Healthcare utilisation and costs associated with different treatment protocols for newly diagnosed childhood acute lymphoblastic leukaemia: a population-based study in Ontario, Canada

Background — Although different treatment protocols for childhood acute lymphoblastic leukaemia (ALL) all achieve high cure rates, their healthcare utilisation and costs have not been rigorously compared.

Methods — Disease, treatment, and outcome data were chart abstracted for all children with ALL in Ontario, Canada, diagnosed 2002-2012. Linkage to population-based databases identified healthcare utilisation. Utilisation-associated costs were determined through validated algorithms. Chemotherapy-associated costs were calculated separately. Healthcare utilisation and costs were compared between patients receiving Children's Oncology Group (COG) versus Dana-Farber Cancer Institute (DFCI)-based treatment.

Findings — Of 802 patients, 146 (18.2%) were treated on DFCI-based protocols. COG patients experienced significantly higher rates of emergency department (ED) visits (adjusted rate ratio [aRR]: 1.3, 95% confidence interval [CI]: 1.1-1.5; p = 0·01), whereas outpatient visit rates were 60% higher among DFCI patients (aRR: 1.6, 95% CI: 1.5-1.7, p < 0.0001). In adjusted analyses, DFCI-associated cost intensity was 70% higher (aRR: 1.7, 95% CI: 1.5-1.9; p < 0.0001), mainly attributable to outpatient visit costs. Total chemotherapy costs were higher among COG-treated patients ($39,400 ± $1100 versus $33,400 ± $2800; p = 0.02). Among PEG-ASNase-treated patients, total chemotherapy costs were highest among DFCI patients (median $54,200 ± $7400; p = 0.003 versus COG patients).

Interpretation — COG and DFCI treatments were associated with higher ED visit rates and higher outpatient visit rates, respectively. Overall utilisation-associated costs were increased in DFCI-treated patients. Administration of some intravenous chemotherapy at home and decreases in PEG-ASNase cost would decrease healthcare utilisation and costs for all patients and mitigate differences between COG and DFCI protocols.