Scholarly article on topic 'Radiotherapy equipment and departments in the European countries: Final results from the ESTRO-HERO survey'

Radiotherapy equipment and departments in the European countries: Final results from the ESTRO-HERO survey Academic research paper on "Economics and business"

CC BY-NC-ND
0
0
Share paper
Academic journal
Radiotherapy and Oncology
OECD Field of science
Keywords
{"Radiotherapy equipment" / "Health economics"}

Abstract of research paper on Economics and business, author of scientific article — Cai Grau, Noémie Defourny, Julian Malicki, Peter Dunscombe, Josep M. Borras, et al.

Abstract Background Documenting the distribution of radiotherapy departments and the availability of radiotherapy equipment in the European countries is an important part of HERO – the ESTRO Health Economics in Radiation Oncology project. HERO has the overall aim to develop a knowledge base of the provision of radiotherapy in Europe and build a model for health economic evaluation of radiation treatments at the European level. The aim of the current report is to describe the distribution of radiotherapy equipment in European countries. Methods An 84-item questionnaire was sent out to European countries, principally through their national societies. The current report includes a detailed analysis of radiotherapy departments and equipment (questionnaire items 26–29), analyzed in relation to the annual number of treatment courses and the socio-economic status of the countries. The analysis is based on validated responses from 28 of the 40 European countries defined by the European Cancer Observatory (ECO). Results A large variation between countries was found for most parameters studied. There were 2192 linear accelerators, 96 dedicated stereotactic machines, and 77 cobalt machines reported in the 27 countries where this information was available. A total of 12 countries had at least one cobalt machine in use. There was a median of 0.5 simulator per MV unit (range 0.3–1.5) and 1.4 (range 0.4–4.4) simulators per department. Of the 874 simulators, a total of 654 (75%) were capable of 3D imaging (CT-scanner or CBCT-option). The number of MV machines (cobalt, linear accelerators, and dedicated stereotactic machines) per million inhabitants ranged from 1.4 to 9.5 (median 5.3) and the average number of MV machines per department from 0.9 to 8.2 (median 2.6). The average number of treatment courses per year per MV machine varied from 262 to 1061 (median 419). While 69% of MV units were capable of IMRT only 49% were equipped for image guidance (IGRT). There was a clear relation between socio-economic status, as measured by GNI per capita, and availability of radiotherapy equipment in the countries. In many low income countries in Southern and Central-Eastern Europe there was very limited access to radiotherapy and especially to equipment for IMRT or IGRT. Conclusions The European average number of MV machines per million inhabitants and per department is now better in line with QUARTS recommendations from 2005, but the survey also showed a significant heterogeneity in the access to modern radiotherapy equipment in Europe. High income countries especially in Northern-Western Europe are well-served with radiotherapy resources, other countries are facing important shortages of both equipment in general and especially machines capable of delivering high precision conformal treatments (IMRT, IGRT).

Academic research paper on topic "Radiotherapy equipment and departments in the European countries: Final results from the ESTRO-HERO survey"

Contents lists available at ScienceDirect

Radiotherapy and Oncology

journal homepage: www.thegreenjournal.com

Radiotherapy

EtOncology

ESTRO-HERO survey

Radiotherapy equipment and departments in the European countries: ((WssMark Final results from the ESTRO-HERO survey

Cai Grau3'*, Noemie Defournyb, Julian Malickic, Peter Dunscombed, Josep M. Borrase, Mary Coffeyf, Ben Slotmang, Marta Boguszh, Chiara Gasparottob, Yolande Lievens1, on behalf of the HERO consortium1

a Aarhus University Hospital, Denmark; b European Society for Radiotherapy and Oncology, Brussels, Belgium;c University of Medical Sciences, Greater Poland Cancer Center, Poznan, Poland; d University of Calgary, Canada;e University of Barcelona, Spain;f Trinity College Dublin, Ireland; g VU Medical Center, Amsterdam, The Netherlands; h Cancer Diagnosis and Treatment Center, Katowice, Poland; i Ghent University Hospital, Belgium

ARTICLE INFO

Article history: Received 19 August 2014 Accepted 21 August 2014 Available online 31 October 2014

Keywords:

Radiotherapy equipment Health economics

ABSTRACT

Background: Documenting the distribution of radiotherapy departments and the availability of radiotherapy equipment in the European countries is an important part of HERO - the ESTRO Health Economics in Radiation Oncology project. HERO has the overall aim to develop a knowledge base of the provision of radiotherapy in Europe and build a model for health economic evaluation of radiation treatments at the European level. The aim of the current report is to describe the distribution of radiotherapy equipment in European countries.

Methods: An 84-item questionnaire was sent out to European countries, principally through their national societies. The current report includes a detailed analysis of radiotherapy departments and equipment (questionnaire items 26-29), analyzed in relation to the annual number of treatment courses and the socio-economic status of the countries. The analysis is based on validated responses from 28 of the 40 European countries defined by the European Cancer Observatory (ECO).

