Supportive Care in Cancer

, 17:1231

Quality of life in survivors of childhood cancer: a systematic review of the literature (2001–2008)

Authors

    • Thames Valley Children’s Centre
  • Miranda Tsonis
    • Elborn CollegeThe University of Western Ontario
Review Article

DOI: 10.1007/s00520-009-0660-0

Cite this article as:
McDougall, J. & Tsonis, M. Support Care Cancer (2009) 17: 1231. doi:10.1007/s00520-009-0660-0

Abstract

Purpose

The purpose of this paper was to provide a comprehensive, contemporary systematic review of studies that have examined the quality of life (QOL) of survivors of childhood cancer in order to extend upon the findings of earlier reviews.

Materials and methods

A review was conducted that used the databases MEDLINE, PubMed, PsycINFO, and CINAHL. Articles were included that were published in English between 2001 and 2008 and used quantitative measures and statistical tests to compare health-related quality of life (HRQL) or QOL of childhood cancer survivors with population norms or matched comparison groups.

Results

Thirteen studies were identified. Findings were contradictory across studies, yet by and large, reflected those of past reviews. In general, survivors’ scores on subscales representing physical, psychological, and social domains of HRQL/QOL were similar to comparisons, with the greatest differences being in physical well-being. Key personal and environmental factors were negatively correlated with the three domains across studies for survivors including: older age at diagnosis, longer time since diagnosis, certain cancer and treatment types, female gender, and a number of socioeconomic factors.

Conclusions

Lack of comparability across studies remains a problem due to wide variation in study focus and designs. Conceptual and methodological issues include: use of numerous HRQOL and QOL measures, lack of distinction between conceptualization and measurement of HRQL and QOL, lack of initial qualitative input from survivors about QOL, little examination of the influence of environmental factors on QOL, little attention to survivors’ satisfaction with life quality, use of small heterogeneous samples, and need for population-based longitudinal studies.

Keywords

Quality of lifeNeoplasmsCancerSurvivorsChildSystematic review

Introduction

Survivors of childhood cancer have become an increasingly important population to study due to improvements in successful treatment outcomes and a corresponding increase in rates of survival [1]. A European study reported survival rates of 75% in 1990, a figure that had increased from 40% in 1970 [2]. Similarly, in 2002, approximately 75% of Canadian children and adolescents with cancer survived beyond 5 years of first diagnosis [3]. By 2008, this number increased to 82% [4]. Also, in the United States, survival rates have been increasing by 1.7% per year for children and 0.9% for adolescents [5]. It is estimated that by 2010, one in every 250 adults between the ages of 20 and 45 in the United States will be a childhood cancer survivor [6].

Although it is encouraging that survival rates have improved in many countries, and are continuing to do so, it is important to realize that survivors of childhood cancer can experience long-term negative consequences related to the disease [7]. Approximately 60% of survivors experience one or more late physical effects [8, 9], which can include cardiopulmonary, renal, endocrine, or pulmonary dysfunction, neurocognitive impairments, the development of secondary cancers, and infertility [4]. Effects can arise as a result of the type of cancer or its treatment [9]. In children, leukemia is the most commonly diagnosed cancer, followed by cancer of the central nervous system (CNS), lymphoma, neuroblastoma, soft tissue cancer, renal tumors, thyroid cancer, melanoma, bone cancer, germ cell and gonadal cancer, retinoblastoma, and hepatic tumors [4].

Survivors who report substantial late physical effects also demonstrate less positive psychosocial outcomes [10]. However, the majority of survivors do not appear to suffer negatively from their illness experience [10]. Past reviews of the health-related quality of life (HRQL) and quality of life (QOL) of survivors of childhood cancer indicate that most survivors are in good physical health and function well psychologically and socially [11, 12]. Similarly, past reviews that focused specifically on the psychological and/or social well-being of childhood cancer survivors [13, 14] have indicated that survivors do not differ in terms of anxiety, depression, or self-esteem when compared with population norms and matched controls [13]. In many cases, survival of cancer can enhance appreciation, and ultimately, QOL [11].

Still, considerable issues remain for survivors of certain types of cancers (e.g., bone tumors and CNS tumors) [1113]. Also, a substantial subgroup of survivors report problems such as pain, fatigue, depression, mood disturbances, tension, anger, confusion, anxiety, lack of friends and involvement in social activities, lower rates of marriage and parenthood, and difficulties obtaining work and health insurance [1114]. In addition, those survivors who report doing well overall still express some concerns about the quality of their lives with respect to certain aspects of physical, psychological, and social well-being [10].

Previous reviews in this area have identified several methodological issues with past studies including: poorly reported medical information (e.g., time of diagnosis, cancer severity) [11, 13, 14], lack of comparability across studies due to varying research designs and measures [1214], the use of heterogeneous samples and proxy reports [11, 13, 14], and a lack of longitudinal work [13, 14]. In order to improve, promote, and maintain the QOL of childhood cancer survivors, it is important to continually review and evaluate new research in this area.

Defining QOL

An important aspect of conducting research in the area of QOL (and a major contributor to the lack of comparability across studies) is to clearly define what this concept represents. This can be problematic as there is no universally accepted definition of QOL [15]. QOL can be thought of and measured in objective terms (e.g., socioeconomic status, presence or absence of illness/disability) or in subjective or perceived terms (i.e., satisfaction with life overall, satisfaction with various aspects of life) [16]. The World Health Organization (WHO) defines QOL as “individuals’ perceptions of their position in life in the context of the culture and value system in which they live and in relation to their goals, standards, and concerns” [17]. Concurringly, Oleson [18] defines QOL as “the subjective perception of happiness or satisfaction with life in domains of importance to the individual.” These definitions reflect a growing sense that QOL may ultimately be best conceived as a subjective phenomenon. Dijkers [15] suggests that, while there may never be complete agreement about how to conceptualize and measure a concept as abstract as QOL, a subjective measurement approach provides a better fit with a client autonomy model where the client is viewed as the expert with respect to his/her own life.

