European Journal of Cancer: Highlights of Issue 37:14

Economics of cancer.

The articles in this special issue were developed from abstracts originally presented at the Second European Conference on the Economics of Cancer, which was arranged by the EORTC Health Economics Unit and took place in Brussels, September 3 - 5, 2000.

The economics of cancer is becoming increasingly important with the development of new treatment modalities, techniques, increased numbers of national screening programmes being set up etc. The balance between cost and efficacy is a delicate one and in some cases where efficacy is the same, cost can be a deciding factor. Cost-effectiveness is one factor that governments consider when deciding which population groups should be routinely screened and how often the invitations should be sent out. Such examples illustrate the importance of economic analyses in oncology.

In this issue, several areas of current interest are covered.

Neymark and colleagues describe a retrospective cost-effective analysis, in the context of the French health care system, of data from the European Organization for research and treatment of cancer (EORTC) randomised trial 22863. This trial randomised patients with locally advanced prostate cancer to receive either radiotherapy (RT) alone (50Gy over 5 weeks with a prostatic boost of 20Gy over 2 weeks) or RT and goserelin (3.6mg subcutaneous injection every 4 weeks for 3 years). Mean survival time was prolonged by approximately one year in the combined treatment group, as assessed by a restricted means method with the time point of restriction determined by statistical criteria and the costs were reduced by approximately 12 700 French Francs. This suggests that the combined therapy should be considered the preferred treatment as a better outcome is obtained at lower costs. However, the exact results (but not the direction) of the economic evaluation are decisively determined by the time point chosen for assessing the difference between the two treatments.

Buijs-van der Woude and colleagues compare the costs of stereotactic large-core-needle biopsy (a less invasive technique) and surgical excision biopsy of breast lesions in the Netherlands. Using cost data from 5 hospitals in an economies of scale analysis and looking at 4 different scenarios of decentralisation, they determined that in a decentralised scenario (all 114 hospitals in the Netherlands having stereotactic equipment) the costs were similar (1186 Euros compared with 1184 (for excisional biopsy). If the stereotactic equipment was centralised in 10 hospitals, a reduction of approximately 50% in costs (572 Euros) would be obtained. They calculated patient travelling costs to be minimal and therefore recommend that the centralisation of stereotactic equipment be considered from an economic perspective.

Bernhard and colleagues study whether a 'reframing' affects the perception of health status for use in utility evaluations using data from the Swiss Group for Clinical Cancer Research (SAKK) 40/93 trial of patients with adenocarcinoma of the colon. The patients had a radical resection and perioperative chemotherapy and then were randomised to either observation, 5-fluorouracil (5-FU) or 5-FU and levamisol adjuvant chemotherapy. The patients estimated their health before surgery, as well as before adjuvant treatment, to be worse than when assessed retrospectively after treatment (the 'pre' score compared with the 'then' score) suggesting patients 'reframe' their assessments. Analysis of the change in scores for the 'then' adjuvant score compared with post-adjuvant treatment score suggested an improvement. The authors state that 'given that patients may change their standards, comparisons of health estimates across different populations and clinical situations are to be interpreted with caution'.

Van der Pol and colleagues estimate the additional demand and subsequent cost increase of extending the invitation to breast cancer screening in Scotland to include those aged 65-67 years old. At present, screening in the UK is three-yearly and restricted to those aged 50-64 years old. Modelling of predicted attendance, using a logistic regression analysis, was successful as most of the variables had the expected negative or positive sign. Independent variables included were the woman's screening history and deprivation score (assessed by the Carstairs deprivation index derived from data from the 1991 Census), as well as screening details ie static site or mobile van and the time of year of the invitation. The authors estimated that an additional 10 829 women would be screened in 2001 with a subsequent increase in costs of £350 000.

On a related theme, Frew and colleagues, in a questionnaire-based study, examine the willingness to pay (WTP) of subjects in the Trent region of east-central England for two types of screening for colorectal cancer; faecal occult blood (FOB) testing and a flexible sigmoidoscopic procedure, 'Flexi-Scope' (FS). Two types of questionnaire were randomly used; an open-ended format where subjects were offered spaces to fill in the amount they were willing to pay for the screening test and a payment scale format where values from £0-£1000 in various scales were listed for the subject to choose from. Approximately 2000 questionnnaires were assessed and regression models suggested that the WTP was affected by many factors including age, gender, income, risk perceptions, illness experiences and general health beliefs. The median WTP was £30 or £50 depending on the type of format used in the questionnaire and these values were similar to the likely costs of the procedures.

