Fax: (612) 626-2815
Education, employment, insurance, and marital status among 694 survivors of pediatric lower extremity bone tumors
A report from the childhood cancer survivor study
Version of Record online: 30 APR 2003
Copyright © 2003 American Cancer Society
Volume 97, Issue 10, pages 2554–2564, 15 May 2003
How to Cite
Nagarajan, R., Neglia, J. P., Clohisy, D. R., Yasui, Y., Greenberg, M., Hudson, M., Zevon, M. A., Tersak, J. M., Ablin, A. and Robison, L. L. (2003), Education, employment, insurance, and marital status among 694 survivors of pediatric lower extremity bone tumors. Cancer, 97: 2554–2564. doi: 10.1002/cncr.11363
- Issue online: 30 APR 2003
- Version of Record online: 30 APR 2003
- Manuscript Accepted: 21 JAN 2003
- Manuscript Revised: 6 JAN 2003
- Manuscript Received: 8 NOV 2002
- National Childhood Cancer Fund Research Fellowship (Arcadia, CA)
- American Society of Clinical Oncology Young Investigator Award (Alexandria, VA)
- Children's Cancer Research Fund (Minneapolis, MN)
- National Institutes of Health (Bethesda, MD). Grant Numbers: U24- CA 55727, T32-CA09607
- late effects;
- pediatric bone tumors;
- Ewing sarcoma
With increasing numbers of childhood cancer survivors, direct sequelae of cancer therapy and psychosocial outcomes are becoming more important. The authors described psychosocial outcomes (education, employment, health insurance, and marriage) for survivors of pediatric lower extremity bone tumors.
The long-term follow-up study of the Childhood Cancer Survivor Study is a multiinstitutional cohort study comprising 14,054 individuals who have survived for 5 or more years after treatment for cancer diagnosed during childhood or adolescence. Baseline demographic and medical information were obtained. Six hundred ninety-four survivors had osteosarcoma or Ewing sarcoma of the lower extremity or pelvis and were classified by amputation status and by age at diagnosis. The median age at diagnosis was 14 years old with a median of 16 years of follow up since diagnosis. Demographic characteristics were used to analyze the rates of psychosocial outcomes.
Amputation status and age at diagnosis did not significantly influence any of the measured psychosocial outcomes. Education was a significant positive predictor of employment, having health insurance, and being currently in their first marriage. Male gender predicted ever being employed and female gender predicted having health insurance and marriage. When compared with siblings, amputees had significant deficits in education, employment, and health insurance.
Overall, no differences between amputees and nonamputees were found. However, gender and education play a prominent role. When compared with siblings, amputees in this cohort may benefit from additional supports. Cancer 2003;10:2554–64. © 2003 American Cancer Society.
Tremendous progress in the treatment of childhood cancer has been made over the last four decades, resulting in increasing numbers of childhood cancer survivors. Clinical research to assess the long-term outcomes, including the medical sequelae and psychosocial outcomes, of these survivors who are now in or entering adulthood has become increasingly important. Assessing the medical sequelae of cancer therapy1–4 is important as it permits timely anticipatory screening, corrective or preventive intervention, and potential modification of future cancer therapies. The assessment of long-term psychosocial outcomes also provides the opportunity to intervene with appropriate supportive services and counseling for those experiencing difficulties in psychosocial functioning.
Medical sequelae are correlated with the type, intensity, and schedule of antineoplastic therapy and host predisposition.5–11 Conversely, psychosocial functioning reflects the complex interaction between cancer treatments, social supports, and the child's emotional and physical maturity at the time of diagnosis and treatment. Accordingly, the literature is more extensive for major health-related outcomes compared with psychosocial status. Nonetheless, the potential impact of psychosocial function (education, employment, health insurance, and marriage) on the overall quality of life among childhood cancer survivors can be profound and also warrants investigation.
Adult survivors of pediatric lower extremity bone tumors represent a specific subgroup of childhood cancer survivors who encounter considerable risks for cancer-related sequelae. Children with bone tumors are not only exposed to significant amounts of chemotherapy, but often also require extensive surgery consisting of amputation or limb-sparing procedures and/or high doses of radiotherapy. Survivors of bone tumors of the lower extremity, the most common site of tumor presentation,12 experience physical consequences of treatment that can be quite noticeable (limb loss or a limp that affects functional status and limits activity) and adversely affect self-image. The potential impact of these treatments is influenced by the emotional and skeletal maturity of the child at the time of diagnosis and treatment.
Young children and adolescents have a wide-ranging and evolving level of maturity that can influence the ability to accept the loss of a limb due to cancer. One study13 reported that adolescents with limb-sparing procedures had more difficulty with emotional rehabilitation compared with amputees who had a quicker acceptance of limb loss. Another study14 revealed that psychosocial adjustment was affected by the age of the patient at diagnosis, as opposed to the type of surgery (amputation or limb-sparing surgery) or resultant physical function. Skeletal maturity is another potential determinant of psychosocial sequelae because of its importance in determining the type of local control surgery (amputation/rotationplasty/expanding prosthesis vs. limb-sparing surgeries) that can alter one's body image and impair self-reliance. Other studies have validated the importance of self-consciousness and body image in adolescents who received multimodal treatment for childhood cancers.15–18 Only very few large studies19, 20 have attempted to describe and analyze psychosocial functioning of survivors of pediatric bone tumors and have shown conflicting results.
