Abstract
Introduction
Despite improved cure rates in childhood cancer, treatment morbidity and decreased quality of life remain concerns. Children and adolescents with cancer frequently report lower quality of life than healthy peers (Bult et al., 2019; Eiser et al., 2017; Fardell et al., 2017; Yağci-Küpeli et al., 2012). Children with cancer have described multiple factors affecting their quality of life, including inability to participate in their preferred activities, isolation from peers due to hospitalizations and infectious risk, sadness and anger at diagnosis, growing closer to family, and a new perspective on life coinciding with persistent worry about disease recurrence (Anthony et al., 2017). Children tend to have poorer quality of life in earlier treatment phases with improvements over time (Dupuis et al., 2016; Pan et al., 2017; Rodgers et al., 2019). Effect of child age on quality of life has been less clear, with some studies finding an association between younger age (Robson et al., 2021) and others of older age (Pan et al., 2017; Pierce et al., 2017; Raybin et al., 2021) with poorer quality of life. Other studies have found no age effect (Rosenberg et al., 2016). Additionally, family factors, such as lower household income (Ilowite et al., 2018; Robson et al., 2021), greater family psychosocial risk (Pierce et al., 2017), maternal chronic disease (Ocak et al., 2021), family conflict (Desjardins et al., 2022), and unhealthy family functioning (Zheng et al., 2018) are also associated with lower quality of life in children with cancer.
Parenting a child with cancer is often highly stressful; parents of children undergoing cancer treatment experience higher than average levels of anxiety and depressive symptoms in comparison to the general population (Katz et al., 2018). This disparity persists beyond the period of active cancer treatment, as parents of children >5 years from their diagnosis continue to report higher rates of anxiety and depression than population norms (Fardell et al., 2021). In one study, three-fourths of parents of children with cancer experienced symptoms of posttraumatic stress disorder with a higher burden in mothers and in cases of relapsed disease (Banienė & Žemaitienė, 2020). Mothers of children with cancer also require more mental healthcare visits than those with well children (van Warmerdam et al., 2020).
Importantly, parent and child well-being are closely tied. A recent meta-analysis of studies assessing distress in pediatric cancer demonstrated that high distress in parents was consistently associated with high distress in their ill children, regardless of whether child distress was self-reported or via parent-proxy (Bakula et al., 2019). Quality of life in children with cancer has been negatively impacted by parents’ chronic stress (Hamner et al., 2015), parenting stress, and illness uncertainty (Mullins et al., 2016). Furthermore, increased caregiver psychosocial distress has been associated with lower child quality of life in children with cancer, both during and after treatment (Bakula et al., 2020; Desjardins et al., 2022; Pierce et al., 2017).
However, associations between parent mental health and child quality of life in children with relapsed/refractory cancer and their parents have not been previously evaluated. The objective of the current study was to examine relationships between parental anxiety and both parent-proxy and child self-reported health-related quality of life (HRQoL) in children with advanced cancer. Children with relapsed cancer face an overall poorer prognosis, added uncertainty, increased symptom burden, and greater potential for mortality. Given these concerns, we expect to have similar findings in our population of children with relapsed/refractory cancer as in prior studies of children during initial treatment or survivorship: Higher parental anxiety will be associated with lower child HRQoL by both parent-proxy and child self-report.
Method
Participants
Children and parents were recruited from a pediatric oncology clinic at a single center in the southeastern United States between February 2019 and May 2021. All patients who attended appointments in the pediatric oncology clinic with a diagnosis of advanced cancer, defined as any stage of relapsed or refractory cancer, were screened for eligibility. Children were eligible if they were aged 5 to 17 years, diagnosed with relapsed or refractory cancer, had no cognitive impairment, and were able to speak and understand English. Parents were aged 18 years or older, had no cognitive impairment, and spoke and understood English. Because the current cross-sectional study was part of a larger study evaluating the benefits of animal-assisted interventions for children with advanced cancer, participants were excluded if they endorsed fear of or allergy to dogs.
After receiving approval from our institution's institutional review board and permission from children's primary oncology providers, trained research staff approached parents of eligible patients during children's routine clinic visits. Researchers described the study to children and parents, ensuring all questions were answered. For those interested, parents gave verbal consent, and children provided verbal assent. Upon this agreement to participate, researchers immediately obtained written consent from parents and written assent from children prior to participation in any study activities.
