Educational attainment in survivors of childhood cancer in Denmark, Finland, and Sweden

The study examines whether and how childhood cancer affects educational attainment, focusing on completion of upper secondary education among survivors in Denmark, Finland, and Sweden. Motivated by improved survival and concerns about late effects, the authors address gaps in prior research (e.g., self-reported outcomes, selection bias, limited follow-up, assessment at a single time-point) and inconsistent findings outside CNS tumours. They investigate if survivors experience only educational delays or persistent deficits into adulthood. The study compares survivors to matched general population controls and siblings, evaluates diagnosis-specific risks, considers age at diagnosis, severity proxies and late effects (hospitalization and hospital contacts), and explores the role and potential effect modification by parental education.

Prior studies consistently show lower educational attainment among survivors of CNS tumours and those exposed to CNS-directed therapy. Findings for other cancer types are mixed, with some reporting worse, equal, or better outcomes than peers. A recent review concluded increased risk of lower educational achievement among childhood and young adult cancer survivors overall, but the evidence was graded very low due to methodological limitations such as self-report, non-participation bias, limited follow-up, and single time-point assessments. Risk factors include diagnostic group and age at diagnosis; however, severity of disease and late effects (somatic and psychiatric) have rarely been incorporated. Parental education strongly influences educational attainment and may confound or modify associations, but effect modification has seldom been evaluated.

Design and data sources: Register-based matched cohort study within the SALICCS programme using nationwide registries from Denmark, Finland, and Sweden linked via unique personal identifiers. Study population: All individuals with a first childhood cancer (including non-malignant CNS tumours) diagnosed at ages 0–14 during 1971–2005 (1971–2003 in Finland), born 1960–1990 (1960–1989 in Finland), alive and not emigrated by the end of the year they turned 25. Survivors were identified from national cancer registries and diagnoses grouped per ICCC: ALL (Ia), other leukaemias (Ib–Ie), lymphomas (II), CNS tumours (III), non-CNS solid tumours (IV–XI). Comparison groups: (1) Five population comparisons per survivor matched on sex, birth year, and country (region in Sweden); (2) all biological/adopted siblings within ≤5 years age difference. Comparisons were cancer-free up to age 20. Individuals with cancer predisposition syndromes (Down syndrome, neurofibromatosis, tuberous sclerosis) were excluded. Reference date: date of survivor’s cancer diagnosis; for siblings, the date when the sibling reached the survivor’s age at diagnosis. Outcomes: Annual highest attained education from national registers (1985–2015; 1985, 1987–2014 in Finland), categorized by ISCED: ≤2 (lower secondary or less), 3–4 (upper secondary or non-tertiary post-secondary), ≥5 (tertiary). In Finland, levels below ISCED 3 are not registered and were considered lower secondary or less. Primary outcome: attainment of upper secondary education (ISCED 3) by age 25. Secondary: attainment of upper secondary without delay (by age 19 in Finland/Sweden; age 20 in Denmark). Covariates: Age at reference date (0–6, 7–11, 12–14), highest parental education (upper secondary or higher vs lower), time spent in hospital during and after diagnosis (average inpatient days/year in the first 5 years post-reference date; dichotomized by median within cancer type, country, and period), somatic hospital contacts at ages 20–24 (none, cancer-related, other), and psychiatric hospital contacts at ages 20–24 (none, any). Statistical analysis: Logistic regression to estimate ORs and 95% CIs for not attaining upper secondary education by 25. Survivors vs population comparisons analyzed with unconditional logistic regression (crude and adjusted for matching factors: country, sex, age, reference year in 10-year intervals). Survivors vs siblings analyzed with conditional logistic regression (within-sibset; crude and adjusted for sex and reference year). Main analyses also run by diagnostic group (ALL, other leukaemias, lymphomas, CNS tumours, non-CNS solid tumours). Stratified analyses by sex, age, calendar period (1971–1989 vs 1990–2005), country, time in hospital, and hospital contacts; additional age-by-period stratification for ALL. Role of parental education assessed via adjustment, stratification, and additive interaction (RERI with 95% CI). Among those who attained by 25, logistic regression compared attainment without delay. Software: SAS 9.4 and Stata 14; alpha=0.05. Sample sizes: 7629 survivors, 35,411 population comparisons, 6114 siblings.

