Childhood Cancer: Stress and Distress Among Moms and Kids
On the sociodemographic CR measure, 32.4% of mothers had zero risk factors; 24.5% had one risk factor, 28.1% had two risk factors, 12.7% had three risk factors, and 2.3% had four risk factors.
Time since diagnosis was not significantly related to any psychological distress or stress variable or CR. Children's new diagnosis versus relapse was also not significantly different on any of these measures. Child age was significantly correlated with mother reports of children's cancer-related stress on the RSQ (r = .24, p < .01) but was not associated with any other measures of mothers' or children's stress, distress, or CR.
Mothers' mean score on the BDI-II (M = 15.2, SD = 10.5) was in the mild range of depressive symptom levels (≥14), with 29% of mothers reporting symptoms in the moderate to severe range (scores ≥20; Beck et al., 1996). Mothers' reports of children's affective symptoms on the CBCL indicated 15.9% of scores fell in clinical range (T ≥ 70); 7.4% of the youths who completed the YSR indicated clinically elevated symptoms (T ≥ 70). On the IES-R, 39.6% of mothers (M = 29.3, SD = 17.6) and 24.8% of youth (M = 23.3, SD = 15.3) scored above the cutoff score of 34, indicating elevated levels of PTSS (Rash, Coffey, Baschnagel, Drobes, & Saladin, 2008). Correlations among measures of stress and distress for mothers and children are presented in Table II.
Bivariate Analyses. Bivariate analyses examining associations of sociodemographic variables to mothers' and children's stress and distress are presented in Table III. Single mothers reported significantly greater depressive symptoms and PTSS than partnered mothers. Single mothers also reported significantly greater levels of cancer-related stress on the RSQ than partnered mothers; however, single versus partnered mothers did not differ on general perceived stress on the PSS. Mothers' race was not significantly associated with self-reported levels of stress or distress. Annual family income was significantly negatively correlated with all four measures of mothers' psychological distress and stress (p < .05), ranging in magnitude from ρ = .12 to ρ = .21. Mothers' education level was also significantly negatively correlated (p < .01) with self-reported depressive symptoms, PTSS, general perceived stress, and cancer-related stress (ranging from r = −.15 to −.24).
Children of single mothers reported significantly greater affective problems on the YSR and PTSS on the IES-R in comparison with children of partnered mothers. In contrast, mothers' reports of children's affective problems on the CBCL did not differ significantly as a function of mothers' relationship status. There were also no significant differences for mother- or self-reported cancer-related stress among these children. Non-White race was significantly positively associated with children's self-reported PTSS and cancer-related stress. Family income was significantly negatively correlated (p < .05) with children's self-reported affective problems on the YSR (ρ = −.20) as well as mothers' reports of children's affective problems on the CBCL (ρ = −.16). Mothers' education level was not significantly correlated with self-reports or mother reports of children's affective problems, PTSS, or cancer-related stress.
Multiple Regression Analyses. In the model predicting mothers' depressive symptoms (Table IV), family income was a significant predictor in Step 1a. Mothers' general perceived stress and cancer-related stress were significant in Step 1b. In Step 2, each type of stress accounted for significant variance in mothers' depressive symptom levels, whereas no other variables were significant predictors. In the model predicting mothers' PTSS (Table IV) single parenthood was a significant predictor in Step 1a and both types of stress were significant in Step 1b. In Step 2, mothers' general stress and cancer-related stress were the only significant predictors.
No sociodemographic variables emerged as significant predictors in any step of the linear regression for youths' self-reported affective problems (Table IV); youths' self-reported cancer-related stress significantly accounted for variance on this measure in Step 1b but was no longer significant in the final step. In the model predicting children's self-reported PTSS (Table IV), mothers' single-parent status significantly predicted greater levels of children's PTSS in Step 1a, and in Step 1b, youths' self-reported cancer-related stress was a significant predictor. In Step 2, mothers' single-parent status and youth's cancer-related stress each significantly accounted for unique variance in youths' PTSS. Finally, in a model predicting mother reports of children's affective problems on the CBCL (Table IV), sociodemographic factors were not significant. Mothers' reports of youths' cancer-related stress was significant in Step 1b, while both mother reports of youth's cancer-related stress and mothers' self-reported depressive symptoms significantly predicted CBCL scores in the final step.
