Question What is the prevalence of and related factors for clinically significant fear of cancer recurrence in adult survivors of childhood cancer?
Findings In this cross-sectional study of 229 North American adults who survived childhood cancer, one-third of survivors reported experiencing elevated fear that their primary cancer will recur or a subsequent malignant neoplasm will develop. Fear of cancer recurrence was associated with chronic health conditions, treatment-related factors, anxiety, depression, and perceived health status.
Meaning These findings suggest that fear of cancer recurrence is common among adult survivors of childhood cancer and targeted interventions are needed.
Importance Fear of cancer recurrence is common among survivors of adult-onset cancer and associated with increased distress, functional impairment, and health care utilization. However, little is known about the prevalence and risk factors of fear of cancer recurrence among adult survivors of childhood cancer who are also at high risk for subsequent malignant neoplasms.
Objective To characterize the prevalence of and risk factors for clinically significant fear of cancer recurrence in adult survivors of childhood cancer.
Design, Setting, and Participants This cross-sectional investigation included participants recruited from the Childhood Cancer Survivor Study, a retrospective cohort study of long-term childhood cancer survivors treated at 31 institutions between 1970 and 1999 across North America. Participants were recruited and completed psychosocial measures via online survey between October 2018 and April 2019. Cancer and treatment-related variables were abstracted from medical records. Data were analyzed from May 2023 to July 2024.
Main Outcomes and Measures Clinically significant fear of cancer recurrence was assessed via the Fear of Cancer Recurrence Inventory-Short Form. Poisson regression models estimated prevalence ratios (PRs) with 95% CIs adjusted for age and sex to examine the associations of demographic, disease, treatment, and psychosocial variables with fear of cancer recurrence.
Results The final sample included 229 adult survivors of childhood cancer (115 female [50.2%]; mean [SD] age, 39.6 [9.9] years; mean [SD] time since diagnosis, 31.7 [8.4] years). Among survivors, 38 (16.6%; 95% CI, 11.6%-21.6%) reported clinically significant fear of cancer recurrence, and an additional 36 (15.7%) reported high fear of cancer recurrence. Clinically significant fear of cancer recurrence was associated with unemployment (PR, 2.5; 95% CI, 1.3-4.8), presence of neurologic chronic health conditions (PR, 3.3; 95% CI, 1.8-6.1), treatment with pelvic radiation (PR, 2.9; 95% CI, 1.5-5.6), and amputation or limb sparing surgery (PR, 2.4; 95% CI, 1.2-4.9). Higher risk of clinically significant fear of cancer recurrence was also associated with having either elevated anxiety or depression (PR, 2.6; 95% CI, 1.2-5.9), having both elevated (PR, 3.2; 95% CI, 1.2-8.4), and perceived poor health status (PR, 3.0; 95% CI, 3.1-9.7).
Conclusions and Relevance Decades following treatment, one-third of childhood cancer survivors in this study reported elevated fear their cancer will recur or a subsequent malignant neoplasm will develop. Findings suggest that fear of cancer recurrence should be routinely screened, and clinically significant symptoms intervened upon as a part of survivorship care.
Fear of cancer recurrence (FCR) is defined as "fear, worry, or concern relating to the possibility that a primary cancer or subsequent malignancy will return or progress to other parts of the body" and is prevalent among survivors of adult-onset cancer. Some vigilance toward pain and associated anxiety is common and likely adaptive in survivors given their increased risk of recurrence of primary cancer and subsequent malignant neoplasms (SMNs). However, clinically significant levels of FCR (CS-FCR; ie, causing functional impairment and/or distress) are associated with several negative outcomes in adult-onset cancer survivors, including anxiety and depression, reduced quality of life, and overuse of outpatient hospital services. Common FCR risk factors in this population include younger current age, physical symptoms (eg, fatigue, pain), female sex, metastatic diagnosis, and treatment factors (eg, chemotherapy). Although FCR has been well characterized among adult-onset cancer survivors, less attention has been paid to FCR among adult survivors of childhood cancer, who are at low risk for recurrence of their primary childhood cancer but at high risk for SMNs(approximately 6-fold increased risk compared with the general population).
