|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 25, No 29 (October 10), 2007: pp. 4610-4615 © 2007 American Society of Clinical Oncology. DOI: 10.1200/JCO.2006.07.6992 Access to Pediatric Cancer Care by Age, Race, and Diagnosis, and Outcomes of Cancer Treatment in Pediatric and Adolescent Patients in the State of Georgia
From the Departments of Pediatrics, Rollins School of Public Health; Department of Epidemiology, Emory University; Children's Healthcare of Atlanta; Aflac Cancer Center and Blood Disorders Service; and the Georgia Center for Cancer Statistics, Atlanta, GA Address reprint requests to Della L. Howell, MD, 2200 Bergquist Dr, Suite 1A, MMNP, Pediatric Hematology/Oncology, San Antonio, TX 78236; e-mail: Della.Howell{at}lackland.af.mil
Purpose There have been concerns among pediatric oncologists that adolescent and minority patients are not getting adequate access to care. This study examines access to cancer care and survival outcomes based on age, race, and type of cancer in patients in Georgia. Patients and Methods We performed a retrospective review of 1,751 cancer patients aged 0 to 19 years, diagnosed between 1998 and 2002, in the Georgia Comprehensive Cancer Registry, which identified patients who were treated at one of five Georgia pediatric cancer centers (Children's Oncology Group [COG] members) at any point in their treatment. Data were further analyzed for age at diagnosis, race, county of residence, and 5-year survival. Results Eighty-seven percent of patients aged 0 to 14 years and 36% of those aged 15 to 19 years were treated at a COG institution. Twenty-five percent of all patients were of African descent, with 75.4% of black versus 70.3% of white patients (age 0 to 19 years) treated at a COG institution (P < .01); 97.1% of other minorities were treated at a COG institution (P < .05). The 5-year actuarial survival rates for more pediatric-specific cancers were significantly lower in all leukemias (75.1% v 46.4%; P = .0015), and acute lymphoblastic leukemia specifically (86.3% v 53.3%; P < .05) for patients not treated at a COG institution. Actuarial survival rates were much lower for blacks than whites in all cancers as a whole (70% v 82%; P < .001) and for many specific subtypes. Conclusion Adolescent-aged patients are less likely to be referred to a COG institution, potentially exposing them to worse outcomes in some cancer subtypes. Reassuringly, minority populations are receiving adequate access to pediatric cancer care; unfortunately their survival rates are lower.
There have been several recent studies looking at access to care for the adolescent-aged cancer patient. Pediatric oncologists in particular have had concerns that adolescents are not getting access to age-appropriate treatments when they are not being seen in pediatric oncology centers. Further concerns stem from recent studies that suggest that by not getting care at a pediatric oncology center, these adolescent patients may also suffer worse outcomes. Several studies have shown that survival for certain pediatric cancers such as acute lymphoblastic leukemia (ALL), lymphoma, and certain solid tumors is better for patients treated at specialized pediatric oncology centers, registered on established treatment protocols.1-6 One item of note is the paucity of enrollment of patients between 15 and 19 years of age onto National Cancer Institute (NCI)-sponsored cooperative group trials. This may be a reflection of many adolescent patients not being seen at pediatric cancer treatment centers. Previous studies have shown that adolescent patients are being seen at pediatric cancer centers at much lower rates than are younger pediatric patients.7-9 Our primary goal was to determine how many new pediatric and adolescent cancer patients are not being seen at centers with pediatric oncologists in pediatric cancer centers. Adolescent-age patients in particular were a focus of this study. We also wanted to analyze access to care in terms of race and ethnicity, to see whether minority populations were getting the same access to pediatric cancer care. A third goal was to analyze the 5-year survival rates for patients not seen at pediatric cancer centers compared with those who were getting access to pediatric oncology care, with subanalysis of survival rates in minority patients.
