Cost-effectiveness analysis (CEA) of IMRT plus C12 boost vs IMRT only in adenoid cystic carcinoma (ACC) of the head and neck

All patients with pathologically confirmed ACC of the head and neck treated with curative intent at Heidelberg University Hospital between 1997 and 2009 were entered into the clinical analysis. Patients undergoing re-irradiation were excluded. The dataset contained information on patient baseline characteristics, initial treatment, time and incidence of locoregional or distant relapse, toxicity, last contact or death, which was previously extracted from individual patient records. Information on further treatment regarding the underlying disease was available at a basic level [13].

Within the German C12 pilot project between 1997 and 2008, C12 was clinically available in three treatment periods (20 days) per year; outside these periods, patients were treated with standard IMRT by the same team of physicians. Selection of patients into either of the cohorts was solely based on availability of C12 and hence time of initial presentation. Both cohorts are similar in their baseline characteristics and offer the unique possibility to analyse two comparable patient groups as a quasi-randomized study population. Treatment in the C12 cohort (58 patients) consisted of 30 fractions photon intensity-modulated radiotherapy (IMRT) plus 6 fractions C12 whereas the other cohort (37 patients) received 32 fractions photon IMRT. Follow-up recommendations were identical for both groups.

The analysis assesses cost-effectiveness from a health care purchaser’s perspective as represented by statutory German sickness funds. Only direct health service costs were evaluated for the duration of available follow-up. Indirect costs such as family/ caretaker leisure time, productivity losses, patient travel or hotel costs are disregarded. Costs are in € at the 2015 price level, price changes over time due to i.e. inflation were not considered.

Cost data were derived from individual patient histories through number and category of events, resource utilization, and attributable unit costs. However, only events related to the initial tumour diagnosis were recorded. Information on supportive treatments (dietician visits, physiotherapy, pain medication) or any non-cancer related problems were not included.

Recorded resource utilization was either associated with the standard follow-up such as diagnostic procedures (CT or MRI) or associated with treatment of either local or distant tumour recurrence, therefore derived costs are limited to tumour-related health care costs.

Individual follow-up examinations were not explicitly recorded. Type of examination, time and frequency of diagnostic follow-up procedures were derived from the institution’s standard follow-up scheme [15]. Information on further treatment was available at a basic level: salvage surgery, re-irradiation and type of radiotherapy, chemotherapy or other treatments and did not include exact details or specifics on these procedures. To attribute costs, procedures were derived from standard operating procedures and expert advice. For salvage surgery, details were derived from initial tumour site and discussed with experienced head and neck as well as maxillofacial surgeons. For re-irradiation, it was assumed that patients underwent a complete second course with adequate radiation dose corresponding to approximately 30 fractions. For palliative radiotherapy, the institutional standard protocols prescribe 30 Gy in 10 fractions using standard techniques both for palliative bone and whole brain irradiation. For chemotherapy, international guidelines [16] recommend only one standard palliative regimen for recurrent or metastatic ACC consisting of a triple drug combination [17]. Patient care during chemotherapy was discussed with an experienced medical oncologist to identify i.e. appropriate number of visits, lab checks, obligatory co-medication and supportive treatment. Items therefore reflect the institution’s current clinical practice. While there are a few experimental second line chemotherapy regimens available, no consensus exists regarding their routine use on failure of first line chemotherapy. No information on second-line treatments was available for patients in this analysis hence only first line chemotherapy was modelled. Any potential second line treatment was disregarded.

Following attribution and detailed characterization of treatments or procedures to individual patient events, unit costs were extracted from official 2015 fee schedules. It was assumed that all treatments except surgical procedures were carried out on outpatient basis where costs are calculated according to uniforme reimbursement (EBM) specialist catalogues 2015 [18, 19]. University outpatient flat fees negotiated with the Association of Statutory Health Insurance Physicians are not considered. Costs for in-hospital procedures were derived from the German DRG reimbursement scheme [20‐22].

C12 radiotherapy reimbursement is subject to negotiation with statutory sickness funds who kindly provided this information [23].

Drug costs for chemotherapy were obtained from AiDKlinik® search engine [24] as official pharmacy prices, potential quantity discounts for large institutions were not taken into account. All costs are reported in 2015 values (€).

Current reimbursement for C12 is a flat fee of €14,000.- [23], together with the IMRT part, the costs of the experimental treatment add to a total of € 16,877.67 vs. € 5675.27 in the standard treatment, cost categories and unit costs identified from individual patient histories are listed in Table 2.

Mean costs were calculated undiscounted and discounted at 3 and 3.5%. Mean difference in total costs is € 16,937 undiscounted vs. €16,035 and € 15,900 discounted at 3 and 3.5% respectively. Heavy censoring was present in the IMRT+C12 group with 55% incomplete cost observations, hence the Lin1 method was used to calculate adjusted mean differences ranging between €21,051 (3.5% discount rate) and €23,624 (undiscounted) (Table 3).

Mean difference in effects is 0.86 years (undiscounted) vs. 0.83 and 0.82 years using an annual discount rate of 3 and 3.5% respectively giving rise to incremental cost-effectiveness ratios of 20,854 €/LY and 22,078€/LY; adjusted values are 26,929 €/LY (3.0% discount) and 26,863 €/LY (3.5% discount).

The net monetary benefit (NMB) as a function of the decision maker’s willingness to pay based on adjusted and discounted ICERs is shown in Fig. 1. When willingness to pay exceeds € 26.863, the NMB turns positive and the experimental treatment becomes acceptable.

In the experimental treatment, re-irradiation on local relapse was more frequently offered as another course of C12 as compared to the standard treatment (Table 1). Costs for re-irradiation with C12 are substantial; hence this potential confounder was addressed. Calculations were performed for two scenarios: the first scenario assumes that all patients undergoing re-irradiation received this treatment as standard treatment (IMRT). The second scenario assumes re-irradiation always as C12. Corresponding costs were replaced in individual patient histories accordingly (Table 4).

ICER calculations suggests that the differing management of local relapse did indeed influence costs and hence the ICER. Both scenarios show a lower ICER with scenario one (re-irradiation as IMRT in all cases) showing the lowest value at 20,638 €/LY (Table 5).