Profile of Daily Life in Children With Brain Tumors: An Assessment of Health-Related Quality of Life
http://www.100md.com
《临床肿瘤学》
the Departments of Neurology, Pediatrics, Neurosurgery, and Radiation Oncology, Stanford University School of Medicine, Palo Alto, CA
Department of Anesthesiology, University of Washington Medical Center, Seattle, WA
Texas A&M University, College Station, TX
ABSTRACT
PURPOSE: The survival of children with CNS tumors approaches 70%, yet health-related quality of life (HRQOL) has not been investigated rigorously in this population. We aimed to show that universal assessment of HRQOL could be obtained easily by using the PedsQL 4.0 and to provide a composite profile of their daily lives.
PATIENTS AND METHODS: The PedsQL was administered to all patients seen in the neuro-oncology clinic at Lucile Packard Children's Hospital (Palo Alto, CA) from December 2001, to September 2002. Patients were compared with healthy controls by using two-sided t tests to evaluate statistically significant differences.
RESULTS: One hundred thirty-four patients (73 male; mean age ± standard deviation, 11.8 ± 5.4 years; 55 had low-grade glioma, 32 had medulloblastoma/primitive neuroectodermal tumor/embryonal tumor, 17 had malignant astrocytoma, nine had germ-cell tumor, and 21 had other types of tumors) were assessed, each in less than 20 minutes. Scores on both child and parent-proxy reports for the total HRQOL, psychosocial, physical, emotional, social, and school-functioning scales were all significantly lower than controls (P < .01). Patients with low-grade glioma were reported to have the highest total HRQOL. Children receiving radiation therapy (XRT) but no chemotherapy had significantly lower total, psychosocial, emotional, and social functioning than those receiving other treatments, including XRT plus chemotherapy.
CONCLUSION: The PedsQL can be used to assess HRQOL rapidly and easily in children with CNS tumors, who have significantly worse HRQOL than healthy children. Children receiving XRT fare worse overall; chemotherapy added to XRT does not seem to worsen HRQOL. Assessment of HRQOL should be included as an outcome in future clinical trials.
INTRODUCTION
Throughout the last decade, the prognosis for children diagnosed with CNS tumors has improved steadily; 5-year survival rates now approach 70%.1 With growing numbers of survivors, increased attention should focus on the quality of life of patients as part of a complete clinical assessment, a variable in making treatment decisions, and an end point for clinical trials.
Past research, although limited, has shown that brain tumors are associated with decreases in quality of life and cognition. The Health Utilities Index survey indicated that 80% of children with brain tumors experience some morbidity, with cognitive deficits affecting the majority.2 Years after diagnosis, these patients have significantly more neurosensory difficulties including increased incidence of late-onset blindness, motor deficits, and chronic seizure disorders.3 As adults, survivors of childhood brain tumors are at increased risk for unemployment, inability to drive, and cognitive, motor, and psychological-emotional impairments.4,5 These patients are more likely than the general population and other cancer survivors to develop psychoses and need psychiatric hospitalization.6 These survivors have also been found to have a lower "health status index" than survivors of other childhood cancers.7 Children diagnosed with medulloblastoma demonstrate significantly lower intelligence quotient (IQ) than patients with low-grade brain tumors.8 Medulloblastoma patients are also likely to never marry or to develop emotional or visual disturbances.4 Whole-brain irradiation before age 7 years correlates with significant reduction in IQ, increases in focal neurological deficits, and need for special education services.9,10
These previous studies have often accrued only small numbers of patients or been restricted to a particular brain tumor pathology.7,11,12 Control groups are inconsistent (often using siblings or other cancer patients), and only the parent or the patient is surveyed.3-6 In addition, these surveys focus on specific aspects of quality of life, such as neurological or cognitive sequelae, rather than health-related quality of life (HRQOL).3,6-8 HRQOL is a multidimensional construct that incorporates both objective and subjective data including, but not limited to, the social, physical, and emotional functioning of the child and, when indicated, his or her family.13 A standard, comprehensive instrument has not been used routinely to assess HRQOL and daily activities simply and quickly in this population.
We sought to demonstrate the feasibility of using the Pediatric Quality of Life Inventory 4.0 and cancer module (PedsQL 4.0)14 to obtain a rapid, complete measurement of HRQOL in a large sample of patients with CNS tumors. Furthermore, we aimed to provide a profile of HRQOL in children with brain tumors, including a composite of their daily life. We hypothesized a priori that variables well established to affect clinical outcomes in childhood brain tumors (age, sex, tumor location, hydrocephalus, pathology, and treatment type) were likely to influence HRQOL.
PATIENTS AND METHODS
All patients 2 years of age followed in the neuro-oncology clinic at Lucile Packard Children's Hospital at Stanford (Palo Alto, CA) for diagnosis of a primary CNS tumor were considered for this study. At Stanford, every child with any brain tumor diagnosis is followed in this clinic at an interval ranging from weekly to annually. From November 2001 to September 2002, all patients and their parents returning for a routine follow-up appointment were consecutively asked to enroll. Patients who were not scheduled for a follow-up appointment during the study period were approached initially by mail and subsequently by telephone. Two patients were surveyed at their homes, and 34 patients were mailed the surveys along with explicit instructions for their completion. Informed consent was obtained from all patients and their parents in accordance with the institutional human subjects review board.
