Poster Presentation NSW State Cancer Conference 2023

Data-driven dietetics dashboards in the digital era: Visualising real-world, evidence-based nutrition care and outcomes for patients with head and neck cancer (#211)

Merran Findlay 1 2 3 4 5 6 , Georgina Kennedy 1 2 , Angela Sita 7 , Tim Churches 2 , Geoff P Delaney 1 2 8 , Katherine Bell 8 , Joanna Fardell 1 2 , Judith D Bauer 9 , Meera Agar 1 2 10
  1. Maridulu Budyari Gumal (SPHERE) Cancer Clinical Academic Group, University of NSW, Sydney, NSW, Australia
  2. South Western Clinical School, University of NSW, Sydney, NSW, Australia
  3. Chris O'Brien Lifehouse, Sydney, NSW, Australia
  4. Cancer Services, Royal Prince Alfred Hospital, Sydney Local Health District, Sydney, NSW, Australia
  5. Cancer Care Research Unit, Susak Wakil School of Nursing and Midwifery, The University of Sydney, Sydney, NSW, Australia
  6. The Daffodil Centre, The University of Sydney - a joint venture with Cancer Council NSW, Sydney, NSW, Australia
  7. Centre for Big Data Research in Health, University of NSW, Sydney, NSW, Australia
  8. Liverpool Cancer Therapy Centre, South Western Sydney Local Health District, Liverpool, NSW, Australia
  9. Department of Nutrition, Dietetics and Food, Monash University, Melbourne, VIC, Australia
  10. Faculty of Health, University of Technology Sydney, Ultimo, NSW, Australia

Purpose

Healthcare dashboards visualise patient-level and aggregate data to guide decision-making and evaluate care delivery and outcomes. Nutritional status and involuntary weight loss negatively impact outcomes yet are not presented in automated cancer dashboards. We aimed to explore the technical feasibility of extracting and visualising near real-time evidence-based nutrition data through development of automated nutrition care dashboards for patients with head and neck cancer (HNC).

Methodology

The Cancer Variation (CaVa) data platform extracts and harmonises data from South Western Sydney Local Health District clinical information systems, including key named entities from free text clinical notes using Natural Language Processing (NLP). Novel nutrition data was assessed for quality, completeness, generalisability and alignment with nutrition outcomes and quality metrics against other prognostic factors including diagnostic and treatment episodes, dietetic resource utilisation and best-practice nutrition care in near real-time.

Results

Comprehensive data were available (2012 to 2021) for patients with a primary diagnosis of HNC (n=942; 74% male, mean age 64yrs). Data captured included: valid TNM stage (92%); treatment modality delivered as radiotherapy (74%) or systemic therapy (42%); weight at median (Q1, Q3) timepoints 13 (7, 20) and at least one consult with a dietitian (79%) or a speech pathologist (85%) providing median (Q1, Q3) episodes of care 9 (5, 13) and 7 (3, 12) respectively. Prototype dashboards created to assess feasibility prior to end-user co-design will be presented.

Conclusion

Timely visualisation of evidence-based nutrition care processes and demonstrated prognostic nutrition outcomes by multidisciplinary teams is feasible. Future iterations will examine nutritional status and outcome variation involving treatment toxicities and unplanned admissions. Adoption of automated nutrition care dashboards in routine care holds potential to inform decision-making and improve patient care and outcomes.