Results: A large variation between countries was found for most parameters studied. There were 2192 linear accelerators, 96 dedicated stereotactic machines, and 77 cobalt machines reported in the 27 countries where this information was available. A total of 12 countries had at least one cobalt machine in use. There was a median of 0.5 simulator per MV unit (range 0.3-1.5) and 1.4 (range 0.4-4.4) simulators per department. Of the 874 simulators, a total of 654 (75%) were capable of 3D imaging (CT-scanner or CBCT-option). The number of MV machines (cobalt, linear accelerators, and dedicated stereotactic machines) per million inhabitants ranged from 1.4 to 9.5 (median 5.3) and the average number of MV machines per department from 0.9 to 8.2 (median 2.6). The average number of treatment courses per year per MV machine varied from 262 to 1061 (median 419). While 69% of MV units were capable of IMRT only 49% were equipped for image guidance (IGRT). There was a clear relation between socio-economic status, as measured by GNI per capita, and availability of radiotherapy equipment in the countries. In many low income countries in Southern and Central-Eastern Europe there was very limited access to radiotherapy and especially to equipment for IMRT or IGRT.

Conclusions: The European average number of MV machines per million inhabitants and per department is now better in line with QUARTS recommendations from 2005, but the survey also showed a significant heterogeneity in the access to modern radiotherapy equipment in Europe. High income countries especially in Northern-Western Europe are well-served with radiotherapy resources, other countries are facing important shortages of both equipment in general and especially machines capable of delivering high precision conformal treatments (IMRT, IGRT).

© 2014 Elsevier Ireland Ltd. Radiotherapy and Oncology 112 (2014) 155-164 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3XI/).

* Corresponding author at: Department of Oncology, Aarhus University Hospital, Norrebrogade 44, Bld. 5, DK-8000 Aarhus C, Denmark.

E-mail address: caigrau@dadlnet.dk (C. Grau).

1 See complete list of HERO consortium co-authors in the online version.

http://dx.doi.org/10.1016/j.radonc.2014.08.029 0167-8140/® 2014 Elsevier Ireland Ltd.

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Evidence-based regimens and novel high precision technology have reinforced the important role of radiotherapy in contemporary multimodality management of cancer. Current data estimate that about 50% of all cancer patients would benefit from radiotherapy during the course of their disease, with many of them requiring several courses of treatment [1,2]. Due to significant technical improvements, it is now possible to cure more patients with fewer side effects. This requires, however, access to modern equipment, including intensity modulated radiotherapy (IMRT), image-guided radiotherapy (IGRT), stereotactic radiotherapy and, most recently, particle therapy [3-11]. The European situation is highly diverse with large differences in demographics, cancer incidence and economic resources among countries. From the ''Radiation Therapy for Cancer: Quantification of Radiation Therapy Infrastructure and Staffing Needs'' (QUARTS) project and more recent analyses, it has been shown that parts of Europe are well-served in radiotherapy resources, whereas others are facing important shortages as well in terms of equipment and infrastructure as of trained personnel [12,13].

The ESTRO initiated HERO-project (Health Economics in Radiation Oncology) has the overall aim to develop a knowledge base and a model for health economic evaluation of radiation treatments at the level of individual European countries [14]. To accomplish these objectives, the HERO project addresses availability, needs, cost and cost-effectiveness of radiotherapy. By providing an updated and validated description of the European radiotherapy landscape in collaboration with the national societies, and through the development of web-based cost and cost-effectiveness models, ESTRO wants to support the European countries and their national radiotherapy societies in benchmarking their position in Europe and computing the cost and cost-effectiveness of radiotherapy in their specific economic context.

This first part of the HERO program is based on a detailed survey providing an inventory of European radiotherapy in terms of resource availability (departments, equipment, and personnel), guidelines and reimbursement. The data will be presented in three simultaneous papers. The two other papers address staffing [15] and guidelines [16], respectively, while the current paper focuses on the distribution of radiotherapy equipment in European countries.

Materials and methods

A web-based questionnaire consisting of 84 questions relating to population and cancer incidence, radiotherapy activity and resources, guidelines and reimbursement was developed and distributed to national scientific and professional radiotherapy societies. The full details of the data collected, the methodological considerations and the practical decisions regarding the data set used for the entire analysis, are described in the Supplementary Documentation.

The current report includes a detailed analysis of radiotherapy departments and equipment (questionnaire items 26-29) in the countries defined by the European Cancer Observatory (ECO), analyzed in relation to the annual number of patients treated in the same countries. Among the 34 ECO countries responding to the questionnaire, 28 countries could be included in the analysis of equipment (Table 1): 24 countries with complete datasets, and 4 countries (Belarus, Germany, Italy and United Kingdom) with partial data. The remaining 12 ECO countries (Romania, Slovakia, Bosnia, Croatia, Cyprus, Latvia, Macedonia, Russia, Serbia, Ukraine, Greece, Republic of Moldova) either provided none or insufficient data, did not submit updates, or did not give their consent to use their previous submission.

From the questionnaire, the number of megavoltage (MV) units was calculated as the sum of linear accelerators (including helical tomotherapy), cobalt-60 and dedicated stereotactic machines. Simulators for radiotherapy imaging and treatment planning were classified as conventional simulators (2D), CT simulators or simulators with a cone-beam CT option (CBCT).