HRQOL was assessed by using the PedsQL 4.0 generic core scales and cancer module. The 23-item generic scales measure physical, emotional, social, and school functioning, and the 27-item cancer module assesses cancer-specific aspects of HRQOL, including treatment and procedural anxiety, cognitive ability, patient worry, physical self-appearance, and communication. Both surveys asked about HRQOL in the past month. Validity of the PedsQL 4.0 was established previously for healthy children and those with acute and chronic conditions by using known-groups comparison and correlations with indicators of morbidity and illness.14,15 Data for healthy children were obtained from a sample of 730 children assessed during well-child checks or by telephone as part of the initial development of the PedsQL 4.0 at the Children's Hospital and Health Center (San Diego, CA).14 Reliability for the PedsQL 4.0 has been demonstrated with Chronbach's coefficients.14 The PedsQL 4.0 cost approximately $7 per patient-parent dyad.
The PedsQL 4.0 was administered as age-appropriate surveys for young children (5 to 7 years of age), children (8 to 12 years of age), and adolescents ( 13 years of age). For children younger than 5 years of age, only a parent completed the PedsQL 4.0. For patients 5 years of age, the PedsQL 4.0 was administered verbally by a research assistant (S.R.B. or T.L.G.) and monitored by the principal investigator (P.G.F.) to ensure that questions were always posed in the same manner. One parent completed a parallel parent-proxy PedsQL 4.0, along with a school and physical-data questionnaire. Both the instrument and the supplemental form were available in English and Spanish. Instructions to those families participating by mail indicated clearly that one parent should complete his or her portion of the survey independently of the child and then read the PedsQL 4.0 to the patient for completion.
To obtain additional data that would provide a comprehensive profile of the daily activities of patients, all parents were asked to complete a school information and physical-data questionnaire. This simple inventory inquired about current school enrollment, special education services, medications, and use of corrective devices (eg, hearing aids or glasses) and also posed questions about patients' abilities to carry out daily activities including dressing, making breakfast, driving, and counting change. Date of birth, sex, date of diagnosis, tumor pathology and location, treatment details, and relapse data were obtained from the Lucile Packard Children's Hospital's neuro-oncology database.
Two-tailed t tests were used to examine differences between patient scores and healthy normals. Differences within patient subgroups were also evaluated by using two-tailed t tests. Pearson correlations were used to assess the association between child self-reports and parent-proxy reports. Analysis of variance was used to elucidate differences by pathology or treatment type. All statistical calculations were conducted by using Statistical Package for the Social Sciences (SPSS) version 10.0 (SSPS Inc, Chicago, IL).
RESULTS
One hundred thirty-four patients (108 patient-parent dyads; 17 parents; nine children) were assessed; their characteristics are listed in Table 1. Twelve children younger than 5 years of age were not surveyed; additionally, five children were low functioning and could not complete the survey. Nine patients came to the clinic unaccompanied by parents or were no longer living with their parents. Only four families refused to take the survey—two for religious reasons and two for personal reasons.
Median age at the time of survey was 11.3 years, whereas the median age at tumor diagnosis had been 7.2 years. Median time from diagnosis was 3.2 years. Tumor pathologic subtypes were similar to those reported for the population.16 Forty-one percent of patients had a low-grade glioma, and 23.9% had medulloblastoma or primitive neuroectodermal tumor. Thirty-one patients had experienced a tumor recurrence or progression before this survey. Prior treatment included surgery (79.9%), radiation therapy (XRT; 53.0%), and chemotherapy (41.8%). Fifty-four percent of patients were white, non-Hispanic; 29.1% were Hispanic; 11.9% were Asian; and 4.5% were African American, similar to controls.14 Sex and age in our sample were not significantly different compared with controls.
Completion of both parent and child surveys required less than 20 minutes in all cases. No difficulties were observed answering the PedsQL 4.0 in children aged 5 to 7 years or those with special education needs, likely because of the instrument's simplicity and brevity. For proxy and self-reports of all age groups, the PedsQL 4.0 generic total and psychosocial Chronbach's coefficients were substantially above 0.70, indicating internal consistency or reliability. For the 5- to 7-year-olds, only the self-report social (0.49) and school-functioning (0.59) scale scores were notably below 0.70, and among 8- to 12-year-olds, scores on the self-report emotional (0.65) and social (0.67) scales were borderline. For children receiving special education services, coefficients exceeded 0.70 for total, psychosocial, physical, and social scale scores and were borderline for the self-report emotional (0.61) and school functioning (0.68) scales. Similarly, for the cancer module, coefficients were generally well above 0.70, except for the 5- to 7-year-old self-reports of pain (0.48), nausea (0.57), treatment (0.52), and perception (0.55). In addition, self-reports of perception in 8- to 12-year-old patients (0.60) and children receiving special education (0.31) were low.
Results from the school and physical-data questionnaire are listed in Table 2. Of the patients sampled, 96.8% were enrolled in school, and 35 children were receiving special education services; 16.4% of the patients were taking medication for headaches, and 9.7% were on an anticonvulsant for seizures. Nine patients wore a hearing aid, and 27.6% of the patients used glasses or contact lenses to correct visual impairments in one or both eyes. Of the 30 patients eligible to drive at the time of survey, 56.7% had a valid license. Ninety-four percent of the patients reported having a best friend. To help demonstrate validity of the PedsQL 4.0 from this questionnaire, comparisons of children on or off medications revealed lower parent and child report of total HRQOL and pain scores (P < .001) when on pain medicines and lower total HRQOL and nausea scores (P < .01) when on antinausea drugs.
