Oral Presentation Clinical Oncology Society of Australia Annual Scientific Meeting 2021

Variations in Chemotherapy Prescribing Practices in New South Wales (#154)

Nasreen Kaadan 1 2 , Geoff Delaney 2 3 , Eugene Moylan 1 , Stephen Della-Fiorentina 4 , Winston Liauw 5 , Pirkko Boyd 4 , Graeme Bell 6 , Brett Ly 7 , Joseph Descallar 2 8 , Rachel Nixon 9
  1. Liverpool Cancer Therapy Centre, Liverpool Hospital, Liverpool, NSW, Australia
  2. Ingham Institute for Applied Medical Research, Liverpool, NSW, Australia
  3. South Western Sydney Local Health District, Liverpool, NSW, Australia
  4. Macarthur Cancer Therapy Centre, Campbelltown Hospital, Campbelltown, NSW, Australia
  5. Cancer Care Centre, St George Hospital, Kogarah, NSW, Australia
  6. Illawarra Shoalhaven Cancer & Haematology Network, Wollongong Hospital, Wollongong, NSW, Australia
  7. Nelune Comprehensive Cancer Centre, Prince of Wales Hospital, Randwick, NSW, Australia
  8. South Western Sydney Clinical School, UNSW, Liverpool, NSW, Australia
  9. District Cancer Services, Western NSW Local Health District, Orange, NSW, Australia

Aims: This study was undertaken following the 2016 parliamentary inquiry into off-protocol prescribing of chemotherapy in NSW Public Hospitals. The aims were to quantify the proportion of cycle one day one chemotherapy drug orders that are varied during routine clinical practice, to identify the reasons for variations and to identify predictors for dose variation.

Methodology: The study was conducted in 11 chemotherapy treating facilities in New South Wales for cycle one day one chemotherapy drug orders given between 01/01/2016 and 31/12/2018. Dose variance was defined as an ordered dose greater than ±5% of the calculated dose. Study data was extracted from the oncology information systems using transact-SQL embedded in SAP® Crystal Reports. Variables analysed included drug dose, drug order year, eviQ protocol, tumour site, stage of disease, age, treatment intent, gender, and indigenous status. Proportions and binomial confidence intervals were used to calculate the dose variation frequency and reasons for variation. Multivariable generalised estimated equation models were used to determine predictors of dose variation.

Results: There were 8,289 patients, 11,182 care plans and 19,747 drug orders in the study dataset. 26.3% of patients had at least one drug dose variance. 16.7% of all drug orders had a dose variation and 42% had a documented reason in a structured data field. Dose variation factors reaching statistical significance included order year (OR=0.77 (95% CI (0.69, 0.86), p < .0001), tumour site (OR=0.33 (0.25, 0.43), p < .0001), stage of disease (OR=0.38 (0.22, 0.66), p < .0006), treatment intent (OR=0.49 (0.44, 0.55), p < .0001) and age group (OR=0.54 (0.42, 0.69), p < .0001).

Conclusion: This is the first Australian study to document the frequency of initial chemotherapy dose reductions using routinely captured data in oncology information systems. Further analysis is warranted to assess impact of dose variation on treatment response and outcomes.