Fax: (323) 664-5512
Parent proxy–reported health-related quality of life and fatigue in pediatric patients diagnosed with brain tumors and acute lymphoblastic leukemia
Article first published online: 23 SEP 2004
Copyright © 2004 American Cancer Society
Volume 101, Issue 9, pages 2116–2125, 1 November 2004
How to Cite
Meeske, K., Katz, E. R., Palmer, S. N., Burwinkle, T. and Varni, J. W. (2004), Parent proxy–reported health-related quality of life and fatigue in pediatric patients diagnosed with brain tumors and acute lymphoblastic leukemia. Cancer, 101: 2116–2125. doi: 10.1002/cncr.20609
- Issue published online: 18 OCT 2004
- Article first published online: 23 SEP 2004
- Manuscript Accepted: 19 JUL 2004
- Manuscript Revised: 12 JUL 2004
- Manuscript Received: 31 MAY 2004
- Kellerman Foundation
- health-related quality of life;
- PedsQL™ fatigue;
- brain tumor;
- acute lymphoblastic leukemia
Pediatric patients with brain tumors (BT) are often excluded from health-related quality of life (HRQOL) studies even though they experience more severe disease and treatment-related sequelae than children with other types of cancer. Parent proxy assessments of HRQOL allow for greater inclusion of children who are developmentally immature, physically ill, or cognitively impaired.
Parents of children ages 2–18 years who were diagnosed at Childrens Hospital Los Angeles and Children's Hospital San Diego with BT (n = 86) or acute lymphoblastic leukemia (ALL; n = 170) evaluated their children's HRQOL over the previous week using the parent-proxy versions of the Pediatric Quality of Life Inventory (PedsQL™) 4.0 Generic Core scales, the PedsQL™ 3.0 Acute Cancer Module, and the PedsQL™ Multidimensional Fatigue scales. Multiple regression analyses were used to determine the independent effect of the child's diagnosis on HRQOL. Separate analyses were conducted for patients receiving treatment, patients who had not received treatment for < 12 months, and patients who had not received treatment for ≥ 12 months.
Patients with BT exhibited more problems than patients with ALL in the physical, social, psychosocial, school, cognitive, and fatigue domains of HRQOL. The Core Physical Health, Core Psychosocial Health, and Fatigue Total scores for patients with BT demonstrated peak improvements for children who had not received treatment for < 12 months and sharp declines for children who had not received treatment for ≥ 12 months. The Core Physical Health and Fatigue Total scores for patients with ALL were highest (better HRQOL) for those who had not received treatment for ≥ 12 months.
Pediatric patients and survivors of BT experienced more fatigue and HRQOL problems than patients with ALL, and HRQOL differed by treatment status. Cancer 2004. © 2004 American Cancer Society.
Approximately 12,400 children and adolescents under the age of 20 are diagnosed with cancer each year in the United States.1 Brain tumors (BT) and acute lymphoblastic leukemia (ALL) are the two most common forms of pediatric cancer.2 From 1990 to 1995, 2200 children were diagnosed with a BT and 2300 were diagnosed with ALL annually.3, 4 Five-year survival rates are now approaching 70% for patients with BT (with the proportion varying by tumor type) and are exceeding 80% for patients with ALL.3, 4 Studies demonstrate that children with a BT experience more severe disease and treatment sequelae than children with other types of cancer, including those diagnosed with ALL.5–7
The assessment of health-related quality of life (HRQOL) has emerged as a powerful method of monitoring medical populations, evaluating treatment outcomes, and identifying patients at risk for adjustment problems.8 Unfortunately, patients with BT are often excluded from quality of life (QOL) studies due to their acute morbidity and frequent presentation of cognitive impairments.9 As a result, there is limited information on the unique experiences of patients with BT and how their long-term sequelae affect HRQOL.9
The Pediatric Quality of Life Inventory (PedsQL™) is a multidimensional instrument designed to evaluate the HRQOL of children.10 The PedsQL™ 4.0 Generic Core scales measure the physical, emotional, social, and role-functioning dimensions of HRQOL, allowing for comparisons across healthy children and patient groups.10 Disease-specific modules enhance the measurement sensitivity of health domains germane to chronic health conditions.10 For example, the PedsQL™ 3.0 Acute Cancer Module focuses on dimensions of HRQOL pertinent to the lives of pediatric oncology patients, whereas the recently developed PedsQL™ Multidimensional Fatigue scales allow investigators to thoroughly evaluate fatigue symptoms.10 Because fatigue has been identified recently as a common problem for oncology patients, the development of the fatigue module is a major advance in the measurement of pediatric HRQOL.10–12
Although a patient's perception of QOL is an important element in the measurement of HRQOL, parent proxy-report is essential when a child cannot participate in the evaluation process, or when the validity of the child's responses is in question.13 Because many children with BT are diagnosed at an early age,3 and experience significant medical, neurologic, cognitive, and psychosocial impairments,5–7 it is often necessary to rely on parent proxy-report to evaluate HRQOL. In oncology populations, the PedsQL™ Measurement Model demonstrates medium to strong concordance between parent proxy-reports and child self-reports.10 These findings indicate that parent proxy-report is a valid method of estimating HRQOL when the child is unable to do so.