There was a significant difference between patient responses and healthy controls for all generic HRQOL subscales (P < .001), as well as a significant difference between parent-proxy reports and normal parent controls (P < .001). Box plots of responses from the study sample and controls are shown in Figure 1. Pearson's correlation coefficients indicated a good reliability between child self-report and parent-proxy report for each subscale (coefficients ranged from 0.344 [communication subscale] to 0.728 [total]; all P < .01). There were no significant differences in PedsQL 4.0 results between children surveyed in the clinic or at home. In addition, none of the measures on the cancer module were significantly different between patients on or off treatment.
Age, time from diagnosis, and sex did not have any significant association with total HRQOL or psychosocial health. However, on subscales, female children reported higher school-functioning scores than male children (P < .05). Among children below the median age at survey, both parent proxies and self-reports revealed greater procedural anxiety (P < .001), and parent-proxy scores showed that children above the median age had worse perceptions of their self-appearance (P < .05). Parent surveys of children beyond the median age at diagnosis revealed that these patients worried about their illness more than younger children (P < .05). Children whose time from diagnosis was above the median expressed more difficulty in communicating about their illness (P < .05), and their parents indicated that these children had lower social functioning (P < .05).
Presence of a shunt was indicative of significantly lower total HRQOL (P < .01 for self-reports; P < .05 for parent reports) and decreased social functioning (P < .05). Patients who had a shunt also indicated lower physical (P < .05) and psychosocial functioning (P < .01).
Tumor location was not significantly associated with total HRQOL for either self-reports or parent-proxy reports. However, children with infratentorial tumors had significantly more problems communicating issues related to their illness than children with supratentorial tumors (P < .05).
Pathology diagnoses were not associated with differences in total HRQOL on child reports (P = .58) but revealed significant differences on parent-proxy reports (P = .02), as displayed in Table 3. Parent-proxy results showed significant overall differences by pathology for the physical (P = .01) and emotional (P = .01) scales. Parents whose children had low-grade gliomas reported higher total HRQOL (P < .05) and better emotional functioning, physical functioning, and communication (all P < .05).
Analysis by treatment type revealed significant differences in total HRQOL for child (P = .03) and parent (P = .001) reports (Table 4). Both children and parents of children receiving XRT without chemotherapy reported lower total HRQOL (P < .05). These patients also had lower perceived physical self-appearance (P < .05). In addition, parent-proxy scores from children who received XRT without chemotherapy demonstrated significantly lower functioning for all subscales except the physical-functioning scale. Patients who received XRT and chemotherapy had higher total HRQOL than those receiving XRT alone (P < .01). Parents of patients who received both XRT and chemotherapy also revealed these children to have better emotional functioning (P < .001) and social functioning (P < .05) than children who received XRT only. Patients who received chemotherapy only had better emotional functioning than those receiving XRT as their sole treatment (P < .05), as indicated by their parents. Children undergoing both XRT and chemotherapy indicated lower physical functioning than those who received surgery alone (P < .05).
DISCUSSION
Assessment of HRQOL in children with brain tumors is feasible and can be accomplished quickly with ease and reliability between patients and their parents. Our study demonstrates that the PedsQL 4.0, with its age-specific nature, is an expedient instrument for investigating HRQOL in this population. Administration of the surveys requires few resources and can be done by nurses or trainees who are already present in the clinic. Rapid assessment is essential to ensure that the attention of children, especially very young patients, can be maintained and to allow for administration of the survey during regularly scheduled clinic visits. Our findings represent an important step toward improved understanding of the lives of children with brain tumors.
Our results indicate that the overall HRQOL of children with brain tumors is highly variable yet significantly lower than normal. These patients experience not only worsened physical health, as expected, but also decreased psychosocial health, emotional functioning, and social functioning. Children with CNS tumors have inferior school functioning and may require special education services. Many patients rely on regular medication for seizures, headaches, or hormonal deficiencies. Only half of those eligible are able to drive. The majority of survivors are able to care for themselves, but many will not be able to become fully independent as adults.
Although some of our findings are predictable, others are more unexpected. Correlation between child reports and parent-proxy reports shows that children and their parents approach HRQOL in a similar fashion; this finding differs from previous studies, which have shown a discord between these two groups.17 Younger patients are more anxious about procedures. Significantly lower HRQOL among children who have a shunt is likely indicative of effects, such as hydrocephalus, that are characteristic of aggressive, larger tumors. Children with brain tumors and their parents recognize the child's impaired school functioning, yet only a minority receive special education services. Children whose age was above the median, namely adolescents, have worse perception of their self-appearance. Among the general childhood cancer-survivor population, older age is associated with functional impairment and limitations in activity.18 Somewhat similarly, our study indicates that as CNS tumor survivors advance further from the time of diagnosis, HRQOL does not improve; in fact, these children demonstrate more difficulty with communication and social functioning, indicating that they may become more socially isolated with time. These findings differ from those observed in other childhood cancer survivors, whose communication and social functioning improve later after diagnosis.14 Regardless, almost all children with brain tumors reported being able to connect with at least one best friend.
Treatment received clearly has a significant impact on HRQOL. Similar to previous observations, our results indicate that children receiving XRT only fare more poorly, with lower physical, social, and school functioning. Patients who received XRT and chemotherapy did not exhibit a lower total HRQOL compared with children who received XRT only. This finding helps reinforce the hypothesis that XRT dosage and fields might be reduced and chemotherapy added without diminishing overall quality of life. Nevertheless, and as expected, children who receive XRT and chemotherapy, with or without surgery, have worse physical functioning than those undergoing surgery only.
