Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 May 1.
Published in final edited form as: Health Psychol. 2021 May;40(5):295–304. doi: 10.1037/hea0001070

Stress and Psychological Adjustment in Caregivers of Children with Cancer

Kyrill Gurtovenko 1,6, Kaitlyn M Fladeboe 1,4,5, Liana R Galtieri 1, Kevin King 1, Debra Friedman 2, Bruce Compas 3, David Breiger 4, Liliana Lengua 1, Madelaine Keim 1, Joy Kawamura 6, Lynn Fainsilber Katz 1
PMCID: PMC9053835  NIHMSID: NIHMS1727618  PMID: 34152783

Abstract

Objective:

To examine effects of stress on caregiver psychological adjustment during the first year of pediatric cancer.

Methods:

Caregivers (N=159) of children with cancer completed monthly questionnaires assessing domains of caregiver psychological adjustment (depression, anxiety, and posttraumatic stress symptoms) and stress (general life stress, treatment related stress, caregiver perceptions of treatment intensity and life threat). Effects of stress were assessed at two levels to examine whether within-person changes in stress predicted concurrent changes in caregiver adjustment, and whether average stress was associated with between-person differences in caregiver adjustment trajectories.

Results:

Overall, higher levels of stress factors were associated with poorer caregiver adjustment at both the between-and within-person levels, with high average levels of treatment related stress and general life stress emerging as leading predictors of worse adjustment.

Conclusions:

Both types of stressors, those directly related as well as unrelated to a child’s cancer, contribute uniquely to caregiver distress. Caregiver distress is impacted by both overall levels of stress over time as well as month to month changes in stress. Implications for informing care for at-risk caregivers are discussed.

Keywords: pediatric cancer, stress, psychological adjustment, parents, caregivers

Pediatric cancer is a significant life stressor for both patients and their families and has potential to significantly impact caregiver psychological adjustment. Caregivers of children with cancer are at increased risk for experiencing a range of difficulties, including heightened levels of depression, anxiety, and posttraumatic stress symptoms (PTSS) (Compas et al., 2015; Kazak et al., 2005; Patenaude & Kupst, 2005; Sharp et al., 2020). Extant studies have identified individual, family-level and illness-related predictors associated with caregiver adjustment (Sultan et al., 2015; Vrijmoet-Wiersma, 2008). However, fewer studies have examined stress as a predictor of maladjustment, including effects of ongoing treatment stress, how various forms of stress affect wellbeing, or whether certain sources of stress are associated with specific symptomatology. Thus, the current study aims to address this gap by longitudinally examining how caregiver psychological adjustment is associated with different domains of stress during the first year of treatment. Addressing these gaps is needed to inform prevention and intervention aimed at mitigating caregiver distress and potential negative downstream effects on the family (Van Schoors et al., 2017).

Caregiver anxiety and depression symptoms are typically elevated around the time of diagnosis (Compas et al., 2015; Katz et al., 2018; Pai et al., 2007), and approximately 68% of mothers and 57% of fathers report PTSS in the moderate to severe range while their children receive active cancer treatment (Kazak et al., 2005). However, there is variability in caregiver adjustment (Sultan et al., 2015), which suggests a need to better understand factors that may predict which caregivers show resilient outcomes and which experience prolonged distress. Literature reviews have identified several predictors of parental psychological adjustment during pediatric cancer treatment (Pai et al., 2007; Sultan et al., 2015; Vrijmoet-Wiersma, 2008). Individual level risk factors associated with poorer psychological adjustment include being a female caregiver, having a lower education level, and lower socio-economic status (SES) (Sultan et al., 2015; Vrijmoet-Wiersma et al., 2008). Characteristics of the child’s illness such as brain tumor diagnoses and intensity of treatment are also associated with poor adjustment (Sultan et al., 2015), as is perceived financial strain, marital strain, and lack of social support (Vrijmoet-Wiersma et al., 2008).

Comparatively, much less is known about the specific influence of stress on caregivers in the context of pediatric cancer, despite its significant role in the development of many psychological problems (Monroe, 2008). Fletcher, Miaskowski, Given, & Schumacher’s (2012) model of stress in the context of cancer caregiving highlights how the stress process can impact the wellbeing of a caregiver, and can be applied in the context of pediatric cancer. First, this model highlights the influence of primary stressors, which in a pediatric context include repeated exposure to children’s illness and treatment related stressors (Hildenbrand et al., 2011; McCaffrey, 2006). Another dimension of this model is secondary stressors, or stressors that arise in aspects of the caregiver’s life outside of the treatment or illness. Secondary stressors represent general life stressors that occur independently of the child’s illness (e.g., losing a job, death in the family) which may be exacerbated by treatment and caregiving. Indeed, total lifetime stressful events, general life stress, and cancer-related stress have been associated with maladjustment in both childhood cancer survivors and their mothers (Barakat et al., 2006; Bemis et al., 2015). Fletcher’s model (2012) also highlights the impact of caregiver appraisals. Individual differences in cognitive appraisals, or the evaluative process through which an individual determines the meaning or significance of a stressor such as a child’s illness (Lazarus & Folkman, 1984), may be another critical factor in predicting which caregivers are most stressed and at risk for psychological maladjustment. For example, perceptions of uncertainty about the safety of their child, appraisal of the strain of the illness, and ability to deal with such strain have each been associated with parental wellbeing (Klassen et al., 2007).

In addition to examining multiple sources of stress, Fletcher et al. (2012) emphasize that a longitudinal perspective is necessary to understand relations between stress and parental psychopathology, as both stressors themselves and caregiver adjustment tend to fluctuate over the cancer trajectory (Houtzager et al., 2004; Katz et al., 2018; Pai et al., 2007). The first year of treatment is an especially important period for examining such relations given that caregiver distress peaks around diagnosis and caregivers may require the most support in the subsequent months (Hoekstra-Weebers et al., 2001; Pai et al., 2007). Moreover, the first year is a critical phase in the child’s treatment and supporting caregivers at this time can not only mitigate their own distress, but also prevent negative downstream effects for the child with cancer and the rest of the family (Katz et al., 2018).

