Fax: (011) 46-8-51773184
Long-term outcomes of childhood cancer survivors in Sweden: A population-based study of education, employment, and income
Article first published online: 19 JAN 2010
Copyright © 2010 American Cancer Society
Volume 116, Issue 5, pages 1385–1391, 1 March 2010
How to Cite
Boman, K. K., Lindblad, F. and Hjern, A. (2010), Long-term outcomes of childhood cancer survivors in Sweden: A population-based study of education, employment, and income. Cancer, 116: 1385–1391. doi: 10.1002/cncr.24840
- Issue published online: 18 FEB 2010
- Article first published online: 19 JAN 2010
- Manuscript Accepted: 15 JUN 2009
- Manuscript Revised: 11 JUN 2009
- Manuscript Received: 22 APR 2009
- childhood cancer;
- central nervous system tumors;
- adult survivors;
- long-term outcomes;
- socioeconomic outcomes;
- social adjustment;
- cohort studies;
Studies of different national populations were indispensable for estimating the impact of illness-related disability on social outcomes in adult childhood cancer survivors. The effects of childhood cancer on educational attainment, employment, and income in adulthood in a Swedish setting were studied.
The study population was a national cohort of 1.46 million Swedish residents, including 1716 survivors of childhood cancer diagnosed before their 16th birthday, followed up in 2002 in registries at >25 years of age. Main outcomes were educational attainment, employment, and net income. Markers of persistent disability were considered, and outcomes were analyzed with multivariate linear and logistic regression models adjusted for age, sex, and socioeconomic indicators of the childhood households.
Non-central nervous system (CNS) cancer survivors had similar education, employment, and income as the general population in adjusted models, whereas survivors of CNS tumors more often had no more than basic (≤9 years) education (relative risk [RR], 1.80 [95% confidence interval (95% CI), 1.45-2.23]), less often attained education beyond secondary school (RR, 0.69 [95% CI, 0.58-0.81]), and less often were employed (RR, 0.85 [95% CI, 0.77-0.94]). Predicted net income from work was lower in CNS tumor survivors (P <.001) than in the general population, even after the exclusion of individuals who received economic disability compensation.
CNS tumor survivors had poorer social outcomes compared with the general population, whereas outcomes for survivors of other childhood cancers were similar to the general population. Established late effects highlighted the importance of improved, safer pediatric CNS tumor treatment protocols and surveillance that identified individual needs for preventive and remedial measures. Cancer 2010. © 2010 American Cancer Society.
Diagnostic and therapeutic advances in the management of childhood cancer have considerably improved survival during the past decades. Continuing progress in treatment and surveillance has led to an understanding that the notable portion of the adult population that has survived a childhood malignancy will continue to increase.1, 2 Illness and/or treatment may affect social and economic outcomes.3-6 This impact varies among diagnoses and over time.6
However, the emergence of intensive multimodal treatments may be associated with significant therapy-related morbidity that adds to the functional disability resulting from the primary disease. Some of these effects may occur late in adult life.7 Ongoing improvements in chemotherapy, surgery, and radiotherapy result in new and altered treatment protocols for which the long-term adverse consequences for essential organ systems are not fully known when treatments are introduced. Therefore, repeated long-term follow-up is mandatory to preserve optimal health and quality of survival.
Childhood cancer has previously been reported to be associated with learning difficulties8-10 and an adverse impact on academic achievement.3, 11-14 Conversely, many long-term survivors appear to catch up with normative standards over time.15-18 Central nervous system (CNS) chemotherapy, as well as neurocranial irradiation, have been found to particularly increase the risk for learning impairment and neurocognitive sequelae,4, 11, 19-23 which might interfere with academic achievement.