Our study does have some limitations. First, it was conducted in a cross-sectional rather than longitudinal fashion. However, the scope of our investigation was to demonstrate the feasibility of HRQOL assessment in this population and to provide an initial composite picture of children with brain tumors. In addition, because this exploratory study undertook many statistical comparisons, differences found, particularly at the .05 significance level, must be interpreted with caution and need to be replicated in future research. Second, although our effort does represent one of the largest patient samples that has been accrued, the number is still small, and its power may account for the lack of some significant differences between sample groups (eg, among patients with differing pathology). Third, heterogeneity in tumor types and variability in therapeutic protocols may confound HRQOL findings. This is a limitation of any study that attempts to survey HRQOL in this diverse population. Finally, results examining internal consistency among particular age groups indicate low reliability in using the PedsQL 4.0. Limitations in reliability results among 5- to 7-year-old patients on both generic and cancer-module subscales indicated that these measures can be used only as secondary, descriptive measures of HRQOL, whereas proxy reports still provide reliable information among this age group. However, for all age groups and those children receiving special education, the total score for self- and proxy reports approached or exceeded 0.90, making the scale relevant for individual assessment. In addition, the PedsQL 4.0 cancer module, a cancer-specific module, did not discriminate subgroups within our brain tumor sample. This could question its validity in this tumor population, although our sample size might have precluded finding different subgroups.
Taken together, our results represent a novel effort toward universal assessment of HRQOL and daily activities among children with brain tumors. Our approach with the PedsQL 4.0 is rapid, inexpensive, and comprehensive. Nevertheless, HRQOL assessment has been conducted sparingly in children with brain tumors until this point, and there are no data that provide a composite of their daily lives. To develop further the understanding of this most important outcome, we suggest that HRQOL measurement be a standard part of all clinical trials in this population as both a baseline assessment of patients commencing treatment and also as an end point. If our approach were adapted prospectively as routine, we might identify specific tumors or therapies that require modifications in management. Future studies should attempt a longitudinal evaluation of HRQOL to provide a better understanding of treatment and ongoing surveillance of children with brain tumors.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Supported by the National Brain Tumor Foundation Pediatric Research Grant.
Presented in part at the 39th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, May 31, 2003.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Jemal A, Murray T, Samuels A, et al: Cancer Statistics, 2003. CA Cancer J Clin 53:5-26, 2003
Barr RD, Simpson T, Whitton A, et al: Health-related quality of life in survivors of tumours of the central nervous system in childhood: a preference-based approach to measurement in a cross-sectional study. Eur J Cancer 35:248-255, 1999
Packer RJ, Gurney JG, Punyko JA, et al: Long-term neurologic and neurosensory sequelae in adult survivors of a childhood brain tumor: Childhood cancer survivor study. J Clin Oncol 21:3255-3261, 2003
Mostow EN, Byrne J, Connelly RR, et al: Quality of life in long-term survivors of CNS tumors of childhood and adolescence. J Clin Oncol 9:592-599, 1991
Lannering B, Marky I, Lundberg A, et al: Long-term sequelae after pediatric brain tumors: Their effect on disability and quality of life. Med Pediatr Oncol 18:304-310, 1990
Ross L, Johansen C, Dalton SO, et al: Psychiatric hospitalizations among survivors of cancer in childhood or adolescence. N Engl J Med 349:650-657, 2003
Foreman NK, Faestel PM, Pearson J, et al: Health status in 52 long-term survivors of pediatric brain tumors. J Neurooncol 41:47-53, 1999
Mulhern RK, Reddick WE, Palmer SJ, et al: Neurocognitive deficits in medulloblastoma survivors and white matter loss. Ann Neurol 46:834-841, 1999
Radcliffe J, Packer RJ, Atkins TE, et al: Three- and four-year cognitive outcome in children with noncortical brain tumors treated with whole-brain radiotherapy. Ann Neurol 32:551-554, 1992
Jenkin D, Danjoux C, Greenberg M: Subsequent quality of life for children irradiated for a brain tumor before age four years. Med Pediatr Oncol 31:506-511, 1998
Mulhern RK, Heideman RL, Khatib ZA, et al: Quality of survival among children treated for brain stem glioma. Pediatr Neurosurg 20:226-232, 1994
Johnson DL, McCabe MA, Nicholson HS, et al: Quality of long-term survival in young children with medulloblastoma. J Neurosurg 80:1004-1010, 1994
World Health Organization: The First Ten Years of the World Health Organization. Geneva, Switzerland, World Health Organization, 1958
Varni JW, Burwinkle TM, Katz ER, et al: The PedsQL in pediatric cancer: Reliability and validity of the Pediatric Quality of Life Inventory generic core scales, multidimensional fatigue scale, and cancer module. Cancer 94:2090-2106, 2002
Varni JW, Seid M, Kurtin PS: The PedsQL 4.0: Reliability and validity of the Pediatric Quality of Life Inventory Version 4.0 generic core scales in healthy and patient populations. Med Care 39:800-812, 2001
Gurney JG, Smith MA, Bunin GR: CNS and miscellaneous intracranial and intraspinal neoplasms, in Ries LAG, Smith MA, Gurney JG, et al (eds): Cancer Incidence and Survival Among Children and Adolescents: United States SEER Program 1975-1995. Bethesda, MD, National Institute of Cancer, SEER Program, 1999, pp 51-63
Achenback TM, McConaughy SH, Howell CT: Child/adolescent behavioral and emotional problems: Implications of cross-informant correlations for situational specificity. Psychol Bull 101:213-232, 1987
Hudson MM, Mertens AC, Yasui Y, et al: Health status of adult long-term survivors of childhood cancer. JAMA 290:1583-1592, 2003(Sundeep R. Bhat, Tress L.)
