Sarcomas are rare tumours (accounting for only 1-2% of all cancers) originating from connective tissue, skin, retroperitoneum, bone and viscera 1 . The rarity of the disease, along with the variety of histological types and locations and the heterogeneity of prognostic factors associated with local or distant spread, mean that physicians have only limited personal experience of managing the disease. Furthermore, outside centres of excellence, there is little graduate or post-graduate medical training in its optimum management. To improve the diagnosis and prognosis of sarcoma, the European Commission funded the Connective Tissue Cancer Network (CONTICANET) aimed at increasing the standardization of diagnostic and therapeutic procedures.

In order to reduce inappropriate medical procedures, Clinical Practice Guidelines (CPGs) were developed by the Fédération Nationale des Centres de Lutte contre le Cancer (the French Federation of Comprehensive Cancer Centres) 2 and by the Italian National Research Council 3 . France and Italy reached a consensus in their CPGs relating to all phases of sarcoma management (initial examination and diagnosis, histopathological report, surgery, chemotherapy, and radiation therapy) except surveillance after therapy (see Annexe 1).

The impact of adherence to CPGs has received some research attention 4 5 6 7 8 9 10 , even in the management of rare cancers 11 12 13 14 . However, little is known about the financial impact of clinicians' adherence to CPGs in general, and the impact of adherence on outcomes and costs of care has only rarely been simultaneously considered 15 . Such assessments are of particular value in the current period of budgetary constraint, which prevents the achievement of improved cancer outcomes through increased health expenditure 16 17 .

We therefore assessed the cost-effectiveness of compliance with CPGs in sarcoma management by investigating the relationship between health outcome and resource consumption in patients treated in the regions of the Rhône-Alpes in France and Veneto in Italy.

Table 3 shows the key characteristics of patients and the health outcome in the whole of the study population (219 cases) and in the CPG + (n = 118) and CPG- (n = 101) groups. Overall, age at diagnosis, ranging from 18 to 94 years, was in average 60 years and the majority of patients studied (55%) were women. There was a variety of histological types, the most frequent being Gastro Intestinal Stromal Tumors(GIST) (26%), liposarcomas (21%) and leiomyosarcomas (12%). Sixtythree percent of tumors were of soft tissue, mostly in the trunk, and 81% deep seated. Tumors had a mean size of 95 mm (range 3 to 450 mm). Tumors were grade II/III in 58% of cases, Thirtytwo patients (15%) had previous cancers. Differences between CPG compliant and non compliant groups were significant for grade (there were more high grade tumors in patients whose management did not comply with CPGs, p = 0.01) and histological type (p = 0.03): CPGs was less likely in liposarcoma and leiomyosarcoma. Relapse free survival was a mean of 2.46 years in the CPG + group and 2.16 years in the non-compliant group (p = 0.04).

Table 4 shows details of the mean costs in CPG + and CPG- groups. The overall (for all 219 cases) cost of sarcoma management per patient, ranging from €1,303 to €107,464, reached €25,296 in average. Surgery (primary and wide surgical excisions), with an average cost of €8,170 represented 32% of the average total cost. Chemotherapy (€6,107) accounted for 24% of the average total cost, follow up (€5,048) for 20% of the average total cost and diagnosis (€3,701) for 15%. Radiotherapy (€2,270) represented 9% of the average total cost. Overall average cost per patient of managing sarcomas (all subtypes included) was somewhat less in the GPG + group than in the CPG- group, but the difference did not reach statistical significance (p = 0.07). Overall average cost was significantly lower in the GPG + group than in the CPG- group for GIST and dermatofibrosarcoma (p < 0.01).

Table 5 shows that compliance with CPGs was less costly and more effective than non compliance. On the basis of the ICER point estimate, clinicians' adherence to CPGs in the management of patients with sarcoma strictly dominates non adherence for relapse free survival. Since the origin of the cost-effectiveness plane presented in Figure 1 was not included in the inner 95% confidence ellipse, clinicians' adherence versus non adherence to CPGs is significantly different in terms of cost and effectiveness. The probability of the ICER belonging to each quadrant of the cost-effectiveness plane is highest (75%) for the southeast quadrant, in which compliance with CPG is both less costly and more effective than non compliance. This is in accordance with the results of Table 5 also concluding for less costly and more effectiveness in compliant group.