To locate all new cancer diagnoses in the state of Georgia between January 1998 and December 2002, patients were identified through the Georgia Comprehensive Cancer Registry (GCCR) in Atlanta, GA. The GCCR is a statewide, population-based cancer registry that has been collecting information on all cancer cases diagnosed among Georgia residents since January 1, 1995. It is a participant in the National Program for Cancer Registries (NPCR), and the North American Association of Central Cancer Registries (NAACCR). In the state of Georgia, all health care providers are required to report information about cancer patients receiving diagnostic evaluations or treatment at their facilities. If a cancer patient is new to a particular facility, it must be reported, regardless of whether it has been reported by another facility.10 The incident capture rate for patients 0 to 19 years of age within the GCCR was 98% or greater for all years analyzed in this study. Patients between the age of 0 and 19 years were used for this study. The GCCR was able to identify which of the patients were ever seen at one of five pediatric cancer centers in the state of Georgia at any time during their treatment. They included centers in Atlanta, Columbus, Augusta, Macon, and Savannah, GA. These centers have all been NCI-cooperative group centers. There are more than 200 COG centers in North America. Specific criteria must be met to be a full COG-member, and this was similar for the legacy groups Pediatric Oncology Group and Children's Cancer Group. Specifics regarding eligibility include having board-certified or board-eligible pediatric oncologists on staff, ready access to pediatric radiation therapy and pediatric surgery, and location within a tertiary care hospital that has full access to appropriate pediatric subspecialties and other pediatric supportive care. Each of these centers is expected to have at least 12 newly diagnosed pediatric oncology cases per year. The GCCR database was used to further analyze data for age at diagnosis, date of birth, type of cancer, race, county of residence, stage of disease, and survival at 5 years. The type of cancer was determined using the International Classification of Diseases for Oncology (ICD-O), as per GCCR requirements. Melanoma was included in our category of "carcinomas" along with various other carcinomas recognized by the ICD-O. We used logistic regression to compare the distribution of age, sex, race, and cancer site for COG and non-COG institutions and to compare the stage distribution of cancer for blacks and whites. We obtained the odds of attending a COG versus a non-COG institution for various levels of these factors, and report odds ratios by defining a referent category for each factor. The use of logistic regression enabled us to estimate these odds ratio and to obtain P values for tests of statistical association.11 We estimated 5-year survival rates using the Kaplan-Meier life-table approach.12 We estimated the hazard ratios (HRs) for blacks versus whites and for COG versus non-COG institutions using the Cox proportional hazards model.13 The HRs for the black/white comparison were interpreted as the case fatality rate among blacks divided by that of whites; whereas for the COG/non-COG comparisons, the HRs were interpreted as the case fatality rate among patients attending a COG institution divided by the case fatality rate of patients attending a non-COG institution. Two-tailed P values less than .05 were considered statistically significant. Logistic regression and the proportionate hazards models were implemented using Stata.14
There were 1,751 new primary malignancies found in Georgia-residence patients aged 0 to 19 years between the years 1998 and 2002. Forty-six patients were missing data regarding their treatment location, and 13 patients were missing race data. Because patients living in "border counties" next to Tennessee, Alabama, and Florida could be going to out-of-state facilities, the data were analyzed with and without these counties, and the results were virtually identical. The GCCR also has a data exchange agreement with the bordering states; if a Georgia resident chooses to receive treatment in a border state, the treatment information is usually relayed to the GCCR via a twice-yearly data exchange. For statistical reasons, those border counties were included in the final results. A total of 1,227 patients (70.1%) were treated at pediatric oncology centers. Table 1 highlights the distribution of patients according to any care at one of the five Georgia pediatric oncology centers by age (P < .0001 for trend), sex (P = .87), race (overall P = .02), and cancer subtype. Only 36% of adolescents aged 15 to 19 years were seen at a pediatric cancer center. Conversely, 75.4% of blacks and 97.1% of other races versus 70.3% of whites were treated at a pediatric cancer center.