The current study compares HRQOL in patients diagnosed with BT with those diagnosed with ALL. We utilized the PedsQL™ parent proxy-report measures to capture HRQOL data from a diverse population that includes children too young or too impaired to provide self-report. Based on the literature, we hypothesized that children with BT would have significantly more impaired HRQOL than children with ALL in the following dimensions: physical health, cognitive functioning, social functioning, and school functioning.
MATERIALS AND METHODS
Subjects and Settings
Eligible participants were English and Spanish-speaking parents of children ages 2–18 years with a diagnosis of BT or ALL. Families were recruited from Childrens Hospital Los Angeles (CHLA) and Children's Hospital and Health Center in San Diego (CHSD). Families whose children were recently diagnosed (< 6 weeks ago) and families in medical crisis were excluded. One parent per family participated in the study. When both parents were present at the interview, the child's primary caretaker completed the forms. Parents were given the option to complete forms in either English or Spanish.
Investigators used clinic rosters to identify eligible patients who visited either hospital between April 2000 and January 2001. Additional patients with BT were recruited at CHLA from August 2002 to February 2003. Informed consent was obtained from all participants. The institutional review boards at CHLA and CHSD approved all study procedures in accordance with requirements established by the U.S. Department of Health and Human Services.
We enrolled 256 families, 215 (84%) from CHLA and 41 from CHSD (16%). The nonparticipation rate at CHLA was 13%. Parents who chose not to participate reported that they did not have enough time or that they did not want to lengthen their clinic visit because their children were tired or not feeling well. The nonparticipation rate for CHSD is not available.
The PedsQL™ 4.0 Measurement Model
The PedsQL™ Measurement Model comprises a set of developmentally appropriate generic core scales and disease-specific modules that measure HRQOL for children and adolescents ages 2–18 years. Both self-report and parent proxy–report versions of the questionnaires are available.10 The parent proxy–report version of the PedsQL™ uses a 5-point Likert scale in which parents rate how much a problem each item has been for their children over the past 7 days (0 = never a problem; 1 = almost never a problem; 2 = sometimes a problem; 3 = often a problem; and 4 = almost always a problem).10 Responses are reverse scored and linearly transformed to a scale ranging from 0 to 100, with higher scores indicating better HRQOL.10
The PedsQL™ 4.0 Generic Core scales: parent proxy report.
The PedsQL™ 4.0 Generic Core scales (23 items) measure physical health (8 items), emotional functioning (5 items), social functioning (5 items), and school functioning (5 items).14 Three summary scores can be calculated: a Physical Health score (8 items), a Psychosocial Health score (15 items), and a Total Core score (23 items).14 Population norms for the PedsQL™ Acute Version of the 4.0 Generic Core scales have been established.10 PedsQL™ scores that are 1 standard deviation (SD) below the population mean indicate at-risk status for poor HRQOL.8
The PedsQL™ 3.0 Acute Cancer Module: parent proxy-report.
The PedsQL™ 3.0 Cancer Module (27 items) assesses 8 dimensions of HRQOL specific to cancer: pain and hurt (2 items), nausea (5 items), procedural anxiety (3 items), treatment anxiety (3 items), worry (3 items), cognitive problems (5 items), perceived physical appearance (3 items), and communication (3 items).10 Scores are calculated for each of the previously named subscales, and a total cancer score represents HRQOL as measured by the PedsQL™ Acute Cancer Module.10
The PedsQL™ Multidimensional Fatigue scales: parent proxy-report.
The PedsQL™ Multidimensional Fatigue scales (18 items) assess general fatigue (6 items), sleep/rest fatigue (6 items), and cognitive fatigue (6 items).10 Each of the three subscales yields a score, and a total fatigue score represents HRQOL as measured by the PedsQL™ Multidimensional Fatigue scales.10 Population norms for the PedsQL™ Multidimensional Fatigue scales have been reported by Varni et al.10
Validity and reliability of PedsQL™ measures
The PedsQL™ 4.0 Generic Core scales, the PedsQL™ 3.0 Acute Cancer Module, and the PedsQL™ Multidimensional Fatigue scales have been tested with pediatric oncology patients and have demonstrated strong internal consistency, reliability, and validity.10 In the current study, Cronbach alpha coefficients15 for the Core, Cancer, and Fatigue scales were > 0.70, which is recommended for group comparisons.16 Total scores on the PedsQL™ 3.0 Acute Cancer Module and the PedsQL™ Multidimensional Fatigue scale were moderately correlated with the PedsQL™ 4.0 Generic Core total score (r = 0.75–0.79), suggesting that the constructs of HRQOL and fatigue are unique but highly related.17
PedsQL™ Family Information Form.