The number of radiotherapy treatment courses, being primary treatments, palliative treatments or retreatments, was recorded in the questionnaire. For the countries where the information about retreatments was unavailable, the primary treatment figures were augmented with 25% [12].

The economic status of the countries was expressed as gross national income per capita (GNl/n) using the Atlas method [17]. In order to identify relatively homogeneous groups of countries based on selected characteristics such as the percentage with IMRT technology, megavoltage machine units per million inhabitants and GNl/n, we used the k-means clustering via principal components analysis using the Hartigan and Wong method [18]. With this method, multidimensional data can be represented on two axes and the cluster centroids (vector of mean values of each variable) could be defined. Germany and ltaly were excluded from this part of the analysis since the data related to IMRT capability were not available. The statistical software R was used to perform this analysis [19].

Results

The validated data on number of radiation treatments, departments and equipment in the 28 ECO countries form the basis of the present analysis (Table 1).

Equipment

A total of 3024 photon therapy units (2705 MV and 319 kV machines) and 7 proton facilities were recorded in the 28 countries. One country did not supply detailed information about MV unit type; in the remaining 27 countries there were 2192 linear accelerators, 96 stereotactic machines and 77 cobalt machines reported. Twelve countries (43%) had at least one cobalt machine in use. Information about equipment for IMRT and IGRT was available for 26 countries; a total of 1327 out of 1915 MV units in 26 countries with this information availble were equipped for IMRT (69%). IGRT equipment was available in 930 of 1915 MV units (49%). ln seven countries (Albania, Belarus, Bulgaria, Hungary, Ireland, Lithuania, Spain) less than half of the MV machines were equipped for delivering lMRT, and in 13 countries (Albania, Montenegro, Hungary, Bulgaria, Belarus, Spain, Lithuania, Switzerland, Czech Republic, Ireland, United Kingdom, Slovenia, Portugal) less than half of the MV units were equipped for IGRT. There was a total of 96 dedicated stereotactic radiotherapy units in 13 countries; the remaining 15 countries did not have such equipment (n = 12) or did not report (n = 3). Simulators for treatment planning were either 2D kilovoltage (kV; n = 220), CT-simulators (n = 592) or 2D kV units with 3D option (CBCT) (n = 62). Overall, 75% of all simulators had 3D capability. ln three countries (Czech Republic, Hungary, Lithuania) less than half of the simulators had 3D capability.

Department infrastructure

A total of 872 facilities were recorded in 27 countries, distributed as 814 departments (93%) and 58 satellites (7%). The highest number of facilities was in France (n = 176) followed by ltaly (n = 165), Spain (n = 112), and the United Kingdom (n = 77); the remaining countries had between 1 and 48 centers each. The

Data on demographics, average number of treatment machines, simulators and departments in 28 European countries included in the HERO database.

Countries

Demographics

Treatment units

Other machines

Simulators

Departments

Population GNI/n RT Ref. year Total Linear Linacs Linacs Dedicated Cobalt Ortho- Proton Carbon Total 2D 2D sim CT Total Departments Satellites

(2011, WB) 2011 Courses equipment MV accelerators with with SRS units voltage facilities ion simulators sim with CT sim facilities

(USD) units IMRT IGRT machines facilities option

Albania 2,829,337 4,050 2,195 2010 4 2 0 0 0 2 1 0 0 2 0 0 2 2 2 0

Austria 8,406,187 48,170 21,481 2010 43 42 35 26 1 0 7 0 0 21 8 1 12 14 14 0

Belarus 9,473,000 6,270 2009 30 8 5 4 0 22 18 0 0 20 9 7 4 23 20 3

Belgium 11,047,744 45,840 34,672 2013 91 87 71 57 3 1 8 0 0 29 8 7 14 36 25 11

Bulgaria 7,348,328 6,640 13,794 2012 13 5 2 1 0 8 10 0 0 6 1 1 4 14 14 0

Czech Republic 10,496,088 18,720 32,630 2009 57 43 29 17 4 10 39 1 0 28 18 0 10 48 36 12

Denmark 5,570,572 60,160 17,680 2010 53 53 50 47 0 0 6 0 0 14 0 0 14 9 7 2

Estonia 1,327,439 15,260 2,122 2012 4 4 4 4 0 0 0 0 0 3 1 0 2 2 2 0

Finland 5,388,272 47,740 13,994 2010 43 41 41 41 2 0 0 0 0 17 2 2 13 13 12 1

France 65,343,588 42,690 187,172 2012 449 421 412 238 28 0 11 2 0 165 26 139 176 172 4

