensification and ongoing self-management to achieve and sustain target glucose levels. However, treatment intensification can be a complex task, with new evidence and treatment options emerging, as well as the need to personalise treatment (based on diabetes and health characteristics, individual preferences, resources and supports). Furthermore, the busy, reactive, time-poor environment of a primary care consultation can be a barrier to implementing high-quality, empowering patient centred care.Clinical decision support (CDS) tools are one way to support general practitioners (GPs) in diabetes clinical management and potentially overcome clinical inertia. A CDS tool based in the electronic medical record (EMR) can automatically draw on data in the EMR to guide high-quality, evidence-based shared treatment decisions in real-time to inform real-world practice.


We undertook a staged design and refinement process to inform the design of an EMR-based CDS tool prototype (GlycASSIST). We scanned literature to identify existing T2D-related CDS tools to inform the initial design. Based on co-design and an evaluation framework for digital health interventions, we engaged health professionals and people with T2D in interviews and focus groups. Initial data collection explored participant’s prior clinical decision-making and reasoning and feedback on the design of GlycASSIST. Following significant refinements, the second stage involved computer simulations of the use of GlycASSIST in mock consultations with GPs (N=8). Focus groups were conducted with people with T2D (N=6), during which they viewed a video recording of the simulated use of GlycASSIST. Data were analysed thematically.


We identified clinical and interactional themes covering issues of: Up-to-date authority and reliability of the tool; integration with prescribing restrictions; integration within the EMR and with existing EMR prescribing software; GlycASSIST as a ‘third actor’ in consultations; making implicit reasoning explicit; supporting shared decision-making. GPs varied in how and the extent to which they used GlycASSIST to engage the person with T2D in shared-decision making. Participants with T2D were surprised at the number of treatment options available and the increased opportunity for choice generated in consultation. An interactive demonstration of the GlycASSIST tool will be presented.


Automated digital real-time decision support will increasingly form part of clinical care. GlycASSIST has potential for scalability and wide-reaching impact, translating rapidly evolving evidence into ‘real world’ clinical practice. While GlycASSIST may support shared-decision making for some, barriers to GP-based person-centred care remain. Following further refinement, GlycASSIST will be tested in a large-scale implementation in Victorian primary care.

Submitted by: 
John Furler
Funding acknowledgement: 
RACGP, Diabetes Australia, Melbourne Networked Society Institute