The PedsQL™ Family Information Form surveys demographic information, including parents' education, marital status, and occupation. These variables were used to calculate the Hollingshead socioeconomic status (SES) index.18
Medical chart data.
Research assistants extracted the following information from each child's medical record: diagnosis, location of BT, date of diagnosis, treatment status, and date of last treatment.
Because adequate response equivalence has been demonstrated for the PedsQL™ Generic Core scales in English and Spanish,10, 14 and identical rigorous PedsQL™ translation standards were used in the development of Spanish language forms for the Core, Cancer, and Fatigue scales,10 responses were pooled across languages for each instrument. HRQOL scores did not differ significantly by either study location, allowing us to combine the data from the two sites.
Descriptive statistics (means, SD, and frequencies) were calculated for all variables. Standard procedures were used to compute PedsQL™ total, summary, and subscale scores. Families were categorized into three SES groups using the Hollingshead SES index.18 We used the patient's date of last treatment to categorize their treatment status as receiving treatment, having not received treatment for < 12 months, or having not received treatment for ≥ 12 months. Chi-square analyses were used to test for group differences between patients with BT and patients with ALL for demographic characteristics and medical/treatment factors.
Analyses of variance were performed to identify demographic and medical/treatment factors related to PedsQL™ scores across both diagnoses. Regression analyses were conducted to determine the independent effect of a child's cancer diagnosis on PedsQL™ total and subscale scores.19 Regression models included child's age at interview, gender, age at diagnosis, and treatment status as covariates. Patients with BT and patients with ALL were stratified by treatment group status (receiving treatment, having not received treatment for < 12 months, or having not received treatment for ≥ 12 months). Tests for linear and quadratic trends were calculated across ordinal categories. t tests were used to determine group differences in HRQOL scores. Pearson correlations were used to estimate the association between PedsQL™ scale scores. All statistical testing used an alpha value of 0.05 (two-tailed). All data analyses were conducted using SAS (Version 9.0; SAS, Cary, NC).
Two hundred fifty-six parents participated, including 86 (34%) with a child diagnosed with BT and 170 (66%) with a child diagnosed with ALL (Table 1). Patients ranged in age from 2 to 18 years and were more often male than female (males, n =149 [58%]; females, n = 107 [42%]). Most patients were diagnosed before age 5 years (n = 122 [48%]) and were receiving active treatment at the time of their assessment (n = 153 [60%]). More than one-half of the participants identified themselves as Latino (n = 141 [55%]). Eighty-eight percent of the parent proxy-reports were completed by female caregivers. Among patients with BT, tumor location was identified as infratentorial (n = 41), supratentorial (n = 20), and midline (n = 21). Information about tumor location was missing for four patients.
|Characteristic||No. of patients (%)||Chi-square P value|
|Total||170 (100)||86 (100)|
|Child's age at evaluation (yrs)|
|Mean (SD)||7.8 (4.1)||9.7 (4.4)|
|2–4||42 (25)||11 (13)|
|5–7||51 (30)||21 (24)|
|8–12||51 (30)||32 (37)|
|13–18||26 (15)||22 (26)||0.03|
|Latino||91 (54)||50 (58)|
|Caucasian||47 (28)||27 (31)|
|Other||32 (19)||9 (10)||0.22|
|Male||101 (59)||48 (55)|
|Female||69 (41)||38 (44)||0.58|
|Family's socioeconomic statusab|
|Mean (SD)||34.5 (16.8)||36.2 (17.5)|
|Low (≤ 24)||67 (40)||30 (35)|
|Moderate (25–44)||53 (32)||22 (26)|
|High (≥ 45)||47 (28)||33 (38)||0.40|
|Child's age at diagnosis (yrs)|
|Mean (SD)||7.1 (4.3)||5 (3.7)|
|< 5||94 (55)||28 (33)|
|5–7||35 (21)||22 (26)|
|8–12||27 (16)||25 (29)|
|13–18||14 (8)||11 (12)||0.005|
|Child's treatment status|
|Receiving treatment||117 (69)||36 (42)|
|No treatment for < 12 mos||26 (15)||26 (30)|
|No treatment for ≥ 12 mos||27 (16)||24 (28)||0.001|
|Mother/stepmother, grandmother||151 (89)||75 (87)|
|Father/stepfather||19 (11)||11 (13)||0.70|
|< High school graduate||65 (38)||31 (36)|
|High school graduate||24 (14)||15 (18)|
|Some college||44 (26)||18 (21)|
|≥ College graduate||35 (21)||21 (25)||0.69|
|< High school graduate||58 (37)||31 (36)|
|High school graduate||21 (14)||12 (14)|
|Some college||36 (23)||18 (21)|
|≥ College graduate||40 (26)||25 (29)||0.94|
Patients with BT were older at the time of their evaluation (P = 0.03), older at diagnosis (P = 0.005), and more likely not to be receiving treatment (P = 0.001) compared with patients who had ALL. Among patients who had not received treatment for ≥ 12 months, the average time without treatment was 6.5 years for patients with ALL (range, 1–13 years) and 5.6 years for patients with BT (range, 1–12 years; t test P = 0.30). Patients with BT and ALL did not differ significantly as to gender, SES, ethnicity, or parents' educational status. Fathers' participation rates were similar for the two groups.