Germany 81,797,673 44,230 2010 450 434 16 103

Hungary 9,971,727 12,840 19,951 2011 36 26 6 2 1 9 4 0 0 19 12 0 7 12 12 0

Iceland 319,014 35,260 595 2010 2 2 2 1 0 1 0 0 1 0 1 0 1 1 0

Ireland 4,576,794 38,960 8,373 2009 32 31 10 10 0 1 2 0 0 12 3 0 9 12 12 0

Italy 59,379,449 35,350 2011 340 165 162 3

Lithuania 3,028,115 13,000 6,268 2011 10 10 3 2 0 0 5 0 0 5 4 0 1 5 4 1

Luxembourg 518,347 77,380 1,180 2010 2 2 2 1 0 0 0 0 0 2 1 0 1 1 1 0

Malta 416,268 19,760 535 2014 2 2 1 1 0 0 1 0 0 1 0 0 1 1 1 0

Montenegro 620,644 6,810 1,500 2011 2 2 0 0 0 0 0 0 0 3 1 1 1 1 1 0

The Netherlands 16,693,074 49,660 55,683 2012 132 132 125 125 0 8 0 0 38 8 0 30 29 21 8

Norway 4,953,088 88,500 13,483 2011 41 40 40 40 1 0 6 0 0 22 11 0 11 9 5 4

Poland 38,534,157 12,340 73,500 2010 120 115 109 77 4 1 5 1 0 76 24 16 36 35 35 0

Portugal 10,557,560 21,420 17,957 2010/12 44 41 30 18 3 0 0 0 0 20 3 5 12 17 17 0

Slovenia 2,052,843 23,940 6,023 2012 8 8 5 3 0 0 1 0 0 3 1 0 2 1 1 0

Spain 46,742,697 30,930 98,525 2011 261 220 56 50 36 5 18 0 0 167 35 132 112 112 0

Sweden 9,449,213 53,530 22,678 2012 63 62 51 44 1 0 4 0 0 21 6 0 15 16 15 1

Switzerland 7,912,398 76,350 19,000 2013 59 52 52 12 6 1 11 2 39 13 26 41 37 4

United Kingdom 63,258,918 37,840 2010/11 314 307 186 109 6 1 50 1 0 140 25 21 94 77 73 4

England 53,012,456 n.a. 121,289 2010 268 261 146 86 6 1 46 1 0 117 19 18 80 68 64 4

Scotland 5,295,000 n.a. 2011 25 25 23 16 0 0 1 0 0 13 4 1 8 5 5 0

Wales 3,063,456 n.a. 6,445 2011 13 13 9 7 0 0 3 0 0 7 2 1 4 3 3 0

Northern Ireland 1,810,863 n.a. 4,180 2010 8 8 8 0 0 0 0 0 0 3 0 1 2 1 1 0

No. entries 28 28 24 26 28 27 26 26 25 26 27 26 25 26 26 23 26 27 27 27

Total 488,012,534 973,640 670,991 2705 2192 1327 930 96 77 319 7 0 874 220 62 592 872 814 58

Median 8,159,293 35,305 15,837 2011 43 41 30 15 1 0 5 0 0 20 5 0 11 14 14 0

Min 319,014 4,050 535 2009 2 2 0 0 0 0 0 0 0 1 0 0 0 1 1 0

Max 81,797,673 88,500 187,172 2014 450 434 412 238 36 22 103 2 0 167 35 21 139 176 172 12

Calculated indicators for availability of radiotherapy equipment in 28 European countries included in the HERO database.

Countries

Indicators

Departments/mil inh MV units/mil inh MV units/dep MV units with IMRT MV units with IGRT Sim/dep Sim/MV unit Sim with 3D Courses/dep Courses/MV