Department of Anesthesiology, University of Washington Medical Center, Seattle, WA
Texas A&M University, College Station, TX
ABSTRACT
PURPOSE: The survival of children with CNS tumors approaches 70%, yet health-related quality of life (HRQOL) has not been investigated rigorously in this population. We aimed to show that universal assessment of HRQOL could be obtained easily by using the PedsQL 4.0 and to provide a composite profile of their daily lives.
PATIENTS AND METHODS: The PedsQL was administered to all patients seen in the neuro-oncology clinic at Lucile Packard Children's Hospital (Palo Alto, CA) from December 2001, to September 2002. Patients were compared with healthy controls by using two-sided t tests to evaluate statistically significant differences.
RESULTS: One hundred thirty-four patients (73 male; mean age ± standard deviation, 11.8 ± 5.4 years; 55 had low-grade glioma, 32 had medulloblastoma/primitive neuroectodermal tumor/embryonal tumor, 17 had malignant astrocytoma, nine had germ-cell tumor, and 21 had other types of tumors) were assessed, each in less than 20 minutes. Scores on both child and parent-proxy reports for the total HRQOL, psychosocial, physical, emotional, social, and school-functioning scales were all significantly lower than controls (P < .01). Patients with low-grade glioma were reported to have the highest total HRQOL. Children receiving radiation therapy (XRT) but no chemotherapy had significantly lower total, psychosocial, emotional, and social functioning than those receiving other treatments, including XRT plus chemotherapy.
CONCLUSION: The PedsQL can be used to assess HRQOL rapidly and easily in children with CNS tumors, who have significantly worse HRQOL than healthy children. Children receiving XRT fare worse overall; chemotherapy added to XRT does not seem to worsen HRQOL. Assessment of HRQOL should be included as an outcome in future clinical trials.
INTRODUCTION
Throughout the last decade, the prognosis for children diagnosed with CNS tumors has improved steadily; 5-year survival rates now approach 70%.1 With growing numbers of survivors, increased attention should focus on the quality of life of patients as part of a complete clinical assessment, a variable in making treatment decisions, and an end point for clinical trials.
Past research, although limited, has shown that brain tumors are associated with decreases in quality of life and cognition. The Health Utilities Index survey indicated that 80% of children with brain tumors experience some morbidity, with cognitive deficits affecting the majority.2 Years after diagnosis, these patients have significantly more neurosensory difficulties including increased incidence of late-onset blindness, motor deficits, and chronic seizure disorders.3 As adults, survivors of childhood brain tumors are at increased risk for unemployment, inability to drive, and cognitive, motor, and psychological-emotional impairments.4,5 These patients are more likely than the general population and other cancer survivors to develop psychoses and need psychiatric hospitalization.6 These survivors have also been found to have a lower "health status index" than survivors of other childhood cancers.7 Children diagnosed with medulloblastoma demonstrate significantly lower intelligence quotient (IQ) than patients with low-grade brain tumors.8 Medulloblastoma patients are also likely to never marry or to develop emotional or visual disturbances.4 Whole-brain irradiation before age 7 years correlates with significant reduction in IQ, increases in focal neurological deficits, and need for special education services.9,10
These previous studies have often accrued only small numbers of patients or been restricted to a particular brain tumor pathology.7,11,12 Control groups are inconsistent (often using siblings or other cancer patients), and only the parent or the patient is surveyed.3-6 In addition, these surveys focus on specific aspects of quality of life, such as neurological or cognitive sequelae, rather than health-related quality of life (HRQOL).3,6-8 HRQOL is a multidimensional construct that incorporates both objective and subjective data including, but not limited to, the social, physical, and emotional functioning of the child and, when indicated, his or her family.13 A standard, comprehensive instrument has not been used routinely to assess HRQOL and daily activities simply and quickly in this population.
We sought to demonstrate the feasibility of using the Pediatric Quality of Life Inventory 4.0 and cancer module (PedsQL 4.0)14 to obtain a rapid, complete measurement of HRQOL in a large sample of patients with CNS tumors. Furthermore, we aimed to provide a profile of HRQOL in children with brain tumors, including a composite of their daily life. We hypothesized a priori that variables well established to affect clinical outcomes in childhood brain tumors (age, sex, tumor location, hydrocephalus, pathology, and treatment type) were likely to influence HRQOL.
PATIENTS AND METHODS
All patients 2 years of age followed in the neuro-oncology clinic at Lucile Packard Children's Hospital at Stanford (Palo Alto, CA) for diagnosis of a primary CNS tumor were considered for this study. At Stanford, every child with any brain tumor diagnosis is followed in this clinic at an interval ranging from weekly to annually. From November 2001 to September 2002, all patients and their parents returning for a routine follow-up appointment were consecutively asked to enroll. Patients who were not scheduled for a follow-up appointment during the study period were approached initially by mail and subsequently by telephone. Two patients were surveyed at their homes, and 34 patients were mailed the surveys along with explicit instructions for their completion. Informed consent was obtained from all patients and their parents in accordance with the institutional human subjects review board.