Association between health-related quality of life and demographic characteristics and disease treatment factors
Univariate results for demographic characteristics and disease/treatment factors are presented in Table 2. Patient's age at interview was inversely related to the Generic Core total score (trend P = 0.05), Core Physical subscale score (trend P = 0.04), and Multidimensional Fatigue total score (trend P = 0.0008). Gender was significantly related to the Core Physical subscale (P = 0.0003) and to the Core total score (P = 0.02), with lower parent proxy–reported scores for female patients. Child's age at diagnosis was inversely related to the Multidimensional Fatigue total score (trend P = 0.003). There was a positive relationship between treatment status and the Acute Cancer total score (trend P = 0.03). Ethnicity, SES, and parents' levels of education were unrelated to PedsQL™ Generic Core, Acute Cancer, or Multidimensional Fatigue scores (data not shown). Among patients with BT, PedsQL™ scores did not differ significantly by tumor location.
|PedsQL™ Core Psychosocial Health score||PedsQL™ Core Physical Health score||PedsQL™ Core Total score||PedsQL™ Acute Cancer Total score||PedsQL™ Multidimentional Fatigue Total score|
|Child's age at interview|
|2–4||73.5 (17.5)||74.5 (20.5)||73.8 (18.1)||77.9 (13.7)||80.0 (14.5)|
|5–7||71.8 (16.5)||69.0 (24.2)||70.7 (17.5)||75.6 (17.3)||77.5 (16.6)|
|8–12||67.4 (16.6)||62.5 (24.4)||65.5 (16.9)||73.3 (16.3)||71.1 (19.6)|
|13–18||67.4 (21.8)||66.7 (26.3)||67.2 (21.7)||71.5 (20.6)||66.5 (24.2)|
|Trend P valuea||0.14||0.04||0.05||0.23||0.0008|
|Male||70.8 (19.0)||72.2 (22.1)||71.3 (18.5)||76.3 (16.3)||75.3 (19.5)|
|Female||68.7 (16.3)||61.3 (25.6)||65.9 (18.1)||72.2 (17.8)||72.0 (19.1)|
|Child's age at diagnos is (yrs)|
|< 5||71.5 (17.7)||71.7 (23.4)||71.6 (18.5)||76.4 (16.0)||77.8 (17.2)|
|5–7||69.0 (17.1)||65.5 (22.6)||67.6 (16.6)||73.9 (17.5)||74.0 (18.9)|
|8–12||68.5 (18.5)||63.2 (26.0)||66.4 (19.1)||73.4 (18.0)||68.3 (22.3)|
|13–18||66.9 (19.6)||61.5 (25.7)||64.9 (20.4)||68.8 (18.5)||65.7 (19.0)|
|Trend P valuea||0.55||0.07||0.17||0.24||0.003|
|Child's treatment status|
|Receiving treatment||69.1 (18.6)||66.1 (25.3)||67.8 (19.3)||72.3 (18.1)||72.4 (20.3)|
|No treatment for < 12 mos||74.2 (15.3)||69.4 (20.9)||72.5 (15.9)||79.0 (13.6)||77.9 (14.5)|
|No treatment for ≥ 12 mos||68.1 (17.7)||70.3 (24.0)||68.9 (18.4)||77.0 (15.8)||74.4 (20.6)|
|Trend P valuea||0.15||0.47||0.30||0.03||0.20|
Health-related quality of life for patients with brain tumors versus acute lymphoblastic leukemia
Results of the multiple regression models that compare HRQOL in patients with BT with patients with ALL are presented in Table 3. Significant group differences are present for the Core Physical Health summary subscale (P = 0.0004; mean difference, −10.7), Core Psychosocial Health summary subscale (P = 0.04; mean difference, −4.2), Core Social Functioning subscale (P = 0.05; mean difference, −5.3), Core School Functioning subscale (P = 0.02; mean difference, −8.2), Core total (P = 0.002; mean difference, −6.8), Cancer Worry subscale (P = 0.05; mean difference, 5.5), Cancer Cognitive subscale (P = 0.005; mean difference, −9.8), Multidimensional Fatigue General Fatigue subscale (P = 0.006; mean difference, −7.2), Multidimensional Fatigue Cognitive Fatigue subscale (P = 0.0003; mean difference, −11.6), and Multidimensional Fatigue total score (P = 0.02; mean difference, −6.0). With the exception of the Acute Cancer worry subscale, parent proxy–reported scores for patients with BT were significantly lower (i.e., indicative of more problems and poorer HRQOL) than parent proxy–reported scores for patients with ALL. No significant group differences were found for the Core Emotional Functioning subscale, the Fatigue Sleep/Rest subscale, or any of the following Acute Cancer Module subscales: Pain, Nausea, Procedural Anxiety, Treatment Anxiety, Perceived Physical Appearance, Communication, and Total.