After excluding patients with leukemia, all of whom are staged as having distant disease, patients treated at pediatric oncology centers were more likely to have more advanced disease than were those treated at other centers. Only 5.8% of the patients were of "unknown" stage for this analysis. When looking at pediatric treatment centers versus nonpediatric centers, percentages for distant, regional, and in situ/localized disease were 21.5% versus 14.7%, 26.4% versus 22.8%, and 45.3% versus 50.2%, respectively. However, after adjustment for age, race, and type of cancer, the distribution of stage for pediatric versus nonpediatric centers is not significantly different (P > .2). The odds ratio for regional disease compared with local disease for pediatric versus nonpediatric centers was 1.3 (95% CI, 0.85 to 1.8), while that for distant disease versus local disease was 1.4 (95% CI, 0.9 to 2.2). The median length of follow-up for all subjects was 25 months. Twenty-five percent of the subjects were followed for 12 or fewer months, whereas 25% were followed for longer than 44 months. The 5-year survival of patients based on race and location of treatment is shown in Table 2. Black race was associated with a worse survival rate in most of the cancer subtypes analyzed, especially Hodgkin's disease, muscle/bone tumors, and CNS tumors, as well as for all cancers together. Figure 1 highlights the fact that blacks and whites had similar stage of disease at time of presentation. Only non-Hodgkin's lymphoma (NHL) showed a statistically significant difference in those with distant disease at time of diagnosis, with whites more likely to present with more advanced disease than blacks.
Patients treated at pediatric oncology centers were also noted to have a better HR for many cancer subtypes that would be deemed mainly "pediatric" types (Table 2). Of note, patients with Hodgkin's disease had a significant lowering of risk of death (HR 0.14; 95% CI, 0.03 to 0.77) when treated at a COG center. Further illustration of the actuarial survival data for the 15- to 19-years old age group is shown in Table 3. Unfortunately, with such low numbers of patients, there are significant limitations to statistical power in comparing the groups, hindering sensitivity to true data differences. HRs (adjusted for race and age) favored treatment at COG centers for patients with leukemia, lymphoma, neuroblastoma, renal tumors, and sarcomas.
We actually had a fairly large number of carcinomas included in our data (n = 156), which was made up of melanoma (32.5%), thyroid carcinoma (27.5%), nasopharyngeal carcinoma (4.2%), adrenocortical carcinoma (2.5%), and other/unspecified carcinomas (32.5%). Within this group of carcinoma patients, when looking at pediatric treatment centers versus nonpediatric centers, percentages for distant, regional, and in situ/localized disease were 18.8% versus 2%, 31.3% versus 19.2%, and 50% versus 67.7%, respectively. The 5-year survival rates in melanoma were 97.9% for nonpediatric center patients and 84.5% for pediatric center patients. Although the numbers are small and may be influenced by stage of disease in terms of a referral bias for pediatric centers, there was no sense that children do better in the care of a pediatric center, and in fact may do better when treated at nonpediatric centers.