The issue of maturity at the time of diagnosis and treatment is important in future anticipatory guidance for these patients and their families. With the establishment of the Childhood Cancer Survivor Study (CCSS),21 a retrospective cohort study initiated to investigate late effects among long-term survivors of childhood cancer, we have the opportunity to explore selected aspects of psychosocial functioning (education, employment, health insurance, and marriage) of a large number of adult survivors of pediatric lower extremity bone tumors. The current investigation focuses on four specific subgroups defined by age at diagnosis and extent of surgery.
MATERIALS AND METHODS
Subject Selection and Contact
The long-term follow-up study of the CCSS is a multiinstitutional study of individuals who have survived for 5 or more years from the diagnosis of cancer, leukemia, tumors, or similar illnesses diagnosed in childhood or adolescence. The methods of the study and cohort characteristics have been described previously.21 Briefly, as of November 1, 2000, 20,276 participants met the cohort inclusion criteria which included 1) diagnosis of one of the following cancers: brain tumor, leukemia, Hodgkin disease, non-Hodgkin lymphoma, cancer of the kidney, neuroblastoma, soft tissue sarcoma, and cancer of the bone; 2) initial treatment at one of the 25 collaborating CCSS institutions; 3) diagnosis date between January 1, 1970 and December 31, 1986; 4) age younger than 21 years at diagnosis; and 5) survival 5 years from diagnosis. This cohort of patients included 1726 survivors of bone tumors. After excluding 328 who refused participation, 219 who were lost to follow-up, 130 who had died before the baseline questionnaire was administered, and 5 who are still pending, 1044 survivors remained. The current study is restricted to a subset of 694 participants who fulfilled the following additional criteria: 1) diagnosis of osteosarcoma or Ewing sarcoma (ICD-O codes: 9180.3, 9181.3, 9182, 9183.3, 9185.3, 9190.3, and 9260.3); 2) tumor located in a lower extremity or the pelvis (ICD-O site codes of c40.2 and c41.4); 3) at least 18 years of age at the time of the baseline questionnaire; and 4) questionnaire completion by the survivor himself/herself.
The CCSS protocol and contact documents were reviewed and approved by the human subjects committee at each participating institution. Baseline data were collected for members of the study cohort, using a 24-page questionnaire (available from URL: www.cancer.umn.edu/ccss). The baseline questionnaire was designed to acquire demographic characteristics, education, income, employment, insurance coverage, marital status, health habits (i.e., smoking, alcohol consumption, physical activity, self-screening examinations), family history, access and utilization of medical care, medication use, frequency of diagnosed medical conditions (i.e., hearing/vision/speech, urinary, hormonal, cardiovascular, respiratory, digestive, brain, and central nervous system disorders), surgical procedures, recurrent cancer, subsequent new neoplasm, and offspring/pregnancy history. If the patient had survived for 5 years but died subsequently, selected information was obtained from a family member, usually a parent.
Cancer Treatment Information
Information on the characteristics of the original cancer diagnosis was obtained on all eligible cases from the treating institution. For all CCSS participants who returned a signed medical record release, information concerning primary cancer therapy was collected, including initial treatment, treatment of disease recurrence, if any, and preparatory regimens for bone marrow transplantation (if applicable) until the time of last contact. Qualitative information was abstracted from the medical record for 42 specific chemotherapeutic agents, for which quantitative information was abstracted for 22. Data were also obtained on tumor site, radiation field, and any surgeries performed from the time of diagnosis. Copies of the abstraction form used in data collection are available for review and downloading from URL: www.cancer.umn.edu/ccss.
Survivors were grouped into one of two categories according to the type of surgical control procedure, i.e., those treated with and without amputation. This information was obtained from the medical record abstraction (ICD-9 codes 8410–8419) and supplemented by the baseline questionnaire. For the current study, radiation exposure was treated as a dichotomous yes/no variable. Twenty-four survivors who did not receive an amputation were excluded because of the lack of a complete medical record abstraction. We were unable to confirm the local control procedure or to determine whether radiation was given. Site codes from the abstraction forms were the primary source for classifying the tumor site. The original medical record abstraction form was reviewed and/or treating institutions were contacted if there was uncertainty in the exact location.
During the collection of baseline data, information concerning siblings was obtained from the cancer survivors. For a random sample of participating survivors, full biologic siblings closest in age to the survivors were invited to participate in the study. If the siblings agreed, they completed an age-specific baseline questionnaire similar to that completed by the survivors. A total of 5857 siblings were selected randomly for participation, 5800 of whom were contacted successfully. Of this number, 3418 siblings remained after 472 refused participation, 1800 were pending completion, and 110 questionnaires were being processed. Of these, 2667 were older than 18 years at the time the baseline questionnaire was administered and were eligible for analysis.