Initial studies examining FCR in childhood cancer survivors show results similar to those in adult-onset cancer. However, significant methodological and conceptual limitations exist. For example, FCR has primarily been examined using unvalidated measures that rely on 1 or 2 items and lack clinical cut-off scores, limiting the ability to fully characterize FCR. The role of chronic pain (ie, pain lasting 3 months or more) has also been overlooked despite the inclusion of acute pain in FCR theoretical models, whereby pain acts as a cancer-related cue, triggering FCR-related thoughts and subsequent anxiety. Thus, repeated daily exposure to pain over a prolonged period (ie, chronic pain) may increase the occurrence of FCR. Recent models of FCR also theorize that intolerance of uncertainty, a personality-based factor associated with anxiety and depression, may play a role in the appraisal of cancer-related cues as individuals with high levels of intolerance of uncertainty are more likely to interpret ambiguous cues as threatening. Despite this, the potential mediating role of intolerance of uncertainty in the childhood cancer context is unclear.
This study aimed to characterize the prevalence of and risk factors for CS-FCR in adult survivors of childhood cancer. Associations between FCR and chronic pain as well as intolerance of uncertainty were examined, followed by an evaluation of the potential mediating role of intolerance of uncertainty in associations between FCR and anxiety and depression. We hypothesized that intolerance of uncertainty would mediate the association between anxiety and FCR as well as depression and FCR.
A random sample of 700 childhood cancer survivors (ie, diagnosed younger than 21 years of age; surviving 5 years or more) enrolled in the Childhood Cancer Survivor Study (CCSS) was recruited for an ancillary study: Exploring Aspects of Survivors Pain (EASE). CCSS is a retrospective cohort study of long-term childhood cancer survivors treated at 31 institutions between 1970 and 1999 across North America. Survivors were recruited to EASE via letters, emails, and phone calls and invited to download the Eureka Research app, where all measures were completed. Inclusion criteria for survivors included being older than age 18 years, able to read and speak English, owning a smartphone, and having access to the internet. Participants were recruited and completed measures via the study app between October 2018 and April 2019. Study recruitment and app support were based at St Jude Children's Research Hospital in Memphis, Tennessee.
Ethics approval was received from St. Jude Children's Research Hospital and Concordia University for secondary data analysis. Written informed consent was obtained for the original ancillary EASE study. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for observational studies.
Demographic variables were ascertained from medical records and/or self-reported on CCSS surveys and included sex, race and ethnicity, age, marital status, annual income, employment status, and educational attainment (eAppendix 1 in Supplement 1). Self-reported categories for race and ethnicity included American Indian, Black, Chinese, Hispanic, and White; data on race and ethnicity were included as a study variable because previous studies have shown associations between race, ethnicity, and mental health outcomes in medical populations. Cancer-related variables included diagnosis, age at diagnosis, cancer recurrence, and SMNs. Treatment variables were abstracted from medical records and included chemotherapy, radiation, and amputation or limb-sparing surgery. Chronic health conditions were self-reported on surveys, categorized, and graded 1 (asymptomatic or mild) to 4 (life-threatening) based on the National Cancer Institute's Common Terminology Criteria for Adverse Events. Only conditions with grades 2 to 4 were included.
FCR was assessed via the 9-item Fear of Cancer Recurrence Inventory-Short Form (FCRI-SF), which defines recurrence as the possibility that cancer could return to the same place or another part of the body (eTable 1 in Supplement 1). Items are rated on a 5-point Likert scale ranging from 0 (not at all) to 4 (a great deal), with higher scores indicating greater FCR. Here a total score below 16 indicates minimal levels, 16 to 21 indicates high levels, and 22 or greater indicates CS-FCR. See eAppendix 2 in Supplement 1 for additional information regarding the psychometric properties of the FCRI-SF and other measures.
Chronic pain was assessed via 2 items (ie, "Do you have any persistent or recurrent pain, more than aches and pains that are fleeting?" and "How long have you been experiencing pain?"). Survivors with pain for 3 months or more were categorized as having chronic pain.
Depressive symptoms were assessed via the Patient Health Questionnaire 8-item (PHQ-8). Items are rated on a 4-point Likert scale ranging from 0 (not at all) to 3 (nearly every day). Higher scores indicate greater symptom severity, and a total score of 10 or higher is the cut-point for elevated levels of depression.
Symptoms of anxiety were assessed via the Generalized Anxiety Disorder 7-item (GAD-7). Items are rated on a 4-point Likert scale ranging from 0 (not at all) to 3 (nearly every day). Higher scores represent greater symptom severity, and a total score of 10 or higher is the cut-point for elevated levels of anxiety.
Self-perceived health was assessed using one item from the 36-item Short Form Health Survey. Participants were asked to rate their health on a 5-point Likert scale: 1 (poor), 2 (fair), 3 (good), 4 (very), or 5 (excellent). For analyses, a dichotomous variable defined as poor or fair vs good, very, or excellent was used.