Thus, the goal of the current study was to longitudinally examine relations between primary stressors (treatment related stress), secondary stressors (general life stress), cognitive appraisal factors (caregiver perceptions of treatment intensity and life threat), and caregivers’ psychological adjustment through the first year of their child’s cancer treatment. To do so, we conceptualized relations between stress and adjustment in two ways. First, given that stress levels often vary across the course of treatment, we examined how month-to-month deviations in stress were associated with concurrent changes in adjustment. We predicted that caregivers’ psychological distress would be higher during months when they experienced higher stress than what is typical for them. Second, as differences in the amount of stress experienced between caregivers may explain differences in adjustment, we examined how cumulative stress over the course of the first year of treatment is associated with one’s adjustment trajectory. We predicted that higher average levels of stress compared to other caregivers would be related to poorer psychological adjustment over time. Finally, to understand which stressors contributed most to adjustment, we assessed the relative strength and unique contributions of treatment related stress, perceptions of life threat and treatment intensity, and general life stress in predicting trajectories of caregiver psychological adjustment. Informed by Fletcher et al.’s (2012) stress process model, we predicted that each source of stress would uniquely explain variability in caregiver psychological adjustment.

Method

Participants

One hundred and fifty-nine families participated in the present study. Children were aged 2–17 (M = 6.3 years, SD = 3.5 years, 49% male) and had been recently diagnosed with cancer. Although the majority of the sample was comprised of children aged 2–10 years old, 19 adolescents aged 11–17 were also included. The majority of children with cancer were identified by caregivers as White/Caucasian (84.1%) with the remaining identified as Black/African-American (5.6%), Asian (.8%), American Indian (.8%), or other (8.8%). 15.1% of participants identified as Hispanic. Children’s cancer diagnoses consisted of leukemia (32.8%), CNS tumors (18.5%), lymphoma (12.6%), sarcoma (10.9%), Wilm’s tumor (11.8%), neuroblastoma (5.0%), or another form of cancer (8.4%).

The primary caregiver in most families was the mother (85.5%), with others being a father (12.0%), grandmother (1.7%), and stepmother (0.9%). Primary caregivers were on average 36.41 years old (SD = 7.9). Among caregivers, 78.3% were married, 7.6% were romantically involved but not married, and 14% were not romantically involved.

Procedures

Participants were recruited as part of a larger study from two urban children’s hospitals in the Northwest and Southeast United States. Families were approached within 2 weeks following the child’s diagnosis. A family was considered eligible if they had a child newly diagnosed with a tumor or cancer aged 2–17 years, English-speaking, the child had no history of developmental delay, and the caregivers at the time of enrollment were the same as before the child’s diagnosis. Children with neurofibromatosis Type I, relapsed cancer, or secondary malignancies were not eligible.

Families were recruited through brochures provided by their physicians or nurses, after which they could consent to be approached in person or by phone by a member of the study team. Brochures stated that the study was examining family functioning in the year following a child’s cancer diagnosis, and provided information regarding confidentiality and payment. Of the 502 families eligible for participation across both sites, 309 (61.6%) were approached, 176 (35.1%) enrolled, with 159 (31.7%) completing at least one study component. Of the families approached who did not enroll, refusal was due to either excessive time required or no reason was given. Institutional review board approval was received from the University of Washington, Seattle Children’s Hospital, and Vanderbilt University for all study procedures.

Data were collected over a twelve-month period beginning with an initial questionnaire packet distributed at the time of consent, followed by monthly questionnaire packets distributed through the mail. Packets were completed by primary caregivers yielding 12 points of measurement. The first packet was received 1.6 (SD = .87) months postdiagnosis on average. Primary caregivers completed 6.8 packets on average (SD = 3.84). After the initial packet, the highest proportion of primary caregivers were retained at Month 6 (67.5%), and the lowest at Month 2 (5.0%). There was marked attrition during Time 2 for several reasons. The time just after the child’s diagnosis is a highly stressful and chaotic period for families. Although efforts were made to begin monthly assessments as close in time to the diagnosis as possible, many participants returned their Time 1 packets very close in time to their Time 2 assessment point. Thus, in an effort to minimize assessment burden, Time 2 packets were skipped by many families who went on to complete their first follow up assessment at Time 3. Number of completed packets was not associated with any demographic variables, and no relation was found between initial symptom scores and amount of missing data over the study period across all adjustment measures.

Measures

Caregiver Psychological Adjustment

Depression.

Depression symptoms were measured using the short form of the Center for Epidemiological Studies-Depression Scale (CESD-10; Andresen, Malmgren, Carter, & Patrick, 1994). This 10-item self-report scale asks caregivers to rate the frequency of each symptom within the past month (e.g., “I had trouble keeping my mind on what I was doing”), from 0 (less than 1 day per week) to 3 (5–7 days per week), yielding a total sum score. Higher scores represent more frequent symptoms. Cronbach’s alpha ranged from .83-.91, with an average of .87 across the 12 time points.

Anxiety.

Anxiety symptoms were measured using the anxiety subscale of the Depression, Anxiety, and Stress Scale (DASS; Lovibond & Lovibond, 1995). This 7-item subscale assesses parent-reported frequency of anxiety symptoms in the past month, with higher scores indicating more frequent symptoms. Cronbach’s alpha in the current sample ranged from .73-.88, with an average of .84 across the 12 time points.

Post-traumatic Stress.

Post-traumatic stress symptoms were measured using the Impact of Events Scale – Revised (IES-R; Weiss & Marmar, 1997). This 22-item self-report scale assesses traumatic stress symptoms. Caregivers were asked to respond to the event specific scale items in reference to their child’s diagnosis and treatment. Participants rated each item on a 5 point scale ranging from 0 (“not at all”) to 4 (“extremely”). The IES-R total score was used in the current study, with higher scores indicating higher posttraumatic stress symptoms. Reliability was high in our sample, with Cronbach’s alpha ranging from .92-.95, with an average of .94 across the 12 time points.

Caregiver Stress

Treatment Related Stress.