The survivors were grouped by age at diagnosis of childhood cancer (12 years or younger and older than 12 years) and by the type of surgery (amputation or limb-sparing surgery), either jointly or separately by the two variables. The age cutoff of 12 years old was chosen because it represents an approximate marker for emotional maturity (preadolescence and adolescence), as well as the age at which significant bone growth is completed. We assessed psychosocial functioning as dichotomous variables (completed high school, completed college, ever had a job, employment during the past year, health insurance coverage, ever had difficulty obtaining health insurance, ever married, and still in first marriage) with demographic and clinical characteristics of the survivors, including the four group age-surgery variables (≤ 12/Amp: those younger than 12 years at diagnosis with an amputation; > 12/Amp: those older than 12 years at diagnosis with an amputation; ≤ 12/LS: those younger than 12 years with limb-sparing surgery; and >12/LS: those older than12 years with limb- sparing surgery). The > 12/LS group was used as the reference group. The associations were quantified by odds ratios (OR) using logistic regression models, first adjusting for current age and time since diagnosis only (age-adjusted analysis) and then adjusting for additional variables (multivariable analysis), with 95% confidence intervals calculated by the Wald method. Multivariable analyses included adjustment for demographic characteristics (age at questionnaire completion and years since diagnosis), tumor type and location, radiotherapy exposure, and other existing psychosocial characteristics (ever married, education, employment). Additional multivariable models assessed the outcomes of interest while individually incorporating specific measures of health status (i.e., self-perceived general health, mental health, and pain as a result of the cancer) derived from responses to the CCSS baseline questionnaire. Adjustment for the health status variables did not change the overall results and, therefore,are not reflected in the multivariable models presented. The definitions of the health status variables and results of the multivariable analyses incorporating health status parameters can be viewed at www.cancer.umn.edu/ccss. The analyses of high school graduation included survivors 18 years old or younger at diagnosis and analyses of college graduation included those over the age of 25 years at questionnaire completion. In addition, each psychosocial functioning outcome was compared between survivors and siblings, with and without stratification of the survivor group by the four-group age-surgery variable, adjusting for current age and gender. To account for potential within-family correlation between a survivor and a sibling, we used the generalized estimating equation.22
Patient characteristics of the 694 bone tumor survivors who fulfilled the eligibility criteria for this analysis are provided in Table 1. Overall, the gender distribution was equal and the predominant site was the long bone of the lower extremity (93%). About 80% of the subjects had osteosarcoma and the most frequent local control surgical procedure was an amputation (67%). As anticipated, based on the natural history of these tumors, there was a predominance of osteosarcoma diagnosed in subjects older than 12 years (65%).
|Parameters||All 694 (100%)||Surgery type||Age at diagnosis (yrs)||Age/surgery group|
|Amp||LS||> 12||≤ 12||> 12/LS||≤ 12/LS||≤ 12/Amp||> 12/Amp|
|471 (67.9%)||223 (32.1%)||451 (65.0%)||243 (35.0%)||152 (21.9%)||71 (10.2%)||172 (24.8%)||299 (43.1%)|
|Male||353 (50.9%)||253 (53.7%)||100 (44.8%)||244 (54.1%)||109 (44.9%)||74 (48.7%)||26 (36.6%)||83 (48.3%)||170 (56.9%)|
|Female||341 (49.1%)||218 (46.3%)||123 (55.2%)||207 (45.9%)||134 (55.1%)||78 (51.3%)||45 (63.4%)||89 (51.7%)||129 (43.1%)|
|Type/site of primary tumor|
|OS: Pelvic||9 (0.3%)||3 (0.6%)||6 (2.7%)||9 (2.0%)||0 (0%)||6 (4.0%)||0 (0%)||0 (0%)||3 (1.0%)|
|OS: Extremity||547 (78.8%)||437 (92.8%)||110 (49.3%)||383 (84.9%)||164 (67.5%)||91 (59.9%)||19 (26.8%)||145 (84.3%)||292 (97.7%)|
|ES: Pelvic||39 (5.6%)||2 (0.4%)||37 (16.6%)||19 (4.2%)||20 (8.2%)||17 (11.2%)||20 (28.2%)||0 (0%)||2 (0.7%)|
|ES: Extremity||99 (14.3%)||29 (6.2%)||70 (31.4%)||40 (8.9%)||59 (24.3%)||38 (25.0%)||32 (45.1%)||27 (15.7%)||2 (0.7%)|
|Radiation use to primary site|
|Limb radiation||88/646 (13.6%)||32/466 (6.9%)||56/180 (31.1%)||47/423 (11.1%)||41/223 (18.4%)||31/129 (24.0%)||25/51 (49.0%)||16/172 (9.3%)||16/294 (5.4%)|
|Pelvic radiation||35/48 (72.9%)||1/5 (20.0%)||34/43 (79.1%)||20/28 (71.4%)||15/20 (75.0%)||19/23 (82.6%)||15/20 (75.0%)||—||1/5 (20.0%)|
|Age at diagnosis||13.5 (3–20)||13.3 (4–20)||14.1 (9–20)||15.7 (13–20)||9.5 (3–12)||16.2 (13–20)||9.7 (3–12)||9.4 (6–12)||15.5 (3–20)|
|Med: 14.0||Med: 13.0||Med: 15.0||Med: 15.0||Med: 10.0||Med: 16.0||Med: 10.0||Med: 10.0||Med: 15.0|
|Mean years from diagnosis to baseline questionnaire||15.8 (6–28)||16.4 (8–27)||14.5 (7–28)||15.7 (7–28)||15.9 (8–27)||14.1 (6–28)||15.3 (7–27)||16.2 (8–26)||16.4 (8–27)|
|Med: 16.0||Med: 16.0||Med: 14.0||Med: 16.0||Med: 16.0||Med: 13.0||Med: 15.0||Med: 16.0||Med: 16.0|
|Age at baseline questionnaire||29.8 (18–45)||30.2 (18–43)||29.1 (18–45)||31.9 (22–45)||25.9 (18–38)||30.8 (20–45)||25.5 (18–37)||26.1 (18–38)||32.5 (22–43)|
|Med: 30.0||Med: 30.0||Med: 29.0||Med: 32.0||Med: 25.0||Med: 30.5||Med: 25.0||Med: 25.0||Med: 33.0|
Among survivors 18 years of age and older at completion of the baseline questionnaire, 93% reported having graduated from high school. Of those older than 25 years, nearly 50% reported having graduated from college (Table 2). In the age-adjusted univariate analysis, all age-surgery groups were less likely to have graduated from high school compared with the reference group but only the ≤ 12/Amp age-surgery groups was significantly less likely to have graduated from high school (OR = 0.2; P = 0.040). No other factors were significant, including amputation status. However, when evaluating subjects who graduated from college, amputees were less likely to have graduated compared with nonamputees (OR = 0.6; P =0.029; Table 2). When evaluating age at diagnosis (older than 12 years or 12 years old or younger), subjects younger than 12 years at diagnosis were no different from subjects who were older at diagnosis in terms of college graduation. However, after adjusting for age at questionnaire completion and years since diagnosis, subjects younger than 12 were 2.5 times more likely to have graduated (P = 0.009).