The Intolerance of Uncertainty Scale 12-item (IUS-12) was used to measure intolerance of uncertainty. Items are rated on a 5-point Likert scale, with higher scores indicating greater intolerance of uncertainty. Continuous scores on the IUS-12 were used.
Sleep disturbances were assessed via the National Institute of Health Patient Reported Outcomes Measurement Information System-Sleep Disturbance (PROMIS-SD)-Short Form 8-item. Items are rated on a 5-point Likert scale, with higher scores indicating more sleep difficulties. Continuous scores on the PROMIS-SD were used.
The percentage of survivors scoring 22 or higher on the FCRI-SF and its corresponding 95% CI was calculated to determine the prevalence of CS-FCR. Comparisons of characteristics between participants and nonparticipants were based on χ or Fisher exact test to assess potential bias. Univariate modified Poisson regression models with robust standard errors were used to investigate the association between CS-FCR and potential risk factors and estimate prevalence ratios (PR) with 95% CIs. Before developing multivariable models, we examined the association between FCR and demographic variables (eg, age at EASE survey, sex) selected a priori, which have been shown to be significantly associated with FCR in previous research, to form our base model. Guided by this model, we adjusted for age at EASE survey completion and sex in multivariable models. Separate multivariable regression analyses were conducted examining associations between CS-FCR and (1) demographic variables, (2) cancer diagnosis, (3) chronic health conditions, (4) treatment exposures, and (5) psychosocial factors. Factors included in the final models were selected stepwise using the Akaike information criterion as an aid to determine the most parsimonious model.
Regarding the mediation model, we followed the recommendations summarized in Hayes. Mediation was defined as occurring if the indirect effect was significantly different from zero. To calculate the indirect effect, we multiplied the effect of elevated anxiety and depression on intolerance of uncertainty (a and a) by the effect of intolerance of uncertainty on CS-FCR (effect b). To meaningfully interpret any mediation effects, we estimated indirect effects (a × b and a × b) with 1000 bootstrap samples, reporting confidence intervals based on the 2.5 and 97.5 percentile of the distribution and testing significance of the estimate utilizing the standard deviation of the distribution. Mediation analyses were adjusted for age at EASE survey completion and sex.
Missing data are summarized (Table 1). If a variable with missing data was used in a particular analysis, only survivors with complete data were used. All assumptions required for statistical analyses were met. Nominal 2-sided P values are shown with a significance level set at P < .05. Within each multivariable regression model, adjusted P values accounting for overall false discovery rate (FDR) are also shown. Analyses were conducted using SAS statistical software, version 9.4 (SAS Institute).
Of 700 invited CCSS survivors, 229 (32.7%) consented and completed study procedures, with a mean (SD) age at diagnosis and study completion of 7.9 (6.1) years and 39.6 (9.9) years, respectively (Table 1; eFigure 1 in Supplement 1). The most common diagnosis was leukemia (79 [34.5%]). Of survivors in our sample, 21 (9.2%) experienced a recurrence of their primary cancer and 17 (7.4%) had been diagnosed with a SMN. Ninety-four survivors (41.0%) reported experiencing chronic pain. Demographic and other descriptive data compared EASE participants and nonparticipating members of the invited CCSS cohort, finding significant differences in marital status, race and ethnicity, and chronic endocrine conditions (eTable 2 in Supplement 1).
The mean (SD) score on the FCRI-SF was 12.0 (9.1) (range, 0-36.0) (Table 2). Thirty-eight participants (16.6%; 95% CI, 11.8-21.4) reported experiencing CS-FCR. A further 36 (15.7%) and 155 (67.7%) participants reported experiencing high and minimal levels of FCR, respectively. Among survivors with chronic pain, 24 (25.5%) and 13 (13.8%) reported experiencing clinical and high levels of FCR, respectively.
Initial univariate models assessed the effect of individual risk factors on CS-FCR (eTable 3 in Supplement 1). In our base model, female participants (PR, 2.0; 95% CI 1.03-3.7) were more likely to experience CS-FCR than male participants, although this result was not significant when FDR-adjusted P values were used (Table 3, Model 0). In a final multivariable model with demographic variables, survivors who were unemployed (PR, 2.5; 95% CI, 1.3-4.8) or who completed some college or an undergraduate degree (PR, 4.9; 95% CI 1.2-19.4) were more likely to experience CS-FCR (Table 3, Model 1). No associations were observed between CS-FCR and recurrence of the primary cancer or development of a SMN (eTable 3 in Supplement 1).