Treatment-related stress was measured using the Treatment-Related Events Questionnaire. This measure was developed for the current study based on qualitative work examining major stressors among parents and children with cancer (McCaffrey, 2006). This scale included 16 items assessing treatment stressors (e.g., “going to the hospital or clinic”, “relapses”, “long hospital stays”) and 8 items assessing treatment procedures (e.g., “chemotherapy”, “blood transfusions”, “lumbar punctures”). Caregivers were asked to rate how frequently each event occurred in the past month. Each item was rated on a scale ranging from 1 = never to 5 = very often. A total treatment related events score was computed for each of the twelve time points. Cronbach’s alpha ranged from .77-.94, with an average of .91 across the 12 time points.

Life Threat and Treatment Intensity.

Perceptions of the child’s life threat and treatment intensity were assessed using the Assessment of Life Threat and Treatment Intensity Questionnaire (ALTTIQ; Stuber et al., 1997). This 4-item scale includes two questions assessing perceived life threat and two questions assessing perceived treatment intensity. Higher scores reflect more perceived life threat or treatment intensity. For life threat, Spearman-Brown reliability coefficients ranged from .66-.84, with an average of .77 across the 12 time points. For Treatment Intensity, coefficients ranged from .52-.91, with an average of .78.

General Life Stress.

General life stress was assessed via primary caregiver report using an adapted version of the Negative Life Events Scale for Children (Sandler, Ramirez & Reynolds, 1986). Eighteen items assessed the occurrence of negative life events (e.g., “you or your partner lost a job”, “you divorced or separated”) in the past month. Items were rated for whether or not it occurred (0 = did not happen, 1 = happened), yielding a measure of summed frequency of stressful life events at each of the twelve time points.

Data Analytic Strategy

Because each caregiver was assessed repeatedly over time, we approached the data as repeated observations nested within persons. We estimated growth curve models with a Multilevel Modeling (MLM) approach using the Maximum Likelihood estimator (ML) in SPSS 18.0 to test hypotheses. Although some families did not have data at each time point, MLM handles missing data by allowing trajectories to be estimated from different numbers of observations per family. Thus, caregivers who were missing data at any time point were still included in the models as long as they had follow up data.

Before testing associations between stress and adjustment outcomes, linear growth models without predictors were estimated to model average trajectories and variability in caregiver psychological adjustment outcomes. Specifically, these models were used to determine whether sufficient variance existed between caregivers in their levels of adjustment or rates of change to test potential predictors (i.e., stress variables) that may account for differences in adjustment over time. The intercept for all growth curve models in the current study were set at the time of diagnosis (time 0).

To test whether stress variables accounted for differences in caregiver psychological adjustment trajectories, we followed the recommendations of Enders & Tofighi (2007) to center predictors at within- and between-person levels (Level 1 and Level 2, respectively). To obtain within-person effects, each observation was subtracted from a given caregiver’s mean level across all observations of that variable over time. This score reflects an individual’s deviation from their own mean level of stress at each time point, and its effect on the outcome explains why an individual might differ from their expected trajectory of adjustment concurrently. To obtain between-person effects, a mean was calculated for each individual for that predictor (i.e., their average across all time points) which was then subtracted from the average of all caregivers’ means, or the grand mean. This score reflects how much each caregiver’s average stress level over time deviated from the sample average, and its effects reflect how this deviation was associated with psychological adjustment at the intercept. We also tested two interactions:

  1. an interaction between the between-person score and time which reflects how differences in average stress levels were related to rates of change over time in psychological adjustment, and

  2. a cross-level interaction, or the interaction between the within- and between-person effects. This score is used to test whether the within-person effect differed depending on an individual’s average level of stress over time.

To assess the independent effect of each type of stress, the within-person score, between-person score, and interaction terms for that stressor (e.g., general life stress) were included together in one model for each outcome (PTSS, depression, and anxiety). To assess the unique contribution of each domain of stress relative to others, all scores for each stress variable were included in a single model for each outcome. These final cumulative models tested the strength and statistical significance of the stress predictors when all other stress variables were also accounted for. For this analytic step variables were standardized for the purpose of directly comparing the relative strengths of each predictor’s association with adjustment outcomes.

Results

Summary of Caregiver Adjustment Trajectories Over Time

Before testing relations between stress predictors and trajectories of caregiver adjustment, the average trajectory of and variability in each adjustment outcome was modeled. Results of these initial growth models are described in earlier work from the larger study from which this data was derived (Katz et al., 2018). For details regarding the general course and variability in caregiver psychological adjustment outcomes in this sample, we refer interested readers to this analysis. For the purposes of the current analysis, sufficient variance was observed in both the intercept and slope (rate of change) to test predictors of these differences.

Analyses were tested with the inclusion of covariates of child and caregiver age, child and caregiver gender, as well as the number of children in the home. We also conducted all analyses with quadratic effects of time to model non-linear rates of change in adjustment outcomes. No substantial differences in the overall patterns of results were found, thus the decision was made to trim covariates and quadratic effects of time from the final models to improve parsimony and preserve power for addressing the substantive research questions.

Associations between stress predictors were examined using Pearson product-moment correlation coefficients. Bivariate correlations were non-significant between general life stress and perceived life threat (r = −.00, p = .99). Small correlations were observed between general life stress and perceived treatment intensity (r = .18, p = .04), with medium correlations found for general life stress and treatment related stress (r = .34, p = .00) and perceived life threat and treatment related stress (r = .33, p = .00). Large correlations were observed between perceived treatment intensity and treatment related stress (r = .46, p = .00) and perceived treatment intensity and perceived life threat (r = .50, p = .00).

Stress Predicting Caregiver Adjustment

Parameter estimates and model fit information for single-predictor models are summarized in Table 1. Single-predictor models were interpreted individually, and then used to inform a final cumulative model for each outcome that included both within- and between-person effects from all four stress predictors to examine the relative contributions of each on caregiver adjustment outcomes (see Table 2).

Table 1.