The results again illustrate how poor access to pediatric oncology care is in the adolescent age range, with only approximately one third of adolescent patients aged 15 to 19 years receiving any care in a pediatric oncology center. Other studies looking specifically at ALL have shown that survival for adolescent patients treated on adult protocols is worse than for those treated on pediatric cooperative group protocols,15,16 and this analysis shows similar findings in more pediatric-specific cancers, such as leukemia and lymphoma. The differences in survival in this analysis are significant only when analyzing all ages together. Some studies looking at solid tumors specifically, such as rhabdomyosarcoma and Ewing's sarcoma, have shown a benefit for adolescents and young adults being treated on pediatric treatment plans1 and at pediatric treatment centers.2 One might assume that adolescent patients are being treated at adult oncologic facilities, which are oftentimes more convenient in location than are pediatric treatment centers. There are many more adult oncology facilities in the state of Georgia than the five pediatric treatment centers. Other factors in treatment location might include patient bias; an adolescent patient may not want to be treated at a "children's" hospital, because socially, they feel that they will be treated as a child, not as a near adult. Our analysis did not include patients aged 20 and 21 years, and these patients would likely be eligible for care in many pediatric centers. One would think that inclusion of these data would only further decrease the percentage of patients receiving access to care at a pediatric oncology center. Younger-aged patients (0 to 14 years of age) also had a percentage (14.7%) of patients not treated at Georgia pediatric cancer centers. A large group (40.1%) of these patients represented "border county" patients, and presumably received care at a neighboring state's institution. Another group of these young patients may have represented those with in situ or localized disease that might not have required specialist intervention. One limitation to this study is that we are not able to ascertain what percentage of a patient's treatment was administered at a COG center. Patients may have had their initial care planned at a nonpediatric center, then later transferred to a pediatric center, or vice versa. This could affect some of the survival outcomes in the patients studied. Another limitation relates to the analysis of disease staging. While the percentage of "unknown stage" is low, we have no way to verify the accuracy of the disease stage that was coded by the various facilities. There could, in fact, be a bias in the stage of cancer if nonpediatric or nonacademic facilities are less detailed in the information provided to the GCCR. However, in cases of advanced-stage disease, these patients might also just be more likely to be referred to a pediatric oncology center with the hopes of getting more specialized care in the most difficult cancer presentations. Both later scenarios are possible, given the greater disease burden of patients treated at pediatric centers. We were gratified to see that minority patients were receiving adequate access to care to childhood facilities. This may be a reflection of patient residence; black patients tend to be living closer to pediatric centers in the more "urban" regions of Georgia in general. This may also be a reflection of socioeconomic status and insurance carriers. The pediatric oncology centers may be more likely to accept a given patient, regardless of their medical insurance status. We hypothesize that private-practice oncology centers would be less likely to care for these uninsured or Medicaid patients compared with state-funded, larger academic oncology centers, but this is conjecture only and cannot be supported by our limited analysis. Studies performed on adult cancer patients looking at survival rates and access to care for patients with private health insurance versus uninsured and those with Medicaid have shown poorer survival for those who are uninsured or with Medicaid.17 On the other hand, it was disappointing to see that patients who were treated at nonpediatric cancer centers had such a difference in survival rates. This held true when analyzed for tumor stage. According to our data, pediatric cancer centers were actually more likely to see patients with more advanced disease than were nonpediatric centers. One can hypothesize that the nonpediatric center patients are not being treated per national standard of care pediatric protocols, but are being treated per institutional or standard adult protocols. Another factor that might