The concept of HRQL emerged in part due to the broadening definition of health from the WHO, which states that health is a “state of complete physical, mental, and social well-being” [19]. At the time, medical interventions did not adopt a broad approach to health, and HRQL emerged as a way of defining QOL in the field of health [20]. Although this development was progressive and extended the aspects of health that were being assessed by the health care field, it has often led to confusion between, and the interchangeable use of, the two concepts of HRQL and QOL.

Anderson and Burckhardt [20], in their review of the conceptualization and measurement of QOL as an outcome variable for health care intervention and research, explain that the concept of QOL has often been confined to, and measured in terms of, health-related concepts, like physical symptoms, functional status, and general health. Both Davis and colleagues [21] and Wallander and colleagues [22] conducted reviews of conceptual and methodological issues related to QOL measurement for children and youth and concluded that HRQL should not be differentiated from a broader notion of QOL. That is, standards for QOL do not depend solely on a child’s health status; rather, physical health is one among other separate but interrelated domains of QOL [23].

The WHO suggests that QOL can be broken down into four core domains: physical, psychological, social relationships, and environmental [24]. Others identify similar domains as core to HRQL. Others extend upon these core domains when describing either HRQL or QOL to include domains such as personal development, spirituality, family, safety/security, and rights [25]. In addition, particular domains have been identified as perhaps more important at certain points in life. For example, autonomy and body image may be especially important in adolescence [26].

For the purpose of this review, the authors take the position that a person’s health status (i.e., presence or absence of disease), other personal factors (e.g., age, sex), and environmental factors (e.g., school or work environment, social support from others) can all impact on three core domains of human functioning or well-being (i.e., physical, psychological [both emotional and cognitive], and social), which, in turn, can provide a picture of a person’s overall QOL. This position is congruent with the WHO’s model of health and disability, as described in the International Classification of Functioning, Disability, and Health [27] and the WHO’s definition of QOL [17].

Purpose

The overall aim of this paper was to provide a comprehensive contemporary systematic review of studies published from 2001 to 2008 that have examined the QOL of survivors of childhood cancer that would extend on previous reviews [1114] in this area that have included studies up until the year 2000. Due to current lack of clarity regarding the difference between the two concepts, studies that stated they were examining either HRQL or QOL were included in the review. The primary purpose of the review was to provide up to date information about this field of study in order to inform families and policy makers regarding the needs of survivors of childhood cancer and to help steer future research directions. Specific questions addressed were: (1) have QOL-related outcomes changed for survivors of childhood cancer in the past decade? and (2) to what extent have study methods in this area, including the measurement of QOL, changed?

Findings of the studies that were reviewed were categorized into three core domains of QOL: physical, psychological, and social well-being; and the personal and environmental factors influencing those domains were discussed.

Materials and methods

The review was conducted using the following databases: MEDLINE, PubMed, PsycINFO, and CINAHL. Both authors conducted separate searches to ensure comprehensive identification of studies. Keywords that were employed included: “quality of life,” “cancer survivor,” and “child.” Articles were chosen on the basis of (1) focus: due to the current interchangeable use of the two terms, studies that were said to specifically focus on either QOL or HRQL in survivors of childhood cancer; (2) year of publication: studies that were published between 2001 and 2008; (3) language: studies that were written in English; (4) method: studies that used standardized measures and conducted statistical tests to compare the results of survivors to population norms or matched comparison groups. Results of the two separate searches were pooled and compared. Identified articles were examined and their reference lists checked for any other relevant studies. Thirteen studies were found to meet the above inclusion criteria.

Results

Participant characteristics and study details

Ten of the 13 studies included participants with several different types of cancer [2837], while two focused exclusively on retinoblastoma [38, 39] and one on neuroblastoma [40]. Six of the studies limited their scope to children and/or adolescents only [29, 30, 32, 33, 38, 40], and seven included adolescents and/or adults [28, 31, 3437, 39]. The studies were conducted in the US, Canada, Austria, Belgium, and the Netherlands. Source of sample and study sample sizes varied widely (from 65 to 7,147). Most studies included comparisons with population norms or samples. Three studies included comparisons with matched groups. Study details are listed in Table 1.
Table 1

Summary of studies included in the review

Study

Sample (n)

Age range

Cancer type

Study origin

Sample origin

Time since diagnosis

Comparison group

Instrument and informant

Results

Blaauwbroek et al. [28]

313

19–50

Leukemia, malignant lymphoma, bone tumor, soft tissue sarcoma, Wilms’ tumor, Langerhans cell histiocytosis, other

Netherlands

Long-term follow-up clinic and past patients of the Department of Pediatric Oncology

m  = 16 years (long-term follow-up group [LF]); m = 26 years (very long-term follow-up groups [VLF])

Age-matched Dutch population sample (n = 772)

RAND-36 Self-report

VFL group had more severe late effects than LF group. Except for bodily pain, LF group did not differ significantly from population sample on RAND-36 subscales. VLF group had significantly lower scores then population sample on physical and social functioning, vitality, and general health. Female gender and late psychosocial effects inversely related to HRQL.