Single Predictor Models of Outcomes Predicted by Treatment Intensity, Life Threat, Treatment, & Life Stress

Variables Intercept B(SE) Time B(SE) Within Effect B(SE) Between Effect B(SE) BetweenXTime Effect B(SE) BetweenX Within Effect B(SE)
Depression
 Perceived Life Threat 12.89*** (.37) −.37*** (.05) .22 (.14) .70*** (.18) .02 (.03) −.07 (.08)
 Treatment Intensity 13.08*** (.47) −.37*** (.05) .21* (.09) .54* (.26) −.01 (.03) −.04 (.06)
 Treatment Related Stress 12.29*** (.48) −.21*** (.05) .05*** (.01) .14*** (.03) .00 (.00) −.00 (.00)
 General Life Stress 12.73*** (.45) −.33*** (.05) .40*** (.12) 1.38** (.36) .08 (.05) −.14 (.09)
PTSS
 Perceived Life Threat 28.59*** (.94) −.99*** (.15) .60 (.36) 1.30** (.46) .10 (.07) −.06 (.21)
 Treatment Intensity 29.50*** (1.31) −1.04*** (.11) .31 (.22) 1.08 (.72) −.03 (.07) .10 (.14)
 Treatment Related Stress 27.66*** (1.30) −.75*** (.13) .09*** (.02) .41*** (.08) −.01 (.01) −.00 (.00)
 General Life Stress 28.81*** (1.30) −.97*** (.11) .58* (.27) 2.91** (1.03) .14 (.10) −.11 (.21)
Anxiety
 Perceived Life Threat 3.31*** (.22) −.12*** (.03) .06 (.08) .23* (.11) .01 (.02) .01 (.05)
 Treatment Intensity 3.42*** (.29) −1.05*** (.03) .04 (.06) .17 (.16) −.00 (.02) −.05 (.04)
 Treatment Related Stress 3.21*** (.29) −.07*** (.03) .02** (.01) .08*** (.02) .00 (.00) .00 (.00)
 General Life Stress 3.24*** (.28) −.09** (.03) .20** (.07) .91*** (.22) .01 (.03) −.01 (.05)
Open in a new tab

Note.

*

= p < .05;

**

= p < .01;

***

= p < .001.

B represents unstandardized effect estimates.

Table 2.

Final Cumulative Growth Models of Caregiver Adjustment Predicted by Life Threat, Treatment Intensity, Treatment Related Stress, & General Life Stress

Model Intercept β(SE) Time β(SE) Life Threat Within Effect β(SE) Life Threat Between Effect β(SE) Treatment Intensity Within Effect β(SE) Treatment Intensity Between Effect β(SE) Treatment Stress Within Effect β(SE) Treatment Stress Between Effect β(SE) Life Stress Within Effect β(SE) Life Stress Between Effect β(SE)
Final Depression Model .23*** (.07) −.03*** (.01) .03 (.02) .23** (.07) .02 (.02) −.14 (.08) .11*** (.02) .27*** (.07) .06** (.02) .27*** (.07)
Final PTSS Model .30*** (.08) −.04*** (.01) .05** (.02) .22** (.08) −.01 (.02) −.18* (.08) .08*** (.02) .28** (.08) .03 (.02) .22** (.07)
Final Anxiety Model .15 (.08) −.01 (.01) .01 (.02) .18* (.08) .01 (.02) −.20* (.08) .06* (.02) .32*** (.08) .07** (.02) .26*** (.07)
Open in a new tab

Note.

*

= p < .05;

**

= p < .01;

***

= p < .001.

β represents standardized effect estimates.

Stress Predicting Depression

At the between-persons level, perceived life threat, perceived treatment intensity, treatment related stress, and general life stress were each associated with levels of depression during this period. Specifically, caregivers who reported a higher average amount in these stressors over time relative to other caregivers tended to have higher levels of depression through the first year of treatment. At the within-person level, perceived treatment intensity, treatment related stress, and general life stress were also associated with depression. Thus, in a given month when caregivers reported an increase in any of these stressors compared to their own average level, they tended to report a concurrent increase in depressive symptoms. No significant interaction effects were found, suggesting that across stressors, between-person differences in average stress levels did not affect the rate of change of depression symptoms over time, and that the strength of the effects representing monthly deviations in stress on depression did not depend on how much average stress one experienced over time.

Stress Predicting PTSS

At the between-persons level, perceived life threat, treatment related stress, and general life stress was associated with levels of PTSS during this period. Caregivers reporting higher average levels of these stressors over time reported higher PTSS through the first year of treatment. At the within-person level, treatment-related stress and general life stress were associated with PTSS. Thus, in a given month when caregivers reported an increase in treatment-related stress and general life stress compared to their own average level, they tended to report a concurrent increase in PTSS. No significant interaction effects were found.

Stress Predicting Anxiety

At the between-persons level, perceived life threat, treatment related stress, and general life stress was associated with levels of anxiety during this period. Caregivers reporting higher average levels of these stressors over time reported higher anxiety through the first year of treatment. At the within-person level, treatment related stress and general life stress were associated with anxiety. Specifically, in a given month when caregivers reported an increase in treatment-related stress and general life stress compared to their own average level, they tended to report a concurrent increase in anxiety symptoms. Similar to results for PTSS and depression, no significant interaction effects were found.

Relative and Cumulative Contributions of Stress on Caregiver Adjustment

Standardized estimates for models assessing the relative and cumulative contributions of each stress variable on each individual caregiver adjustment outcome are shown in Table 2. Given that the interaction terms in the earlier models were non-significant and did not improve model fit, they were trimmed from these final models for parsimony.