|Parameters||Graduated high school (≤ 18 yrs old at diagnosis)||Graduated college (> 25 yrs old at questionnaire completion)||Ever had a job||Employed past year|
|No. of responses (% graduated)||OR (95% CI)||No. of responses (% graduated)||OR (95% CI)||No. of responses (% employed)||OR (95% CI)||No. of responses (% employed)||OR (95% CI)|
|Female||312 (93.3)||1.0 (—)||237 (48.5)||1.0 (—)||340 (95.3)||1.0 (—)||320 (78.1)||1.0 (—)|
|Male||299 (92.6)||0.9 (0.48–1.66)||260 (48.5)||1.0 (0.68–1.39)||350 (98.3)||2.5 (0.95–6.50)||343 (88.3)||2.1 (1.39–3.26)b|
|Age at diagnosis (yrs)|
|>12||382 (94.0)||1.0 (—)||386 (47.9)||1.0 (—)||227 (98.2)||1.0 (—)||436 (83.9)||1.0 (—)|
|≤12||229 (91.3)||0.5 (0.16–1.43)||111 (50.5)||2.5 (1.26–4.92)c||243 (94.2)||1.0 (0.22–4.30)||227 (82.4)||0.8 (0.38–1.63)|
|No||193 (95.3)||1.0 (—)||156 (57.7)||1.0 (—)||467 (98.2)||1.0 (—)||214 (87.4)||1.0 (—)|
|Yes||418 (91.9)||0.6 (0.26–1.21)||341 (44.3)||0.6 (0.44–0.96)c||223 (96.2)||0.4 (0.13–1.27)||449 (81.5)||0.7 (0.41–1.07)|
|>12/LS||125 (96.8)||1.0 (—)||124 (57.3)||1.0 (—)||152 (98.7)||1.0 (—)||148 (86.5)||1.0 (—)|
|≤12/LS||68 (92.7)||0.3 (0.05–1.43)||32 (59.4)||2.6 (0.95–7.07)||71 (97.2)||1.3 (0.13–13.28)||66 (89.4)||1.1 (0.37–3.43)|
|≤12/Amp||161 (90.7)||0.2 (0.05–0.93)c||79 (46.8)||1.6 (0.69–3.60)||172 (93.0)||0.5 (0.07–3.23)||161 (79.5)||0.5 (0.21–1.28)|
|>12/Amp||257 (92.6)||0.4 (0.13–1.22)||262 (43.5)||0.7 (0.45–1.11)c||295 (98.0)||0.5 (0.10–2.63)||288 (82.6)||0.8 (0.44–1.38)|
|Ewings sarcoma||122 (94.3)||1.0 (—)||86 (53.5)||1.0 (—)||138 (97.1)||1.0 (—)||133 (87.2)||1.0 (—)|
|Osteosarcoma||489 (92.6)||0.7 (0.30–1.67)||411 (47.5)||0.8 (0.47–1.21)||552 (96.7)||0.9 (0.18–1.81)||530 (82.5)||0.7 (0.38–1.19)|
|Extremity||571 (92.5)||1.0 (—)||461 (48.4)||1.0 (—)||—||—||—||—|
|Pelvic||40 (100)||N/A||36 (50.0)||1.1 (0.54–2.12)||—||—||—||—|
|No||573 (92.7)||1.0 (—)||466 (49.1)||1.0 (—)||—||—||—||—|
|Yes||38 (97.4)||2.9 (0.39–21.90)||31 (38.7)||0.7 (0.31–1.38)||—||—||—||—|
|High school graduate|
|No||—||—||—||—||44 (84.1)||1.0 (—)||35 (54.3)||1.0 (—)|
|Yes||—||—||—||—||614 (98.2)||10.5 (3.64–0.27)b||597 (86.1)||5.2 (2.57–0.58)b|
|No||—||—||—||—||369 (95.1)||1.0 (—)||350 (77.4)||1.0 (—)|
|Yes||—||—||—||—||289 (100)||N/A||282 (92.9)||3.9 (2.30–6.55)b|
In the multivariable analysis of high school and college graduation, after adjustment for tumor type, site, gender, age at questionnaire completion, and years since diagnosis until questionnaire completion, there were no statistically significant associations identified with the age-surgery subgroups (Table 3). However, the OR estimates were similar to those of the age-adjusted analysis.