|Mean (SD)||P valuea|
|ALL (n = 170)||BT (n = 86)|
|PedsQL™ Generic Core scales|
|Total||71.3 (18.1)||64.5 (18.5)||0.002|
|Physical Health||71.2 (22.5)||60.5 (25.8)||0.0004|
|Psychosocial Health||71.3 (17.9)||67.1 (17.7)||0.04|
|Emotional Functioning||67.6 (20.0)||67.2 (20.0)||0.34|
|Social Functioning||77.0 (20.4)||71.7 (20.2)||0.05|
|School Functioning||68.0 (22.5)||59.8 (24.7)||0.02|
|PedsQL™ Acute Cancer Module|
|Total||75.1 (16.3)||73.4 (18.5)||0.16|
|Pain||75.4 (25.4)||76.7 (26.6)||0.78|
|Nausea||79.0 (21.5)||81.7 (21.7)||0.98|
|Procedural Anxiety||58.8 (32.3)||65.7 (32.7)||0.64|
|Treatment Anxiety||72.5 (29.5)||74.3 (28.9)||0.63|
|Worry||77.7 (29.1)||83.2 (21.6)||0.05|
|Cognitive||74.5 (21.8)||64.7 (24.8)||0.005|
|Perceived Physical Appearance||79.8 (23.0)||79.3 (24.0)||0.98|
|Communication||76.2 (25.0)||70.8 (33.8)||0.12|
|PedsQL™ Multidimensional Fatigue scales|
|Total||75.9 (17.8)||69.9 (21.6)||0.02|
|General Fatigue||73.8 (20.3)||66.6 (25.4)||0.006|
|Sleep/rest Fatigue||74.4 (21.4)||75.9 (21.8)||0.39|
|Cognitive Fatigue||79.4 (21.5)||67.8 (28.4)||0.0003|
Health-related quality of life by treatment group status
To further explore the differences between patients with BT and ALL, we examined the two PedsQL™ Generic Core summary scales, (i.e., Physical Health and Psychosocial Health), PedsQL™ Acute Cancer Module total scores, and PedsQL™ Multidimensional Fatigue scale total scores for patients who were receiving treatment and patients who had not received treatment for < 12 and ≥ 12 months, respectively (Table 4).
|Mean (SD)||Linear trend P value||Quadratic trend P value|
|Receiving treatment||No treatment for < 12 mos||No treatment for ≥ 12 mos|
|PedsQL™ Generic Core scales|
|BT||50.3 (26.0)||73.5 (18.8)||61.7 (26.4)||0.09||0.01|
|ALL||71.0 (23.1)||65.4 (22.4)||77.9 (19.1)||0.03||0.13|
|BT||64.8 (18.1)||73.9 (16.2)||63.2 (17.3)||0.80||0.04|
|ALL||70.4 (18.7)||74.5 (14.8)||72.4 (17.3)||0.21||0.41|
|PedsQL™ Acute Cancer Module|
|BT||68.5 (20.4)||78.5 (16.1)||75.2 (16.8)||0.07||0.15|
|ALL||73.4 (17.3)||79.5 (11.0)||78.5 (16.1)||0.008||0.21|
|PedsQL™ Multidimensional Fatigue scales|
|BT||64.9 (22.9)||79.0 (14.6)||67.7 (23.6)||0.37||0.06|
|ALL||74.6 (18.9)||76.8 (14.6)||80.2 (15.5)||0.009||0.93|
Physical health for patients with brain tumors and acute lymphoblastic leukemia.