Final Depression Model

The final cumulative model for depression showed significant between-person effects for perceived life threat, treatment related stress, and general life stress. In addition, within-person effects estimates remained statistically significant for treatment related stress and general life stress. These final cumulative model effects estimates were comparable to the previous models, suggesting that the effects for each predictor were stable and relatively independent of one another. Results indicated that each of the stress categories (except perceived treatment intensity) predicted unique variance in caregiver depression over and above others. Moreover, variability in stress at the between-person as well as within-person level made unique contributions to predictions of caregiver depression. In other words, both cumulative amounts of stress over time and monthly fluctuations in stress appear to contribute to caregiver depression. A comparison of standardized coefficients indicated that the between-person effects for general life stress (β = .27) and treatment related stress (β = .27) were the strongest predictors of depression. For example, this suggests that relative to those at the mean, caregivers who were 1 SD above the mean on general life stress over time reported .27 SD higher depression through the first year of treatment. The next strongest predictor of depression was the between-person effect for perceived life threat (β = .23), followed by within-person effects of treatment related stress (β = .11) and life stress (β = .06). Figure 1a & 1b illustrates between and within person effects of general life stress on depression, controlling for effects of other stressors. Figure 1b shows a model based estimate of how caregiver depression is predicted to deviate from the average trajectory as a function of fluctuations in general life stress. For example, if a caregiver was to experience a 1 SD increase in general life stress at month 3, Figure 1b shows how this individual’s depression would be predicted to concurrently increase. Although the effect is modeled at the 3 month timepoint in Figure 1b, such fluctuations could occur in any given month during this time period.

Figure 1. Between and Within Person Effects of General Life Stress with Caregiver Adjustment Outcomes.

Figure 1

Open in a new tab

Note. Figures represent general life stress effects estimates derived from final cumulative models controlling for the effects of other stressors (see Table 2); Figures 1b, 1d, and 1f show a model based estimate of how adjustment outcomes are predicted to deviate from their average trajectory as a function of fluctuations in general life stress. Although these effects are modeled at time 3 above, such fluctuations could occur at any month during the first year post-diagnosis; Within person effect for general life stress & PTSS above is not statistically significant.

Final PTSS Model

The final cumulative model for PTSS showed significant between-person effects for all four stress predictors. All effects were in the expected direction, except the between-person effect of perceived treatment intensity. This suggested that higher average treatment related stress, general life stress, and perceived life threat relative to other caregivers were each uniquely associated with higher PTSS. However, higher average perceived treatment intensity relative to other caregivers was associated with lower reported PTSS over this period. Within-person effects estimates remained statistically significant for perceived life threat and treatment related stress. A comparison of standardized coefficients indicated that the between-person effect for treatment related stress (β = .28) was the strongest predictor of PTSS. This suggests that relative to those at the mean, caregivers who were 1 SD above the mean on treatment related stress over time reported .28 SD higher PTSS through the first year of treatment. The next strongest predictors of PTSS were between-person effects of life threat (β = .22), general life stress (β = .22), and perceived treatment intensity (β = −.18), followed by within person effects of treatment related stress (β = .08) and perceived life threat (β = .05). Figure 1e & 1f illustrates between and within person effects of general life stress on PTSS, controlling for effects of other stressors.

Final Anxiety Model

The final cumulative model for anxiety showed significant between-person effects for all four stress predictors. Similar to the final PTSS model, all effects were in the expected direction except perceived treatment intensity, suggesting higher average treatment related stress, general life stress, and perceived life threat relative to other caregivers were each uniquely associated with higher anxiety, though higher average perceived treatment intensity relative to other caregivers was associated with fewer anxiety symptoms over this period. Within-person effects estimates remained statistically significant for treatment related stress and general life stress. A comparison of standardized coefficients suggested that the between-person effect for treatment related stress (β = .32) was the strongest predictor of anxiety symptoms. Thus, relative to those at the mean, caregivers who were 1 SD above the mean on treatment related stress over time reported .32 SD higher anxiety through the first year of treatment. The next strongest predictors of anxiety were between-person effects of general life stress (β = .26), treatment intensity (β = .20), and life threat (β = .18), followed by within person effects of general life stress (β = .07) and treatment related stress (β = .06). Figure 1c & 1d illustrates between and within person effects of general life stress on anxiety, controlling for effects of other stressors.

Discussion

In line with Fletcher et al.’s (2012) conceptual model of the cancer caregiving stress process, the current study adopted a longitudinal trajectory approach to investigating how primary stressors (treatment related stress), secondary stressors (general life stress), and appraisal factors (perceived life threat and treatment intensity) are associated with depression, anxiety, and PTSS in caregivers over the first year of pediatric cancer treatment. To capture the nuanced relationship between stress and adjustment over time, we examined the effect of these four stress variables on caregiver adjustment in two ways: 1) the degree to which fluctuations in a caregiver’s stress levels from their own typical level was associated with their concurrent adjustment, and 2) how individual differences in stress between caregivers (i.e., average amount of total stress experienced during this period, relative to other caregivers) was associated with their adjustment.

In support of our hypotheses, results showed that these four domains of stress were associated with both within and between person variability in the trajectory of caregiver adjustment over the first year of treatment. Our initial single-predictor models showed that caregivers who experienced more general life stress, treatment related stress, perceived life threat, and treatment intensity reported higher depressive symptoms over time. In addition, experiencing increases in general life stress, treatment related stress, or perceived treatment intensity in any given month was associated with concurrent increase in depressive symptoms. Similar patterns were observed in regard to caregiver anxiety and PTSS. Higher cumulative levels of general life stress, treatment related stress, and perceived life threat were associated with higher anxiety and PTSS, and increases in stress from general life stressors and treatment related stressors in a given month were associated with higher levels of anxiety and PTSS concurrently. Our findings suggest that caregivers’ psychological adjustment is impacted by not only how much of these stressors they experience in sum over the first year, but also by any fluctuations in these stress factors in any given month.

The current study also examined which stressors had the strongest association with caregiver psychological adjustment. Our final cumulative models which accounted for all stress variables showed that overall frequency of general life stress and treatment related stress emerged as the strongest predictors of caregiver psychological adjustment. This suggests that assessing the total frequency of treatment related events (e.g., lumbar punctures) and negative life events (e.g., parent losing a job; death of a pet) may be the best stress related indicator of caregivers’ psychological adjustment during the first year of their child’s cancer treatment. Repeatedly watching one’s child go through painful treatments with uncertainty about their outcomes is emotionally taxing. The additional stress of general life stress only adds to a high burden at a time when resources may be depleted.

