Discussion
These findings confirm our hypothesis that age differences exist in QOL. Older children reported worse quality of life, which is clinically seen with adolescents tolerating cancer treatment worse than toddlers. Our findings contrast with two studies, one in children with leukemia,2 and in children with advanced cancer,6 in which younger children had worse reported QOL. In the study of children with leukemia, perhaps this difference was related to a younger age range in the study, or because leukemia is more common in young children, whereas our study examined children with all cancer diagnoses. Compared to the findings in the study of children with advanced cancer, our study’s findings may reflect higher symptom burden in adolescents in the first year of chemotherapy due to treatment toxicity.
Older children also had less erect posture, which is confirmed by the literature and fits with clinical observations that young children sit and stand straighter than adolescents. However, our hypothesis that sex differences exist in posture was not confirmed and differs from published literature. Further study is warranted to explore whether sex differences truly exist in posture.
Pain and nausea are the most prevalent distressing symptoms reported by children with cancer.5,44 We also found that age had a moderate relationship with pain and nausea and therefore both should be targeted for symptom management especially in older children. Potential strategies that may appeal to adolescents include integrative modalities that have been shown to improve both pain and nausea in children with cancer.45,46
In our study, worry had the strongest relationship with age. This suggests the need for improved treatment to address anxiety through medication management and cognitive behavioral interventions,47 particularly for older children. Integrative modalities such as aromatherapy or acupuncture also are known to have usefulness in children with cancer and may be better tolerated in adolescents.48
We did not find any other demographic associations with QOL, posture, or Faces Scale scores, which diverged from our hypotheses and from the literature. Other studies have found that girls report worse QOL, but our results showed no differences between girls and boys on either the subjective or the objective QOL measures. Although disparities for minorities have also been reported in QOL for children with cancer, we did not find any differences based on race or ethnicity. However, we collapsed the racial and ethnic subsets into a single dichotomous “minority status” variable due to sample size limitations, a data analysis strategy that could mask results for specific races if these exist.
We found moderately strong relationships between posture, PedsQL, and the Faces Scale, which confirm the idea that all three of these metrics are related to a single underlying construct, QOL. These findings suggest that children who report that they feel better, also stand up straighter, and pick a happier face on the scale. Posture was particularly related to the more physical aspects of the PedsQL such as pain and nausea, which is consistent with the theory of embodiment in which the body stature reflects symptoms and emotions.49 Children’s report of pain had the strongest relationship with the Faces Scale, which aligns with the original intent of the Faces Scale for pain assessment.26 However, the total PedsQL scores for both child and parent report also had moderate relationships with the Faces Scale, a finding that supports further study of this simple tool as a way to measure global QOL in children with cancer.
Likewise, posture and the Faces Scale were moderately related and thus both measures show promise as simple methods to reflect the overall sense of how a child with cancer is feeling. If these relationships can be confirmed and shown to be responsive to change over time, posture and the simplified Faces Scale could potentially be used as surrogates for the more time-consuming QOL questionnaires. Furthermore, the Faces Scale was the only measure that had 100 percent completion which exemplifies the feasibility of this simplified measure in children under stress during cancer treatment.
Additionally, the posture measure could represent a biologic and objective technique that would obviate the subjective limitations of QOL questionnaires.50,51 Posture measures have the added benefit of not causing increased pain or distress as seen with blood draws or obtaining cerebrospinal fluid. A posture assessment is even less invasive than obtaining a saliva sample, except the child must feel well enough to stand, which may have been the reason that this measure had the lowest completion rate in our study (68%). A sitting posture could also be explored for very sick patients. The posture measure could also obviate the limitations of other objective systemic biomarkers such as cortisol that are influenced by the cancer or its treatment.
As a retrospective secondary analysis, this study was powered for the original aims and no correction was used for inflated alpha. Therefore, the current results must be viewed with caution. Although the measures were previously validated and their psychometric characteristics are generally strong, the subscale for cognition on the PedsQL may warrant further examination because it was the only subscale that reflected a significant difference based on sex (parent report) and SES (child report). Selection bias may have existed because children and their parents who chose to participate in the study may have been more likely to do art or were perhaps more interested in QOL. Additionally, the ages of children in this study were skewed toward younger children. Replication of the current results with a larger and more diverse sample of children with cancer is therefore needed.
This study advances the field of symptom science in children with cancer. By adding to the literature on demographic associations with QOL, future intervention studies can now target the specific high-risk group of older children. More in-depth study of the relationship of sex and age with posture in children with cancer will also further the investigation of posture as a potential biomarker of QOL. The correlation of the Faces Scale to the PedsQL also supports further study of simplified measures of overall QOL in children with cancer. A prospective multi-site longitudinal study is warranted to further examine posture as an objective measure of QOL in children with cancer.