Survival after childhood cancer: patterns, temporal trends and inequalities with a focus on Germany

Background: Although the incidence of cancer in children is low, it constitutes a matter of considerable public health relevance. Childhood cancer is the leading disease-related cause of death among children in high-income countries (HICs). As primary preventive measures are not available, improving survival probabilities remains the primary goal. Although survival has increased markedly over the past decades, prognosis still differs across diagnoses and is influenced by a range of clinical factors. Moreover, a growing body of studies observed poorer survival among children from families of lower socioeconomic position (SEP), even within HICs. Since evidence from Germany is largely lacking, this PhD project sought to comprehensively assess childhood cancer survival patterns in Germany and, as a complement and for comparison purposes, childhood leukaemia survival patterns in Sweden. The objectives of this PhD project were addressed through four independent, yet content-related studies, which together form the basis of this cumulative dissertation.
Methods: We used data from the German Childhood Cancer Registry to analyse temporal patterns of overall survival (OS) from childhood cancer as well as to assess potential survival inequalities in association to area-based SEP, we assigned calendar-year-specific values of the German Index of Socioeconomic Deprivation (GISD), developed and published by the Robert-Koch Institute, to the cancer diagnoses via municipality code. The association between SEP and survival from childhood leukaemia in Sweden was assessed at both the individual and area-based level. Individually linked data on leukaemia cases and parental SEP were obtained through the use of the Swedish civil register infrastructure.
Using Joinpoint Regression Program we performed time trend analyses of OS estimates. Univariable and multivariable Cox proportional hazard models were fitted to assess potential survival inequalities in both Germany and Sweden. For Sweden in particular, we considered individual-level and area-based covariates separately to reveal their individual contribution to survival in multivariable models as well as period-specific patterns.
Results: For Germany, we observed increasing five-year OS for all cancer types combined reaching 86.5 % in 2011-2016, with strongest survival improvement for acute myeloid leukaemia (AML). Still, five-year OS from malignant CNS tumours (among others) showed only modest improvements and recently plateaued at below 80 %. We found little evidence for social inequalities in childhood cancer survival in Germany. However, while our findings indicated survival disadvantages for children with AML residing in more deprived areas, we found the opposite for children with CNS tumours. Results from Sweden revealed survival inequalities for lymphoblastic leukaemia in relation to both, individual and area-based SEP. Notably, most associations were only evident in more recent years.
Conclusions: Substantial enhancements, including the development of highly standardised treatment protocols for most cancer types, have likely contributed to overall improvements in survival. The observed null association between area-based SEP and childhood cancer survival in Germany was contrary to our initial expectations. This finding may, in part, reflect the dense network of treating hospitals across Germany, which could compensate for any socioeconomic inequalities. Moreover, the use of the GISD may have contributed to inconclusive results, as this composite score is limited by the municipalities’ comparatively coarse and disproportionate geographical resolution. German municipalities vary considerably in size, extending from small villages to major cities such as Berlin or Hamburg, which have populations in the millions and exhibit extensive social diversity. On the contrary, we observed survival inequalities among children with leukaemia in Sweden, where analyses were based on both individual-level and high-resolution area-based SEP information. In order to identify potential survival differences that may also exist in Germany, access to individual data for epidemiological research continues to be urgently required. The observed period-specific pattern of more pronounced survival inequalities in relation to SEP may be related to the increasing social and cultural diversity in both Sweden and Germany, which clearly underscores that healthcare provision for children with cancer will remain a matter of high public health relevance in the future.