Poor academic achievement is likely to influence patients' chances of reaching subsequent vocational and economic goals. Prior North American studies24, 25 indicate that adult survivors of childhood cancer are at increased risk for unemployment. A similar risk has generally not been demonstrated in the few European studies that have been conducted, as indicated in a meta-analysis by de Boer et al in which national variation, in addition to cancer type, were identified as a determinant of unemployment.26 A large Dutch study, however, found the percentage of those employed to be lower among survivors than in a comparison group of the community.3
In addition to employment, a detailed study of work-related outcomes of survivors should be combined with related informative measures such as economic status to determine the effect on final socioeconomic outcomes.27 To the best of our knowledge, such outcomes have rarely been studied, but a Norwegian study found that lower mean income was reported among young adult CNS tumor survivors.28
Large–scale studies on social consequences of childhood cancer are indispensable to inform a strategy for targeting preventive measures, remedial measures, or both, to diminish the impact of illness and treatment-related disability on social outcomes in adult life. Yet, to our knowledge, very few large studies outside the North American continent have included multiple social outcomes or ascertained such in adult survivors at an age when they are old enough for final outcomes to be reliably evaluated.
Social and economic outcomes are expectedly more sensitive than medical and psychologic sequelae to the cultural context and national differences in societal remedial programs. In this study, we present population-based data on educational attainment, employment, and income from a Swedish cohort of adult childhood cancer survivors in comparison with general population data. In within-group analyses, we examine the relations between diagnostic subgroups of childhood cancer and social outcomes.
MATERIALS AND METHODS
This study was based on Swedish national registers held by the National Board of Health and Welfare and Statistics Sweden linked through each individual's unique personal identification number. The study population included all individuals born in Sweden between 1963 and 1976 who were still registered as residents in Sweden in December 2002, with the exclusion of the 11,000 patients who had received a first cancer diagnosis after their 16th birthday according to the Swedish Cancer Register. In this study population of 1,457,805 individuals, there were 1716 childhood cancer survivors who had received a cancer diagnosis before their 16th birthday according to the Swedish Cancer Register. Brain tumors and acute lymphatic leukemia (ALL) comprised approximately half of these cases, with a wide variety of cancers noted in the remaining half (Table 1).
|No.||Leukemia||Lymphoma||CNS Tumors||Bone Tumors||Other||All Cancers||No Cancer|
|Mean age in 2002, y||30.3||31.4||32.1||32.3||31.9||31.6||32.7|
|Residency, %||Large city||25.3||34.0||23.1||39.5||30.5||28.2||27.0|
|SES of the childhood household, %||Unclassified||18.0||13.5||20.3||14.8||18.6||18.2||20.2|
|Maternal country of birth, %||Sweden||91.4||90.0||89.9||92.6||90.0||91.4||91.4|
|Other European country||1.4||3.0||3.2||3.7||3.2||2.7||2.5|
|Economic compensation due to disability in 2002, %a||Yes||9.0||4.5||28.3||21.0||9.7||15.3||3.0|
|Student in 2002, %||Yes||11.8||7.5||6.3||9.9||11.7||9.4||9.8|
Educational, Employment, and Income Outcomes
The highest completed education as of December 2002 was derived from the Swedish Educational Register. Education was categorized as basic if the study subject had completed no more than the compulsory 9 years of primary school (≤9 years), and as postsecondary if at least 1 educational level had been completed after secondary school (≥14 years).
The Swedish National Tax Board collects information regarding Swedish residents from many sources based on the personal identification number. Employers each month report the salaries received by their employees, income from social insurances is reported directly from the government agencies responsible for these, and each individual taxpayer reports other income in their yearly declaration of income. This information is summarized in a large number of different variables in the annual Total Enumeration Income Survey. In this study, we used 2 variables from this survey of 2002: 1) employment, indicating having an income from employment in November 2002; and 2) net salary, salaried, or self-employed income before tax deductions, but excluding all income from social insurances.