|Characteristics||OR (95% CI)|
|High school||College||Ever had a job||Work the past year||Health insurance||Health insurance problems||Ever married||Currently in first marriage|
|> 12/LS (reference)||1.0||1.0||1.0||1.0||1.0||1.0||1.0||1.0|
|≤ 12/LS||0.2 (0.04–1.20)||2.7 (0.99–7.48)||10.9 (0.63–191.22)||1.8 (0.52–6.54)||2.9 (0.43–19.50)||1.2 (0.40–3.44)||1.4 (0.61–3.37)||1.3 (0.43–3.87)|
|≤ 12/Amp||0.2 (0.06–1.11)||1.5 (0.65–3.53)||2.4 (0.22–25.72)||0.5 (0.19–1.4)||0.8 (0.21–2.92)||1.3 (0.50–3.17)||1.0 (0.47–2.19)||1.6 (0.62–3.97)|
|> 12/Amp||0.5 (0.16–1.46)||0.7 (0.43–1.08)||0.7 (0.11–4.26)||0.6 (0.31–1.17)||1.0 (0.42–2.20)||0.9 (0.50–1.51)||1.3 (0.72–2.00)||1.2 (0.70–2.15)|
|High school/college||—||—||15.5bc (4.65–51.87)||6.0bc (2.88–12.60)||5.7de (2.56–12.63)||0.7e (0.44–1.03)||—||2.9be (1.88–4.53)|
|Males vs. females||0.9 (0.48–1.71)||1.0 (0.67–1.40)||3.8d (1.14–12.57)||2.4b (1.48–3.78)||0.4d (0.20–0.72)||1.0 (0.63–1.50)||0.5b (0.35–0.74)||0.9 (0.57–1.33)|
|Pelvic vs. extremity||N/A||0.7 (0.36–1.60)||0.3 (0.02–2.89)||0.4 (0.16–1.02)||0.8 (0.20–3.15)||1.1 (0.50–2.59)||0.9 (0.43–2.04)||3.0d (1.03–8.62)|
|Currently married||—||—||—||—||2.1b (1.06–4.30)||0.6 (0.35–1.04)||—||—|
Greater than 97% of the participants in this cohort reported ever having had a job and 83% of participants reported having worked in the past year. In the age-adjusted analysis, age-surgery groups with an amputation were less likely to have ever had a job compared with the >12/LS group, but these findings were not statistically significant (Table 2). Higher educational attainment and male gender were associated significantly with having been employed. In the multivariable analysis (Table 3), high school graduation (OR = 15.5; P < 0.001) and male gender (OR = 3.8; P = 0.029) remained significant when simultaneously considering site of tumor, age at questionnaire completion, and years since diagnosis until questionnaire completion.
Gender and educational status were the only independent predictors of recent employment (past year). High school graduates (OR = 5.2; P < 0.001), college graduates (OR = 3.8; P < 0.001), and male gender (OR = 2.1; P < 0.001) were more likely to have worked in the previous year (Table 2). In multivariable analysis, educational attainment and male gender remained significant (Table 3).
At the time of questionnaire completion, 87% of the participants reported having health insurance, whereas almost 30% reported having experienced difficulty in obtaining health insurance. The likelihood of currently having insurance was associated significantly with being a female (OR = 2.0; P = 0.003), being a nonamputee (OR = 1.7; P = 0.032), completing college (OR = 3.3; P < 0.001), and currently being married (OR = 2.4; P = 0.003; Table 4). These all remained significantly predictive in the multivariable analysis, except for amputation status (Table 3). No predictive factors of having difficulty in obtaining health insurance were found (Tables 3, 4).