Measures
Parents first completed a demographic form, which included date of their child's cancer diagnosis and type of cancer. Parents then completed the State-Trait Anxiety Inventory-Trait (STAI-T) form for adults, which is a self-report anxiety measure that has been commonly used in studies of parents of children with cancer (Cakiroglu et al., 2021; Santacroce, 2002). The State-Trait Anxiety Inventory consists of a total of 40 items with 20 assessing current or “state” anxiety and 20 assessing typical baseline or “trait” anxiety. For this study, parents completed the 20-item trait form only, which typically required 5 to 10 min. Example items include, “I worry too much over something that doesn’t really matter” and “I feel nervous and restless.” Each item has four response options, ranging from not at all to very much so. Scores range from 20 to 80 and higher scores are indicative of greater anxiety. The STAI has demonstrated high internal consistency reliability (Cronbach's α > 0.9) in prior studies evaluating anxiety in parents of children with cancer (Santacroce, 2002). The reliability of the scores in this study was 0.94 (Cronbach's α).
Parents and children also completed age-specific measures of HRQoL, which included the Pediatric Quality of Life Inventory (PedsQL) 4.0 Generic Core Scales and PedsQL 3.0 Cancer Module. The PedsQL Generic Core consists of 23 total items across four dimensions: physical function (eight items), emotional function (five items), social function (five items), and school function (five items). The psychosocial health score is the sum of emotional, social, and school functioning scales, while the physical health score includes only the physical functioning scale (Varni, 2017). The PedsQL Cancer Module is comprised of 27 items evaluating the following eight dimensions: pain and hurt (two items), nausea (five items), procedural anxiety (three items), treatment anxiety (three items), worry (three items), cognitive problems (five items), perceived physical appearance (three items), and communication (three items). Both the PedsQL Generic Core and Cancer Module ask the participant to recall how much of a problem each item has been for the child in the last month and take approximately 10 min to complete. For children aged 8 to 17 years and parents, a 5-point Likert scale (from 0 = never a problem to 4 = almost always a problem) is utilized, while a simpler 3-point Likert scale is used for children aged 5 to 7 years. Items are reverse-scored such that a higher score on each scale means better functioning within that dimension. The PedsQL Generic Core has shown excellent internal consistency reliability (total score Cronbach's α = 0.93 parent, 0.88 child) in children with cancer, while the PedsQL Cancer Module was reliable in parents but more variable in younger children (average score Cronbach's α = 0.87 parent, 0.72 child; Varni et al., 2002). The validity of the PedsQL has also been demonstrated with good discriminability between healthy children and those with cancer (Varni et al., 2002).
Data Analysis
We used IBM SPSS Statistics (Version 28) for all analyses. Continuous data distributions were summarized using either mean and standard deviation (child age) or median and interquartile range (IQR) due to skewness of the distribution. Categorical data were summarized using frequency distributions. We used Spearman's rho coefficients to assess the correlations between parent-reported and child-reported child HRQoL, as well as correlations between the total parent STAI-T score and both parent-reported and child-reported QoL scales. If the p-value was <.05, the correlation was considered statistically significant.
Results
The sample consisted of 20 children and their 20 parents (20 dyads). As shown in Table 1, children were a mean age of 9.5 years old, with half of the sample being female, and slightly more than half (55%) having extracranial solid tumors. A majority of children and parents identified as White (≥55%) and 90% of parents were female. Most children were receiving cancer-directed therapies with either curative or palliative intent and no patients died during the study period.
Demographic and Clinical Characteristics (N = 20 Dyads)
Note. CNS = central nervous system.
The median parent STAI-T score for this sample was 42 (IQR = 33, 50), which is in the middle portion of the range of possible scores, yet a score of 40 or higher has been used to define the presence of high anxiety in other studies (Guido et al., 2021; Hart & McMahon, 2006). Summaries of PedsQL scores are shown in Table 2. With the exception of cognitive problems, median child self-reported PedsQL scores were higher than parent scores on all other scales of both the Cancer Module and Generic Core, often by 10 points or more. We observed the strongest correlations between the parent and child PedsQL reports for school functioning, procedural anxiety, and worry. None of the other correlations were statistically significant and several were quite small (e.g., psychosocial health other than school, rs < .15; treatment anxiety, rs = .11; nausea, rs = −.06; see Table 2).
