Best Abstracts at the 10th Health Services and Policy Research Conference


As part of the peer review process for the Conference the scientific committee identified the highest ranked abstract overall and from an emerging researcher.

Best Abstract Overall was awarded to Thomas Elliott Senior Research Assistant, QIMR Berghofer Medical Research Institute for his abstract:

Using a hybrid discrete-event /agent-based model to examine improved triaging of GPreferred patients to colonoscopies
Thomas Elliott1, Anton Lord1, Lisa Simms1, Graham Radford-Smith2, L.Gordon1
1QIMR Berghofer Medical Research Institute; 2Royal Brisbane and Women’s Hospital

Thomas Elliott

Background: Demand for colonoscopy services is placing significant pressure on health resources in Australia with over 700,000 procedures annually and costing over $1.1 billion. Current triaging processes do not effectively discriminate patients who will or will not show significant bowl disease (SBD). A new risk assessment tool may improve system efficiency and enhance patient care.

Methods: We constructed a hybrid discrete event/agent-based model in AnyLogic software to simulate the GP-referral patterns, triaging, appointment booking, waiting times and costs for colonoscopies at the Royal Brisbane and Womens’ Hospital (RBWH). Using data from a cohort study of 466 GP-referred patients with lower abdominal symptoms, we utilised a newly designed clinical risk assessment tool (RAT) to model the current inefficiency in treatment urgency categorisation. The RAT had a specificity of 98%, showing its appropriateness for use as a rule-in test. We examined the RAT’s effect on waiting times and costs through simulation modelling at the hospital system

Results: Approximately 2,200 patients referred to the colonoscopy clinic at the RBWH are treated each year. The RAT improved triaging for 9% of moderate urgent and 4% of least urgent patients. Over a year approximately 89 patients with SBD could receive more timely care. Moderately urgent patients with SBD received care 121 days earlier, while least urgent patients received care 114 days earlier.

Conclusion: Our preliminary analyses indicate the RAT has significant potential to improve the coordination of patient care, improve patient outcomes, spare some patients from colonoscopies, while also saving substantial hospital expenditure.


Best Abstract from an Early Career Researcher was awarded to Grace Kyoon-Achan University of Manitoba for

Journeys to healthcare transformation: Partnering for change through Community based Participatory Research with Manitoba First Nations
Grace Kyoon-Achan1, Josée G. Lavoie1, Kathi Avery Kinew2, Stephanie Sinclair2
1University of Manitoba; 2First Nations Health and Social Secretariat of Manitoba

Grace Kyoon-Achan


Background: Community-based participatory research can help support communities’ efforts in the formulation and implementation of innovative and transformative change. We share a collaborative journey in our work with eight (8) First Nations in Manitoba Canada who were seeking to improve the health of their communities. Participating communities wanted to produce evidence to move toward good health as they define it and to inform changes in healthcare based on community input on how to improve the health of their respective communities.

Methods: Indigenous knowledges and processes guided data gathering and analysis. Community based participatory methods were used for the conceptualization of the study, data collection, analyses and knowledge translation. A variety of methods were used including: qualitative interviews, administrative health data analyses, surveys and case studies depending on the needs of the studies and the interests of different First Nation stakeholders, health directors and researchers.

Results: Research relationships built upon strong ethics and locally recognized protocols enhance mutual commitment to support community-driven transformation. Collaborative and respectful relationships are platforms for defining and strengthening First Nations health and healthcare priorities. This is an important step in building long term changes that will be sustained long after research projects come to an end.

Conclusion: This study yielded a blueprint for respectful community-based participatory research
involving First Nations peoples, communities, University-based researchers and First Nations
community-based researchers. The result is a locally owned yet widely recognized process that is
effective and sustainable while also having fulfilled researchers and funding obligations.

Highly Commended  

Tim Tenbensel   University of Auckland       

Gaming the New Zealand Emergency Department target. How did it vary between hospitals, and why?
Tim Tenbensel1, Peter Jones2, Linda Chalmers2, Sarah Appleton-Dyer1, Lisa Walton1
1University of Auckland’ 2Auckland District Health Board

Background: New Zealand’s Shorter Stays in Emergency Departments target was implemented from 2009 to address ED crowding. The target required 95% of patients to be seen, treated or discharged within six hours. ‘Gaming’ is commonly regarded as an undesirable side-effect of using targets to drive health service improvement.

Aims: Our research sought to identify the presence and extent of gaming after the target’s implementation, and explain variation in gaming across four case-study sites.

Methods: We adopted a mixed-methods approach. To track the extent of gaming, ED utilisation data was collected from four case study hospitals from 2007 (two years pre-target) to 2012. Rates of pretarget ‘spikes’ and ‘digit preference bias’ were derived from this data. We conducted 68 semistructured interviews about target implementation with ED and hospital clinicians and managers in 2011 and 2012.

Results: Significant ‘spikes’ of the proportion of ED discharges for the time interval immediately prior to the six hour target appeared in all sites after the target’s introduction. At all sites, patients were taken off the target stream (eg moved to short-stay units) to avoid target breaches. However, lower levels of gaming were associated with sites that adopted a ‘whole-of-hospital’ approach to target implementation. Gaming was more prevalent where implementation focused solely on ED staff and processes, without additional resources.

Conclusion: While gaming occurred in all sites, our research demonstrates that a myopic focus on avoiding target breaches within ED accentuates gaming, whereas more holistic approaches to target implementation and additional resources help to reduce it.