We found a reverse pattern with respect to physical functioning when BT group and ALL group scores were plotted for patients receiving treatment and for patients who had not received treatment for < 12 months and ≥ 12 months, respectively (Fig. 1). For patients with ALL, there was a slight reduction in physical functioning for patients who had not received treatment for < 12 months compared with patients receiving treatment (P = 0.26; mean difference, −5.6) followed by a significant increase in scores (indicative of better physical functioning) when patients with ALL who had not received treatment for ≥ 12 months were compared with their counterparts who had not received treatment for < 12 months (P = 0.02; mean difference, 12.5). With physical functioning, there was a significant linear trend for patients with ALL across the treatment groups (linear trend P = 0.03). For patients with BT, we found a quadratic trend, not a linear trend, across the treatment groups (Physical Health: quadratic trend P = 0.01). Physical Health scores peaked (better physical functioning) among patients with BT who had not received treatment for < 12 months. There was a marked decrease in Physical Health scores for patients with BT who had not received treatment for ≥ 12 months compared with patients who had not received treatment for < 12 months (Physical Health: P = 0.07; mean difference, 11.8).
Psychosocial health for patients with brain tumors and acute lymphoblastic leukemia.
Although patients with BT and ALL demonstrated similar patterns in terms of psychosocial functioning across the three treatment groups, the pattern was more pronounced for the patients with BT (Fig. 2). For patients with BT, we again found a significant quadratic trend across the treatment groups (quadratic trend P = 0.04). Psychosocial functioning scores were highest for patients with BT who had not received treatment for < 12 months and lowest (poorer psychosocial functioning) for patients who had not received treatment for ≥ 12 months. In contrast, we found that psychosocial functioning in patients with ALL remained rather flat across the three treatment groups.
Acute Cancer Module total scores for patients with brain tumors and acute lymphoblastic leukemia.
With the Acute Cancer total score, we found a linear pattern for patients with BT and ALL across the treatment groups (ALL: linear trend P = 0.009; BT: linear trend P = 0.07). Patients with BT and ALL who were receiving treament experienced more cancer-specific problems than patients who were not receiving treatment.
Multidimensional Fatigue total scores for patients with brain tumors and acute lymphoblastic leukemia.
The fatigue patterns across treatment groups also varied for patients with BT and ALL (Fig. 3). For patients with ALL, there was a gradual decrease in fatigue symptoms (higher fatigue scores) across the treatment groups (linear trend P = 0.008). In contrast, patients with BT demonstrated the familiar quadratic pattern with fatigue scores being highest (less fatigue) in patients who had not received treatment for < 12 months compared with patients who were receiving treatment and patients who had not received treatment for ≥ 12 months (quadratic trend P = 0.06). Patients with BT who had not received treatment for ≥ 12 months demonstrated a trend toward more fatigue (lower scores) when compared with patients who had not received treatment for < 12 months (P = 0.07; mean difference, 11.3).
Health-related quality of life in patients with brain tumors and acute lymphoblastic leukemia compared with healthy children
To further understand HRQOL in patients with BT and ALL, we compared data from patients who had not received treatment for ≥ 12 months (i.e., long-term cancer survivors) with normative data collected from parents of healthy children.10, 14 On the PedsQL™ Generic Core scale, we found that 63% of the patients with BT and 37% of the patients with ALL had a total score > 1 SD below the normative value (n = 157; mean, mean parent proxy-report total Core score for healthy children, 84.5; SD, 15.7). On the PedsQL™ Multidimensional Fatigue scales, the mean fatigue total score for patients with ALL fell within the normal range, whereas the mean total Fatigue score for patients with BT was nearly 2 SD below the norm (N = 102; mean parent proxy–reported total Fatigue score for healthy children, 89.6; SD, 11.4).10 Patients with BT were significantly more fatigued than their peers (P = 0.0001).
In the current study, parent proxy-reports were used to assess HRQOL in children diagnosed with a BT and ALL. Specific dimensions of HRQOL, including physical health, psychosocial health, and fatigue, were charted for patients who were receiving treatment and also for patients who had not received treatment for < 12 months and ≥ 12 months, respectively. The use of parent-proxy assessments of HRQOL allowed for greater inclusion of children who were developmentally immature, physically ill, or cognitively impaired. For example, 95 (35%) of the 256 parents who participated reported that their children could not provide self-report. This methodology, which results in a more representative sample, is important because children who cannot provide self-report often have the greatest need for clinical services.