 

De Clercq et al. [29]

67

8–14

Lymphoblastic and acute nonlymphoblastic leukemia, Hodgkin’s disease, NHL, solid tumors

Belgium

Oncology department of university hospital’s database

Free of symptoms for 3  years

Age- and sex-matched comparison group without cancer from two schools in Antwerp (n = 81)

Pediatric Quality of Life Inventory 4.0 (PedsQL) Self-report and parent-report

Survivors reported higher QOL overall and for all subscales except social functioning than comparison group. Correspondence between mothers and children was higher for survivors than comparison group.

 

Grant et al. [30]

84

15–19

Leukemia (ALL, AML, CML), Hodgkin’s disease, Burkitt’s lymphoma, ependymoma, astrocytoma, medulloblastoma, neuroblastoma, Wilms’ tumor, renal carcinoma, malignant bone tumors, soft tissue sarcomas, gonadal neoplasms, unspecified malignant neoplasms

Canada

McMaster Children’s Hospital

m = 8  years

Population norms (n = 129)

Health Utilities Index System (HUI2 HUI3) Self-report

Overall HRQL scores lower for both males and females than general population. Single attribute differences were not statistically significant between leukemia and brain tumor survivors. Females reported greater burden of morbidity in attributes of emotion and cognition.

 

Pemberger et al. [31]

78

18+

Hematological neoplasms, bone and soft tissue sarcoma, CNS tumors, tumors of internal organs

Austria

St. Anna Children’s Hospital

m = 14.7  years

Population norms (n = 484)

SF-36 Health Survey (SF-36) and Nottingham Health Profile (NHP) Self-report

Survivors showed a higher rating of the various areas of QOL than general population. There was 64% agreement between medical records and self-assessment. A small subgroup of survivors who reported severe impairment reported the worst assessment of QOL.

 

Shankar et al. [32]

162

8–13

Leukemia, lymphoma, solid tumors, brain tumors

US

12 U.S. institutions

m = 5.9  years for cancer survivors; m = 1 year for those undergoing treatment

Age-matched US population sample (n = 481)

Manchester Quality of Life—Youth Form (MMQL-YF) Self-report

Compared with general population, survivors reported a higher overall QOL. However, children still undergoing treatment reported lower overall QOL than general population.

 

Speechley et al. [33]

800

6–16

Leukemia, CNS, Wilms’ tumor, neuroblastoma, lymphoma, soft tissue sarcoma, retinoblastoma, germ cell, liver, bone, carcinoma

Canada

12 Canadian pediatric oncology centers and provincial cancer registries

m = 9.5  years

Age- and sex-matched Canadian population sample (n = 923)

Child Health Questionnaire (CHQ) Parent-report

HRQL was lower for both physical and psychosocial health for survivors than general population. Survivors of CNS tumors, lymphoma, and leukemia, and those receiving cranial radiation were reported to have the lowest HRQL.

 

Stam et al. [34]

353

18–30

Leukemia, lymphoma, solid tumor, brain tumor

Netherlands

The Emma Children’s Hospital/Academic Medical Center in Amsterdam

m = 17.2  years

Age- and sex-matched comparison group from general practitioners’ registry lists with no history of cancer (n = 507)

RAND-36 Self-report

Survivors reported lower HRQL than peers. There was a strong positive correlation between predictive coping and HRQL.

 

Zeltzer et al. [35]

7,147

18–54

Leukemia, CNS malignancy, Hodgkin’s disease, NHL, kidney cancer, neuroblastoma, soft tissue sarcoma, malignant bone tumor

US and Canada

Childhood Cancer Survivor Study

m = 23  years

388 siblings and US population norms (no number of persons provided)

SF-36 Health Survey (SF-36) Self-report

Compared with population norms, survivors and their siblings report better psychological adjustment. Survivors have increased risk of poor HRQL in physical and social domains compared with population norms. Survivors of CNS tumors, lymphoma, bone and soft tissue sarcomas have the lowest HRQL. Risk factors for poor HRQL were female gender, lower educational attainment, unmarried status, lower household income, unemployment, lack of medical insurance, having a major medical condition, longer time since diagnosis, and treatment with cranial radiation.

 

Langeveld et al. [36]

400

16–49

Leukemia/NHL without CRT, Leukemia/NHL with CRT, solid tumor, brain, CNS tumor

Netherlands

The Emma Kinderzieke-nhuis Academic Center

m = 16  years

Age-matched comparison group recruited through survivors’ general practitioners (n = 560)

Medical Outcome Study Scale (MOS-24) Self-report

QOL not significantly different from comparison group. Female gender, older age at follow-up, unemployment, fewer years since completion of therapy, severe late effects, and self-esteem explained variation in QOL only to a certain extent.

 

Maunsell et al. [37]

1,334

15–37

Leukemia, lymphoma, CNS tumors, carcinoma, soft tissue cancer, bone cancer, germ cell and other gonadal cancer, kidney, other

Canada

Canadian Childhood Cancer Surveillance Program

Most were diagnosed over 10 years previous

Age- and sex-matched Canadian population sample (n = 1,477)

SF-36 Health Survey (SF-36) Self-report

QOL differences between survivors and general population were small. Three clinical characteristics: having had bone or CNS cancer, more than one treatment series, and two or more dysfunctional organs at treatment end were independently associated with lower QOL in the physical domain. Only those with two or more dysfunctional organs at treatment end had lower QOL in both the physical and psychosocial domains.

 

Van Dijk et al. [38]

65

8–18

Retinoblastoma

Netherlands

Dutch national RB register

m = 11  years

Age- and sex-matched Dutch population sample (exact number not provided)

KIDSCREEN Self-report and parent-report

Survivors reported better HRQL than general population on dimensions of “mood and emotions” and “autonomy.” Survivors with normal visual acuity scored higher on physical well-being than visually impaired survivors. Parents judge HRQL of children to be poorer than do children.