affect the overall survival of these patients is supportive care. Although the adult protocols likely address supportive care issues, they may not be geared toward the pediatric- or adolescent-aged patient. Pediatric cancer centers may also be more likely to provide inpatient versus outpatient care for these patients, which may allow for more timely treatment of life-threatening complications. Looking at ALL specifically, adult cancer centers are actually less likely to encounter this disease entity and may not have the well-defined leukemia programs that pediatric centers use, and adult programs may be less likely to adhere to strict protocol regimens.18 Although our data are too limited to make a definitive statement, it appears that the 15- to 19-year-old population has better survival when receiving cancer treatment at centers that commonly manage their specific cancer subtype. They seem to be better served in the setting of leukemia by receiving care in a pediatric cancer center. An example of a tumor likely to be treated at either type of center is that of germ cell tumors; these patients appear to do equally as well in either type of center. Finally, in the case of carcinomas and other more "adult" cancers, these patients actually appear to do better at nonpediatric centers, although small numbers still limit the ability to declare statistical significance. Minority patients have shown decreased survival rates for the majority of tumors analyzed. This is despite having access to care at pediatric cancer treatment facilities. One might hypothesize that this is a result of delay in care for minority patients; however, our data for solid tumors show that blacks and whites have very similar disease burden at time of diagnosis, and in the case of NHL, blacks presented with less severe disease, but a worse survival outcome. This may be related to specific genetic heterogeneity that is more likely to occur in some subpopulations than in others, thus predisposing a patient to a worse outcome despite similar care and treatments. Previous studies have supported genetic or other associated racial or ethnic differences in survival.19 The reason for the survival difference is unclear, but several components may be involved, from differences in mercaptopurine metabolism to nonadherence to treatment protocols.20 In diseases such as acute myelogenous leukemia (AML), studies have shown an increase in the GST.T1 null genotype in blacks, which has also been associated with worse AML outcome.21 A recent study looking at AML survival difference by race and ethnicity showed that worse survival in minority patients must stem from factors other than chemotherapy compliance, such as potential pharmacogenetic differences in chemotherapy metabolism or genetic differences within the cancer itself, since AML's treatment plan is almost exclusively inpatient.22 Our data differ from Baker et al,23 which showed no significant difference in failure-free survival for rhabdomyosarcoma when analyzed by race. This may be a result of the difference of end point (failure-free v 5-year survival) or a more complete adjustment for a higher number of risk categories with higher patient numbers in their study than we had available for this limited analysis. The Surveillance, Epidemiology, and End Results monograph for childhood cancers 1975 to 1995 supports our findings for survival by race in the majority of cancer subtypes studied, including rhabdomyosarcoma.24 We have demonstrated that adolescent patients are not getting access to care at COG centers, which may influence their overall survival, especially for traditionally pediatric hematopoietic malignancies. In contrast, adolescent patients with more "adult" malignancies, such as carcinoma and melanoma, may be better served at a nonpediatric center. Once we have data on more patients, we can do further analysis to get more information regarding survival so we can educate providers about appropriate referral patterns for these specific diseases. We have also shown that minority patients are getting access to care at COG centers with similar stage of disease but with much worse outcomes. Further study should focus on biologic differences or subtle differences in the care of these patients.
The author(s) indicated no potential conflicts of interest.
Conception and design: Della L. Howell, John L. Young, William G. Woods Provision of study materials or patients: Kevin C. Ward, John L. Young Collection and assembly of data: Della L. Howell, Kevin C. Ward, John L. Young Data analysis and interpretation: Harland D. Austin Manuscript writing: Della L. Howell, William G. Woods Final approval of manuscript: Della L. Howell, William G. Woods