Summaries of PedsQL Scores and Correlations Between Parent and Child Reports (N = 20 Dyads)
Note. PedsQL = Pediatric Quality of Life Inventory; IQR = interquartile range; rs = Spearman's rho; Mdn = median.
n = 16 dyads.
*p < .05. **p < .01.
Correlations of parent STAI-T scores with parent-reported and child self-reported PedsQL scores are presented in Table 3. We observed statistically significant inverse correlations between parent trait anxiety and parent-reported child psychosocial health (rs = −.54, p = .013), emotional functioning (rs = −.49, p = .027), and school functioning (rs = −.45, p = .046) on the PedsQL Generic Core, indicating that higher levels of parental anxiety were associated with their own perceptions of lower child quality of life with regard to overall psychosocial health, emotional functioning, and school functioning. The only statistically significant correlation between parental anxiety and their perception of cancer-specific HRQoL was for child pain and hurt (rs = −.45, p = .048). Again, the inverse direction indicated that higher levels of parental anxiety were associated with their own perceptions of lower child quality of life with regard to pain. All other correlations of parental anxiety with the PedsQL Cancer Module scales were smaller and not statistically significant. Similarly, none of the correlations of parent STAI-T scores with child-reported HRQoL on either the PedsQL Generic Core or Cancer Module scales were statistically significant (rs < ±.40, p > .10).
Correlations of Parent STAI-T With Child QoL (N = 20 Dyads)
Note. STAI-T = State-Trait Anxiety Inventory; QoL = quality of life; rs = Spearman's rho; PedsQL = Pediatric Quality of Life Inventory.
n = 16 dyads.
*p < .05.
Discussion
Children with advanced cancer are at risk for reduced HRQoL (Rosenberg et al., 2016; Wolfe et al., 2000, 2015); thus, it is important to identify factors that may impact quality of life in this vulnerable population. The objective of our study was to examine relationships between parental anxiety and both parent-reported and child-reported HRQoL in children with advanced cancer. We found that greater parent anxiety was associated with lower parent-reported psychosocial HRQoL with negative impacts on emotional and school functioning. These results are similar to previous findings from a general pediatric oncology population which indicated that parental anxiety and stress are negatively associated with parent-reported HRQoL (Hamner et al., 2015; Link & Fortier, 2016). However, in looking at the specific PedsQL scales, there were some differences between our current study and prior studies, which not only found a significant association with emotional functioning but also with physical and social functioning (Hamner et al., 2015; Link & Fortier, 2016). In contrast to our findings, neither demonstrated a statistically significant association between parent stress/anxiety and school functioning. Of note, our study was limited to children experiencing cancer relapse or refractory disease, a time when many children are not able to attend school. This is different from the populations in the aforementioned studies, which were comprised of a significant proportion of children with acute lymphoblastic leukemia (ALL) or who were off therapy; children almost universally attend school during the maintenance phase of ALL therapy and after completion of therapy. Thus, we suspect that parents in our study were in a unique situation as compared to parents in the prior studies, as they may not have been able to answer the school functioning questions as objectively due to having to imagine how their children would be doing if they were in school rather than observing it. Prior work has shown that increased maternal anxiety is associated with more catastrophizing, particularly in ambiguous situations (Lester et al., 2012). This may have resulted in parents with greater anxiety answering questions more pessimistically (perceiving that their children would not be able to keep up with schoolwork or pay attention if they were in school) while less anxious parents may have been more optimistic, feeling that their children were still capable of these tasks despite not currently attending school. Thus, the ambiguity about school for our study population may have exaggerated the effect of parental anxiety on perceived school functioning.
With regard to pain, our findings are consistent with prior research in children during and after cancer treatment, which found a positive correlation between parent anxiety and parent perception of child's pain frequency and severity with no significant correlation between parent anxiety and child-reported pain (Link & Fortier, 2016). Prior qualitative analyses in children after cancer treatment have shown that parents frequently continue to perceive their child's pain as a threat, associating it with many of their child's most traumatic medical experiences, while children tend to be more objective or look to their parents to determine the appropriate level of worry for their pain (Tutelman et al., 2019). This difference between parents and children in their assessments of the meaning of pain may help explain the why there was an association between parent anxiety and parent perception of the child's pain but not between parent anxiety and child self-reported pain.
