gms | German Medical Science

16. Deutscher Kongress für Versorgungsforschung

Deutsches Netzwerk Versorgungsforschung e. V.

4. - 6. Oktober 2017, Berlin

Number of radiotherapy centres in the population and cancer mortality: A global perspective

Meeting Abstract

Suche in Medline nach

  • Daniel Medenwald - Martin-Luther-Universität Halle-Wittenberg, Halle (Saale), Germany
  • Christian Dietzel - Martin-Luther-Universität Halle-Wittenberg, Halle (Saale), Germany

16. Deutscher Kongress für Versorgungsforschung (DKVF). Berlin, 04.-06.10.2017. Düsseldorf: German Medical Science GMS Publishing House; 2017. DocP095

doi: 10.3205/17dkvf255, urn:nbn:de:0183-17dkvf2551

Veröffentlicht: 26. September 2017

© 2017 Medenwald et al.
Dieser Artikel ist ein Open-Access-Artikel und steht unter den Lizenzbedingungen der Creative Commons Attribution 4.0 License (Namensnennung). Lizenz-Angaben siehe



Objectives: To assess the association between the number of radiotherapy centres (RTC) in the population and incidence adjusted cancer mortality on a global scale.

Methods: Data on cancer incidence and mortality were obtained from the GLOBOCAN projects, while we extracted country-specific information on number of RTC from the DIRAC database. The World Bank database was taken into account for remaining data used.

We used linear regression models to assess the association between RTC per 10,000,000 inhabitants (logarithmized) and the log-transformed mortality/incidence ratio considering also cubic splines to model non- linearity (only used when superior based on information criteria and visual assessment).

Models were adjusted for health care expenditure, number of surgical procedures, neonatal mortality rate and GDP per capita. To assess bias due to unobserved confounders, mortality from kidney cancer was considered as a negative control. A sensitivity analysis considered only high quality data (C3 or higher). A significance level of 5% was assumed.

Results: We found an inverse linear relation between RT centres in the population and the cancer mortality to incidence ratio for prostate cancer (9% per log(RTC per 10,000,000 inhabitants), 95% CI: 5%-14%), female breast (5%, 95% CI: 2%-8%) and colorectal cancer (5%, 95% CI: 2%-7%). No bias was apparent after covariate adjustment. Results were robust when only countries with high quality data were considered.

Conclusion: The country-specific number of RTC seems a considerable predictor of cancer mortality. The absence of a low plateau (non-linear relation) indicates no oversupply with RTC.