Albania 0.7 1.4 2.0 0% 0% 1.0 0.5 100% 1098 549

Austria 1.7 5.1 3.1 81% 60% 1.5 0.5 62% 1534 500

Belarus 2.1 3.2 1.5 17% 13% 1.0 0.7 55%

Belgium 2.3 8.2 3.6 78% 63% 1.2 0.3 72% 1387 381

Bulgaria 1.9 1.8 0.9 15% 8% 0.4 0.5 83% 985 1061

Czech Republic 3.4 5.4 1.6 51% 30% 0.8 0.5 36% 906 572

Denmark 1.3 9.5 7.6 94% 89% 2.0 0.3 100% 2526 334

Estonia 1.5 3.0 2.0 100% 100% 1.5 0.8 67% 1061 531

Finland 2.2 8.0 3.6 95% 95% 1.4 0.4 88% 1166 325

France 2.6 6.9 2.6 92% 53% 1.0 0.4 84% 1088 417

Germany 5.5

Hungary 1.2 3.6 3.0 17% 6% 1.6 0.5 37% 1663 554

Iceland 3.1 6.3 2.0 100% 50% 1.0 0.5 100% 595 298

Ireland 2.6 7.0 2.7 31% 31% 1.0 0.4 75% 698 262

Italy 2.7 5.7 2.1

Lithuania 1.3 3.3 2.5 30% 20% 1.3 0.5 20% 1567 627

Luxembourg 1.9 3.9 2.0 100% 50% 2.0 1.0 50% 1180 590

Malta 2.4 4.8 2.0 50% 50% 1.0 0.5 100% 535 268

Montenegro 1.6 3.2 2.0 0% 0% 3.0 1.5 67% 1500 750

The Netherlands 1.3 7.9 6.3 95% 95% 1.8 0.3 79% 2652 422

Norway 1.0 8.3 8.2 98% 98% 4.4 0.5 50% 2697 329

Poland 0.9 3.1 3.4 91% 64% 2.2 0.6 68% 2100 613

Portugal 1.6 4.2 2.6 68% 41% 1.2 0.5 85% 1056 408

Slovenia 0.5 3.9 8.0 63% 38% 3.0 0.4 67% 6023 753

Spain 2.4 5.6 2.3 21% 19% 1.5 0.6 79% 880 377

Sweden 1.6 6.7 4.2 81% 70% 1.4 0.3 71% 1512 360

Switzerland 4.7 7.5 1.6 88% 20% 1.1 0.7 67% 514 322

United Kingdom 1.2 5.0 4.3 59% 35% 1.9 0.4 82%

England 1.2 5.1 4.2 54% 32% 1.8 0.4 84% 1895 453

Scotland 0.9 4.7 5.0 92% 64% 2.6 0.5 69%

Wales 1.0 4.2 4.3 69% 54% 2.3 0.5 71% 2148 496

Northern Ireland 0.6 4.4 8.0 100% 0% 3.0 0.4 100% 4180 523

No. entries 27 28 27 26 26 26 26 26 24 24

Median 1.7 5.3 2.6 73% 45% 1.4 0.5 72% 1173 419

Min 0.5 1.4 0.9 0% 0% 0.4 0.3 20% 514 262

Max 4.7 9.5 8.2 100% 100% 4.4 1.5 100% 6023 1061

■a' m

Bulgaria Belarus Czech Republic Switzerland Albania Estonia Montenegro Luxembourg Malta Iceland Italy Spain Lithuania Portugal European median France Ireland Hungary Austria Poland Finland Belgium Sweden United Kingdom Netherlands Denmark Slovenia Norway

Fig. 1. Histogram showing the average number of radiotherapy treatment machines (MV units) per department in 27 European countries.

average number of MV units per department ranged from fewer than two (Belarus, Bulgaria, Czech Republic, Switzerland) to more than six (Denmark, The Netherlands, Norway, Slovenia); the

median being 2.6 MV units per department (Table 2 and Fig. 1). There was a median of 1.4 simulators per department (range 0.4-4.4) and 0.5 simulators per MV unit (range 0.3-1.5).

Albania Bulgaria Estonia Poland Belarus Montenegro Lithuania Hungary Luxembourg Slovenia Portugal Malta United Kingdom Austria European median Czech Republic Germany Spain Italy Iceland Sweden France Ireland Switzerland Netherlands Finland Belgium Norway Denmark

Fig. 2. Histogram showing the average number of radiotherapy treatment machines (MV units) per million inhabitants in 28 European countries.

v ♦

.......♦....."

♦ ......

♦ ♦ ♦ ♦

♦ ♦

♦ ♦ ♦ ♦

♦ ♦

_ 2 1 0

0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 100,000

GNI per capita (USD)

Fig. 3. Diagram showing the relationship between economic status (GNI per capita) and the average number of radiotherapy treatment machines (MV units) per million inhabitants in 26 European countries.

Fig. 4. Diagram showing the relationship between economic status (GNI per capita) and the percentage of radiotherapy treatment machines (MV units) capable of delivering IMRT in 27 European countries.

Demographic and economic indicators

The availability of radiotherapy services in relation to key demographic and economic parameters is presented in Table 2. The average number of departments per million inhabitants ranged from 0.5 to 4.7 (median 1.7). This variation was to some extent reflected in a similarly large variation in the average size of the departments, as shown in Fig. 1.

Fig. 2 shows the average number of MV units per million inhabitants; the median number was 5.3. There was a sevenfold varia-

tion in this parameter, ranging from very low availability of less than 2 MV units per million in Albania and Bulgaria to more than 8 MV units per million in Belgium, Denmark and Norway.

There was a significant influence of national economic status on the availability of radiotherapy services in the 28 European countries. Fig. 3 shows the correlation between GNI per capita and the number of MV units per million inhabitants. High-income countries had more machines per million than countries with lower GNI per capita. Countries with GNI per capita above USD

Fig. 5. Diagram showing the relationship between economic status (GNI per capita) and the average number of treatment courses per radiotherapy treatment machines (MV units) in 27 European countries.

Table 3

Correlation between variables included in the k-means clustering via principal components analysis.

IMRT (%) GNI/n MV units/mil inh

IMRT (%) GNI/n MV units/mil inh 1 0.68 0.59 0.68 1 0.76 0.59 0.76 1

30,000 in general had more than 4 MV units per million inhabitants, the only exception being Luxembourg (3.9 MV/mil). The eight countries with GNI per capita below USD 16,000 all had less than four MV units per million inhabitants.

The economic status also influenced access to machines with the ability to deliver highly conformal treatments (IMRT). Fig. 4 shows the relationship between GNI per capita and percentage of MV units capable of IMRT. As can be seen, more than 75% of machines could deliver IMRT in countries with GNI/n above USD 40,000, compared to less than 25% in the four lowest income countries. Poland and Estonia were exceptions, having a high percentage of IMRT capability despite a relatively low GNI per capita.