 

Van Dijk et al. [39]

87

18–35

Retinoblastoma

Netherlands

Dutch national RB register

m = 24.4  years

Age-matched Dutch population sample (n = 701)

SF-36 Health Survey (SF-36) Self-report

Survivors have a good but slightly decreased QOL compared to the general population. Survivors report more problems with mental health. Bullying in childhood and greater impairment are main predictors of lower QOL.

 

Gurney et al. [40]

137

8–17

Neuroblastoma

US

Children enrolled previously in Children’s Cancer Group studies

m = 10.7  years

Population norms (exact number not provided)

Pediatric Quality of Life Inventory 4.0 (PedsQL) Self-report

No differences were found between survivors and general population on any domain or overall QOL scores. A clear pattern of poorer self-reported QOL among children with parent-reported academic and psychosocial problems compared to those without such problems.

 

Measures

Nine different generic measures were used to assess HRQL or QOL across the 13 studies. Three measures were used to assess QOL and included the Pediatric Quality of Life Inventory 4.0 (PedsQL) [41], the SF-36 Health Survey (SF-36) [42], and the Medical Outcome Study Scale (MOS-24) [43]. The PedsQL was used in two studies, the MOS-24 in one study, and the SF-36 in two studies to assess QOL. The SF-36 was also used in two more studies to assess HRQL (see Table 1).

The remaining seven measures were used to assess HRQL and included: the KIDSCREEN [44], the Health Utilities Index (HUI) system, specifically the HUI2 and the HUI3 [45], the RAND-36 [46], the Child Health Questionnaire (CHQ) [47], the Minneapolis–Manchester Quality of Life—Youth Form (MMQL-YF) [48], and the Nottingham Health Profile [49]. The RAND-36 was used in two studies. One study used the HUI2 and HUI3 in combination. Another study used the NHP in combination with the SF-36. The CHQ and the MMQL-YF were used in one study each (see Table 1).

Although some instruments were used to measure HRQOL and others to measure QOL (with the SF-36 used to measure both), all included subscales that assessed some aspect of physical, psychological, and social functioning or well-being, with the exception of the HUI2 and the HUI3 that included no subscales for social well-being. A summary of the domains of QOL included in each instrument can be found in Table 2.
Table 2

Instruments used to assess HRQL and QOL in childhood cancer survivors

Instrument

Assessment of HRQL or QOL

Domains covered

Child Health Questionnaire (CHQ)

HRQL

General behavior, mental health, self-esteem, bodily pain, general health, role limitations due to physical health problems, and role limitations due to emotional health problems physical functioning

Health Utilities Index System

HUI2

HRQL

Sensation, mobility, self-care, emotion, cognition, pain, and fertility

 

HUI3

HRQL

Vision, hearing, communication, ambulation, dexterity, emotion, cognition, and pain

KIDSCREEN

HRQL

Physical and psychological well-being, moods and emotions, self-perception, autonomy, parent relations and home life, peer relations and social support, cognitive and school functioning, bullying and social acceptance, and perceived financial opportunities

Minneapolis–Manchester Quality of Life Form—Youth Form (MMQL-YF)

HRQL

Activity, mobility, relationship with family members, moods and feelings, hearing, speech and sight, and outlook on life/family dynamics

Pediatric Quality of Life Inventory 4.0 (PedsQL)

QOL

Physical, emotional, social, and school functioning

RAND-36

HRQL

Physical functioning, social functioning, role limitations due to physical health problems, role limitations due to emotional problems, general mental health, vitality, bodily pain, and general health perceptions

SF-36 Health Survey (SF-36)

QOL and HRQL

Functional health, well-being, and physical and mental health

Nottingham Health Profile (NHP)

HRQL

Energy, pain, role emotional, sleep, social isolation, and physical mobility

Medical Outcome Study Scale (MOS-24)

QOL

Physical functioning, role functioning, social functioning, mental health, bodily pain, general health, and vitality

Overall HRQL/QOL

Four of 13 studies provided an overall summary score for either HRQL or QOL [29, 30, 32, 38]. One study found that the overall HRQL of survivors of childhood cancer was significantly poorer than that of the general population [30]. This study was based on self-reports of a sample of 84 adolescents with many different types of cancer. Another study that included a sample of 65 children and adolescent survivors of neuroblastoma found very few significant differences between survivors and a population-based sample, with overall high ratings of QOL reported for both groups [38]. Parent proxy reports revealed that survivors judged their QOL to be higher than did their parents.

Two studies found significantly higher overall HRQL/QOL among survivors when compared to children who had not experienced childhood cancer [29, 32]. Both of these studies were based on self-reports of medium-sized samples (67 and 162) of children and adolescents with varying cancer types. One of the studies included parent proxy reports that showed high agreement with self-reports [29].

Physical well-being

Differences in physical well-being between survivors and comparison groups ranged from large to none across most studies [28, 30, 3240]. Two studies reported survivors as scoring their physical well-being as better than the general population [29, 31]. Speechley et al. [33] reported that 800 children aged 16 years of age or younger who, on average, had survived cancer for close to 10 years, had significantly poorer HRQL ratings than an age- and sex-matched population sample in the areas of physical functioning, general health, and an overall physical summary score. Effect sizes across subscales ranged from small (physical functioning) to large (general health), with the overall physical summary score indicating a medium effect size. Likewise, using a sample of 353 adolescents and adults aged 18 to 30 years, Stam et al. [34] found that survivors showed worse HRQL than an age- and sex-matched comparison group with respect to physical functioning and an overall physical summary score. However, all significant differences were small.

