Addition of Docetaxel to First-line Long-term Hormone Therapy in Prostate Cancer (STAMPEDE): Modelling to Estimate Long-term Survival, Quality-adjusted Survival, and Cost-effectiveness

Abstract

Background: Results from large randomised controlled trials have shown that adding docetaxel to the standard of care (SOC) for men initiating hormone therapy for prostate cancer (PC) prolongs survival for those with metastatic disease and prolongs failure-free survival for those without. To date there has been no formal assessment of whether funding docetaxel in this setting represents an appropriate use of UK National Health Service (NHS) resources.

Objective: To assess whether administering docetaxel to men with PC starting long-term hormone therapy is cost-effective in a UK setting.

Design, setting, and participants: We modelled health outcomes and costs in the UK NHS using data collected within the STAMPEDE trial, which enrolled men with high-risk, locally advanced metastatic or recurrent PC starting first-line hormone therapy.

Intervention: SOC was hormone therapy for ≥2 yr and radiotherapy in some patients. Docetaxel (75mg/m²) was administered alongside SOC for six three-weekly cycles.

Outcome measurements and statistical analysis: The model generated lifetime predictions of costs, changes in survival duration, quality-adjusted life years (QALYs), and incremental cost-effectiveness ratios (ICERs).

Results and limitations: The model predicted that docetaxel would extend survival (discounted quality-adjusted survival) by 0.89 yr (0.51) for metastatic PC and 0.78 yr (0.39) for nonmetastatic PC, and would be cost-effective in metastatic PC (ICER £5514/QALY vs SOC) and nonmetastatic PC (higher QALYs, lower costs vs SOC). Docetaxel remained cost-effective in nonmetastatic PC when the assumption of no survival advantage was modelled.

Conclusions: Docetaxel is cost-effective among patients with nonmetastatic and metastatic PC in a UK setting. Clinicians should consider whether the evidence is now sufficiently compelling to support docetaxel use in patients with nonmetastatic PC, as the opportunity to offer docetaxel at hormone therapy initiation will be missed for some patients by the time more mature survival data are available.

Patient summary: Starting docetaxel chemotherapy alongside hormone therapy represents a good use of UK National Health Service resources for patients with prostate cancer that is high risk or has spread to other parts of the body.

Keywords: Cost-effectiveness analysis; Docetaxel; Prostate cancer.

Fig. 1

Model structure. Patients start treatment in the hormone-sensitive health state and then progress to the castrate-resistant prostate cancer (CRPC) states. At treatment failure, patients enter the CRPC state that reflects their worst previous disease event (with the worst event being visceral metastases, then bone metastases with history of a skeletal-related event [SRE], then bone metastases without an SRE, then CRPC with no metastases or only lymph-node metastases). Further events can cause movement to more severe health states. Death due to prostate cancer or non–prostate cancer is possible from any of the health states (not shown for parsimony).

Fig. 2

Predicted patient prognosis over time. (A) Overall survival for patients with M0 disease. (B) Proportion of patients with M0 disease receiving standard of care (SOC) by health state. (C) Difference in proportion of patients with M0 disease in each health state (SOC + docetaxel (Doc) minus SOC). (D) Overall survival for patients with M1 disease. (E) Proportion of patients with M1 disease receiving SOC by health state. (F) Difference in proportion of patients with M1 disease in each health state (SOC + Doc minus SOC). CRPC = castrate-resistant prostate cancer; SRE = skeletal-related event, M0 = nonmetastatic, M1 = metastatic. The grey shaded area denotes the duration of patient follow-up in STAMPEDE.

Fig. 3

Impact of baseline characteristics, health state, and treatment allocation on patient health-related quality of life (HRQOL) in STAMPEDE. This graph presents the results of an analysis of EQ-5D data obtained from STAMPEDE adjusted for baseline characteristics, treatment allocation, and current health state. Data were collected at baseline and at follow-up visits: every 6 wk for the first 6 mo, then every 12 wk up to 2 yr, every 6 mo up to 5 yr, and annually thereafter. The impact of each covariate on HRQOL is shown relative to a reference patient with nonmetastatic disease, World Health Organisation class 0, age ≤60 yr, and node-negative in their first year of standard of care. Positive values indicate better and negative values indicate worse HRQOL relative to the reference patient. CRPC = castrate-resistant prostate cancer; SRE = skeletal-related event, M0 = nonmetastatic, M1 = metastatic.