In Pearson bivariate correlations, CR was significantly positively correlated (p < .01) with mothers' self-reported depressive symptoms, PTSS, general stress, and cancer-related stress (r = .13–.23). CR was also significantly correlated with children's self-reported affective problems (r = .16, p = .05), PTSS (r = .29, p < .01), and cancer-related stress (r = .28, p < .01) but not with mothers' reports of their child's affective problems or cancer-related stress.
Multiple regression analyses were used to predict mothers' and children's emotional distress from CR (Step 1a), stress (Step 1b), as well as the relative contributions of these variables while controlling for family members' symptoms (Step 2; see Table IV). In the model predicting mothers' depressive symptoms, CR was a significant predictor in Step 1a but did not remain significant in Step 2 when general stress and cancer-related stress were added to the model. The pattern for the model predicting mothers' PTSS was similar. CR was again a significant predictor when entered on its own, but only general stress and cancer-related stress were significant in Step 2 (Table IV). Consistent with the regression models above, all relevant stress variables were also significant when entered on their own (Step 1b).
In the model for children's self-reported affective problems (Table IV), CR did not account for significant variance at any step; only mothers' depressive symptoms was significant in the final step (Step 2) of the model. CR was also not significant in the model for mother reports of children's affective problems (Table IV), although mother reports of children's cancer-related stress was a significant predictor along with mother's depressive symptoms in Step 2 of this model. In the model predicting children's self-reported PTSS (Table IV), CR was a significant predictor in Step 1a. In the second step of this model, children's self-reported cancer-related stress and mothers' PTSS each accounted for significant variance; CR was not significant when stress and mothers' symptoms were added. Again, in all models children's cancer-related stress was also significant when entered on its own (Step 1b).
Exploratory analyses examined whether CR would moderate the association between stress and psychological distress. Following the same multiple regression method described above, a regression model was run for each psychological distress outcome that used the centered CR and stress variables and included the interaction terms in Step 2. In each of these models, neither the CR × general stress nor the CR × cancer-related stress interaction terms were significant predictors of mothers' self-reported depressive symptoms or PTSS, children's self-reported affective problems or PTSS, or mothers' reports of their child's affective problems.
Results
Preliminary Analyses
On the sociodemographic CR measure, 32.4% of mothers had zero risk factors; 24.5% had one risk factor, 28.1% had two risk factors, 12.7% had three risk factors, and 2.3% had four risk factors.
Time since diagnosis was not significantly related to any psychological distress or stress variable or CR. Children's new diagnosis versus relapse was also not significantly different on any of these measures. Child age was significantly correlated with mother reports of children's cancer-related stress on the RSQ (r = .24, p < .01) but was not associated with any other measures of mothers' or children's stress, distress, or CR.
Mothers' mean score on the BDI-II (M = 15.2, SD = 10.5) was in the mild range of depressive symptom levels (≥14), with 29% of mothers reporting symptoms in the moderate to severe range (scores ≥20; Beck et al., 1996). Mothers' reports of children's affective symptoms on the CBCL indicated 15.9% of scores fell in clinical range (T ≥ 70); 7.4% of the youths who completed the YSR indicated clinically elevated symptoms (T ≥ 70). On the IES-R, 39.6% of mothers (M = 29.3, SD = 17.6) and 24.8% of youth (M = 23.3, SD = 15.3) scored above the cutoff score of 34, indicating elevated levels of PTSS (Rash, Coffey, Baschnagel, Drobes, & Saladin, 2008). Correlations among measures of stress and distress for mothers and children are presented in Table II.
Independent Associations of Each Sociodemographic Factor With Distress and Stress
Bivariate Analyses. Bivariate analyses examining associations of sociodemographic variables to mothers' and children's stress and distress are presented in Table III. Single mothers reported significantly greater depressive symptoms and PTSS than partnered mothers. Single mothers also reported significantly greater levels of cancer-related stress on the RSQ than partnered mothers; however, single versus partnered mothers did not differ on general perceived stress on the PSS. Mothers' race was not significantly associated with self-reported levels of stress or distress. Annual family income was significantly negatively correlated with all four measures of mothers' psychological distress and stress (p < .05), ranging in magnitude from ρ = .12 to ρ = .21. Mothers' education level was also significantly negatively correlated (p < .01) with self-reported depressive symptoms, PTSS, general perceived stress, and cancer-related stress (ranging from r = −.15 to −.24).