The average number of treatment courses delivered per year per department varied from fewer than 700 (Switzerland, Ireland, Iceland, Malta) to more than 2000 (Poland, Denmark, The Netherlands, Norway, Slovenia); the median number was 1173 treatment courses. The median number of treatment courses per MV unit in 24 countries was 419, with a large variation from fewer than 300 courses per MV in Ireland, Iceland and Malta to more than 700 in Bulgaria, Montenegro and Slovenia. There was a clear correlation with the socio-economic status, with more treatments delivered per machine in the countries with the lowest GNI per capita (Fig. 5).

Cluster analysis

Table 3 shows correlations between the three variables (%IMRT, MV units/mil, GNI/n) used in the clustering analysis. GNI was

Fig. 6. The four clusters found through the k-means clustering via principal components analysis.

highly correlated with the percentage of IMRT capability (r = 0.68) and MV machines per million inhabitants (r = 0.76), however these last two variables showed a slightly lower correlation between them (r = 0.59). In the subsequent clustering analysis it was found that these three variables could be graphically depicted using two axes which represented 92.7% of the total variability. Using these two components, we found 4 clusters of countries (Fig. 6 and Table 4). Cluster 1 defined by Luxembourg, Norway and Switzerland showed high equipment inventories per million inhabitants and the highest GNI/n. Countries in cluster 2 (Austria, Belgium, Denmark, Finland, The Netherlands) had lower GNI/n but higher equipment inventories per million inhabitants than that of cluster 1. Cluster 3, formed by France, Iceland, Ireland, Spain and

Centroids of the clusters identified in the k-means clustering via principal components analysis.

Cluster IMRT (%) GNI/n MV units/mil inh

1 Luxembourg, Norway, Switzerland 95.3 80.7 6.6

2 Austria, Belgium, Denmark, Finland, The Netherlands, Sweden 87.3 50.8 7.6

3 France, Iceland, Ireland, Spain, UK 60.6 37.1 6.2

4 Albania, Belarus, Bulgaria, Czech Republic, Estonia, Hungary, Lithuania, Malta, Montenegro, Poland, Portugal, Slovenia 41.8 13.4 3.4

UK showed centroid values lower than those of cluster 2. Cluster 4 was formed by former Eastern European countries, Malta, and Portugal, presented the lowest GNI/n, equipment and IMRT technology.

Discussion

This is the initial report on the outcome of the ESTRO-HERO database, based on validated national data, collected in collaboration with the national radiotherapy societies. The main findings confirm a large variation between countries for all types of equipment and availability indicators studied.

The first parts of the HERO project build on the experience of the QUARTS-project, the first real attempt to arrive at estimates for the appropriate level of radiotherapy infrastructure and staffing in Europe [12,20]. The two main equipment indicators from QUARTS were MV units per million inhabitants and number of treatments per MV unit. Official guidelines for the number of linear accelerators and personnel were available in about 40% of the countries in the QUARTS analysis. For accelerators, the analysis of guidelines came to a recommendation of an average of 5.5 MV units/million in the high, 3.5 MV units/million in the medium and 2.0 MV units/million in the low resource countries [20]. These numbers have not changed significantly over the last decade, as illustrated in the update of guidelines survey performed as part of the HERO project [16]. In the second part of the QUARTS analysis [12], the best available evidence on radiotherapy indications in 23 main cancer types was combined with epidemiological data from all 25 EU countries at that time and with published benchmarks for accelerator throughput. A large variation in crude cancer incidence observed within the analyzed EU countries translated into a similarly large variation in the estimated number of required linear accelerators per million inhabitants (between 4.0 and 8.1 linear accelerators/million inhabitants), hovering around a European average of 5.9. In the current HERO data set, the actual number was 5.3 MV units/million, meaning that the overall European average is now close to the QUARTS guideline standard for high income countries, albeit with a significant sevenfold difference from the highest to the lowest coverage. The lowest income countries all have less than 4 MV units per million inhabitants, and the highest income countries have close to 10 MV units per million. This socioeconomic disparity was also highlighted in the recent IAEA DIRAC study [13]. Despite some discussion about potential problems with the DIRAC data previously addressed [21,22], the two studies thus reach the same overall conclusions about the heterogeneity in radiotherapy equipment availability between European countries. Whether the actual need is still 5.9 MV units/million will be refined in the ongoing HERO project by calculating the needs based on actual cancer incidence in the individual countries and using evidence-based indications for radiotherapy.

Although the high correlations observed between GNI and MV units per million inhabitants and the percentage with IMRT could suggest the conclusion that economic resources available at country level are the main determinants of the radiotherapy

equipment, the cluster analysis showed that the situation is not so straight forward. The countries included in the cluster with the highest GNI (Luxemburg, Norway, and Switzerland) do not have the highest number of MV per million, although they have the most technological updated equipment. From this perspective, it could be suggested that health policy decisions at country level matters.