Our hypotheses were partially supported by the finding that most stress domains were uniquely associated with psychological adjustment over and above other stress variables. When accounting for other stressors, perceptions of treatment intensity no longer contributed to differences in depression and showed unexpected effects for PTSS and anxiety. These unexpected findings suggested that caregivers who reported higher overall perceived treatment intensity experienced lower PTSS and anxiety symptoms. One possible explanation for these findings may be that caregivers exposed to more frequent and intense treatments may be experiencing lower anxiety and PTSS as a function of habituation or other exposure based mechanisms (Benito & Walther, 2015). That is, caregivers who have no choice but to be frequently exposed to their child’s intensive treatment learn to tolerate and manage their anxiety and PTSS better than caregivers who have intermittent or less intensive exposure. Caregivers whose children experience higher intensity treatment may also feel an increased sense of control over the medical outcome of their child’s illness. Engagement with more intensive treatments may function as a form of primary control coping, i.e. coping with stress by directly targeting the source of the stress (treating the illness) (Compas et al., 2015; Sharp et al., 2020). A third reason for these unexpected findings may be due to statistical suppression among the independent variables, particularly given the significant associations between perceived treatment intensity and other stress predictors. These are just a few possible explanations for these unexpected findings and should be interpreted with caution. It will be important to determine whether these results replicate or whether they are unique to our sample.

Cumulative stress models suggested that general life stressors showed relatively stronger associations with depression and anxiety symptoms compared to PTSS. Many of the effect sizes for the relation between stress and adjustment were small. This highlights the importance of considering other factors besides stress which may account for caregiver psychological adjustment. It is also important to consider the clinical significance of these findings. Given that the negative influence of stress cuts across a range of adjustment outcomes, we believe that efforts to mitigate this risk still has clinically significant implications despite these smaller effects. Even small decreases in caregiver depression, anxiety, and PTSS can lead to significantly improved wellbeing and functioning during this critical period in the child’s treatment.

Results from the current study replicate and extend findings from prior studies showing that stress negatively affects psychological adjustment among caregivers of children with cancer and other cancer populations (Golden-Kreutz et al., 2004; Varni et al., 1992). Similar to our results, Bemis et al. (2015) identified general life events and cancer related events as robust predictors of adjustment in mothers of children with cancer. Our results further support this association by demonstrating that both experiencing high frequency of such events over the course of treatment and experiencing higher levels during a given period of time than what is generally typical for a caregiver (regardless of what amount is typical) may contribute to poor adjustment. Another recent study with mothers of children with cancer found that perceived stress at diagnosis was correlated with depression at different timepoints, though it did not predict trajectory patterns of depression (Sharp et al., 2020). These authors noted that more proximal and ongoing measures of stress throughout treatment may better predict caregiver adjustment, and our current findings support this view.

Our findings advance clinical efforts for this population by informing both screening and intervention practices. As Sharp et al. (2020) have also recently suggested, strengthening mental health support for caregivers early on in the child’s treatment is needed. Prevention efforts delivered around the time of diagnosis may include brief psychoeducation about mental health and the potential sources and impacts of stress during this critical time period. Caregivers could be provided with brief stress management skills training early on, including being taught how to monitor their own stress levels and where to seek additional support for managing periods of heightened stress. Caregivers’ stress experiences could also be assessed through clinical interviews or brief validated questionnaires as a part of the child’s treatment and follow up care.

We emphasize the importance of attention to both primary and secondary stressors. Our results suggest that healthcare providers should monitor how frequently caregivers are experiencing stressful events both related and unrelated to their child’s treatment. Furthermore, providers should closely attend to not only whether a caregiver is experiencing high levels of stress for an extended period, but also if their stress levels have recently increased relative to their typical levels. For example, the caregiver of a child whose treatment requires frequent procedures and hospital stays over the course of many months may be at risk for maladjustment, but so may be the caregiver of a child who typically has a mild treatment but encounters treatment complications and has a sudden increase in these events. In addition, the robust nature of effects at both of these levels suggest that caregiver adjustment in the first year of treatment is a dynamic rather than a static outcome. A caregiver who shows little to no psychological distress in one month may show increases in depression, anxiety, or PTSS in the next month, in part as a function of increases in treatment and general life stressors. With close monitoring of caregivers’ stress levels, particularly in the domains of treatment-related stressors and general life stressors, providers may better identify which families need more support and provide more timely preventative intervention.

Interventions that target stress management techniques may provide caregivers with skills to cope with stressful events and improve adjustment. Recent work testing a novel stress-management intervention for adolescents and young adults with cancer and their caregivers has shown promising results (Rosenberg et al., 2018; Yi-Frazier et al., 2017). Findings from the current study also highlight important associations between cognitive appraisal factors and caregiver psychological adjustment. Specifically, relations between perceived life threat and caregiver PTSS suggest that stress-management interventions targeting cognitions and stress appraisal may be particularly effective. Enhancing adaptive coping and emotion regulation skills that focus on cognitive change strategies such as reappraisal may be an effective way to mitigate unhelpful appraisals of stressors (Tomaka, Blascovich, Kelsey, & Leitten, 1993; Tomaka, Blascovich, Kibler, & Erns, 1997). Taken together, it is critical for health care providers to identify families experiencing frequent stressful events and promote stress-management and coping skills to prevent or ameliorate poor psychological adjustment.

This study has several notable strengths. To our knowledge, this is the first study of caregiver adjustment to pediatric cancer that has used growth modeling to predict individual differences in caregiver adjustment over time, as well as the first to examine how intra-individual changes in specific domains of stress may influence trajectories of psychological adjustment. In addition, no studies outside of our research group have examined stress or adjustment at a monthly level, which has allowed for a more nuanced examination of changes in stress over time and effects on adjustment. These findings advance contemporary research efforts in this area by suggesting that the links between caregiver stress and psychological adjustment are dynamic and change over time, and thus should be investigated using more nuanced longitudinal research. Further identifying contextual factors that may exacerbate or account for within-person fluctuations in stress may be one important next step in this line of work. Additionally, studying whether the association between caregiver stress and adjustment fluctuates over time, or varies by time since diagnosis or treatment phase is an important next step. Such questions could help further identify specific high risk periods of when caregivers are most susceptible to stressors.