Maunsell et al. [37] compared 1,334 survivors aged 15 to 37 years to an age- and sex-matched comparison group and found small significant differences with respect to general health and a physical summary score. Zeltzer et al. [35], also using a large population-based sample that included 7,147 cancer survivors aged 18 to 54 years, found that survivors had an increased risk of lower HRQL than the general US population regarding physical functioning, general health, vitality, and an overall physical summary score. But, again, with the exception of vitality (with a large effect size), effect sizes were small. Somewhat conversely, a study by Langeveld et al. [36] reported that, while 400 survivors, aged 16 to 49 years, scored significantly lower for physical functioning than an age-matched comparison group, they scored higher on vitality and general health.

The effects of age at diagnosis and time since diagnosis on physical functioning were investigated in several studies. Blaauwbroek et al. [28] found that 129 long-term survivors (≤20 years) did not differ significantly compared to an age-matched population sample on any physical functioning subscales. On the bodily pain subscale, they even scored significantly higher. However, very long-term survivors (>20 years) (n = 314) had significantly lower scores on the subscales of physical functioning, vitality, and general health. Similarly, Speechley et al. [33] showed that the longer the time since diagnosis, the greater the survivor–comparison group differences and effect sizes with respect to physical functioning. Stam et al. [34], using a sample of 353 adolescents and adults aged 18 to 30 years, reported that the older a survivor at diagnosis, the worse the mean score on the physical component of HRQL. Zelzter et al. [35] found that both older age at diagnosis and longer time since diagnosis were associated with poorer physical function, bodily pain, general health, and an overall physical summary score. In contrast, Shankar et al. [32] did not find older age or time since diagnosis to significantly predict poorer physical functioning in 162 survivors, aged 8 to 13 years.

The relationship between type of cancer that children were originally diagnosed with and physical well-being was examined in a number of studies. Speechley et al. [33] reported that seven survivor diagnostic groups had considerably poorer mean physical summary scores compared to an age- and sex-matched population sample with medium to large effect sizes: CNS tumors, bone tumors, lymphoma, leukemia, soft tissue sarcoma, liver tumor, and Wilm’s tumor. Zeltzer et al. [35] also found that survivors of CNS tumors, lymphoma, bone tumors, and soft tissue sarcoma had the lowest mean HRQL scores in most domains pertaining to physical functioning. In addition, although Maunsell et al. [37] found small differences between 1,334 adolescent and adult survivors and an age- and sex-matched population sample on physical functioning subscales and overall physical functioning, it was noted that those persons who had survived CNS tumors or bone cancer had the lowest QOL scores on overall physical functioning domain. Similarly, Blauuwbroek et al. showed [28] no significant differences between diagnoses except that bone tumor patients scored significantly lower on a physical functioning subscale. Stam et al. [34] reported that diagnosis did not contribute to physical health summary scores.

In addition to age at, and time of, diagnosis and type of cancer, treatment type was indicated to have some effect on later physical well-being of cancer survivors. In the Blaauwbroek et al. [28] study, those survivors aged 19 to 50 who were treated with both chemotherapy and radiation therapy had more severe late physical effects than those treated with chemotherapy alone. Also, leukemia patients who were treated with cranial radiation had more severe late physical effects than those who did not receive this type of therapy. Speechley et al. [33] observed that adjusted survivor–comparison group differences were largest among survivors treated with a combination of radiation, surgery, and chemotherapy, followed by those with radiation and surgery, and radiation only. Effect sizes for all treatment groups were large. In addition, a clinically meaningful survivor–comparison group difference for the physical domain of HRQL was found when survivors were examined by number of treatment series. Effect sizes moved from medium to large for those who had at least two treatment series. Maunsell et al. [37] found that, the greater number of treatment series, the poorer the scores for general health, bodily pain, and physical function.

Age at diagnosis, time since diagnosis, and type of treatment all appear to contribute simultaneously to later physical effects for cancer survivors. Zeltzer et al. [35] found that, after adjusting for gender, age at diagnosis and survival time and treatments with surgery, chemotherapy, and cranial radiation therapy were each associated with poorer HRQL in the physical domains.

In a number of studies, gender was investigated in relation to survivor’s physical well-being. Grant et al. [30] reported that females had significantly lower mean utility scores (i.e., greater pain) than males related to pain. Similarly, female cancer survivors were five times more likely to report poor HRQL in the physical functioning domains than males in a study by Shanker et al. [32]. A study by Stam et al. [34] showed that the mean HRQL was worse for women than men regarding a physical summary score. Langeveld et al. [36] observed that female gender was significantly associated with lower physical functioning and vitality and more bodily pain. Speechley et al. [33] found that the adjusted survivor–comparison group difference in the mean HRQL physical summary score was slightly larger among females than males. However, when individual domains were examined, survivor–comparison group differences indicated only minimal variations by gender. Pemberger et al. [31] found no significant gender differences when survivors were divided into four groups by perceived health status and extent of cancer and treatment sequelae.

Zeltzer et al. [35] also identified a number of other personal and environmental factors in addition to gender that contributed to poorer physical HRQL for survivors of childhood cancer including: older age, unemployment, lower annual household income, lower educational attainment, being unmarried, being Hispanic, and lack of health insurance. Langeveld et al. [36] also identified lower educational attainment and unemployment as contributing to poorer physical HRQL. In addition, van Dijk et al. [38] found that having been bullied in school was related to poorer physical functioning for survivors of retinoblastoma.

Psychological well-being

Most studies found small to no significant differences between survivors and comparison groups with respect to psychological well-being [28, 30, 33, 34, 36, 37, 39, 40]. Five studies reported higher psychological well-being for survivors [29, 31, 32, 35, 38]. The Speechley et al. study [33] showed a small effect size for 800 children aged 16 years of age or younger with respect to mental health and a psychosocial summary score. Stam et al. [34] found that, among a sample of 353 adolescents and adults aged 18 to 30 years, none of the subscale scores related to psychological functioning or a mental health summary component was significantly different from those of an age- and sex-matched comparison group. Likewise, Maunsell et al. [37], who compared 1,334 survivors aged 15 to 37 years to an age- and sex-matched comparison group, found no significant differences with respect to psychological functioning on any subscales or on an overall psychological summary score.