|Parameters||Currently have health insurance||Difficulty obtaining health insurance||Ever married||Currently in first marriage|
|No. of respondents (% insured)||No. of respondents (% insured)||No. of respondents (% insured)||No. of respondents (% insured)||No. of respondents (% married)||OR (95% CI)||No. of respondents (% married)||OR (95% CI)|
|Female||335 (91.0)||1.0 (—)||337 (30.9)||1.0 (—)||341 (71.3)||1.0 (—)||243 (66.7)||1.0 (—)|
|Male||347 (83.3)||0.3 (0.30–0.78)b||344 (28.2)||0.9 (0.63–1.22)||353 (63.2)||0.5 (0.35–0.73)c||216 (66.2)||0.9 (0.66–1.44)|
|Age at diagnosis (yrs)|
|>12||442 (86.9)||1.0 (—)||239 (31.8)||1.0 (—)||461 (75.8)||1.0 (—)||336 (68.5)||1.0 (—)|
|≤12||240 (87.5)||1.1 (0.51–2.54)||442 (28.3)||1.5 (0.85–2.77)||243 (51.0)||1.0 (0.53–1.84)||123 (61.0)||1.4 (0.70–2.85)|
|No||219 (90.4)||1.0 (—)||222 (26.6)||1.0 (—)||223 (65.5)||1.0 (—)||143 (68.5)||1.0 (—)|
|Yes||463 (85.5)||0.6 (0.32–0.95)c||459 (30.9)||1.3 (0.87–1.82)||471 (67.9)||0.9 (0.62–1.36)||316 (65.5)||0.9 (0.60–1.42)|
|> 12/LS||149 (88.4)||1.0 (—)||151 (26.5)||1.0 (—)||152 (70.4)||1.0 (—)||105 (70.5)||1.0 (—)|
|≤ 12/LS||70 (94.6)||2.1 (0.55–7.67)||71 (26.8)||1.4 (0.62–3.15)||71 (54.9)||1.4 (0.62–3.30)||38 (63.2)||1.6 (0.57–4.37)|
|≤ 12/Amp||170 (85.4)||0.7 (0.25–1.72)||168 (33.9)||2.0 (0.97–4.10)||172 (49.4)||1.0 (0.45–2.00)||85 (60.0)||1.3 (0.56–3.17)|
|> 12/Amp||293 (85.9)||0.7 (0.37–1.29)||291 (29.2)||1.2 (0.77–1.90)||299 (78.6)||1.1 (0.69–1.85)||231 (67.5)||1.0 (0.59–1.66)|
|Ewings sarcoma||137 (91.2)||1.0 (—)||136 (27.2)||1.0 (—)||646 (67.2)||1.0 (—)||428 (65.4)||1.0 (—)|
|Osteosarcoma||545 (86.1)||0.6 (0.31–1.11)||545 (30.1)||1.2 (0.76–1.79)||48 (66.7)||1.0 (0.49–2.00)||31 (80.6)||2.2 (0.86–5.44)|
|Extremity||635 (86.8)||1.0 (—)||633 (29.4)||1.0 (—)||138 (62.3)||1.0 (—)||84 (66.7)||1.0 (—)|
|Pelvic||47 (91.5)||1.6 (0.57–4.70)||48 (31.3)||1.1 (0.58–2.06)||556 (68.4)||0.9 (0.57–1.42)||375 (66.4)||0.9 (0.52–1.46)|
|No||367 (82.0)||1.0 (—)||364 (31.3)||1.0 (—)||373 (66.2)||1.0 (—)||242 (57.9)||1.0 (—)|
|Yes||384 (93.7)||3.3 (1.91–5.79)b||285 (26.3)||0.8 (0.55–1.10)||289 (68.2)||0.7 (0.51–1.10)||196 (80.1)||2.8 (1.84–4.41)b|
|Worked past year|
|No||107 (86.0)||1.0 (—)||105 (35.2)||1.0 (—)||—||—||—||—|
|Yes||548 (87.4)||1.2 (0.64–2.13)||547 (28.4)||0.7 (0.47–1.14)||—||—||—||—|
|Current marital status|
|Not married||78 (25.6)||1.0 (—)||77 (41.6)||1.0 (—)||—||—||—||—|
|Married||368 (88.9)||2.4 (1.30–4.51)c||368 (28.8)||0.9 (0.65–1.35)||—||—||—||—|
|No||—||—||—||—||651 (67.3)||1.0 (—)||432 (66.0)||1.0 (—)|
|Yes||—||—||—||—||43 (65.1)||1.0 (0.47–2.07)||27 (74.1)||1.5 (0.60–3.57)|
Two-thirds of the participants reported being married or living as married. Females (OR = 2.0; P < 0.005) were also significantly more likely than males to have ever been married (Table 4). In the multivariable model including age-surgery group, site, gender, age at questionnaire completion, and years since diagnosis until questionnaire completion, only gender remained significant with having been married being twice as likely for females compared with males (P = 0.004; Table 3).
Of those who reported having ever been married, 66% were currently in their first marriage (34% had been divorced or separated). College graduation was the only predictor of being in a first marriage in age-adjusted analysis (Table 4) whereas college completion (OR = 2.8; P < 0.001) and pelvic tumors (OR = 3.0; P = 0.045) were statistically significant independent predictors in multivariable analysis (Table 3).
Compared with the bone tumor survivors, the sibling cohort was slightly younger (mean, 29.0 years; range, 18–56 years) at completion of the baseline questionnaire and was more likely to be female (53.1% vs. 49.1%). Therefore, analyses comparing survivors to the sibling cohort are adjusted for age and gender (Table 5). Only the > 12/Amp age-surgery group was less likely to graduate from high school (OR = 0.6; P = 0.047) and college (OR = 0.8; P = 0.037) compared with siblings, whereas ever having a job was significantly less likely among the amputee age-surgery groups (≤ 12/Amp [OR = 0.2; P < 0.001] and > 12/Amp [OR = 0.2; P = 0.003]). Bone tumor survivors were significantly more likely to report insurance problems overall and in all age-surgery groups compared with the sibling cohort. Only the >12/Amp age-surgery group (OR = 0.7; P = 0.022) was statistically less likely to currently have health insurance. The survivors were statistically less likely to report ever being married or living as married. Survivors 12 years old or younger at diagnosis were less likely to be in their first marriage if ever married and the ≤ 12/Amp age-surgery group was significantly less likely to be in their first marriage (OR = 0.6; P = 0.033).