The number of patients treated per MV unit is often used as a measure of machine productivity. From the guidelines, QUARTS estimated a European benchmark of 450 patients per machine per year, with an estimate of 400-450 patients per year accounting for increasing complexity. The recent HERO update of available guidelines showed no major change in this recommendation over the last decade [16]. In the current analysis of actually available equipment, the benchmark is reached, as the median number of treatment courses was 419 per MV unit. As shown in Fig. 5 there was a striking variation in machine throughput, which seemed to be related to the socio-economic status of the country, with high values in low income countries. In many of these low income countries, the equipment is being utilized for extended hours. In e.g. Slovenia, where the average throughput is over 700 patients per linear accelerator, all machines are used in two shifts per day. It can be seen from Table 2 that these countries also have the least advanced technology available, i.e. fewer linear accelerators capable of IMRT and IGRT, and relatively more cobalt units. The throughput is also dependent upon other factors, including referral base, complexity of treatment and the age of the equipment. Increasing use of high quality advanced conformal treatments and daily imaging will in many situations be more time-consuming and thus put a limit to how many treatments a machine can deliver. This is further discussed in the HERO staffing paper [15].

The type of radiotherapy equipment available in the European countries differed significantly. In total, 69% of all MV units were equipped for IMRT and 49% for IGRT, with much higher rates in high income countries. A number of Eastern European countries still have cobalt machines, which are not able to deliver modern conformal treatments. This skewed distribution was also noted in the DIRAC study [13].

The size of radiotherapy departments is an indicator of the organizational infrastructure. In the Nordic countries, Poland, The Netherlands and Slovenia, radiotherapy services are centralized in departments treating on average more than 1600 patients per year. Such large departments enable subspecialization of the staff. In the Dutch model, each radiation oncologist has 2-3 areas of expertise; to cover all tumor sites it is necessary to have at least eight specialists, which again require 1600-2000 patients per year to be efficient [23]. In most other European countries, however, facilities are small with only one or two machines. An average of fewer than 1,000 courses per year per department, as seen in eight countries (Malta, Switzerland, Iceland, Ireland, Spain, Czech Republic and Bulgaria), may indicate a fragmentation of radiotherapy services, which potentially influence treatment quality and might have negative effect on productivity. The major cost components of radiotherapy are buildings, equipment and staff. Due to the

economics of scaling, the average cost per patient decreases with increasing department size and optimal utilization of resources. Earlier studies concluded that the costs per patient substantially decrease when a department is treating more than 1000 patients annually [24] or 1000-1200 annually [25]. Similar calculations have been done for Canada [26], finding a threshold about 1600 patients annually.

This HERO study has several limitations. We used national averages of institutional data, collected by individuals in the national societies over several years. This will inevitably introduce bias and variation in the data. Although we have tried to validate the data by repeated updates and comparison to other sources, and exclusion of the least reliable datasets, there will be some uncertainty in the presented data. This uncertainty specifically applies to data on complexity of the equipment (IMRT, IGRT, SRS), which are also sensitive to the time of data collection, since the field is quickly evolving. The survey also collected equipment capability only and did not collect data to understand if the equipment was actually being used to its full capability. This is important as equipment can be operational without the new technologies actually being implemented, e.g. due to limited workforce availability and constraints related to skills or resources. Finally, the interpretation in the various countries of specific entities such as 'dedicated stereotactic equipment' and 'intra-operative linacs' may have been different in different countries. Since these units are only a very small fraction of the total machines, such variations will have little impact on the overall conclusions.

Collecting and validating the data has been a huge task for many of the representatives from the national societies. It is our hope that the experience obtained and the collegial network established through this project will be valuable not only for future updates of the HERO database but also forming the basis for more qualified discussions within the national societies[21]. The next step in the HERO framework is to benchmark the data to the equipment needs in the individual countries, based on cancer incidence and stage mix and performed in cooperation with the Collaboration for Cancer Outcomes, Research and Evaluation (CCORE) in Australia. The data will also be used in developing the HERO costing model for European countries, in order to provide budgetary estimates of the radiotherapy optimization process in various jurisdictions.

In conclusion, the results of this survey document a significant heterogeneity in the access to modern radiotherapy equipment in Europe. Although the European average number of MV machines per million inhabitants and per department is now better in line with QUARTS recommendations from 2005, there is still a significant heterogeneity in the access to radiotherapy equipment in Europe. While high income countries especially in Northern-Western Europe are well-served with radiotherapy resources, other countries are facing important shortages of both equipment in general and especially machines capable of delivering high precision con-formal treatments (IMRT, IGRT).

Conflict of interest

The authors have no conflict of interest.

Funding sources

This project was supported by the European Society for Radiotherapy and Oncology.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.radonc.2014. 08.029.

References

Delaney G, Jacob S, Featherstone C, Barton M. The role of radiotherapy in cancer treatment: estimating optimal utilization from a review of evidence-based clinical guidelines. Cancer 2005;104:1129-37.

Barton MB, Jacob S, Shafiq J, Wong K, Thompson SR, Hanna TP, et al. Estimating the demand for radiotherapy from the evidence. A review of changes from 2003 to 2012. Radiother Oncol 2014;112:140-4.

