ORAHS 2015

HEC Montréal, July 19 - 24, 2015

/system/images/000/000/306/logos3_default.png

ORAHS2015

HEC Montréal, July 19 — August 31, 2015

Schedule Authors My Schedule
Cal add eabad1550a3cf3ed9646c36511a21a854fcb401e3247c61aefa77286b00fe402

TuD3 Disease Modeling and Policy 1

Jul 21, 2015 04:00 PM – 05:30 PM

Location: Banque Scotia

Chaired by Martin L. Puterman

3 Presentations

  • Cal add eabad1550a3cf3ed9646c36511a21a854fcb401e3247c61aefa77286b00fe402
    04:00 PM - 04:30 PM

    Predictive capabilities in hierarchical node-based clustering of flu time series

    • Hootan Kamran Habibkhani, presenter, University of Toronto
    • Dionne M. Aleman, University of Toronto
    • Michael Carter, University of Toronto
    • Kieran Moore, Queens University

    Time series are sometimes expected to exhibit local behavior patterns. Depending on the features of signals, a particular clustering may yield the most significant predictive capabilities. We focus on the predictive capabilities that arise from identifying significantly lagged effects between clusters, and use a flu dataset from 103 hospitals across Ontario to identify clustering schemes with maximum predictive capabilities.

  • Cal add eabad1550a3cf3ed9646c36511a21a854fcb401e3247c61aefa77286b00fe402
    04:30 PM - 05:00 PM

    Managing the Additional Demand for Colonoscopy Services following the Introduction of Average Risk Colorectal Cancer Screening: A Comparison of Two Strategies

    • Leslie Anne Campbell, presenter, Dalhousie University
    • John Blake, Dalhousie University
    • George Kephart, Dalhousie University

    Population-level average risk colorectal cancer screening may be an important strategy for the control of colorectal cancer. However, the introduction of even two-step screening substantially increases the overall demand for colonoscopy services both to provide screening follow-up colonoscopies for positive stool tests as well as for ongoing surveillance. A discrete event simulation model, the Simulation of Cancer Outcomes for Planning Exercises (SCOPE) Model, was constructed to compare the effects of various colorectal cancer screening decisions on demand for colonoscopy services, crude colorectal cancer incidence, and cumulative colorectal cancer mortality. Total demand for follow-up screening, diagnostic, high-risk screening and surveillance colonoscopy services increased by 33% to 54%, depending on the screening stool test and uptake rate. This additional demand was not offset by reductions in demand for diagnostic colonoscopies due to lower disease incidence with screening. Inadequate colonoscopy resources led to poor outcomes for both average- and high-risk populations. Two strategies aimed at mitigating the additional demand were compared for their effect on colonoscopy services, colorectal cancer incidence and mortality. The first scenario consisted of following up an initial positive stool test with a repeat test prior to referral for colonoscopy. The demand for screening follow-up colonoscopies was reduced by 90%. In the second scenario, selection of a stool test with a higher positivity threshold reduced the demand for screening follow-up colonoscopies by 65%. In both scenarios, many of the benefits of screening were maintained, with reductions in both crude colorectal cancer incidence and cumulative colorectal cancer mortality after 15 years of follow-up. Screening programs may therefore wish to consider strategies to mitigate the additional demand for colonoscopy services to take advantage of the potential benefits of screening without overwhelming colonoscopy services and causing unintended harm.

  • Cal add eabad1550a3cf3ed9646c36511a21a854fcb401e3247c61aefa77286b00fe402
    05:00 PM - 05:30 PM

    Improving Access to Cancer Treatment

    • Martin L. Puterman, presenter, Sauder School of Business, UBC
    • Claire Ma,
    • Leah Weber,
    • Emma Liu,
    • Antoine Sauré,
    • Scott Tyldesley,

    Operations research methods have been used extensively to address challenges facing British Columbia Cancer Agency management when striving to deliver high quality and timely care to cancer patients. Reducing wait times for initial oncologist consultations required accounting for demand variability, the multiplicity of cancer types and urgency levels, oncologist specialization mix and downstream demand generated by initial appointments. Optimization and simulation models showed the impact of managerial levers on relevant performance metrics and produced concrete capacity management recommendations to reduce wait times. A key finding was that without adequate capacity, enhanced appointment scheduling rules had little impact on wait times for first consultation.

Back