HRQOL was assessed by using the PedsQL 4.0 generic core scales and cancer module. The 23-item generic scales measure physical, emotional, social, and school functioning, and the 27-item cancer module assesses cancer-specific aspects of HRQOL, including treatment and procedural anxiety, cognitive ability, patient worry, physical self-appearance, and communication. Both surveys asked about HRQOL in the past month. Validity of the PedsQL 4.0 was established previously for healthy children and those with acute and chronic conditions by using known-groups comparison and correlations with indicators of morbidity and illness.14,15 Data for healthy children were obtained from a sample of 730 children assessed during well-child checks or by telephone as part of the initial development of the PedsQL 4.0 at the Children's Hospital and Health Center (San Diego, CA).14 Reliability for the PedsQL 4.0 has been demonstrated with Chronbach's coefficients.14 The PedsQL 4.0 cost approximately $7 per patient-parent dyad.
The PedsQL 4.0 was administered as age-appropriate surveys for young children (5 to 7 years of age), children (8 to 12 years of age), and adolescents ( 13 years of age). For children younger than 5 years of age, only a parent completed the PedsQL 4.0. For patients 5 years of age, the PedsQL 4.0 was administered verbally by a research assistant (S.R.B. or T.L.G.) and monitored by the principal investigator (P.G.F.) to ensure that questions were always posed in the same manner. One parent completed a parallel parent-proxy PedsQL 4.0, along with a school and physical-data questionnaire. Both the instrument and the supplemental form were available in English and Spanish. Instructions to those families participating by mail indicated clearly that one parent should complete his or her portion of the survey independently of the child and then read the PedsQL 4.0 to the patient for completion.
To obtain additional data that would provide a comprehensive profile of the daily activities of patients, all parents were asked to complete a school information and physical-data questionnaire. This simple inventory inquired about current school enrollment, special education services, medications, and use of corrective devices (eg, hearing aids or glasses) and also posed questions about patients' abilities to carry out daily activities including dressing, making breakfast, driving, and counting change. Date of birth, sex, date of diagnosis, tumor pathology and location, treatment details, and relapse data were obtained from the Lucile Packard Children's Hospital's neuro-oncology database.
Two-tailed t tests were used to examine differences between patient scores and healthy normals. Differences within patient subgroups were also evaluated by using two-tailed t tests. Pearson correlations were used to assess the association between child self-reports and parent-proxy reports. Analysis of variance was used to elucidate differences by pathology or treatment type. All statistical calculations were conducted by using Statistical Package for the Social Sciences (SPSS) version 10.0 (SSPS Inc, Chicago, IL).
RESULTS
One hundred thirty-four patients (108 patient-parent dyads; 17 parents; nine children) were assessed; their characteristics are listed in Table 1. Twelve children younger than 5 years of age were not surveyed; additionally, five children were low functioning and could not complete the survey. Nine patients came to the clinic unaccompanied by parents or were no longer living with their parents. Only four families refused to take the survey—two for religious reasons and two for personal reasons.
Median age at the time of survey was 11.3 years, whereas the median age at tumor diagnosis had been 7.2 years. Median time from diagnosis was 3.2 years. Tumor pathologic subtypes were similar to those reported for the population.16 Forty-one percent of patients had a low-grade glioma, and 23.9% had medulloblastoma or primitive neuroectodermal tumor. Thirty-one patients had experienced a tumor recurrence or progression before this survey. Prior treatment included surgery (79.9%), radiation therapy (XRT; 53.0%), and chemotherapy (41.8%). Fifty-four percent of patients were white, non-Hispanic; 29.1% were Hispanic; 11.9% were Asian; and 4.5% were African American, similar to controls.14 Sex and age in our sample were not significantly different compared with controls.
Completion of both parent and child surveys required less than 20 minutes in all cases. No difficulties were observed answering the PedsQL 4.0 in children aged 5 to 7 years or those with special education needs, likely because of the instrument's simplicity and brevity. For proxy and self-reports of all age groups, the PedsQL 4.0 generic total and psychosocial Chronbach's coefficients were substantially above 0.70, indicating internal consistency or reliability. For the 5- to 7-year-olds, only the self-report social (0.49) and school-functioning (0.59) scale scores were notably below 0.70, and among 8- to 12-year-olds, scores on the self-report emotional (0.65) and social (0.67) scales were borderline. For children receiving special education services, coefficients exceeded 0.70 for total, psychosocial, physical, and social scale scores and were borderline for the self-report emotional (0.61) and school functioning (0.68) scales. Similarly, for the cancer module, coefficients were generally well above 0.70, except for the 5- to 7-year-old self-reports of pain (0.48), nausea (0.57), treatment (0.52), and perception (0.55). In addition, self-reports of perception in 8- to 12-year-old patients (0.60) and children receiving special education (0.31) were low.
Results from the school and physical-data questionnaire are listed in Table 2. Of the patients sampled, 96.8% were enrolled in school, and 35 children were receiving special education services; 16.4% of the patients were taking medication for headaches, and 9.7% were on an anticonvulsant for seizures. Nine patients wore a hearing aid, and 27.6% of the patients used glasses or contact lenses to correct visual impairments in one or both eyes. Of the 30 patients eligible to drive at the time of survey, 56.7% had a valid license. Ninety-four percent of the patients reported having a best friend. To help demonstrate validity of the PedsQL 4.0 from this questionnaire, comparisons of children on or off medications revealed lower parent and child report of total HRQOL and pain scores (P < .001) when on pain medicines and lower total HRQOL and nausea scores (P < .01) when on antinausea drugs.
There was a significant difference between patient responses and healthy controls for all generic HRQOL subscales (P < .001), as well as a significant difference between parent-proxy reports and normal parent controls (P < .001). Box plots of responses from the study sample and controls are shown in Figure 1. Pearson's correlation coefficients indicated a good reliability between child self-report and parent-proxy report for each subscale (coefficients ranged from 0.344 [communication subscale] to 0.728 [total]; all P < .01). There were no significant differences in PedsQL 4.0 results between children surveyed in the clinic or at home. In addition, none of the measures on the cancer module were significantly different between patients on or off treatment.
