Methods

Design, study population and research setting

The Swiss Childhood Cancer Survivor Study (SCCSS) is a nationwide follow-up study of all patients registered in the Swiss Childhood Cancer Registry (SCCR), who were diagnosed with cancer between 1976 and 2010 before the age of 21 years, and who survived at least 5 years after diagnosis.21 The SCCR centrally registers all children and adolescents, who are diagnosed with leukaemia, lymphoma, central nervous system (CNS) tumours, malignant solid tumours, or Langerhans cell histiocytosis in Switzerland.22,23As part of the SCCSS, we sent questionnaires to adolescent survivors of childhood cancer aged 16-19 years at study between 2007 and 2017. The SCCSS questionnaires were developed based on the questionnaires used in North American and British childhood cancer survivor studies to increase international comparability.24,25We added questions on socio-economic characteristics adapted to the Swiss context.26,27We asked survivors for consent to contact their siblings as comparison group. Adolescent siblings of the same age range (16-19 years) received the same questionnaire between 2009 and 2012 without cancer-related questions. Ethical approval of the SCCR and the SCCSS was granted by the Ethics Committee of the Canton of Bern (166/2014; 2021-01462).

Outcome measures – body image

We assessed the body image of survivors and siblings with three statements in the questionnaire: i) I am satisfied with my body image , ii) I would like to change a few things regarding my body , and iii) I would like to change many things regarding my body . These questions have been previously used in healthy adolescents in Switzerland as part of the Swiss multicentre adolescent survey on health in 2002 (SMASH-2002).28Survivors and siblings were asked to indicate their level of agreement with these statements on a 4-point Likert scale (0=completely agree; 1=slightly agree; 2=slightly disagree; 3=completely disagree). For analysis purposes we reverse coded statements ii) and iii) with higher scores indicating a more negative body image.

Socio-demographic characteristics

For survivors and siblings, we assessed the following socio-demographic characteristics in the questionnaire: age at study (16-17 years; 18 years; 19 years), sex, language region in Switzerland (German; French or Italian), migration background, and whether they currently have a boyfriend or girlfriend (no; yes). We considered survivors as having a migration background if they were not living in Switzerland since birth, were not Swiss citizens since birth, or have another nationality than or in addition to the Swiss nationality.

Clinical characteristics and chronic health conditions

We obtained the following clinical characteristics from the SCCR: age at diagnosis (<5 years; 5-10 years; >10 years), cancer diagnosis according to the International Classification of Childhood Cancer - Third edition (ICCC-3)29, treatment, time since diagnosis (5-10 years; 10-15 years; >15 years), and history of relapse (no; yes). For analyses, cancer diagnoses were categorized into leukaemia (ICCC-3 Group I), lymphoma (II), CNS tumour (III), bone tumour and soft tissue sarcoma (VIII, IX), and other tumours (IV, V, VI, VII, X, XI, XII, and Langerhans cell histiocytosis). Treatment modalities were coded hierarchically into surgery only, chemotherapy (may have had surgery), radiotherapy (may have had surgery or chemotherapy), and haematopoietic stem cell transplantation (HSCT; may have had surgery, chemotherapy, or radiotherapy).
In the questionnaire we collected information on chronic health conditions involving the cardiovascular, pulmonary, and endocrine system, hearing and vision problems, and musculoskeletal, renal or digestive, and neurological conditions. Chronic health conditions were asked using questions from the North American24 and British25 Childhood Cancer Survivor Studies. Survivors had to indicate whether they suffer from symptoms/ diseases involving the respective body systems. Survivors were classified as having the chronic health condition if at least one of the corresponding symptoms/ diseases was reported. If information on symptoms/ diseases was missing, we assumed the condition is not present or at least not serious, as done previously.30 We then created a sum score based on the number of body systems affected by chronic health conditions for each survivor.

Statistical analysis

We used descriptive statistics to describe the study population and chi-squared tests to compare clinical characteristics between participating and non-participating survivors. To increase the comparability between survivors and siblings, we standardized siblings for age at study and sex according to the distribution in survivors as previously done.31-33We used multivariable logistic regression with being a sibling as outcome to calculate appropriate weights. The weight for survivors was set to 1 and all subsequent analyses were based on weighted siblings.
First, we used chi-squared tests to compare the level of agreement with the three body image statements between survivors and weighted siblings. We applied ordered logistic regression to identify potential differences across ordered categories. Among survivors, we then fitted ordered logistic regression models to identify associations between the level of agreement with the three body image statements and socio-demographic characteristics, clinical characteristics, and chronic health conditions. Separate models were created for each body image statement. We a priori decided to adjust each model for the potential confounding factors age at study and sex based on previous literature.34 In addition, we separately analysed the association between body image and growth hormone deficiency due to its relatively high prevalence and direct impact on the body.8 All analyses were performed using Stata version 15.1 (StataCorp. 2017. Stata Statistical Software: Release 15 . College Station, TX: StataCorp LLC).