Analysis of parent proxy-reports found that overall, children with a BT experienced more HRQOL problems than children with ALL. According to parents, patients with BT had significantly poorer HRQOL than patients with ALL in the dimensions of physical and psychosocial health, and cognitive, social, and school functioning. These findings are consistent with literature indicating that patients with BT often experience problems with physical, intellectual, and psychosocial functioning that fail to resolve over time.7, 20–24
Based on our findings, patients with ALL appear to worry significantly more than patients with BT. The PedsQL™ Acute Cancer worry subscale items evaluate the parent's perception of their child's worry about 1) the side effects of treatment, 2) the efficacy of medical treatments, and 3) disease recurrence. Although the cognitive impairments of patients with BT may limit their insight and worry, it is also possible that the clinic experience itself may induce more worry for patients with ALL compared with patients with BT. For example, patients with ALL undergo disease evaluations and medical procedures with each clinic visit, including blood counts and/or lumbar punctures and bone marrow aspirations. Disease evaluations for patients with BT are less frequent, and tend to be less invasive (e.g., imaging studies). The frequent medical procedures experienced by ALL patients have the potential to affect HRQOL ratings when assessments are obtained in the clinical setting. It is also noteworthy that the standard treatment for patients with ALL occurs over a longer period of time than many treatments for patients with BT. Therefore, it is plausible that the length of time of active treatment is positively correlated with feelings of uncertainty and worry.
Although the current study is not longitudinal, the cross-sectional view of HRQOL scores among patients receiving treatment, patients who had not received treatment for < 12 months, and patients who had not received treatment for ≥ 12 months suggests that patients with BT and ALL may follow different trajectories once they complete their therapies. Psychosocial Health scores for patients with ALL remained relatively consistent across the three treatment groups. However, with respect to Physical Health, patients with ALL who had not received treatment for ≥ 12 months had significantly higher scores than did those who had not received treatment for < 12 months, suggesting an improvement over time. In contrast, the Physical Health and Psychosocial Health scores of patients with BT demonstrated an inverted U (quadratic) pattern, with peak improvements for children who had not received treatment for < 12 months and sharp declines for children who had not received treatment for ≥ 12 months.
Patients who have not received treatment for < 12 months are in a period of transition, and it appears that patients with BT and patients with ALL experience this transition phase, or early phase of survivorship, differently. BT and ALL treatment regimens may help to explain these differences. For instance, patients with BT often receive intensive treatments throughout their entire course of therapy. Only when treatment ends do these patients have the opportunity to fully recover from the acute toxic effects of surgery, radiotherapy, and chemotherapy. Thus, HRQOL scores for patients with BT are apt to improve dramatically once treatment has been discontinued. In contrast, the final phase of treatment for patients with ALL (i.e., maintenance therapy) lasts for a long period of time, and is rather mild with minimal side effects. Patients with ALL make gradual progress toward recovery long before the end of treatment. Thus, HRQOL changes are apt to be minimal for patients with ALL who are receiving treatment compared with those who have not received treatment for < 12 months. Monitoring patients as they proceed through the various phases of treatment and survivorship should help to identify factors that facilitate or impede transitions to maximum health.
Comparisons between parent proxy–reported total Core scale scores for healthy children and for patients who had not received treatment for ≥ 12 months revealed that approximately two-thirds of the patients with BT and one-third of the patients with ALL were > 1 SD below the healthy norms.14 This finding is consistent with extant literature indicating that survivors of a BT are at high risk of developing significant physical, cognitive, and social problems,7, 20–24 whereas most survivors of ALL return to normal levels of functioning.25, 26 In the current study, HRQOL findings among long-term survivors illustrate the higher prevalence of problems among patients with BT, but also demonstrate the need for ongoing support and clinical services for both populations. Longitudinal HRQOL studies are needed for the early identification of patients at greatest risk for adjustment difficulties to maximize coping skills and minimize HRQOL problems.
Although adult patients identify fatigue as the most common and distressing cancer-related symptom,27 little is known about fatigue in pediatric patients with cancer. Only recently have valid, developmentally sensitive, multidimensional measures been available to assess fatigue in children with cancer.10
As one of the first studies to evaluate fatigue in pediatric patients with BT, our data suggest that the pattern of fatigue over the course of treatment differs for patients with BT and ALL. For patients with ALL, fatigue symptoms decrease as time without treatment increases, and comparisons with healthy norms indicate that long-term survivors of ALL experience no more fatigue than their healthy peers. These findings are consistent with another recent study that has examined fatigue in long-term survivors of ALL.25 Among patients with BT, fatigue follows the familiar quadratic pattern, with patients receiving treatment and patients who have not received treatment for ≥ 12 months experiencing similar levels of fatigue. In contrast to survivors of ALL, long-term survivors of BT were significantly more fatigued than their healthy peers.