A weakness of the present study is that its retrospective nature did not allow assessment of outcome in cost per Quality Adjusted Life Year (QALY), as recommended by the National Institute for Health and Clinical Excellence (NICE) 24 25 . Notwithstanding this, freedom from relapse undoubtedly contributes to quality of life in cancer patients.

Another weakness was the lack of information on potentially confounding variables such as the initial performance status of the patient, the cumulative volume for the surgeon and the hospital, the type of provider (academic cancer centre, non academic cancer centre, non academic non cancer centre) 26 27 28 that could not be controlled for in the statistical analysis. Any resulting bias, however, would have tended to cause an underestimation rather than an overestimation of the true effect of compliance, since in the CPG + group the costs of diagnosis were higher (Table 4) despite a lower grade of tumours (Table 3).

Interestingly, the higher overall cost involved in diagnosing patients according to CPGs was accounted for almost entirely by increased hospitalisation (rather than by greater use of imaging or biopsy or external consultations) (Table 4). The average cost of hospitalization was €4,097 in CPG + and €2,317 in CPG- groups. The difference corresponds to 2.3 additional days of hospital stay at the prevailing daily cost of €760 per day (Table 2). Adherence to CPGs requires that decisions be made within a multidisciplinary committee, and it is likely that patients' hospitalization was prolonged in order to schedule the multidisciplinary meeting needed to reach a consensus on treatment.

A greater adherence to CPGs during diagnosis appears to decrease all subsequent costs, notably those of surgery and chemotherapy, probably by reducing the need for surgical re-intervention or more intense use of antineoplastic drugs (Table 4). Surgery and chemotherapy were the main factors driving the cost of initial treatment, representing about 66% of the overall cost in the CPG + group ((7,397 + 5,164)/19,175) and 76% ((9,075 + 7,207)/21,502) in the CPG- group.

In contrast to other studies that have shown it may be cheaper in the short term to deviate from CPGs 29 , the present study found that compliance with CPGs strictly dominates for relapse-free survival, meaning that management of sarcoma according to CPGs is less costly and more effective. This evidence should encourage health providers to promote adherence to guidelines since adopting this approach achieves a better quality of care and is a more efficient allocation of resources. It should also encourage compliance among physicians, some of whom perceive CPGs as limiting their freedom to make diagnostic and therapeutic decisions.

Compliance with CPGs in sarcoma management has increased since 2001. A retrospective study of the medical records of sarcoma patients in the University Hospital of Lyon and the Cancer Centre Léon Bérard found that between 1999 and 2001 initial clinical management had been consistent with the CPG in only 32% of cases 12 . In this study, the compliance rate reached 54%. Compliance with guidelines was greater for chemotherapy than for other aspects of sarcoma management since detailed protocols are more common than in diagnosis, surgery, radiotherapy or follow up. Compliance with CPGs was also more frequent for low-grade than for high-grade sarcomas (p = 0.01); and it has been suggested that physicians are less likely to adhere to CPGs if they believe that compliance will not improve outcome 30 . Compliance was also better for GIST than for other histologies, probably because imatinib and related drugs have led to a remarkable improvement in management on which there is global consensus 31 .

The present findings should encourage physicians' efforts to increase their compliance with CPGs and encourage healthcare administrators to invest in implementing CPGs in the management of sarcoma.

ANOVA: Analysis of variance; CEA: Cost-effectiveness analysis; CPG: Clinical practice guidelines; CPG +: Clinicians' adherent to practice guidelines; CPG-: Clinicians' non adherent to practice guidelines; CONTICANET: Connective tissues cancers network; GIST: Gastro Intestinal Stromal Tumours; HAS: French National Authority for Health (Haute Autorité de Santé); ICER: Incremental cost-effectiveness ratio; NICE: National Institute for Health and Clinical Excellence; NOS: Not otherwise specified; QALY: Quality adjusted life year

The authors declare that they have no competing interests.

LP, AB, GM designed the study, acquired and interpreted the clinical and cost data, undertook the statistical analysis, and prepared the manuscript. AV, PS, FD participated in clinical data acquisition and analysis. FNG carried out surgical data acquisition and analysis, PB chemotherapy data acquisition and analysis, DRV, AVD, APDT, JMC, and PT diagnosis data acquisition and analysis, and CB the analysis of data from young patients. CS participated in the statistical analysis. JYB, CRR and IRC participated in general CONTICANET coordination as well as study design and compliance supervision. All authors read and approved the final manuscript.