Children of single mothers reported significantly greater affective problems on the YSR and PTSS on the IES-R in comparison with children of partnered mothers. In contrast, mothers' reports of children's affective problems on the CBCL did not differ significantly as a function of mothers' relationship status. There were also no significant differences for mother- or self-reported cancer-related stress among these children. Non-White race was significantly positively associated with children's self-reported PTSS and cancer-related stress. Family income was significantly negatively correlated (p < .05) with children's self-reported affective problems on the YSR (ρ = −.20) as well as mothers' reports of children's affective problems on the CBCL (ρ = −.16). Mothers' education level was not significantly correlated with self-reports or mother reports of children's affective problems, PTSS, or cancer-related stress.
Multiple Regression Analyses. In the model predicting mothers' depressive symptoms (Table IV), family income was a significant predictor in Step 1a. Mothers' general perceived stress and cancer-related stress were significant in Step 1b. In Step 2, each type of stress accounted for significant variance in mothers' depressive symptom levels, whereas no other variables were significant predictors. In the model predicting mothers' PTSS (Table IV) single parenthood was a significant predictor in Step 1a and both types of stress were significant in Step 1b. In Step 2, mothers' general stress and cancer-related stress were the only significant predictors.
No sociodemographic variables emerged as significant predictors in any step of the linear regression for youths' self-reported affective problems (Table IV); youths' self-reported cancer-related stress significantly accounted for variance on this measure in Step 1b but was no longer significant in the final step. In the model predicting children's self-reported PTSS (Table IV), mothers' single-parent status significantly predicted greater levels of children's PTSS in Step 1a, and in Step 1b, youths' self-reported cancer-related stress was a significant predictor. In Step 2, mothers' single-parent status and youth's cancer-related stress each significantly accounted for unique variance in youths' PTSS. Finally, in a model predicting mother reports of children's affective problems on the CBCL (Table IV), sociodemographic factors were not significant. Mothers' reports of youths' cancer-related stress was significant in Step 1b, while both mother reports of youth's cancer-related stress and mothers' self-reported depressive symptoms significantly predicted CBCL scores in the final step.
Associations of Sociodemographic CR With Distress and Stress
In Pearson bivariate correlations, CR was significantly positively correlated (p < .01) with mothers' self-reported depressive symptoms, PTSS, general stress, and cancer-related stress (r = .13–.23). CR was also significantly correlated with children's self-reported affective problems (r = .16, p = .05), PTSS (r = .29, p < .01), and cancer-related stress (r = .28, p < .01) but not with mothers' reports of their child's affective problems or cancer-related stress.
Multiple regression analyses were used to predict mothers' and children's emotional distress from CR (Step 1a), stress (Step 1b), as well as the relative contributions of these variables while controlling for family members' symptoms (Step 2; see Table IV). In the model predicting mothers' depressive symptoms, CR was a significant predictor in Step 1a but did not remain significant in Step 2 when general stress and cancer-related stress were added to the model. The pattern for the model predicting mothers' PTSS was similar. CR was again a significant predictor when entered on its own, but only general stress and cancer-related stress were significant in Step 2 (Table IV). Consistent with the regression models above, all relevant stress variables were also significant when entered on their own (Step 1b).
In the model for children's self-reported affective problems (Table IV), CR did not account for significant variance at any step; only mothers' depressive symptoms was significant in the final step (Step 2) of the model. CR was also not significant in the model for mother reports of children's affective problems (Table IV), although mother reports of children's cancer-related stress was a significant predictor along with mother's depressive symptoms in Step 2 of this model. In the model predicting children's self-reported PTSS (Table IV), CR was a significant predictor in Step 1a. In the second step of this model, children's self-reported cancer-related stress and mothers' PTSS each accounted for significant variance; CR was not significant when stress and mothers' symptoms were added. Again, in all models children's cancer-related stress was also significant when entered on its own (Step 1b).
Exploratory Analyses
Exploratory analyses examined whether CR would moderate the association between stress and psychological distress. Following the same multiple regression method described above, a regression model was run for each psychological distress outcome that used the centered CR and stress variables and included the interaction terms in Step 2. In each of these models, neither the CR × general stress nor the CR × cancer-related stress interaction terms were significant predictors of mothers' self-reported depressive symptoms or PTSS, children's self-reported affective problems or PTSS, or mothers' reports of their child's affective problems.
Source...