Zeltzer et al. [35], using a sample of 7,147 cancer survivors aged 18 to 54 years, found that survivors had higher scores than the general US population on a mental health subscale and did not differ from the norm on a mental health component summary. Zeltzer et al. [35] used another measure to more closely examine mental health and found that depression, anxiety, somatization, and global distress scores were all significantly better for survivors than the general population. Pemberger et al. [31] also explored mental health more closely and had similar findings. Using smaller samples, De Clercq [29], Shankar et al. [32], and van Dijk et al. [38] all found that survivors had higher emotional functioning than general population samples.

Most studies that examined the relationships between age at diagnosis and time since diagnosis and psychological functioning had no significant findings to report. Blaauwbroek et al. [28] did not find either long-term or very long-term survivors to significantly differ in terms of psychological functioning from a population sample. Similarly, Zeltzer et al. [35] found no significant relationships between either age at diagnosis or time since diagnosis and any of the psychological subscales for HRQL. However, Shanker et al. [32] observed that survivors who were more than 6 years from diagnosis at the time of the study were more likely to report higher psychological functioning.

In terms of the type of cancer survivors had overcome, Shanker et al. [32] found that survivors of non-neurologic solid tumors were more likely to report higher psychological functioning compared to an age-matched population sample. Speechley et al. [33] showed that means scores in mental health for survivors of CNS tumors, neuroblastoma, and leukemia were significantly lower than an age- and sex-matched population sample. Zeltzer et al. [35] found that survivors of CNS tumors, lymphoma, bone tumors, and soft tissue sarcoma had the lowest mean HRQL scores in the domain of mental health, but not an overall mental health component summary. Maunsell et al. [37] also showed that CNS survivors reported lower scores in the psychosocial domains, but effects sizes were small. van Dijk et al. [39] reported clinically significant mental health problems for retinoblastoma survivors, who experienced anxiety, depression, and feelings of loss of control. Both Pemberger et al. [31] and Stam et al. [34] reported that diagnosis did not contribute to mental health index scores.

Type of treatment was also examined with respect to its impact on psychological well-being. In the study of Speechley et al. [33], the largest mean score differences between survivors and a population sample were, for survivors treated with radiation only, followed by those treated with radiation and surgery. Effect sizes for both treatment groups were large. Maunsell et al. [37] did not find significant relationships between number of treatment series and mental health.

With respect to gender, similar results were found for psychological well-being as with physical well-being. Langeveld et al. [36] observed that female gender was significantly associated with lower mental health. Females were more likely to experience poorer emotional and cognitive functioning than males in the study of Grant et al. [30]. In addition to female gender, Zeltzer et al. [35] found that lower educational attainment, unmarried status, lower annual income, and having a major medical condition were all associated with a higher risk of psychological distress for survivors.

Social well-being

Most studies found small to no significant differences between survivors and comparison groups in the area of social well-being [28, 29, 3137, 39, 40]. Only one study of survivors of retinoblastoma aged 8 to 11 years old [38] reported significantly better “parent relations and home life” for these children than an age-matched comparison group.

Speechley et al. [33], Stam et al. [34], Maunsell et al. [37], and Zeltzer et al. [35] all reported that survivors had significantly poorer ratings than population samples or matched comparison groups with respect to role limitations due to physical functioning. The studies of Speechley et al. [33] and Zeltzer et al. [35] also reported significantly poorer ratings for survivors regarding role limitations due to emotional functioning. Three studies [28, 34, 35] showed that survivors’ scores on subscales representing social well-being were significantly poorer for survivors than comparisons. However, effect sizes/clinically significant differences in most studies were small.

Zelzter et al. [35] found that both older age at diagnosis and longer time since diagnosis were associated with greater role limitations due to physical functioning. Similarly, Blaauwbroek et al. [28] reported that long-term survivors had significantly lower scores on a social functioning subscale than the general population. Other studies had similar findings with respect to years of survival and social well-being [36, 37].

Type of cancer was found to have some relationship to social well-being in the study of Zeltzer et al. [35]. Survivors of CNS tumors, lymphoma, soft tissue, or bone malignancies, Hodgkin’s disease, and non-Hodgkin’s lymphoma (NHL) all reported more problems in social function and role limitations due to physical problems than the general population [35]. Stam et al. [34] reported that survivors who had been treated otherwise than with surgery alone had better mental health scores than those treated with surgery alone. Maunsell et al. [37] found that, the greater the number of treatment series, the poorer the scores for role limitations due to physical functioning.

With respect to gender, Gurney et al. [40] observed that male survivors had poorer scores on both a social functioning and school functioning subscale than female survivors. However, in the study of Zelzter et al. [35], female gender, older age, lower educational attainment, unemployment, lower annual income, and lack of health insurance were all associated with poorer social well-being. Negative attitudes of, and relationships with, peers within the school environment also appear to have a significant effect on survivors’ social well-being. van Dijk [39] reported that having experienced bullying at school was an independent predictor of role limitations due to physical functioning, role limitations due to emotional functioning, and social functioning for survivors of neuroblastoma.

Interrelationships among physical, psychological, and social well-being

None of the studies provided correlations between the various subscales measuring aspects of QOL. However, the study by Blaauwbroek et al. [28] provided some indication that the various components of QOL impact each other. Findings indicated that physical (i.e., gastrointestinal) late effects were negatively related to the subscales physical functioning, vitality, general health, and social functioning. In addition, psychosocial late effects were negatively related to social functioning, mental health, vitality, and bodily pain.