Langendijk JA, Lambin P, De Ruysscher D, Widder J, Bos M, Verheij M. Selection of patients for radiotherapy with protons aiming at reduction of side effects: the model-based approach. Radiother Oncol 2013;107:267-73. Allen AM, Pawlicki T, Dong L, Fourkal E, Buyyounouski M, Cengel K, et al. An evidence based review of proton beam therapy: the report of ASTRO's emerging technology committee. Radiother Oncol2012;103:8-11. Hoyer M, Roed H, Traberg Hansen A, Ohlhuis L, Petersen J, Nellemann H, et al. Phase II study on stereotactic body radiotherapy of colorectal metastases. Acta Oncol 2006;45:823-30.

De Ruysscher D, Lodge MM, Jones B, Brada M, Munro A, Jefferson T, et al. Charged particles in radiotherapy: a 5-year update of a systematic review. Radiother Oncol 2012;103:5-7.

Nishi T, Nishimura Y, Shibata T, Tamura M, Nishigaito N, Okumura M. Volume and dosimetric changes and initial clinical experience of a two-step adaptive intensity modulated radiation therapy (IMRT) scheme for head and neck cancer. Radiother Oncol 2013;106:85-9.

Uhl M, van Triest B, Eble MJ, Weber DC, Herfarth K, De Weese TL. Low rectal toxicity after dose escalated IMRT treatment of prostate cancer using an absorbable hydrogel for increasing and maintaining space between the rectum and prostate: results of a multi-institutional phase II trial. Radiother Oncol 2013;106:215-9.

Combs SE, Adeberg S, Dittmar JO, Welzel T, Rieken S, Habermehl D, et al. Skull base meningiomas: long-term results and patient self-reported outcome in 507 patients treated with fractionated stereotactic radiotherapy (FSRT) or intensity modulated radiotherapy (IMRT). Radiother Oncol 2013; 106:186-91.

Leclerc M, Maingon P, Hamoir M, Dalban C, Calais G, Nuyts S, et al. A dose escalation study with intensity modulated radiation therapy (IMRT) in T2N0, T2N1, T3N0 squamous cell carcinomas (SCC) of the oropharynx, larynx and hypopharynx using a simultaneous integrated boost (SIB) approach. Radiother Oncol 2013;106:333-40.

Mortensen HR, Jensen K, Aksglaede K, Behrens M, Grau C. Late dysphagia after IMRT for head and neck cancer and correlation with dose-volume parameters. Radiother Oncol 2013;107:288-94.

Bentzen SM, Heeren G, Cottier B, Slotman B, Glimelius B, Lievens Y, et al. Towards evidence-based guidelines for radiotherapy infrastructure and staffing needs in Europe: the ESTRO QUARTS project. Radiother Oncol 2005;75:355-65.

Rosenblatt E, Izewska J, Anacak Y, Pynda Y, Scalliet P, Boniol M, et al. Radiotherapy capacity in European countries: an analysis of the Directory of Radiotherapy Centres (DIRAC) database. Lancet Oncol 2013;14:79-86. Lievens Y, Grau C. Health economics in radiation oncology: introducing the ESTRO-HERO project. Radiother Oncol 2012;103:109-12. Lievens Y, Coffey M, Defourny N, Malicki J, Dunscombe P, Borras JM, et al. on behalf of the HERO consortium. Radiotherapy staffing in the European countries: Final results from the ESTRO-HERO survey. Radiother Oncol 2014;112:178-86.

Dunscombe P, Grau, C., Defourny, N., Malicki, J., Borras, J.M., Coffey, M., Bogusz, M., Gasparotto, C., Slotman, B., Lievens, Y., on behalf of the HERO consortium. Guidelines for equipment and staffing of radiotherapy facilities in the European countries: Final results of the ESTRO-HERO survey. Radiother Oncol 2014;112:165-77.

World Bank Database 2011 [Feb 2014]. Available from: <http://data. worldbank.org/>.

Hartigan JA, Wong MA. Algorithm AS 136: a K-Means clustering algorithm. J R Stat Soc Ser C (Appl Stat) 1979;28:100-8.

R Core Team. A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2013. Available from:< http://www. R-project.org/>.

Slotman BJ, Cottier B, Bentzen SM, Heeren G, Lievens Y, van den Bogaert W.

Overview of national guidelines for infrastructure and staffing of radiotherapy.

ESTRO-QUARTS: work package 1. Radiother Oncol 2005;75:349-54.

Grau C, Borras JM, Malicki J, Slotman B, Dunscombe P, Coffey M, et al.

Radiotherapy capacity in Europe. Lancet Oncol 2013;14:196-8.

Senan S, Slotman BJ. Outcomes research radiotherapy capacity in Europe-time

to even things out?. Nat Rev Clin Oncol 2013;10:188-90.

[23] Slotman BJ, Vos PH. Planning of radiotherapy capacity and productivity. Radiother Oncol 2013;106:266-70.

[24] Lievens Y, van den Bogaert W, Kesteloot K. Activity-based costing: a practical model for cost calculation in radiotherapy. Int J Radiat Oncol Biol Phys 2003;57:522-35.

[25] Kesteloot K, Lievens Y, van der Schueren E. Improved management of radiotherapy departments through accurate cost data. Radiother Oncol 2000;55:251-62.

[26] Dunscombe P, Roberts G, Walker J. The cost of radiotherapy as a function of facility size and hours of operation. Br J Radiol 1999;72:598-603.