The cost of cancer care is not related to its outcomes

Solid tumours make up 90% of all proliferative diseases and the main action for cure remains surgery, removing the visible tumour as well as the surrounding tissue. Radiotherapy is an added value for eliminating local microscopic as well as regional disease. Systemic treatment has a small impact on the outcome but has a cost, which is as much as all the other actions such as diagnostic tools and treatments.


Background and introduction
The cost of health care is still increasing and the top of the bill is still not in view. New treatments, mainly that of pharmaceutical agents, are exploding [1] and are only effective at a very low level without a clear impact on the population [2].

Methods
Hospital bills of individual incoming patients (n = 637) at the Oncologic Centre UZ Brussel between 01 January 2006 and 30 June 2006 were individually analysed according to the different actions described and summed up to avoid too much detail: consultation, surgery, hospitalisation, day hospital, pharma, medical material, radiology, anatomopathology, clinical biology, radiotherapy, nuclear medicine.
All bills were followed for five years or until the patients died, taking also into account the cost of relapses and palliation. Individual costs such as travelling, work loss etc. were not taken into account. Table 1 shows the individual described costs in Euros of the summed up activities.The main cost is pharmaceutical agents and together with the administration of them (D Hosp) it makes up 49.7% of the total cost. Surgery being the most important curable factor, at least for solid tumours, is only 4.9% and radiotherapy 9.7%. Diagnostic procedures covering radiology, anatomopathology (which is the confirmation of the exact tumour origin), clinical biology and nuclear medicine makes up 17% of the total bill. The initial cost (first year) is 53.4% of the total five-year cost and that for the fifth year is only 8.4%, the latter corresponds mostly to palliative care for relapsing patients. Our patient distribution in order of localisation and percentage is: breast 22%, brain (including metastatic locations 16%; lung 9%, prostate 9%, rectum 6%, head and neck 5%, skin 3%, … The work-up as well as the treatment was done according to international guidelines. The outcomes of our data in breast- [4][5][6], rectum- [7][8], head and neck- [9], lung- [10], and prostate cancer [11][12] were published in several peer reviewed papers and also at the European collaborative level [7,13].

Discussion
The cost of health care is rapidly increasing, mainly in the development of new drugs which have increasingly been focused on 'personalised' medicine in recent years [1]. In the period analysed, there was no major new drug reimbursed by social security and this analysis is certainly no representative of the present day increasing cost mainly because those drugs use immunomodulatory activity [17]. The positive impact of systemic treatment is limited [2]. As for example say in breast cancer,when we observe patients with taxanes and herceptine, accepted as part of routine treatment, in node positive patients on top of hormone treatment like oestrogen positive patients and also with agressive treatments in triple negative patients from 2000 on, the impact on the population is not that obvious at five years (90.1-91%) [3]. If we evaluate precision medicine thus far in many patients (18,000) who have undergone sequencing in the past decade, the number of reported cases are rare and only a few showed a complete response [14]. Moreover, the good responders were already good responders to chemotherapy.
The only proofs of the bonus of 'precision or targeted' medicine are randomised trials and the SHIVA trial. We see treating patients according to their identified mutations versus selected treatment by the physician showed no difference in progression-free survival (PFS) (2.3 veruss 2.0 months) [15]. Even if we have some long living patients like in melanoma, we need to wait longer for the impact to be seen on five year survival. However, the major outcome bonus is related to shift in stage at diagnosis to screening (50%shifted from SII to SI in breast cancer between 1989 and 1999 SEER DATA). The same is observed for colorectal cancer and prostate cancer: since 2000 no bonus has been observed (data not shown). In non-small cell lung cancer NSCLC, a bonus of 3% has been observed since 2000, but this is perhaps because of better diagnostic tools such as PET-CT scan which is now mandatory before starting treatment. Previously, it was described that the bonus of cytotoxic chemotherapy was about 2.1-2.3% [2] and we are now confronted with new molecules at astronomic amounts for a single treatment, of which the bonus is difficult to evaluate on population outcome so far [1,16].

Recently the US Food
For most people arguing for new drugs even by trials outcome, it should be worthwhile to read the paper by L Saltz: 'The Value of Considering Cost, and the Cost of Not Considering Value' [16] and his conclusion 'We need to understand, and we must help our patients, our partners in industry, and our elected officials to understand, the meaning and importance of value, and that there must be upper limits to cost. We must recognise and then work to remove the perverse incentives that impede the healthy functioning of the cancer drug market'.
We here end with Prasad's recent thought on precision medicine: 'We can ask if rhetoric so far outpaces the reality that we risk fooling even ourselves' [18].

Conclusions
Cost effectiveness in cancer care is inversely proportional to outcome, mainly in solid tumours, which from 2000 on has had nearly no impact on the population outcome. Politicians should take this into account when reimbursing health care for cancer.

Conflicts of interest
No conflicts of interest.