The current study also has several limitations. Some caregivers did not complete all measures at each of the twelve time-points. Attrition was especially high at time 2, in part because this may have been an especially stressful period for families. These missing data may have impacted the results. Statistical modeling was used to account for missing data, however, unexamined systematic differences may have existed between those with complete data versus those with missing data. In addition, there may have been differences between families who participated in the study and those who turned down participation. For example, families who refused to participate may have been those experiencing higher stress at the time of diagnosis. Another limitation is related to the correlation among the stress predictors, which increases the likelihood that multi-collinearity or statistical suppression may have influenced the results. Additionally, the caregiver sample in this study was comprised of mostly White educated mothers, thus the generalizability of the findings may be limited. Future research should include more diverse samples and investigate changes in fathers’ stress and adjustment. Measurement also relied entirely on self-report, and single-reporter bias may have had some influence on the results; for example, caregivers experiencing poor psychological adjustment may be more likely to attend to stress in their environment or overestimate the frequency of stressful experiences. Future research on these questions should aim to mitigate this limitation by adopting a multi-measure, multi-method approach. In addition, although our measure of treatment related stress was informed by prior literature, it was developed for the purposes of this study and has not been previously validated. Thus, future studies are needed to establish the psychometric properties of this measure, or other valid and reliable measures which can usefully assess this critical variable should be developed and studied. Finally, although the treatment related stress measure assessed the frequency of treatment related events, frequency does not equate to the severity of a medical event. Some events such as a lengthy admission to the ICU may be more impactful than more routine treatment events. Future research which considers the relative severity of different treatment events is needed to further understand the impact of this domain of stress on caregiver psychological adjustment.

In summary, when a child has cancer, stressful experiences both related and unrelated to cancer may contribute to caregiver distress. Providers should thus make efforts to continuously screen for stressors, including those unrelated to cancer treatment, and attend to both cumulative stressors over time and sudden fluctuations in stress to best identify at-risk caregivers and intervene appropriately. Since caregivers play an essential role in caring for and supporting patients during treatment, monitoring the mental health of caregivers and assessing factors that confer risk for maladjustment are essential to comprehensively supporting families and patients.

Acknowledgments

This research was supported by Grant 5R01CA134794 from the National Cancer Institute to Lynn Fainsilber Katz.