- Attainment by age 25: 6127 survivors (80.3%) vs 29,880 population comparisons (84.4%) and 5135 siblings (84.0%) attained upper secondary education by 25. - Overall risk of not attaining by 25: adjusted OR 1.32 (95% CI: 1.23–1.40) comparing survivors to population comparisons; OR 1.57 (95% CI: 1.40–1.77) vs siblings. - By diagnosis: Strongest deficits among CNS tumour survivors (ORsurvivors vs population comparisons 2.05, 95% CI: 1.83–2.29; ORsurvivors vs siblings 2.72, 95% CI: 2.19–3.39). ALL showed modest elevation (OR 1.15, 95% CI: 1.00–1.33 vs population; OR 1.27, 95% CI: 0.98–1.65 vs siblings). No clear associations for other leukaemias, lymphomas, or non-CNS solid tumours vs population comparisons. - Effect modifiers: Associations varied by sex, age at diagnosis, period, and country. For CNS tumours, the strongest deficits were among those diagnosed before school age (0–6 years). For ALL, lower attainment was most pronounced with diagnosis at ages 12–14 and largely confined to 1971–1989 for younger ages; ages 12–14 showed deficits in both periods. - Hospitalization and late effects: Longer time in hospital during/after diagnosis was associated with higher odds of not attaining by 25 (OR 1.61, 95% CI: 1.48–1.76 vs population comparisons), consistent across diagnostic groups. Hospital contacts at ages 20–24 increased risk, especially psychiatric contacts (OR 4.00, 95% CI: 3.26–4.90). Survivors with shorter hospitalization and no specified hospital contacts had odds similar to population comparisons in several non-CNS groups. - Parental education: Higher parental education was associated with higher attainment for both survivors and comparisons. The difference between survivors and comparisons was more pronounced among those with higher parental education; no significant additive interaction (RERI not statistically significant). Among individuals with parents of lower education, survivors of leukaemia and non-CNS solid tumours attained upper secondary education at least as often as population comparisons. - Educational delay: Among those who attained by 25, completion without delay occurred in 4361 survivors (71.3%) vs 18,801 population comparisons (77.2%) and 3326 siblings (77.7%). Adjusted ORs for completion without delay were 0.75 (95% CI: 0.70–0.80) vs population comparisons and 0.72 (95% CI: 0.64–0.81) vs siblings. This pattern was consistent across diagnostic groups and countries. - Tertiary education by age 30: A smaller proportion of survivors attained tertiary education by 30 compared with comparisons, though differences diminished when restricting to those who had attained upper secondary by 25; survivors of non-CNS solid tumours showed similar tertiary attainment to peers in this restricted group.

The study shows that childhood cancer survivors are overall slightly less likely to complete upper secondary education by age 25 compared with peers, with the deficit primarily driven by CNS tumour survivors. Survivors of non-CNS cancers largely catch up by young adulthood, indicating that educational delays rather than permanent deficits are common outside CNS involvement. The elevated risks associated with longer hospitalization and hospital contacts in early adulthood, especially psychiatric contacts, underscore the impact of treatment burden, complications, and late effects on educational trajectories. The lower attainment in ALL survivors diagnosed in earlier decades likely reflects historical use of cranial radiotherapy, with heightened vulnerability in younger children and persistent risk in those diagnosed at ages 12–14. Parental education strongly influences outcomes; notably, relative differences between survivors and comparisons were larger among those with higher-educated parents, suggesting that childhood cancer may be particularly disruptive to otherwise advantaged educational trajectories or that relative differences appear larger where non-completion is rarer. The findings emphasize the need for tailored educational support, particularly for CNS tumour survivors, those with substantial hospitalization, and those with psychiatric morbidity, and they support ongoing surveillance of educational outcomes in survivorship care.

By leveraging comprehensive registry data across three Nordic countries with comparisons to both the general population and siblings, the study demonstrates that most childhood cancer survivors experience educational delays but many catch up by age 25. Exceptions include CNS tumour survivors, who have persistently lower attainment of upper secondary education. Identified risk groups include survivors with longer hospitalization around diagnosis and those with somatic or psychiatric hospital contacts in early adulthood. Parental education is an important determinant and modifier of survivor–peer differences. These insights inform guidelines for surveillance and highlight the need for targeted educational support during and after treatment. Future research should incorporate detailed treatment data (e.g., cranial irradiation), assess more recent treatment eras, explore mechanisms behind psychiatric morbidity’s impact, and evaluate interventions to mitigate educational disruption.

- Lack of detailed treatment information (e.g., cranial radiation), limiting ability to link specific therapies to educational outcomes. - No data on reasons for delayed graduation, which may not always reflect educational difficulties. - Potential heterogeneity across countries’ systems and over calendar time; while matched and stratified analyses were performed, generalizability beyond the Nordic context may be limited. - Long study period with evolving treatments; findings may not fully apply to contemporary cohorts. - Limited statistical power for rarer, more specific cancer subtypes and some subgroup analyses.