In the UK, mass population screening for cervical cancer has been implemented since the 1960s. The current national screening programme recommends withdrawal from the programme at 65 years of age. It has been suggested that where there is a favourable previous screening history associated with a low risk ie three previous negative screens and/or the subject is human papilloma virus (HPV)-negative at 50 years of age, this age of withdrawal could be lowered. Philips and colleagues examine this idea by means of Monte Carlo simulations to determine the cost effectiveness ratios for various scenarios of earlier withdrawal. The authors conclude that their 'findings offer little support for the early withdrawal of subjects from the cervical cancer screening programme'.

Van Agthoven and colleagues compare the costs of patients randomised (in a 2:1 ratio) to receive either peripheral blood stem cell transplantation (PBSCT) or autologous bone marrow transplantation (ABMT) for refractory/relapsed non-Hodgkin's lymphoma (NHL) or Morbus Hodgkin's (MH) disease. They examined the costs from the induction of chemotherapy to 3 months after discharge from hospital in 91 patients (62 PBSCT and 29 ABMT). Patients' quality of life (QoL) was also compared between the two groups using the EuroQoL, Rotterdam Symptom Checklist (RSCL) and SF-36 instruments. The costs were found to be significantly lower in the PBSCT group (by approximately 15% or approximately 21% if the costs were considered from the harvesting phase). Using the RSCL, the QoL was also significantly better for the PBSCT patients suggesting that 'PBSCT is the treatment of choice for patients with refractory or relapsed NHL/MH'.

Johnson and colleagues have modelled the future breast cancer and non-breast cancer costs associated with screening. Patients were stratified using the Nottingham Prognostic Index (NPI) into 5 groups: ductal carcinoma in situ (DCIS), poor, moderate, good and excellent prognosis groups, and a Markov model was used to estimate the future breast cancer and non-breast cancer associated costs in each group. As might be expected, the breast cancer costs increased with poorer prognosis, whereas the reverse pattern was seen for the non-breast cancer associated costs. However, the average total future costs were similar for the prognostic groups at approximately £10 000 to £11 000. As a percentage of the screening costs, there was a 20.9% saving in future costs, but this saving disappeared if the non-breast cancer costs were included. It is still undecided whether such non-breast cancer costs should be included especially as the authors state that the 'inclusion of such costs implies it is almost always more cost-effective to do nothing rather than screen for breast cancer and attempt cure'. The debate is certain to continue!

Lung cancer has a very poor prognosis if diagnosed at a late stage. 5-year survival rates of 2% have been reported for patients with distant disease compared with 50% for patients with localised disease. Screening in high-risk populations could therefore potentially allow patients to be identified with disease at an earlier stage with a subsequent impact upon survival. However, screening tests using sputum and chest X-ray techniques have not proven effective. The development of a new technique, low dose helical computed tomography, is a promising and relatively low cost modality for the early detection of lung cancer. Marshall and colleagues used a decision analysis model to look at the cost-effectiveness of screening a high-risk cohort aged 60-74 years yearly for 5 years. Using combined data from the Surveillance, Epidemiology and End Results (SEER) registry public-use database and the Early Lung Cancer Action Project (ELCAP), they demonstrated that such a programme would be cost-effective at approximately US $19 000 per life year saved. Adjusting for biases increased the costs somewhat to US$ 61 723, but this was still relatively cost-effective when compared with figures in the literature for breast, colorectal and prostate cancer screening. Sensitivity analyses changing the base-line assumptions for the incidence rates and the costs of screening and follow-up had little effect on the cost-effectiveness ratio. The authors suggest screening such cohorts may be feasible under optimal conditions, but state that further study is needed.

Finally, Will and colleagues report on the use of Canada's population health model (POHEM) as a tool for economic assessment of interventions to control cancer. They discuss three examples where POHEM has been implemented; in a cost-effectiveness analysis of chemotherapy for advanced non-small cell lung cancer, in assessing the impact of a reduced length of stay in hospital following surgery for breast cancer and in a cost-effectiveness analysis of the provision of preventive tamoxifen to women at a high risk of breast cancer. As the authors discuss, the POHEM data from the examples listed above has influenced several policy decisions such as the funding of vinorelbine and gemcitabine for advanced lung cancer treatment in the province of Ontario. Thus, this paper emphasises the concept behind all the papers in this issue that, despite the drawbacks of some modelling techniques, economic evaluations of health care interventions against cancer is an important additional tool that can help decision-making in oncology.

Back . . .