Socioeconomic Disability Index
To identify individuals who received economic compensation from society because of disability, we created a summarized index based on the following: 1) economic compensation because of disability assistance, indicating the need for a personal assistant at least 4 hours daily (identified in the National Social Insurance Board register); 2) sickness pension, indicating lifelong pension because of longstanding illness or disability; and 3) handicap allowance, indicating a permanent disability (2 and 3 identified in the Total Enumeration Income Survey 2002). More detailed analyses of these indicators have been published in a previous article.5
Socioeconomic Indicators of the Childhood Household
The mothers of the individuals in the study were identified in the Swedish Multigeneration Register. The socioeconomic status (SES) of the household of the mother was identified in the last Swedish Population and Housing Census of 1980, 1985, or 1990 before the study subjects had their 19th birthday. Socioeconomic groups were defined according to a classification created by Statistics Sweden, which is based on occupation but also takes the level of education, type of production, and position at work of the head of the household into account.29 Residency was defined by the characteristics of the community as large city (metropolitan areas of Sweden's 3 largest cities Stockholm, Gothenburg, and Malmö), town (other predominately urban communities), and rural. Maternal country of birth was categorized into 4 geographical groups: Sweden, Nordic countries, rest of Europe, and the rest of the world.
The dichotomized outcome variables of education and employment described above were used as dependent variables in multivariate analysis. Because of the high frequency of some of our outcomes, we used logistic regression on the log scale to calculate estimates equivalent to relative risk (RR).30 All individuals were included in the analysis of education.
Employment was analyzed twice: first, only students (those identified by having received student support or loan in 2002) were excluded; and second, individuals receiving economic compensation due to disability also were excluded. Year of birth was entered as a continuous variable in the regression models. Missing data were entered as a separate category in the analytic models. Model 1 in the logistic regression analysis was adjusted for sex and age only. Socioeconomic variables were added as confounders in Model 2. We calculated 95% confidence intervals using the test-based method.
Net income was analyzed in a linear regression model adjusted for age and sex only, first excluding students, and second also excluding individuals receiving economic compensation due to disability. In all linear regression analyses, we excluded individuals with a net income >1,000,000 Swedish krona (SEK), which is equivalent to 111,265 US dollars (USD).
All statistical analyses were performed using the SAS 9.0 software package for Windows (SAS Institute Inc, Cary, NC).
Sociodemographic indicators of the study population by childhood cancer status are presented in Table 1. The childhood cancer survivors, and particularly those who had survived blood or lymph cancer, had a lower mean age (P <.001) at follow-up than the rest of the population, but sociodemographic background indicators of the childhood household, including parental country of birth, were similar to the general population. Survivors of CNS and bone tumors particularly more often received economic compensation due to disability.
Crude rates of educational attainment and employment and mean net income are presented by diagnostic category of childhood cancer in Table 2. Survivors of CNS tumors had the lowest educational attainment, the lowest rate of employment, and the lowest net income.
|No.||Leukemia||Lymphoma||CNS Tumors||Bone Tumors||Other||All Cancers||No Cancer|
|Highest attained education|
|Basic (≤9 y)||9.7||7.0||15.5||3.7||9.4||10.8||8.8|
|Postsecondary (≥14 y)||32.5||39.5||29.5||39.0||41.4||34.7||36.8|
|Mean net income 2002a||154,324||219,495||141,528||205,172||172,196||166,651||191,751|
Table 3 presents the multivariate analysis of education and employment. Survivors of brain tumors more often had attained a basic education only. Despite their higher mean age, survivors of brain tumors less often (RR, 0.69) had a postsecondary education and less often (RR, 0.85) were employed. Adjusting the analysis to potential socioeconomic confounders had marginal or no effect on these estimates. However, when the analysis was restricted to survivors without disability compensation, the risk ratios of employment became very similar to the general population.