|Subgroups||OR (95% CI)|
|High school||College||Ever had a job||Work the past yr||Currently have health insurance||Health insurance problems||Ever been married||Currently in first marriage|
|All cases (n = 694)||0.8 (0.56–1.08)||0.9 (0.77–1.13)||0.3b (0.17–0.51)||0.3b (0.26–0.43)||0.8 (0.60–1.01)||12.7b (9.63–16.56)||0.8c (0.65–0.99)||0.8 (0.58–1.01)|
|> 12/LS (n = 152)||1.6 (0.57–4.91)||1.2 (0.82–1.77)||0.5 (0.10–2.18)||0.4b (0.26–0.71)||0.9 (0.51–1.43)||10.9b (7.10–16.74)||0.7 (0.48–1.13)||1.1 (0.68–1.60)|
|≤ 12/LS (n = 71)||1.0 (0.39–2.63)||1.0 (0.98–1.01)||0.7 (0.18–2.77)||N/A||2.1 (0.75–5.86)||10.0b (5.16–19.54)||1.1 (0.66–1.77)||0.6 (0.31–1.21)|
|> 12/Amp (n = 299)||0.6c (0.34–0.95)||0.8c (0.59–0.99)||0.2b (0.07–0.58)||0.3b (0.21–0.43)||0.7c (0.46–0.94)||12.0b (8.51–16.92)||0.8 (0.73–1.32)||1.0 (0.73–1.32)|
|≤ 12/Amp (n = 172)||0.7 (0.42–1.31)||0.9 (0.60–1.41)||0.2b (0.11–0.43)||0.2b (0.16–0.37)||0.7 (0.47–1.12)||19.3b (12.93–28.57)||0.7 (0.52–1.01)||0.6c (0.40–0.96)|
As more survivors of pediatric cancers enter adulthood, the assessment of their functioning and psychosocial status is becoming increasingly important in determining whether these survivors require additional resources and supportive services to enhance psychosocial outcomes (e.g., educational achievement, vocational opportunities, employment status, and interpersonal relationships). Several authors have explored psychosocial outcomes in survivors of pediatric cancers and have found deficits in marriage, employment, education, and health insurability.23–32 Our study focused on survivors of pediatric and adolescent lower extremity bone tumors because they face unique challenges and may have more difficulty adapting to the challenges of adulthood.
The few studies of survivors of pediatric bone tumors that exist generally have either included a small number of study subjects, short follow-up, or do not focus specifically on lower extremity bone tumor patients.19, 20 The two small sibling–control studies of adult survivors of childhood bone tumors (Ewing sarcoma and osteosarcoma) had a mean age at diagnosis and a mean age at follow up of approximately 15 and 30 years, respectively. Novakovic et al.19 did not observe any difference in educational attainment among 89 adult survivors of childhood bone tumors compared with sibling controls. However, bone tumor survivors had statistically significant deficits in employment and marriage rates and significantly higher rates of health insurance problems and divorce. In contrast, Nicholson et al.20 observed that employment and marriage rates did not differ from sibling controls in 111 long-term survivors of pediatric bone tumors. However, survivors did report both physical and emotional disabilities significantly more frequently than sibling controls. Nicholson et al. concluded that there were few adverse psychosocial outcomes in this group of cancer survivors whereas Novakovic et al. reported that there were significant issues regarding employment and marriage that warranted further investigation. Smaller studies indicated no differences in education, occupational status, interpersonal/social interaction, and self-image between patients treated with amputation and limb-sparing procedures.15, 33, 34 None of these studies focused on specific tumor sites (e.g., the lower limb). The conflicting conclusions may result from the limited number and diverse presentations and treatments of bone tumors.
Using the CCSS cohort, we explored the issues of education, employment, health insurance, and marriage in a large number of pediatric lower extremity bone tumor survivors. This cohort was established as a resource to assess the broad spectrum of health-related and psychosocial outcomes of long-term survivors of childhood and adolescent cancer. Due to the large size, the CCSS cohort permits evaluation of particular subsets of survivors with specific demographic and clinical features who otherwise would only have been studied in either small numbers or in case series. In the current study, the cohort of lower extremity bone tumors is consistent with the epidemiologic features of osteosarcoma and Ewing sarcoma and also with data from the Surveillance, Epidemiologic, and End Results program.12 There is a slightly lower than expected proportion of patients with pelvic tumors (48 of 694), which typically would account for 25% of Ewing sarcoma and 5% of osteosarcoma presentations.35 This may be due to the poorer prognosis of of subjects with pelvic tumors at the time this cohort was diagnosed. The number of amputees in this cohort is also much higher than current number of amputees. This is attributable, in part, to changes in surgical management over the past three decades.36, 37 This change in practice to more limb-sparing surgeries has not resulted in more adverse outcomes (lower disease-free survival or higher local disease recurrence rates).38–41
Survivors of lower extremity bone tumors had high employment (97%), graduation (high school, 93%; college, 50%), and marriage (67%) rates. Unfortunately, the proportion of survivors who report health insurance problems (30%) and the proportion who have health insurance (87%) are of concerning, given the potential for further medical needs of this group of survivors. Potential medical complications range from difficulties arising from their surgical procedure to the long-term effects of chemotherapy and radiation.