Survivors with a neurological condition (PR, 3.3; 95% CI, 1.8-6.1) were more likely to experience CS-FCR (Table 3, Model 3). Survivors who underwent pelvic radiation (PR, 2.9; 95% CI, 1.5-5.6) or limb-sparing or amputation surgery (PR, 2.4; 95% CI, 1.2-4.9) were at higher risk of CS-FCR (Table 3, Model 4).
Regarding psychosocial factors, survivors with either elevated depression or anxiety (PR, 2.6; 95% CI, 1.2-5.9) or both (PR, 3.2; 95% CI, 1.2-8.4) were more likely to experience CS-FCR (Table 3, Model 5). Survivors who rated their health as poor or fair (PR, 3.0; 95% CI, 1.6-5.9) were also at greater risk of CS-FCR compared with good, very good, or excellent (Table 3, Model 5). Although significant in univariate analyses, chronic pain (PR, 1.2; 95% CI, 0.6-2.4) and intolerance of uncertainty (PR, 1.1; 95% CI, 0.9-1.3) were not associated with CS-FCR in this model.
Multiple mediation models with indicators representing various combinations of anxiety and depression dichotomized were calculated. Results discussed are from Model C (eTable 4 in Supplement 1). In examining whether intolerance of uncertainty mediated the association between FCR and anxiety, a direct effect of elevated anxiety on CS-FCR, with or without elevated depression, was found (c' = 1.37; 95% CI, 0.57-2.17) (Figure). After bootstrapping, a significant indirect effect was also observed (a × b = 0.42; 95% CI, 0.02-0.82). There was no statistically significant direct effect of elevated depression, without anxiety, on CS-FCR (c' = 0.98; 95% CI, -0.03 to 2.00; P = .06) and the indirect effect (a × b = 0.11; 95% CI -0.08 to 0.29) was not significant. These findings suggest intolerance of uncertainty is a partial mediator of the association between FCR and anxiety.
In this novel study using a validated and comprehensive measure, we identified approximately 1 in 3 adult survivors of childhood cancer experienced FCR at a severity likely to affect their functioning and where psychological intervention would be recommended. These findings underscore the substantial psychological and functional burden of FCR and suggest health care professionals should routinely assess FCR as a part of providing comprehensive care to long-term survivors. The FCRI-SF and newly developed single-item screeners of FCR are well suited for rapid administration and interpretation within clinical settings.
Similar to studies in young childhood cancer survivors, no associations emerged between FCR and age at survey completion. Unemployment was associated with CS-FCR. As previous studies have shown associations between elevated FCR and functional impairment, survivors with elevated FCR may be experiencing similar impairment and associated difficulties working.
No associations emerged between cancer type and FCR. Previous research in this area is mixed, with no consistent associations observed across studies. Contrary to previous findings among childhood cancer survivors, no associations emerged between FCR and recurrence of the primary cancer or a SMN. Given the limited number of survivors with a primary recurrence or a SMN, it is possible the analyses were underpowered to detect a significant effect.
Consistent with our prior work, survivors with either elevated anxiety or depression or both were at higher risk of CS-FCR. Some chronic neurological conditions are associated with symptoms that could be misinterpreted as a sign of cancer such as pain and weakness. Thus, misinterpretation of these types of symptoms may lead to FCR.
Limb-sparing surgery or amputation and pelvic radiation were also associated with increased risk of CS-FCR. The resulting effects of limb-sparing surgery and amputation may activate survivors' FCR-related thoughts due to more visible and constant reminders of the cancer experience. As pelvic radiation is associated with long-term adverse effects (eg, diarrhea, abdominal pain) that can mimic symptoms of bowel and urinary tract cancers, survivors may misinterpret these symptoms as a sign of a new malignant neoplasm.
Consistent with prior research, survivors with either elevated anxiety or depression or both were at higher risk of CS-FCR. Given the anxiety-based nature of FCR and its high comorbidity with anxiety disorders, as well as between anxiety and depression, it is not surprising associations between these variables were found. Consistent with previous studies, survivors who rated their health as poor or fair were more likely to experience CS-FCR. Chronic pain and intolerance of uncertainty were significant predictors in univariate analyses but did not remain significant in the multivariable model. As nearly 40% of survivors with chronic pain also experienced high or greater levels of FCR, these 2 constructs appear to be intricately linked. However, findings from the multivariable model suggest that psychological factors likely play a stronger role in FCR.