Age, time from diagnosis, and sex did not have any significant association with total HRQOL or psychosocial health. However, on subscales, female children reported higher school-functioning scores than male children (P < .05). Among children below the median age at survey, both parent proxies and self-reports revealed greater procedural anxiety (P < .001), and parent-proxy scores showed that children above the median age had worse perceptions of their self-appearance (P < .05). Parent surveys of children beyond the median age at diagnosis revealed that these patients worried about their illness more than younger children (P < .05). Children whose time from diagnosis was above the median expressed more difficulty in communicating about their illness (P < .05), and their parents indicated that these children had lower social functioning (P < .05).
Presence of a shunt was indicative of significantly lower total HRQOL (P < .01 for self-reports; P < .05 for parent reports) and decreased social functioning (P < .05). Patients who had a shunt also indicated lower physical (P < .05) and psychosocial functioning (P < .01).
Tumor location was not significantly associated with total HRQOL for either self-reports or parent-proxy reports. However, children with infratentorial tumors had significantly more problems communicating issues related to their illness than children with supratentorial tumors (P < .05).
Pathology diagnoses were not associated with differences in total HRQOL on child reports (P = .58) but revealed significant differences on parent-proxy reports (P = .02), as displayed in Table 3. Parent-proxy results showed significant overall differences by pathology for the physical (P = .01) and emotional (P = .01) scales. Parents whose children had low-grade gliomas reported higher total HRQOL (P < .05) and better emotional functioning, physical functioning, and communication (all P < .05).
Analysis by treatment type revealed significant differences in total HRQOL for child (P = .03) and parent (P = .001) reports (Table 4). Both children and parents of children receiving XRT without chemotherapy reported lower total HRQOL (P < .05). These patients also had lower perceived physical self-appearance (P < .05). In addition, parent-proxy scores from children who received XRT without chemotherapy demonstrated significantly lower functioning for all subscales except the physical-functioning scale. Patients who received XRT and chemotherapy had higher total HRQOL than those receiving XRT alone (P < .01). Parents of patients who received both XRT and chemotherapy also revealed these children to have better emotional functioning (P < .001) and social functioning (P < .05) than children who received XRT only. Patients who received chemotherapy only had better emotional functioning than those receiving XRT as their sole treatment (P < .05), as indicated by their parents. Children undergoing both XRT and chemotherapy indicated lower physical functioning than those who received surgery alone (P < .05).
DISCUSSION
Assessment of HRQOL in children with brain tumors is feasible and can be accomplished quickly with ease and reliability between patients and their parents. Our study demonstrates that the PedsQL 4.0, with its age-specific nature, is an expedient instrument for investigating HRQOL in this population. Administration of the surveys requires few resources and can be done by nurses or trainees who are already present in the clinic. Rapid assessment is essential to ensure that the attention of children, especially very young patients, can be maintained and to allow for administration of the survey during regularly scheduled clinic visits. Our findings represent an important step toward improved understanding of the lives of children with brain tumors.
Our results indicate that the overall HRQOL of children with brain tumors is highly variable yet significantly lower than normal. These patients experience not only worsened physical health, as expected, but also decreased psychosocial health, emotional functioning, and social functioning. Children with CNS tumors have inferior school functioning and may require special education services. Many patients rely on regular medication for seizures, headaches, or hormonal deficiencies. Only half of those eligible are able to drive. The majority of survivors are able to care for themselves, but many will not be able to become fully independent as adults.
Although some of our findings are predictable, others are more unexpected. Correlation between child reports and parent-proxy reports shows that children and their parents approach HRQOL in a similar fashion; this finding differs from previous studies, which have shown a discord between these two groups.17 Younger patients are more anxious about procedures. Significantly lower HRQOL among children who have a shunt is likely indicative of effects, such as hydrocephalus, that are characteristic of aggressive, larger tumors. Children with brain tumors and their parents recognize the child's impaired school functioning, yet only a minority receive special education services. Children whose age was above the median, namely adolescents, have worse perception of their self-appearance. Among the general childhood cancer-survivor population, older age is associated with functional impairment and limitations in activity.18 Somewhat similarly, our study indicates that as CNS tumor survivors advance further from the time of diagnosis, HRQOL does not improve; in fact, these children demonstrate more difficulty with communication and social functioning, indicating that they may become more socially isolated with time. These findings differ from those observed in other childhood cancer survivors, whose communication and social functioning improve later after diagnosis.14 Regardless, almost all children with brain tumors reported being able to connect with at least one best friend.
Treatment received clearly has a significant impact on HRQOL. Similar to previous observations, our results indicate that children receiving XRT only fare more poorly, with lower physical, social, and school functioning. Patients who received XRT and chemotherapy did not exhibit a lower total HRQOL compared with children who received XRT only. This finding helps reinforce the hypothesis that XRT dosage and fields might be reduced and chemotherapy added without diminishing overall quality of life. Nevertheless, and as expected, children who receive XRT and chemotherapy, with or without surgery, have worse physical functioning than those undergoing surgery only.