Conclusions and future research directions

The purpose of this review was to provide an overview of the current literature from 2001 to 2008 written in English with respect to QOL for survivors of childhood cancer. Studies were included that used standardized measures and statistical tests to compare the findings of survivors to population norms or matched comparison groups. Specific questions addressed were: (1) have QOL-related outcomes changed for survivors of childhood cancer in the past decade? and (2) to what extent have study methods in this area, including the measurement of QOL, changed?

QOL-related outcomes

Like past reviews [1114], this study found inconsistent and contradictory findings with respect to overall QOL and the three primary domains of QOL for survivors of childhood cancer. Studies in this review variously reported lower, higher, or similar ratings for overall QOL and for physical, psychological, and social well-being in survivors when compared with population norms and matched comparison groups. However, despite inconsistencies, like past reviews, the studies, as a whole, indicated that most survivors were doing well overall and in the three primary domains of QOL. The domain for which survivors were fairing significantly poorer than comparisons was physical well-being, with population-based studies reporting moderate to large effect sizes for the subscales of general health [33] and vitality [35] and for an overall physical functioning summary score [33]. Surprisingly, a number of studies [29, 31, 32, 35, 38] reported survivors as having significantly better psychological functioning than comparisons, including a population-based study that reported a large effect size for a mental health subscale [35]. Past reviews also included studies with findings of this nature [1214]. However, the consistency and strength of the findings in this review appear to provide even stronger evidence of this phenomenon.

There are a number of explanations for why survivors of childhood cancer might report higher overall QOL or better psychological well-being than individuals who have not had childhood cancer. One suggestion is that survivors who rate their QOL as equal or higher than their peers are demonstrating self-deception response bias, which is a tendency to deny problems [50]. This would cause survivors to exaggerate the positive aspects of their life and understate the negative aspects [51]. It is apparent that not all survivors display self-deception response bias [50], which would also account for why some report high and some poor QOL. A second explanation suggests that survivors perceive their life as particularly valuable due to the difficulties they have overcome, and as a result, impairments have a lower impact on their perceived QOL [52]. Along the same lines, van Dijk et al. [38] suggest that survivors experience shifting values, a greater appreciation for life, and an increased ability to cope with issues due to the illness experience. Survivors have also been found to have high adaptability [53]. Clearly, further research is required to examine the plausibility of the various explanations for why cancer survivors would report better psychological functioning than comparisons [38].

Consistent with the findings of past reviews, this review identified a number of correlates of QOL for survivors of childhood cancer, including older age at diagnosis, greater time since diagnosis, certain types of cancer (i.e., CNS tumors, lymphoma, bone tumors, and soft tissue sarcoma), aggressive treatment types and length of treatment series, and female gender. Some authors suggest that the tendency for female survivors to report poorer QOL difference is due to the tendency for females to discuss problems more openly than males [38]. Still, others point out that the sex differences observed in survivors is not an effect specific to having had childhood cancer and indeed are reflective of findings for the general population [37]. Researchers have interpreted this as providing further evidence that most survivors do indeed resemble the general population with respect to QOL [37].

Two studies [35, 36] using large samples identified a number of socioeconomic factors as being correlated with the three domains of QOL, with lower educational attainment and lower household income being correlated with all three domains. One study [39] using a smaller sample examined the relationships between environmental factors of an interpersonal or relational nature and QOL and found that having experienced bullying at school was significantly correlated with poorer physical and social well-being for adult survivors of childhood cancer. These findings emphasize the importance of not only identifying disease-related factors and socioeconomic factors, like age, sex, and income level, but of also identifying other personal and environmental factors, like self-esteem, supportive care needs, and societal attitudes, that may be impacting the lives of childhood cancer survivors. Findings also support the position that multidisciplinary, holistic models of follow-up care are needed to address the array of factors that may affect various aspects of QOL for survivors of childhood cancer [26].

Study methods

The findings of this review reflect those of earlier reviews that included studies conducted up to 2000 [1114]. Like past reviews, this review identified a wide variety of studies. Studies varied substantially regarding comparison groups selected and characteristics of the study samples (e.g., type of cancer and treatment, age of, and time since, diagnosis), making comparability across studies difficult. Studies were also characterized by a large degree of heterogeneity regarding their focus or primary research question, sample size, and the HRQL/QOL measure used. Some of the studies with smaller sample sizes focused on very specific topics, like investigating coping abilities [34], self-esteem [36], or academic problems [40] in relation to QOL for cancer survivors. Larger, population-based studies tended to focus on identifying correlates of QOL for survivors [33, 35, 36]. Certainly, there is a need for both types of studies. What may be most useful is for more focused studies with smaller samples to use homogeneous samples and study more closely, for example, why older survivors or survivors with certain types of cancer or treatments experience poorer QOL. Only three of the studies with smaller samples included in this review focused on survivors with a specific diagnosis [3840].

This review focused on studies that were quantitative in nature. Qualitative research examining QOL in childhood cancer survivors has begun to emerge and offers another important approach for understanding this population [54, 55]. The type of data collected using qualitative research can highlight areas not included in standardized measures [13]. One qualitative study found that survivors exhibited thriving due to their illness experience, which is a form of living fully, consciously, and richly [56]. Thriving is a complex process and results from deep changes in the way survivors understand themselves [56]. Another qualitative study reported difficulties for survivors, including uncertainty surrounding secondary cancers and feelings of being different than others [57]. Despite these issues, the survivors stated plans for moving on and creating a future [57]. It was also noted in this study that survivorship is unique and depends on many variables [57].