Contrary to our hypothesis, correlations of parent anxiety with all dimensions of child-reported quality of life were not statistically significant. This may be explained, in part, by the fact that correlations between child and parent reports on most of the PedsQL scales were small and not statistically significant. While children generally reported better HRQoL than parents, child-reported PedsQL Generic Core scores in our study were slightly lower than those previously reported by children with incurable brain cancer and comparable to those reported by children with relapsed/refractory ALL prior to chimeric antigen receptor T-cell therapy (Laetsch et al., 2019; Mandrell et al., 2016). However, PedsQL Cancer Module scores were higher than reported previously by children with advanced cancer (Akard et al., 2021). The discrepant findings between parent and child perspectives have been noted in other studies evaluating quality of life in childhood cancer and other chronic childhood illnesses, primarily with parents reporting worse HRQoL, as we found in the current study (Galloway et al., 2019; Lam et al., 2021; Mack et al., 2020). While we did not evaluate parents' own HRQoL, poorer parent quality of life has been associated with a greater difference between parent and child reports of child HRQoL (Mack et al., 2020). Similarly, prior work in pediatric oncology has shown that parental stress mediates the relationship between child-reported symptom burden and parent-reported child quality of life, but not child self-reported quality of life (Lam et al., 2021). However, it could also be that the PedsQL Generic Core and Cancer Module do not fully capture what is most meaningful to patients with relapsed cancer. In prior qualitative research, children and adolescents with cancer have identified that psychological growth and relationships are important for their quality of life but these are often not evaluated on quantitative quality of life measures, while physical functioning and cognitive concerns may be overrepresented (Anthony et al., 2017).
Our study's limitations include the relatively small sample size, as is common in other studies of children with advanced cancer, and cross-sectional design. Additionally, while the STAI-T is used to identify parents' typical baseline anxiety, if parents were to recall a recent anxiety-provoking situation (such as a conversation about their child's disease relapse), their STAI-T score may reflect a heightened level of anxiety from their true baseline. Furthermore, due to the cross-sectional approach, we were only able to evaluate correlation and not causation. Thus, it is not possible to know whether parental anxiety drives perception of quality of life or perceived poor quality of life causes increased anxiety. For example, for children enrolled in a Phase 1 study for treatment of diffuse intrinsic pontine glioma, lower parent-reported brain tumor-specific HRQoL, increased cognitive problems, and greater procedural anxiety predicted lower mental health HRQoL in parents (Mandrell et al., 2016). This suggests that parents who witness a decline in their child's quality of life as their disease progresses may develop worsening distress. While parent mental health and child quality of life are very intertwined and the relationship may vary significantly within and among individual families, better understanding the direction of the relationship and whether one more commonly tends to give rise to the other would promote development of the most appropriate intervention.
Implications for Research and Clinical Practice
Given the need to better understand the nuances and directionality of the relationship between parent anxiety and parent-reported child quality of life, future research should be qualitative, exploring factors contributing to parental anxiety and perception of their child's quality of life more comprehensively. Furthermore, evaluating parent quality of life in addition to anxiety in this population would allow for a broader understanding of parental well-being and review of other parental factors that may be associated with quality of life in children with advanced cancer. We do not yet know if an intervention targeting parent anxiety would result in improved quality of life; an intervention study to evaluate this would also be beneficial. With knowledge of the current findings, all families of children with advanced cancer should receive, at minimum, a screening mental health assessment and be provided access to appropriate mental health resources. Routine involvement of interdisciplinary psychosocial team members, such as psychologists, child life specialists, and social workers, is essential for provision of the best care with optimization of quality of life for children with advanced cancer and their families. Additionally, it is critical that all children facing cancer relapse have access to palliative care services to ensure symptoms of disease and its treatment are managed, the family is supported, and goal-concordant care is provided.
Footnotes
Acknowledgments
The authors are grateful to the families who graciously participated in the study. Special thanks to Rebecca Brisman for her work in data collection.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Human-Animal Bond Research Institute (HABRI, grant number: HAB18-010) and National Center for Advancing Translational Science/National Institutes of Health (grant number: UL1 TR000445).