|Diagnostic Category||No.||Model 1a RR (95% CI)||Model 2b RR (95% CI)|
|Postsecondary education (≥14 y)||Leukemia/lymphoma||489||0.93 (0.80-1.08)||0.92 (0.79-1.07)|
|CNS tumors||537||0.69 (0.58-0.81)||0.69 (0.58-0.81)|
|Other cancers||690||1.09 (0.97-1.22)||1.09 (0.97-1.22)|
|Basic education only (≤9 y)||Leukemia/lymphoma||489||1.03 (0.76-1.40)||1.07 (0.79-1.45)|
|CNS tumors||537||1.78 (1.44-2.21)||1.80 (1.45-2.23)|
|Other cancer||690||1.04 (0.81-1.35)||1.05 (0.82-1.36)|
|Employmentc||Leukemia/lymphoma||440||0.98 (0.89-1.08)||0.98 (0.89-1.08)|
|CNS tumors||503||0.85 (0.77-0.94)||0.85 (0.77-0.94)|
|Other cancers||611||0.95 (0.87-1.03)||0.95 (0.87-1.03)|
|Employmentd||Leukemia/lymphoma||406||1.02 (0.92-1.13)||1.03 (0.93-1.13)|
|CNS tumors||355||0.98 (0.88-1.10)||0.98 (0.88-1.09)|
|Other cancers||545||0.99 (0.90-1.08)||0.99 (0.90-1.08)|
Table 4 demonstrates the linear regression analysis of net salary. The mean net salary of the comparison population without a previous cancer diagnosis was 203,175 SEK and the median net salary was 211,200 SEK, which was equivalent to ∼ 22,613 and ∼ 23,507 USD, respectively, at that date. All diagnostic subcategories of cancer survivors had lower mean net salaries than the comparison population, even after disabled individuals had been excluded. When the analysis was compensated for age and sex in a linear regression, however, only survivors of brain tumors were found to have a statistically significant lower net salary compared with the noncancer population.
|Diagnostic Category||No.||Mean, SEKb||Median, SEKb||B, SEKb||P|
There were no indications in interaction analyses that the effects of childhood cancer outcomes differed by the SES of the household.
In this national population-based cohort study of 1716 survivors of childhood cancer, followed in national registers at the ages of 26 to 39 years, survivors of CNS tumors were found to have poorer educational and employment outcomes than a general population of 1.46 million Swedish noncancer survivors of the same ages. Net income remained significantly lower among survivors of CNS tumors, after the exclusion of individuals with indications of disability. Survivors of other types of childhood cancer presented educational, employment, and income outcomes similar to the general population.
The findings of the current study regarding education are in line with what has previously been demonstrated in a Danish register study13 and a Dutch self-report questionnaire study.3 A Canadian clinical questionnaire study4 identified significant educational deficits for all diagnostic subgroups, with survivors of CNS tumors, leukemia, and neuroblastoma reporting the worst outcomes. Other clinical questionnaire studies of survivors of different pediatric cancers have also reported that young adult survivors in general fall short in education compared with siblings.20
In the Canadian study referred to earlier,4 survivors who had received cranial irradiation had poorer educational outcomes. It is well known that the treatment of patients with ALL and non-Hodgkin lymphoma using cranial irradiation (with the possible addition of intrathecal methotrexate for patients with ALL) carries risks for cognitive dysfunction19 and a poor academic career.11 Cranial irradiation has previously been identified as the strongest independent prognostic factor of educational outcome in survivors.3 Neurodevelopmental impairment caused by cranial radiotherapy23, 31 is most likely a partial explanation for our findings regarding the vulnerability of CNS tumor survivors.
Employment outcomes also demonstrated a worse prognosis for the survivors of CNS tumors. In a recent meta-analysis by de Boer et al regarding survivor employment, the risk of unemployment was almost doubled for survivors in general,26 whereas survivors of CNS tumors had a nearly 5-fold risk of unemployment (idem). Given this large risk according to such findings, those of the current study may appear unexpectedly favorable. The exclusion of individuals aged <25 years, who supposedly are at a higher risk for unemployment, may partly explain our findings. Effective national societal remedial measures aimed at compensating disability may be an additional explanation. The study of De Boer et al demonstrated that survivors in general in the United States were at an overall 3-fold risk of becoming unemployed, compared with no such risk for European survivors. A recent North American study24 appears to confirm that societal programs play a role for final employment outcomes of survivors. Adequate and ambitious welfare measures may prevent disability from turning into a handicap. Such a notion appears to gain at least partial support from the findings of the current study, because employment deficits among CNS tumor survivors disappeared when individuals with economic disability compensation were excluded. Hence, a full understanding of socioeconomic outcomes of survivors needs to consider the context provided by the welfare system of the actual society.