These results show a unique perspective that can arise from qualitative research and that more studies of this nature would be valuable for understanding the QOL of survivors of childhood cancer. Indeed, these types of studies could be useful in helping to explain the paradox of why survivors often report higher psychological well-being than the general population. In addition, qualitative studies can help in the formulation of theories and models that are grounded in the experiences of survivors of childhood cancer [54]. Such theories and models could not only help guide decisions about which data to collect but also how to interpret data [54].

Population-based cross-sectional studies that allow comparisons across diagnoses, smaller focused studies using homogeneous samples, and qualitative studies can all be useful for identifying the types of variables to include in population-based, longitudinal studies, which are required to examine the long-term impact of having survived cancer on individuals’ QOL. For example, as mentioned, one smaller quantitative study [39] identified bullying in school as an important correlate of QOL. It may be highly beneficial to include parsimonious measures of this and other interpersonal level environmental factors, such as social support and family functioning in population-based, longitudinal studies to be able to examine their direct and indirect impact on survivors’ QOL [58].

In addition to different foci and the use of different research designs and heterogeneous samples, inconsistencies and contradictions among study findings with respect to survivors’ QOL may also be due to the numerous measures used to assess QOL. An abundance of both generic and disease-specific QOL measures are available to researchers. Indeed, just the increased number of pediatric cancer-related QOL instruments alone resulted in a recommendation for a moratorium in 1999 on developing new instruments [54]. Despite the plethora of measures available, only recently has research using QOL instruments been published [58]. Prior to 1995, studies in the area of cancer survivorship tended to focus on single attributes of QOL, such as emotional or social well-being [58]. It was encouraging that almost all of the instruments used in studies included in the present review contained subscales that measured at least the three primary domains of QOL, which allowed gross comparisons across studies in these domains. In addition, the measures used in the studies meet appropriate psychometric requirements.

However, the interchangeable use of the concepts of HRQL and QOL to define available measures is one problem that has not been helpful in moving the field of QOL research forward. When looking at the instruments used in the studies included in this review, it became evident that there was no difference in the constructs being measured regardless of whether researchers stated they were examining HRQL or QOL. Further conceptual work is required to clarify the distinction, if indeed one should exist, between HRQL and QOL. It is hoped that, as more research is done in this area, the most appropriate and psychometrically sound measures that include the most credible constructs of QOL will be identified and consistently used in order to allow greater comparability across studies.

It was encouraging to note that all but one of the studies included in the review, even most that included younger children, used self-report measures. To properly assess and study QOL, it is essential to understand the patient’s point of view by obtaining data from their individual perspectives [59]. Studies included in this review showed that using parents as proxy reporters for their children sometimes results in different interpretations of the child’s experience. Other reviews in this area have had similar findings. For example, Calamus and Kiebert [12] observed that parent proxies tended to report fewer problems in psychological well-being, whereas they reported more problems in physical well-being. Research has indicated that children and youth can self-report on their health and well-being [60] and that it is important for them to do so because the subjective feelings of children and youth about themselves and their life quality can only be known to them [61]. However, obtaining the perspectives of both children and parents is important for making decisions related to interventions [62] since each is likely to have unique knowledge about, and place different values on, various life dimensions [63].

Another point worth making is that, although QOL data was primarily collected from the point of view of the survivor, data reflecting satisfaction with overall life quality was rarely collected as part of any of the QOL measures used or through the use of a separate measure. It seems that it would be important for future quantitative studies to collect information about satisfaction with overall life quality along with data about various aspects of well-being to be able to compare the relationships among these constructs [64]. For example, a survivor may report their physical functioning as poor, but at the same time, may be quite satisfied with life in general.

Recommendations for future research

To summarize, as past reviews have also concluded, on average, survivors of childhood cancer appear to enjoy a QOL that is comparable to the general population. However, certain subgroups appear to experience very poor well-being in one or more primary domains of QOL and some survivors report surprisingly high psychological well-being. To fully understand the QOL for all survivors, it is recommended that:
  • Qualitative research should be conducted to advance theory and model development, to identify the variables that are most important to QOL from the perspective of survivors, and to help understand why survivors may be enjoy better QOL than those who have not experienced childhood cancer.

  • Quantitative studies should employ rigorous study methods and conceptually sound, reliable, and validated measures of QOL, involve multiple informants, and collect and report important information (e.g., respondent rate, time since diagnosis, effect sizes). Such efforts will help inform clinical practice, aid in the interpretation of findings and replication of studies, and help advance this area of research.

  • Focused quantitative studies using homogeneous samples should be conducted to provide insight into specific questions pertaining the QOL for certain subgroups of survivors (e.g., what factors are related to poorer physical well-being for survivors with certain types of cancers or treatments?).

  • Population-based or multisite cross-sectional studies that include important socioeconomic, personal, and environmental variables related to QOL for survivors should continue to be carried out to identify the primary correlates of QOL and to be able to compare reports of QOL across various diagnoses. International collaboration is recommended to be able to make cultural comparisons across studies.

  • In addition, it would be beneficial for such studies to collect information about survivors’ perceptions of satisfaction with their overall life quality to be able to study the relationships among life satisfaction, important socioeconomic, personal, and environmental variables, and physical, psychological, and social well-being.

  • Finally, large, population-based or multisite longitudinal studies using advanced statistical techniques that can examine both group and individual differences within data will need to be conducted to be able to identify the most important predictors and mediators of changes in important QOL domains, overall QOL, and/or satisfaction with life quality for survivors of childhood cancer over time and to examine interrelationships among these various constructs over time.

Copyright information

© Springer-Verlag 2009