Our findings also indicate intolerance of uncertainty is a partial mediator of the association between FCR and anxiety. Consistent with theoretical models of FCR, intolerance of uncertainty may act as a personality-based factor that colors survivors' experience of anxiety in relation to FCR. Notably, when our mediational analyses are considered in light of our multivariable psychosocial model, anxiety and depression appear to have a stronger effect on FCR than intolerance of uncertainty. Thus, anxiety and depression may play a more substantial role in the development and maintenance of FCR than intolerance of uncertainty.
Regarding treatment, psychological interventions such as cognitive behavioral therapy and acceptance and commitment therapy, as well as mind-body interventions, are effective in reducing FCR in survivors of adult-onset cancer; and may also be effective in childhood cancer survivors. Interventions such as cognitive bias modification have also shown promise. Furthermore, survivors may benefit from less intensive psychological interventions (eg, psychoeducation, peer support groups) to prevent FCR from transitioning to clinical levels.
Study limitations should be noted. As adult survivors are at low risk for recurrence of their primary childhood cancer but at high risk for SMNs, it is unclear to what degree each component is driving the overall prevalence. Studies specifically examining fear around the primary cancer recurring and of SMNs are needed. Nonetheless, that one-third of survivors have elevated or CS-FCR is striking. Although only 32% of contacted survivors participated, minimal differences were observed between participants and nonparticipants in the larger CCSS cohort. Only 1 item was used to assess self-perceived health, limiting the measurement of this variable. A cross-sectional design was used; therefore, inferences regarding causation or directionality cannot be made. Ideally, mediation would be evaluated using longitudinal data to provide stronger evidence of a causal relationship. As previously shown, a cross-sectional mediation analysis could lead to erroneous conclusions. Therefore, our results showing the effect of anxiety on FCR is mediated in part by intolerance of uncertainty should be considered hypothesis generating and be confirmed in a longitudinal analysis with data collected at 3 waves. Survivors were eligible for the current study if they had internet and smartphone access, which may have biased our sample. Future studies should integrate alternative survey completion methods (eg, paper and pencil). The racial and ethnic diversity of the sample was also limited, and future work with more diverse populations of survivors are needed.
Despite being decades from the completion of cancer treatments, one-third of adult survivors of childhood cancer experience elevated fear their primary cancer will recur or a SMN will develop. In addition to providing one of the first characterizations of FCR among adult survivors of childhood cancer, the current study demonstrated the robust role of anxiety, depression, and survivors' perception of their health as risk factors for CS-FCR. Moreover, our findings help lay the groundwork for an improved understanding of FCR among long-term survivors, which is vital to informing psychological screening and intervention efforts.
Corresponding Author: Nicole M. Alberts, PhD, Department of Psychology, Concordia University, 7141 Sherbrooke St West, Montréal, QC H4B 1R6 (nicole.alberts@concordia.ca).
Author Contributions: Drs Alberts and Leisenring had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Pizzo, Leisenring, Alschuler, Olgin, Armstrong, Alberts.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Pizzo, Leisenring, Stratton, Lamoureux, Flynn, Alschuler, Stinson, Alberts.
Critical review of the manuscript for important intellectual content: Pizzo, Leisenring, Flynn, Alschuler, Krull, Jibb, Nathan, Olgin, Armstrong, Alberts.
Administrative, technical, or material support: Lamoureux, Flynn, Alberts.
Supervision: Alschuler, Olgin, Stinson, Armstrong, Alberts.
Conflict of Interest Disclosures: Dr Leisenring reported receiving grants from National Institutes of Health (NIH) during the conduct of the study. Dr Alschuler reported receiving grants from Patient-Centered Outcomes Research Institute, NIH, National Multiple Sclerosis Society, and International Progressive MS Alliance outside the submitted work. Dr Armstrong reported grants from NIH during the conduct of the study. No other disclosures were reported.
Funding/Support: This study was supported by the Childhood Cancer Survivor Study Career Development Award (Dr Alberts, principal investigator), and the National Cancer Institute (grant No. CA55727; Dr Armstrong, principal investigator). Support to St. Jude Children's Research Hospital was also provided by the Cancer Center Support (grant No. CA21765) and the American Lebanese Syrian Associated Charities (ALSAC). This research was also undertaken, in part, thanks to funding from the Canada Research Chairs Program. Support for the Eureka digital research platform was provided by the NIH (grant No. U2CEB021881; Dr Olgin, principal investigator).
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Meeting Presentation: Data were presented, in part, at the International Society for Pediatric Oncology World Congress; Ottawa, Canada; October 2023.