In addition, previous studies have typically addressed groups comprising younger survivors than shown in the current study.3, 18, 24, 26, 32 The effects of a potential catch up later in life15, 17 or of late development of health-related sequelae regarding certain outcomes cannot be evaluated in the young members of such groups, despite the finding that some health-related sequelae still may develop during adulthood33 and counter such a catch-up. However, our exclusion of those patients aged <25 years was done to avoid a bias due to age-related difficulty in evaluating final social goals in adulthood and may partly explain the differences between some prior findings and those of this study.
Measures of income reflect complementary aspects of vocational conditions and are most likely more sensitive than the employment measures in uncovering consequences related to, for example, less qualified jobs and part-time employment. A lower net income than in the noncancer population was found for survivors of CNS tumors. Although this difference was attenuated when survivors with a disability were excluded from analyses, it remained statistically significant. This notable finding indicates that established subtle emotional and psychosocial late effects34 and neurocognitive impairment35 still play a significant role in cases for which obvious expressions of disability are absent.
Even if long-lasting emotional reactions should not be understated, the average social adaptation after non-CNS cancer demonstrates an impressive restorative capacity after exposure to a combined severe medical and psychologic childhood trauma. The parents, although themselves at risk for experiencing high levels of disease-related distress,36 appear to have been able to provide their child with sufficient support and emotional shelter. The fact that medical care for children is free in Sweden, and available to all in a fairly equitable manner, is most likely an important factor that reduces an additional burden of financial strain, and facilitates for the family to sustain their child.
An advantage of the national register study approach is that it allows for the investigation of entire populations of survivors. Compared with clinical studies and self-reported data, attrition is not a problem, and the validity of the Swedish Cancer Register data is high, with, for example, approximately 99% of the cancer cases being morphologically verified.37 Conversely, an essential general limitation of the register approach concerns the dependency on data that are already collected. Because subjects remain anonymous to the researchers, refined information that would shed further light on the study questions cannot be collected.
In conclusion, non-CNS tumor survivors of childhood cancer do not present with educational, vocational, or income-related attainments that are significantly below what applies for same age general population comparisons. However, survivors of CNS tumors run a significant risk of adverse socioeconomic consequences in adult life. These findings highlight the importance of safer treatment protocols for children diagnosed with CNS malignancies and individualized surveillance and supportive measures during follow-up to reduce the risk for adverse sequelae in this vulnerable subgroup of survivors. The efficiency of national welfare systems determines the extent to which survivors can be compensated and assisted to optimal social adjustment and quality of survival in cases in which late sequelae remain inevitable.
CONFLICT OF INTEREST DISCLOSURES
Supported by The Swedish Childhood Cancer Foundation, The Cancer and Traffic Injury Fund Sweden, and Karolinska Institutet/Stockholm County Council ALF grants. The research position of Dr. Lindblad was financed by the Swedish Council for Working Life and Social Research.
- 1HewittM, SusanL, WeinerSL, SimoneJV, eds. Childhood cancer survivorship: improving care and quality of life. Washington DC: National Academic Press; 2003.
- 2Late effects of childhood cancer. London, UK: Hodder Education; 2003., .
- 16Educational, occupational, and insurance status of childhood cancer survivors in their fourth and fifth decades of life. J Clin Oncol. 1992; 10: 1394-1406., , , et al.
- 29Statistics-Sweden. Socioeconomic Classification (SEI). Stockholm: Statistics Sweden; 1982.
- 37Cancer Incidence in Sweden 2004. Stockholm, Sweden: The National Board of Health and Welfare, Centre for Epidemiology; 2006.