Many of the significant predictors in the age-adjusted univariate analyses lost their significance in multivariable analysis. The remaining significant factors included higher education, which was a strong predictor for all outcomes (employment, health insurance, and marriage). In addition, male gender was predictive of employment whereas female gender was predictive of health insurance coverage and ever being married. Site and marital status were also predictive for marriage and health insurance-related issues, respectively. In all multivariable analyses, the age-surgery groups that account for the age at which surgical procedures were done (preadolescence or adolescence) were not significant. This finding conflicted with our original hypothesis based on current surgical practices. We believed that patients treated with a limb-sparing procedure (≤ 12/LS and > 12/LS) would have higher rates of employment, marriage, and educational attainment compared with patients with an amputation (≤ 12/Amp and >12/Amp).
The significance of gender in many of the outcomes is not surprising and has been described in other cancers. Males in this study were less likely to ever have been married, a finding that may be related to the physical sequelae of their cancer therapy and potentially to societal standards placed on males as described by Byrne et al.29 and Rauck et al.30 Having ever been employed and currently working the past year may also relate to societal standards that identify males as the financial head of household.28 In addition, females are more likely to have been married. Therefore, it is not surprising that they would have a higher likelihood of having health insurance (through their spouse).27
All age-surgery groups were significantly more likely to experience difficulty obtaining health insurance compared with siblings. The > 12/Amp group has the most difficulties compared with siblings, particularly with education attainment and employment. The lack of differences between the other groups (≤ 12/LS, ≤ 12/Amp, and > 12/LS) and siblings may be due to the limited number of patients in the groups. The ≤ 12/Amp and ≤ 12/LS age-surgery groups encounter similar deficits in health insurance and the ≤ 12/Amp group has further deficits in employment and marriage. The referent group, > 12/LS, experiences difficulties in working the past year and has health insurance problems. The lack of deficits experienced by the limb-sparing groups (≤ 12/LS and > 12/LS) may be related to the varying nonamputation procedures, which vary from knee or hip replacement to radiation-only treatments. Even after adjusting for measures of current health status, there was no substantial impact on the OR of our outcomes of interest.
One limitation of the study includes ascertainment bias because the cohort includes only survivors who agreed to participate in the CCSS. This may underestimate deficits by excluding survivors who are having more difficulty adjusting and are unable or unwilling to participate. Other limitations include the limited number of participants in the subgroups (e.g., the ≤ 12/LS age-surgery group) and the inaccuracies and limited detail related to self-reported data. The potential bias of self-reported data were minimized because the questions were piloted before being administered to the cohort. The limited number of respondents in the subgroups is related to the time frame when these surgical procedures were done. Most of the procedures were done during a time when amputations were performed more frequently because of the limited experience with limb-salvaging procedures. This may not appropriately reflect today's surgical treatments given the improvements in cancer treatments, surgical techniques, and supportive care. However, uniform evaluation of a population of this size still provides a better view of the long-term consequences of treatment than previous studies. This project will lead to future prospective studies of these issues and it will provide patients and clinicians with a better understanding of the limitations of the surgical options, as well as the impact of age, radiation use, and other factors on the function and quality of life.
Overall, these survivors still require improvement in health insurance availability and barriers to health insurance need to be explored. The findings also identify other groups at risk for deficits (the > 12/Amp age-surgery group, females in terms of employment, males in terms of marriage and insurance and those with less education). When evaluating amputation status (amputees vs. nonamputees), amputees do as well as nonamputees, but there are deficits when amputees are compared with siblings. The lack of difference between amputees and nonamputees may reflect the improved adaptation of amputees over time or may represent the worsening health status of nonamputee survivors over time due to chronic issues with the nonamputated extremity. Long-term prospective studies with sequential assessments may help to resolve this issue.
The identification of groups at risk for adverse outcomes permits better allocation of resources (e.g., funding support for higher education and health insurance policy changes to provide better access for childhood cancer survivors and to ensure appropriate coverage for prosthesis replacements and additional surgeries) and the development of interventions (e.g., long-term social work follow-up and support, development of peer support groups for cancer survivors, and education of physicians of adult childhood cancer survivors regarding the unique issues facing survivors). This information will also allow clinicians, patients, and families to have a better understanding of the surgical options and be able to address the specific needs of survivors of pediatric lower extremity bone tumors and their barriers to a healthy adjustment to adulthood.
- 12Cancer incidence and survival among children and adolescents: United States SEER program 1975-1995. NIH Pub. No. 99-4649. Bethesda: National Cancer Institute; 1999., , , et al.
- 13Use of denial in adolescents with bone cancer. Health Social Work. 1976; 1: 70–87..
- 35Common malignant bone tumors. In: SimonM, SpringfieldD. Surgery for bone and soft-tissue tumors. Philadelphia: Lippincott-Raven, 1998: 265–298..
- 38Primary chemotherapy and delayed surgery for nonmetastatic osteosarcoma of the extremities. Results in 164 patients preoperatively treated with high doses of methotrexate followed by cisplatin and doxorubicin. Cancer. 1993; 72: 3227–3238., , , et al.
- 39Complications and surgical indications in 144 cases of nonmetastatic osteosarcoma of the extremities treated with neoadjuvant chemotherapy. Clin Orthop Rel Res. 1993; 295: 226–238., , , et al.