In our study, fatigue was inversely related to the child's overall HRQOL (PedsQL™ Generic Core total score) (P = 0.0001), a finding that has been reported frequently among adult patients with cancer.28, 29 Fatigue appears to be a significant issue for patients with BT, including those who have not received treatment for years. Additional research is needed to determine if factors such as tumor location, treatment modalities, physical disabilities, endocrine abnormalities, and/or neurocognitive impairments play a role in the development of fatigue. Now that pediatric assessment measures such as the PedsQL™ Multidimensional Fatigue Module are available, investigators are in a position to better define fatigue and its impact on HRQOL.
Several factors must be considered when interpreting our results. The current study utilized a convenience sampling methodology, with no information available on nonparticipating patients. The wide variability in HRQOL scores for patients with BT and ALL suggests that a broad range of subjects were recruited from both patient populations, and the subjects included are representative of the population from which they were drawn. However, when data are gathered in a clinic setting, the sickest children (who may be hospitalized or in hospice care) and healthy survivors (who are not experiencing problems) are both less likely to visit the clinic, and may not be included in the study. Selection bias, if present, is most likely nondifferential in nature because recruitment procedures were identical for patients with BT and ALL.
Although parent proxy–reported HRQOL assessments provide important clinical information about a child's adjustment to illness, parent reports may be subject to inflation due to parental stress. A study by Johnston et al.30 suggests that parents of patients with cancer and healthy children tend to rate their children's HRQOL poorer than children report for themselves. It is noteworthy that although parents may overreport HRQOL problems, a recent study by Carpentieri et al.13 suggests that the self-reports of patients with BT may underestimate HRQOL problems because of “psychologically-based or neurologically-based denial, defensiveness, or lack of insight” (p. 284).13 Given that both types of reports have strengths and weaknesses, parent proxy-reports and self-reports presumably provide unique and complementary information concerning the child's HRQOL, and both can be useful in clinical decision-making.10
To summarize, parent perceptions of child HRQOL are important because they often drive pediatric health care decisions and program development.31 Parent proxy-reports also allow investigators to assess HRQOL for young children and extremely ill or cognitively delayed children who are unable to complete self-report forms. The findings indicate that patients with BT are at higher risk for HRQOL problems than patients with ALL, and that patients with BT and ALL experience stages of treatment differently. In the current study, the majority of survivors of BT have significant physical, psychosocial, and fatigue problems. Although most long-term survivors of ALL appear to return to normal functioning, a subset of the ALL population continues to struggle with psychosocial problems. A principal clinical implication of our research is that psychosocial support interventions for patients with BT and ALL need to be appropriately timed in relation to the stage of the patient's treatment and recovery. A clear understanding of the types of problems experienced during distinct phases of treatment can be useful in the development of HRQOL screening programs, psychosocial interventions for patients with BT and ALL, and long-term follow-up care programs. Longitudinal studies using both self-reporting and parent proxy-report are needed to identify factors that facilitate or impede transitions to maximum health, and to determine whether child-reported patterns of HRQOL are similar to the patterns reported by parents.
- 1Introduction. In: RiesLA, SmithMA, GurneyJG, et al., editors. Cancer incidence and survival among children and adolescents: United States SEER program 1975–1995 [monograph online]. Bethesda: National Cancer Institute, 1999: 1–17. Available from URL: http://seer.cancer.gov/publications/childhood [accessed April 5, 2004]., , .
- 2RiesLA, EisnerMP, KosaryCL, et al., editors. SEER cancer statistics review, 1975–2000. Bethesda: National Cancer Institute, 2003. Available from URL: http://seer.cancer.gov/csr/1975_2000 [accessed April 5, 2004].
- 3CNS and miscellaneous intracranial neoplasms. In: RiesLA, SmithMA, GurneyJG, et al., editors. Cancer incidence and survival among children and adolescents: United States SEER program 1975–1995. Bethesda: National Cancer Institute, 1999: 51–64. Available from URL: http://seer.cancer.gov/publications/childhood [accessed March 17, 2004]., , .
- 4Leukemia. In: RiesLA, SmithMA, GurneyJG, et al., editors. Cancer incidence and survival among children and adolescents: United States SEER program 1975–1995. Bethesda: National Cancer Institute, 1999: 17–34. Available from URL: http://seer.cancer.gov/publications/childhood [accessed March 17, 2004]., , , .
- 6Brain tumors in children. In: HollandJF, editor. Cancer medicine. Baltimore: Williams and Wilkins, 1997: 2975–2989., , .
- 16Psychometric theory. 3rd ed. New York: McGraw-Hill, 1994., .
- 17Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum and Associates, 1998..
- 18Four factor index of social status. New Haven: Department of Sociology, Yale University, 1975..
- 19Statistical methods in cancer research. I. The analysis of case-control studies. Lyon: International Agency for Research on Cancer, 1980., .
- 30Parent and child reporting of negative life events: discrepancy and agreement across pediatric samples. J Pediatr Psychol. 2001; 42: 661–667., , , .