Supported by the Aflac Cancer Center and Blood Disorders Service. Presented in poster format at the 18th Annual Meeting of the American Society of Pediatric Hematology/Oncology (ASPHO), May 13 to 16, 2005, Washington, DC. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Ferrari A, Dileo P, Casanova M, et al: Rhabdomyosarcoma in adults: A retrospective analysis of 171 patients treated at a single institution. Cancer 98:571-580, 2003[CrossRef][Medline] 2. Paulussen M, Ahrens S, Juergens HF: Cure rates in Ewing tumor patients aged over 15 years are better in pediatric oncology units. Results of GPOH CESS/EICESS studies. Proc Am Soc Clin Oncol 22:816, 2003 (abstr 3279) 3. Kramer S, Meadows AT, Pastore G, et al: Influence of place of treatment on diagnosis, treatment, and survival in three pediatric solid tumors. J Clin Oncol 2:917-923, 1984[Abstract] 4. Meadows AT, Kramer S, Hopson R, et al: Survival in childhood acute lymphocytic leukemia: Effect of protocol and place of treatment. Cancer Invest 1:49-55, 1983[Medline] 5. Lennox EL, Stiller CA, Jones PH, et al: Nephroblastoma: Treatment during 1970-3 and the effect on survival of inclusion in the first MRC trial. BMJ 2:567-569, 1979 6. Wagner HP, Dingeldein-Bettler I, Berchthold W, et al: Childhood NHL in Switzerland: Incidence and survival of 120 study and 42 non-study patients. Med Pediatr Oncol 24:281-286, 1995[Medline] 7. Ross JA, Severson RK, Pollock BH, et al: Childhood cancer in the United States, a geographical analysis of cases from the Pediatric Cooperative Clinical Trials Groups. Cancer 77:201-207, 1996[CrossRef][Medline] 8. Bleyer WA, Tejeda H, Murphy SB, et al: National cancer clinical trials: Children have equal access; adolescents do not. J Adolesc Health 21:366-373, 1997[CrossRef][Medline] 9. Liu L, Krailo M, Reaman G, et al: Childhood cancer patients' access to cooperative group cancer programs: A population-based study. Cancer 97:1339-1345, 2003[CrossRef][Medline] 10. Georgia Department of Human Resources, Division of Public Health: Georgia Comprehensive Cancer Registry, Policy and Procedure Manual for Reporting Facilities, http://health.state.ga.us/pdfs/chronic/cancer/PolicyandProceduresManual.pdf 11. Hosmer DW, Lemeshow S: Applied Logistic Regression. New York, NY, John Wiley & Sons, 1999 12. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-482, 1958[CrossRef] 13. Kalbfleisch JD, Prentice RL: The Statistical Analysis of Failure Time Data. New York, NY, John Wiley & Sons, 1980 14. StataCorp: Stata Statistical Software: Release 9.0. College Station, TX, Stata Corporation, 2005 15. Stock W, Sather H, Dodge RK, et al: Outcome of adolescents and young adults with ALL: A comparison of Children's Cancer Group (CCG) and Cancer and Leukemia Group B (CALGB) regimens. Blood 96:467a, 2000 (suppl; abstr) 16. Boissel N, Auclerc M, Lheritier V, et al: Should adolescents with acute lymphoblastic leukemia be treated as old children or young adults? Comparison of the French FRALLE-93 and LALA-94 trials. J Clin Oncol 21:774-780, 2003 17. McDavid K, Tucker TC, Sloggett A, et al: Cancer survival in Kentucky and health insurance coverage. Arch Intern Med 163:2135-2144, 2003 18. Schiffer CA: Differences in outcome in adolescents with acute lymphoblastic leukemia: A consequence of better regimens? Better doctors? Both? J Clin Oncol 21:760-761, 2003 19. Pollock BH, DeBaun MR, Camitta BM, et al: Racial differences in the survival of childhood B-precursor acute lymphoblastic leukemia: A Pediatric Oncology Group Study. J Clin Oncol 18:813-823, 2000 20. Bhatia S: Influence of race and socioeconomic status on outcome of children treated for childhood acute lymphoblastic leukemia. Curr Opin Pediatr 16:9-14, 2004[CrossRef][Medline] 21. Davies SM, Robison LL, Buckley JD, et al: Glutathione S-transferase polymorphisms and outcome of chemotherapy in childhood acute myeloid leukemia. J Clin Oncol 19:1279-1287, 2001 22. Aplenc R, Alonzo TD, Gerbing RB, et al: Ethnicity and survival in childhood acute myeloid leukemia: A report from the Children's Oncology Group. Blood 108:74-80, 2006 23. Baker KS, Anderson JR, Lobe TE, et al: Children from ethnic minorities have benefited equally as other children from contemporary therapy for rhabdomyosarcoma: A report from the Intergroup Rhabdomyosarcoma Study Group. J Clin Oncol 20:4428-4433, 2002 24. Ries LAG, Smith MA, Gurney JG, et al: (eds): Cancer Incidence and Survival among Children and Adolescents: United States SEER Program 1975-1995. NIH Pub. No. 99-4649. Bethesda, MD, National Cancer Institute, SEER Program, 1999 Submitted May 31, 2006; accepted March 16, 2007.
Related Editorial
Related Correspondence
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