Our study does have some limitations. First, it was conducted in a cross-sectional rather than longitudinal fashion. However, the scope of our investigation was to demonstrate the feasibility of HRQOL assessment in this population and to provide an initial composite picture of children with brain tumors. In addition, because this exploratory study undertook many statistical comparisons, differences found, particularly at the .05 significance level, must be interpreted with caution and need to be replicated in future research. Second, although our effort does represent one of the largest patient samples that has been accrued, the number is still small, and its power may account for the lack of some significant differences between sample groups (eg, among patients with differing pathology). Third, heterogeneity in tumor types and variability in therapeutic protocols may confound HRQOL findings. This is a limitation of any study that attempts to survey HRQOL in this diverse population. Finally, results examining internal consistency among particular age groups indicate low reliability in using the PedsQL 4.0. Limitations in reliability results among 5- to 7-year-old patients on both generic and cancer-module subscales indicated that these measures can be used only as secondary, descriptive measures of HRQOL, whereas proxy reports still provide reliable information among this age group. However, for all age groups and those children receiving special education, the total score for self- and proxy reports approached or exceeded 0.90, making the scale relevant for individual assessment. In addition, the PedsQL 4.0 cancer module, a cancer-specific module, did not discriminate subgroups within our brain tumor sample. This could question its validity in this tumor population, although our sample size might have precluded finding different subgroups.
Taken together, our results represent a novel effort toward universal assessment of HRQOL and daily activities among children with brain tumors. Our approach with the PedsQL 4.0 is rapid, inexpensive, and comprehensive. Nevertheless, HRQOL assessment has been conducted sparingly in children with brain tumors until this point, and there are no data that provide a composite of their daily lives. To develop further the understanding of this most important outcome, we suggest that HRQOL measurement be a standard part of all clinical trials in this population as both a baseline assessment of patients commencing treatment and also as an end point. If our approach were adapted prospectively as routine, we might identify specific tumors or therapies that require modifications in management. Future studies should attempt a longitudinal evaluation of HRQOL to provide a better understanding of treatment and ongoing surveillance of children with brain tumors.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Supported by the National Brain Tumor Foundation Pediatric Research Grant.
Presented in part at the 39th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, May 31, 2003.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Jemal A, Murray T, Samuels A, et al: Cancer Statistics, 2003. CA Cancer J Clin 53:5-26, 2003
Barr RD, Simpson T, Whitton A, et al: Health-related quality of life in survivors of tumours of the central nervous system in childhood: a preference-based approach to measurement in a cross-sectional study. Eur J Cancer 35:248-255, 1999
Packer RJ, Gurney JG, Punyko JA, et al: Long-term neurologic and neurosensory sequelae in adult survivors of a childhood brain tumor: Childhood cancer survivor study. J Clin Oncol 21:3255-3261, 2003
Mostow EN, Byrne J, Connelly RR, et al: Quality of life in long-term survivors of CNS tumors of childhood and adolescence. J Clin Oncol 9:592-599, 1991
Lannering B, Marky I, Lundberg A, et al: Long-term sequelae after pediatric brain tumors: Their effect on disability and quality of life. Med Pediatr Oncol 18:304-310, 1990
Ross L, Johansen C, Dalton SO, et al: Psychiatric hospitalizations among survivors of cancer in childhood or adolescence. N Engl J Med 349:650-657, 2003
Foreman NK, Faestel PM, Pearson J, et al: Health status in 52 long-term survivors of pediatric brain tumors. J Neurooncol 41:47-53, 1999
Mulhern RK, Reddick WE, Palmer SJ, et al: Neurocognitive deficits in medulloblastoma survivors and white matter loss. Ann Neurol 46:834-841, 1999
Radcliffe J, Packer RJ, Atkins TE, et al: Three- and four-year cognitive outcome in children with noncortical brain tumors treated with whole-brain radiotherapy. Ann Neurol 32:551-554, 1992
Jenkin D, Danjoux C, Greenberg M: Subsequent quality of life for children irradiated for a brain tumor before age four years. Med Pediatr Oncol 31:506-511, 1998
Mulhern RK, Heideman RL, Khatib ZA, et al: Quality of survival among children treated for brain stem glioma. Pediatr Neurosurg 20:226-232, 1994
Johnson DL, McCabe MA, Nicholson HS, et al: Quality of long-term survival in young children with medulloblastoma. J Neurosurg 80:1004-1010, 1994
World Health Organization: The First Ten Years of the World Health Organization. Geneva, Switzerland, World Health Organization, 1958
Varni JW, Burwinkle TM, Katz ER, et al: The PedsQL in pediatric cancer: Reliability and validity of the Pediatric Quality of Life Inventory generic core scales, multidimensional fatigue scale, and cancer module. Cancer 94:2090-2106, 2002
Varni JW, Seid M, Kurtin PS: The PedsQL 4.0: Reliability and validity of the Pediatric Quality of Life Inventory Version 4.0 generic core scales in healthy and patient populations. Med Care 39:800-812, 2001
Gurney JG, Smith MA, Bunin GR: CNS and miscellaneous intracranial and intraspinal neoplasms, in Ries LAG, Smith MA, Gurney JG, et al (eds): Cancer Incidence and Survival Among Children and Adolescents: United States SEER Program 1975-1995. Bethesda, MD, National Institute of Cancer, SEER Program, 1999, pp 51-63
Achenback TM, McConaughy SH, Howell CT: Child/adolescent behavioral and emotional problems: Implications of cross-informant correlations for situational specificity. Psychol Bull 101:213-232, 1987
Hudson MM, Mertens AC, Yasui Y, et al: Health status of adult long-term survivors of childhood cancer. JAMA 290:1583-1592, 2003(Sundeep R. Bhat, Tress L.)