References

  1. Andresen EM, Malmgren JA, Carter WB, & Patrick DL (1994). Screening for depression in well older adults: Evaluation of a short form of the CES-D. American Journal of Preventive Medicine, 10(2), 77–84. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/8037935 [PubMed] [Google Scholar]
  2. Barakat LP, Alderfer MA, Kazak AE (2006) Posttraumatic growth in adolescent survivors of cancer and their mothers and fathers. Journal of Pediatric Psychology, 413–419. doi: 10.1093/jpepsy/jsj058 [DOI] [PubMed]
  3. Bemis H, Yarboi J, Gerhardt CA, Vannatta K, Desjardins L, Murphy LK, Compas BE (2015). Childhood cancer in context: Sociodemographic factors, stress, and psychological distress among mothers and children. Journal of pediatric psychology, 40(8), 733–743. doi: 10.1093/jpepsy/jsv024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Benito KG & Wlalther M. (2015). Therapeutic process during exposure: Habituation model. Journal of Obsessive-Compulsive and Related Disorders. 1(6), 147–157. doi: 10.1016/j.jocrd.2015.01.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Compas BE, Bemis H, Gerhardt CA, Dunn MJ, Rodriguez EM, Desjardins L, Preacher KJ, Manring S, & Vannatta K. (2015). Mothers and fathers coping with their child’s cancer: Individual and interpersonal processes. Health Psychology, 34, 783–793. doi: 10.1037/hea0000202 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Enders CK, & Tofighi D. (2007). Centering predictor variables in cross-sectional multilevel models: A new look at an old issue. Psychological Methods, 12(2), 121–138. doi: 10.1037/1082-989X.12.2.121 [DOI] [PubMed] [Google Scholar]
  7. Fletcher BS, Miaskowski C, Given B. & Schumacher K. (2012). The cancer family caregiving experience: An updated and expanded conceptual model. European Journal of Oncology Nursing, 16, 387–398. doi: 10.1016/j.ejon.2011.09.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Francis JL, Moitra E, Dyck I, & Keller MB (2012). The impact of stressful life events on relapse of generalized anxiety disorder. Depression and Anxiety, 29(5), 386–391. doi: 10.1002/da.20919 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Golden-Kreutz DM, Browne MW, Frierson GM, & Andersen BL (2004). Assessing stress in cancer patients: A second-order factor analysis model for the perceived stress scale. Assessment, 11(3), 216–223. doi: 10.1177/1073191104267398 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Hildenbrand AK, Clawson KJ, Alderfer MA, & Marsac ML (2011). Coping with pediatric cancer: Strategies employed by children and their parents to manage cancer-related stressors during treatment. Journal of Pediatric Oncology Nursing, 28(6), 344–354. DOI: 10.1177/1043454211430823 [DOI] [PubMed] [Google Scholar]
  11. Hoekstra-Weebers JEHM, Jaspers JPC, Kamps WA, & Klip EC (2001). Psychological adaptation and social support of parents of pediatric cancer: A prospective longitudinal study. Journal of Pediatric Psychology, 26(4), 225–235. doi: 10.1093/JPEPSY/26.4.225 [DOI] [PubMed] [Google Scholar]
  12. Houtzager BA, Oort FJ, Hoekstra-Weebers JEHM, Caron HN, Grootenhuis MA, & Last BF (2004). Coping and family functioning predict longitudinal psychological adaptation of siblings of childhood cancer patients. Journal of Pediatric Psychology, 29(8), 591–605. doi: 10.1093/jpepsy/jsh061 [DOI] [PubMed] [Google Scholar]
  13. Katz LF, Fladeboe K, King K, Gurtovenko K, Kawamura J, Friedman D, … Stettler N. (2018). Trajectories of child and caregiver psychological adjustment in families of children with cancer. Health Psychology, 37(8), 725–735. doi: 10.1037/hea0000619 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Kazak AE, Boeving CA, Alderfer MA, Hwang WT, & Reilly A. (2005). Posttraumatic stress symptoms during treatment in parents of children with cancer. Journal of Clinical Oncology, 23(30), 7405–7410. doi: 10.1200/JCO.2005.09.110 [DOI] [PubMed] [Google Scholar]
  15. Klassen A, Raina P, Reineking S, Dix D, Pritchard S, & O’Donnell M. (2007). Developing a literature base to understand the caregiving experience of parents of children with cancer: A systematic review of factors related to parental health and wellbeing. Supportive Care in Cancer, 15(7), 807–818. doi: 10.1007/s00520-007-0243-x [DOI] [PubMed] [Google Scholar]
  16. Lazarus RS, & Folkman S. (1984). Stress, appraisal, and coping. New York, NY: Springer. [Google Scholar]
  17. Lovibond PF, & Lovibond SH (1995). The structure of negative emotional states: Comparison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and Anxiety Inventories. Behaviour Research and Therapy, 33(3), 335–343. doi: 10.1016/0005-7967(94)00075-u [DOI] [PubMed] [Google Scholar]
  18. McCaffrey CN (2006). Major stressors and their effects on the well-being of children with cancer. Journal of Pediatric Nursing, 21(1), 59–66. doi: 10.1016/j.pedn.2005.07.003 [DOI] [PubMed] [Google Scholar]
  19. Monroe SM (2008). Modern approaches to conceptualizing and measuring human life stress. Annual Review of Clinical Psychology, 4, 33–52. doi: 10.1146/annurev.clinpsy.4.022007.141207 [DOI] [PubMed] [Google Scholar]
  20. Pai AL, Greenley RN, Lewandowski A, Drotar D, Youngstrom E, & Peterson CC (2007). A meta-analytic review of the influence of pediatric cancer on parent and family functioning. Journal of Family Psychology, 21(3), 407. doi: 10.1037/0893-3200.21.3.407 [DOI] [PubMed] [Google Scholar]
  21. Patenaude AF, & Kupst MJ (2005). Psychosocial functioning in pediatric cancer. Journal of pediatric psychology, 30(1), 9–27. doi: 10.1093/jpepsy/jsi012 [DOI] [PubMed] [Google Scholar]
  22. Rosenberg AR, Bradford MC, McCauley E, Curtis JR, Wolfe J, Baker KS, & Yi-Frazier JP (2018). Promoting resilience in adolescents and young adults with cancer: results from the PRISM randomized controlled trial. Cancer, 124(19), 3909–3917. doi: 10.1002/cncr.31666 [DOI] [PubMed] [Google Scholar]
  23. Sandler I N, Ramirez R, Reynolds K D, (1986). Life stress for children of divorce, bereaved and asthmatic children. Paper presented at the American Psychological Association Convention; Washington, DC. [Google Scholar]
  24. Sharp KMH, Fisher RS, Clark OE, Zackery DO, Dunnells ZDO, Murphy LK, Prussien KV, Vannatta K, Compas BE, & Gerhardt CA (2020). Long-term trajectories of depression symptoms in mothers of children with cancer. Health Psychology, 39, 89–98. 10.1037/hea0000826 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Stuber ML, Kazak AE, Meeske K, Barakat L, Guthrie D, Garnier H, Pynoos R, & Meadows A. (1997). Predictors of posttraumatic stress symptoms in childhood cancer survivors. Pediatrics, 6: 958–964. doi: 10.1542/peds.100.6.958 [DOI] [PubMed] [Google Scholar]
  26. Sultan S, Leclair T, Rondeau É, Burns W, & Abate C. (2016). A systematic review on factors and consequences of parental distress as related to childhood cancer. European Journal of Cancer Care, 25(4), 616–637. doi: 10.1111/ecc.12361 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Tomaka J, Blascovich J, Kelsey R M,. Leitten C L, (1993). Subjective, physiological, and behavioral effects of threat and challenge appraisal. Journal of Personality and Social Psychology, 65(2), 248–60. doi: 10.1037/0022-3514.65.2.248 [DOI] [Google Scholar]
  28. Tomaka J, Blascovich J, Kibler J, Ernst J M, (1997). Cognitive and physiological antecedents of threat and challenge appraisal. Journal of Personality and Social Psychology, 73(1), 63–72. doi: 10.1037//0022-3514.73.1.63 [DOI] [PubMed] [Google Scholar]
  29. Van Schoors M, Caes L, Knoble NK, Goubert L. Verhofstadt LL, & Alderfer MA (2017). Systematic review: Associations between family functioning and child adjustment after pediatric cancer diagnosis: A meta-analysis. Journal of Pediatric Psychology, 42(1), 6–18. doi: 10.1093/jpepsy/jsw070 [DOI] [PubMed] [Google Scholar]
  30. Varni JW, Katz ER, Colegrove R, Dogin M. (1992) Perceived stress and adjustment of longterm survivors of childhood cancer. Journal of Psychosocial Oncology, 12, 1–16. doi: 10.1300/J077V12N03_01 [DOI] [Google Scholar]
  31. Vrijmoet-Wiersma CMJ, van Klink JMM, Kolk AM, Koopman HM, Ball LM, & Egeler RM (2008). Assessment of parental psychological stress in pediatric cancer: A review. Journal of Pediatric Psychology, 33(7), 694–706. doi: 10.1093/jpepsy/jsn007 [DOI] [PubMed] [Google Scholar]
  32. Weiss DS, & Marmar CR (1997). The Impact of Event Scale-Revised. In Wilson JP & Keane TM (Eds.), Assessing Psychological Trauma and PTSD, New York: Guilford, 399–411. [Google Scholar]
  33. Yi-Frazier JP, Fladeboe K, Klein V, Eaton L, Wharton C, McCauley E, & Rosenberg AR (2017). Promoting resilience in stress management for parents (PRISM-P): An intervention for caregivers of youth with serious illness. Families, Systems & Health. 35(3), 341–351. doi: 10.1037/fsh0000281. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES