Predictors of Suboptimal Follow-up in Pediatric Cancer Survivors

Transcription

1 ORIGINAL ARTICLE Predictors of Suboptimal Follow-up in Pediatric Cancer Survivors Leana May, DO, MPH,* David D. Schwartz, PhD,w Ernest Fruge, PhD,z Larry Laufman, EdD,y Suzanne Holm, PhD,8 Kala Kamdar, MD, MS,z Lynnette Harris, PhD,w Julienne Brackett, MD,z Sule Unal, MD,z Gulsah Tanyildiz, MD,z Rosalind Bryant, CNP, PhD,z Hilary Suzawa, MD, MS,z Zoann Dreyer, MD,z and M. Fatih Okcu, MD, MPHz Summary: Attendance to follow-up care after completion of cancer treatment is an understudied area. We examined demographic, clinical, and socioeconomic predictors of follow-up by pediatric cancer patients at a large center in 442 newly diagnosed patients using multivariable logistic regression analyses. Patients who did not return to clinic for at least 1000 days were considered lost to follow-up. Two hundred forty-two (54.8%) patients were lost. In multivariable analyses, the following variables were independent predictors of being lost to follow-up: treatment with surgery alone (odds ratio [OR] = 6.7; 95% confidence interval [CI], ), older age at diagnosis (reference, 0 to 4; ages, 5 to 9: OR = 1.8, 95% CI, 1.1-3; ages, 10 to 14: OR = 3.3; CI, ; and ages, 15 and above: OR = 4.8; CI, ), lack of history of stem cell transplantation (OR = 2, 95% CI, ) and lack of insurance (OR = 3.4; CI, ). Hispanic patients had the best follow-up rates (53.7%) compared to whites and blacks (P = 0.03). Attendance to long-term follow-up care is suboptimal in childhood cancer survivors. Predictors that were associated with nonattendance can be used to design targeted interventions to improve follow-up care for survivors of pediatric cancer. Key Words: childhood cancer, late effects, clinic attendance, survivorship (J Pediatr Hematol Oncol 2017;39:e143 e149) Since 1975, overall pediatric cancer 5-year survival rates have improved from 58% to over 80%, leading to a Received for publication June 2, 2016; accepted November 15, From the *Section of Emergency Medicine, Children s Hospital of Colorado, Aurora, CO; wdepartment of Pediatrics, Section of Psychology; zdepartment of Pediatrics, Section of Hematology- Oncology, Texas Children s Cancer Center; ydepartment of Medicine, Section of General Medicine, Baylor College of Medicine, Houston; and 8Department of Psychiatry, Pauline Allen Gill Center for Cancer and Blood Disorders, Dallas, TX. L.M. was supported by a cancer prevention fellowship funded by/ through NCI training grant R25 CA57730, Robert M. Chamberlain, PhD, Principal Investigator, The University of Texas MD Anderson Cancer Center. Presented at the 24th Annual Meeting of the American Society of Pediatric Hematology Oncology: Brackett J, Holm S, Laufman L, et al. Predictors of adherence to long-term follow-up care in pediatric cancer survivors; Baltimore, MD, April, The authors declare no conflict of interest. Reprints: M. Fatih Okcu, MD, MPH, Department of Pediatrics, Section of Hematology-Oncology, Texas Children s Cancer Center, Baylor College of Medicine, Clinical Care Center, Suite , 6621 Fannin, CC , Houston, TX ( mfokcu@txch.org). rapidly growing population of childhood cancer survivors in the United States estimated in 2010 to be >379,000. 1,2 Childhood cancer survivors are often burdened by lingering sequelae of the illness or treatment (eg, physical or cognitive disabilities), and have an increased risk of future morbidities (eg, recurrence or second cancer). Many have no symptoms immediately after completion of treatment and most are not aware of the potential future risk for multiple chronic disorders. It has been estimated that 60% to 95% of cancer survivors will develop one or more subsequent health conditions, many of which will be severe, resulting in a significantly greater likelihood of disability and early death. 3 5 Consequently, for the past decade, risk-based long-term follow-up of pediatric cancer patients has been considered vital in order to provide screening for late effects, appropriate health counseling, anticipatory guidance, and education on how to reduce risks. 6,7 The Institute of Medicine has called for lifelong risk-based screening and surveillance to help ensure early identification of emergent complications, 8,9 and the Children s Oncology Group has developed evidence-based guidelines for risk-based followup care. 10,11 Long-term follow-up is especially important for patients at high risk for late effects such as cardiopulmonary disease, secondary malignancies, and neurocognitive deficits associated with specific treatments (eg, radiation, certain chemotherapy agents). 2,4 Despite the recent development of risk-based guidelines, follow-up for long-term survivorship care remains suboptimal. Fewer than 1 in 5 patients in the Childhood Cancer Survivor Study (CCSS) reported having a medical visit directly related to late-effect surveillance in the preceding 2 years, and as many as 1 in 10 reported that they did not visit any health care provider at all. 6,12 In the British CCSS, only 36.4% of survivors reported regular hospital follow-up care related to their cancer diagnosis. 13 Adherence to risk-based screening recommendations such as echocardiogram, mammography, and colonoscopy is even lower. 14,15 There is limited but growing literature on potential reasons why cancer survivors do not attend follow-up care. Lower income, lack of insurance, and being male are likely risk factors for lower clinic attendance, 6 as are deficits in survivors cognitive function, and their understanding of their cancer diagnosis, treatment effects, and long-term cancer-related health risks. 11,12,16,17 There is evidence that even pediatric oncologists may not always be aware of and follow guidelines for late effects surveillance. 18 However, our knowledge of factors associated with lower clinic J Pediatr Hematol Oncol Volume 39, Number 3, April e143

2 May et al J Pediatr Hematol Oncol Volume 39, Number 3, April 2017 attendance remains limited. For example, risk factors for poor clinic attendance such as lack of patient understanding and provider awareness of long-term risks may be especially salient for patients seen in community settings, where the majority of survivors receive their long-term care. 6 Conversely, risk for nonattendance may be attenuated in cancer centers that offer continued care to long-term survivors, or different risk factors may emerge as salient in these settings. The primary aim of this study was to examine predictors of nonattendance to long-term followup clinic appointments at a large urban cancer center with a dedicated long-term survivorship specialty clinic. MATERIALS AND METHODS Study Center Description We performed a retrospective chart review of pediatric cancer survivors diagnosed and treated at the Texas Children s Cancer Center (TCCC). The TCCC is one of the largest pediatric cancer centers in the Unites States, treating over 500 new cancer patients diagnosed from infancy to age 21 years each year. Patients who have completed therapy and have no current evidence of disease are followed by their primary pediatric oncology team for 2 to 5 years (depending on cancer diagnosis), after which they are referred to the multidisciplinary long-term survivorship clinic for lifelong cancer-related follow-up care. This clinic is located within the TCCC and is staffed by a physician boarded in internal medicine and pediatrics, a nurse practitioner, a psychologist and a social worker. Patients are seen annually or less often based on their diagnosis, previous therapy, and the risk-based Children s Oncology Group Long-Term Follow-Up Guidelines. 10,11 Patients are followed through this clinic into adulthood with no upper age limit, while they also continue routine health care with a primary care provider. Patients are mailed a reminder card to schedule an appointment 2 months prior to their due visit. At their first survivorship clinic visit, the patient (or parent for minor patients) is provided with a written summary of his or her treatment history and a plan for future risk-based medical follow-up. Study Population This study was approved by the Baylor College of Medicine Institutional Review Board. We reviewed institutional databases and identified a total of 1086 patients seen for a new office visit between January 1, 1998 and December 31, 2001 aged 21 years or younger at the time of the first visit. Inclusion criteria were: (1) a diagnosis of a neoplasm with completion of any required treatment including surgical resection, radiation therapy, or chemotherapy; (2) no evidence of disease at last visit; and (3) documented recommendation for further follow-up through TCCC. Patients were excluded if they had died due to any cause (n = 414), were seen for a 1-time visit (consultation or second opinion) (n = 148), or were referred to another institution for care (n = 82), leaving a total of 442 patients available for analysis. Deaths were identified by indication in the chart or the national social security death index database search. Data Collection Abstraction of medical and appointment history data was completed in April, Data on the following variables were abstracted: date of diagnosis and last follow-up visit; age at diagnosis (categorized into 5-year bands: 0 to 4, 5 to 9, 10 to 14, and 15 y and above); sex; race/ethnicity (non-hispanic white, black, Hispanic, and other); cancer type (leukemia, lymphoma, central nervous system [CNS] tumors, and solid tumors); treatment modality (chemotherapy [C] only, radiation [R] only, surgery [S] only, C + R, C + S, R + S, and C + R + S); history of stem cell transplantation; and health insurance at last visit (active or lacking). Patients were considered lost to follow-up if they had not attended a recommended appointment through the TCCC (either with their primary oncology team or through the long-term survivor clinic) for at least 1000 days past their last recommended follow-up appointment. If patients do not schedule or miss a scheduled appointment, 3 attempts are made to contact them by phone, after which a certified letter is sent recommending that they contact us to schedule. No patient classified as lost in this study had an appointment scheduled for a future date. Our active followup group was defined as having had an appointment within 1000 days of data extraction in January 2014 thus, their last recorded appointment could not have occurred prior to April 2011 for them to still be considered active. Given that our cohort was diagnosed between 1998 and 2002, the youngest a patient could be, and still be considered in active follow-up in 2011, was B9 years old. Our selection criteria therefore artificially excluded the possibility of patients having their last appointment before age 9 and still being considered in active follow-up. Because of this artifact of our selection criteria, we were unable to look at age at last follow-up as a variable in our multivariable analyses. Statistical Analyses Descriptive statistics were calculated using standard methods for the total study sample and compared between patients in active treatment versus patients lost to follow-up using 2-sample t test for continuous and univariate logistic regression for categorical variables. Variables that were statistically significant in univariate analyses were coded using indicator coding and entered into a multivariable logistic regression model to compute the relative odds of being lost to follow-up for Z1000 days. The final model included only the variables that remained statistically significant. Analyses were conducted using PAS Statistics version 23 for Windows (IBM SPSS Inc., Chicago, IL). All comparisons used a 2-sided test, and a P-value of <0.05 was considered statistically significant. RESULTS Descriptive Characteristics Descriptive study variables are shown in Table 1. The sample was 56.1% female and ethnically diverse, with minorities comprising almost 55% of participants. Mean age at diagnosis was 6.8 (SD = 5.2; median = 4.1) years, and mean age at last follow-up was 13.4 (SD = 5.7; median = 12.8) years. Mean time since diagnosis was 13.4 (SD = 1.2, median = 13.4) years at last clinic visit. Leukemia was the most common diagnosis (39.8%), and chemotherapy alone was the most common treatment modality (36.7%). Treatment type was associated with cancer diagnosis 88% of patients who received chemotherapy alone had leukemia. Conversely, nearly all patients (96.7%) who had surgery (either alone or in combination with another therapy) had a CNS or solid tumor. Most e144

3 J Pediatr Hematol Oncol Volume 39, Number 3, April 2017 Long-Term Follow-up in Childhood Cancer Patients TABLE 1. Univariate Comparison of Characteristics of Patients in Active Follow-up and Patients Lost to Follow-up After Therapy Completion n (%) P All Patients (N = 442) Active Follow-up (n = 200) Lost to Follow-up (n = 242) Age at diagnosis (mean) (y) < Age at diagnosis (categorical) (y) (48.0) 120 (60.0) 92 (38.0) < (24.9) 49 (24.5) 61 (25.2) (18.6) 23 (11.5) 59 (24.4) (inclusive) 38 (8.6) 8 (4.0) 30 (12.4) Time since diagnosis (mean) (y) Sex 0.1 Female 248 (56.1) 110 (55.0) 138 (57.0) Male 194 (43.9) 90 (45.0) 104 (43.0) Race/ethnicity White 196 (45.1) 86 (43.4) 110 (46.4) Black 49 (11.3) 13 (6.6) 36 (15.2) Hispanic 180 (41.4) 96 (48.5) 84 (35.4) Other 10 (2.3) 3 (1.5) 7 (3.0) Insurance 0.01 Yes 415 (93.9) 194 (97.0) 221 (91.3) No 27 (6.1) 6 (3.0) 21 (8.7) Diagnosis < Leukemia 176 (39.8) 103 (51.5) 73 (30.2) Lymphoma 43 (9.7) 17 (8.5) 26 (10.7) Solid tumor 149 (33.7) 55 (27.5) 94 (38.8) CNS tumor 74 (16.7) 25 (12.5) 49 (20.2) Treatment < Chemotherapy alone 162 (36.7) 95 (47.5) 67 (27.7) Radiation alone 6 (1.4) 1 (0.5) 5 (2.1) Surgery alone 60 (13.6) 11 (5.5) 49 (20.2) Chemotherapy + radiation 66 (14.9) 30 (15.0) 36 (14.9) Chemotherapy + surgery 66 (14.9) 29 (14.5) 37 (15.3) Radiation + surgery 14 (3.2) 2 (1.0) 12 (5.0) Chemotherapy + radiation + surgery 68 (15.4) 32 (16.0) 36 (14.9) Stem cell transplant 0.03 No 378 (85.5) 163 (81.5) 215 (88.8) Yes 64 (14.5) 37 (18.5) 27 (11.2) P-values indicate significance of univariate predictors of follow-up status (active vs. lost). Global P-value is calculated by logistic regression for categorical variables and 2-sample t test for continuous variables. CNS indicates central nervous system. patients (93.9%) had health insurance at the time of their last clinic visit. Univariate Analyses More than half of the patients (n = 242, 54.8%) were lost to follow-up, at an average of 6.9 years (range, 0.1 to 14.6 y) from diagnosis. Patients who were in active followup were diagnosed at an earlier age on average than patients who were lost (5.5 vs. 7.8 y, P < 0.001). The groups did not differ in time since diagnosis (13.3 vs y, P = 0.17) or in sex composition (55.0% vs. 57.0% female, P = 0.1). Hispanics had the lowest proportion of lost to follow-up patients (46.7%) compared to others (P = 0.007; Table 1). In univariate analyses, the following variables were associated with being lost to follow-up: a diagnosis of lymphoma (odds ratio [OR] = 2.8; 95% confidence interval [CI], ; P < 0.001), solid tumor (OR = 2.4; CI, ; P < 0.001), or CNS tumor (OR = 2.2; CI, ; P = 0.02) compared to leukemia; treatment with surgery alone (OR = 6.3; CI, ; P < 0.001), surgery plus radiation (OR = 8.5; CI, ; P < 0.001), or surgery plus chemotherapy (OR = 1.8; CI, ; P = 0.04) compared to chemotherapy alone; black race/ethnicity (OR = 2.2; CI, ; P = 0.03) compared to white; older age at diagnosis (ages, 5 to 9: OR = 1.6, CI, ; ages, 10 to 14: OR = 3.3, CI, ; and ages, 15 and above: OR = 4.9, CI, ; P < 0.001) compared to age 0 to 4 years; and lack of insurance (OR = 3.1; CI, ; P = 0.01). Collapsing across treatment types, patients treated with any surgery (alone or in combination) were more likely to be lost to follow-up than patients who did not receive surgery (64.4% vs. 46.2%; OR = 2.1; CI, ; P < 0.001). In contrast, patients treated with chemotherapy (alone or in combination) were less likely to be lost to follow-up than patients who did not receive any chemotherapy (48.6% vs. 82.5%; OR = 0.2; CI, ; P < 0.001). There was no effect of radiation therapy in either direction. Multivariable Analyses The following 21 categorical variables were entered into a multivariable logistic regression: diagnosis (leukemia, e145

4 May et al J Pediatr Hematol Oncol Volume 39, Number 3, April 2017 lymphoma, CNS tumor, solid tumor), treatment (chemotherapy, radiation, surgery, chemotherapy + radiation, chemotherapy + surgery, radiation + surgery, chemotherapy + radiation + surgery), stem cell transplant, race/ethnicity (white, black, Hispanic, other), age group at diagnosis (0 to 4, 5 to 9, 10 to 14, 15 to 20 y), and insurance status. As can be seen in Table 2, the following variables remained statistically significant predictors of loss to follow-up: age at diagnosis (P < 0.001); race/ethnicity (P = 0.03); lack of insurance (P = 0.02); cancer treatment modality (P = 0.014); and lack of treatment with stem cell transplantation (P = 0.04). Patients who were older at diagnosis were more likely to be lost to follow-up, with the odds of being lost increasing significantly in each 5-year age band. Patients without insurance at their last clinic visit were also more likely to be lost (OR = 3.4; 95% CI, ). Treatment with surgery alone was associated with a >6-fold increase in odds for being lost compared to chemotherapy alone (OR = 6.7; CI, ). Although cancer type was a strong predictor of follow-up in univariate analyses, it was no longer statistically significant in the multivariable model. Time from diagnosis was not a statistically significant variable and the results did not change when this variable was included in the model. DISCUSSION It is concerning to report that over half of the study patients were lost to follow-up on average 6.9 years after diagnosis at a tertiary care children hospital with a dedicated survivorship clinic integrated into standard followup. Although this follow-up rate is suboptimal, it is higher than the cancer-focused care follow-up rates reported in large multi-institution studies, including the CCSS, 6,19 the British CCSS, 13 and the Swiss CCSS, 20 suggesting that having specialty care available may help increase clinic follow-up. A study of multiple survivorship clinics in Canada reported a follow-up rate similar to ours (43.3%). 21 In contrast, follow-up in our study was generally lower than what was reported in other single institution studies with a dedicated long-term survivor clinics Only 15% of patients did not attend scheduled follow-up appointments in a study at St. Jude Children s Research Hospital, 22 one of the premier childhood cancer centers in the United States. St. Jude provides significant financial support for medical care, lodging, and transportation to patients, reducing many of the barriers likely faced by patients in our cohort and most pediatric cancer centers in the United States. Similar to St. Jude, nonattendance was only 18% at the Yorkshire Regional Centre for Pediatric Oncology and Hematology Long Term Follow-up Clinic in the United Kingdom, 23 but substantial differences between the US and UK health care systems makes institutional comparison difficult. Thus, both of these institutions may not be representative of the care available to most pediatric cancer survivors in the United States. At the Children s Hospital of Philadelphia, 24 25% of patients were not seen for survivorship care within 5 years of diagnosis. The authors found that distance to the hospital was a strong predictor of follow-up. Although we did not have data on distance to clinic for most patients, our center has a very wide catchment area, potentially elevating our number of patients who were lost to follow-up. In contrast, a study of the survivorship clinic at Yale found that 72.2% of patients did not have a long-term survivorship visit. 25 This high rate of nonattendance may have reflected the high proportion of uninsured patients in their cohort (28.8%, vs. only 6.1% in our cohort), especially as insurance status was the single sociodemographic predictor of follow-up in their study. Taken together, findings across studies are quite concerning for follow-up patient care. At the same time, the differences TABLE 2. Multivariable Analyses for Being Lost to Follow-up after Therapy Completion in Pediatric Cancer Patients Variables Lost to Follow-up (n [%]) OR (95% CI) P Age at diagnosis (y) < (referent) 212 (43.4) (55.5) 1.8 (1.1-3) (72.0) 3.3 ( ) (inclusive) 38 (78.9) 4.8 (2-11.7) Race/ethnicity 0.03 White (referent) 196 (56.1) Black 49 (73.5) 2.0 ( ) Hispanic 180 (46.7) 0.7 ( ) Other 10 (70.0) 2.0 ( ) Insurance 0.02 Yes (referent) 415 (53.3) No 27 (77.8) 3.4 ( ) Treatment < Chemotherapy alone (referent) 162 (41.4) Radiation alone 6 (83.3) 3 (0.3-28) Surgery alone 60 (81.7) 6.7 ( ) Chemotherapy + radiation 66 (54.5) 1.7 ( ) Chemo + surgery 66 (56.1) 1.4 ( ) Radiation + surgery 14 (85.7) 5.6 ( ) Chemotherapy + radiation + surgery 68 (52.9) 1.6 ( ) Stem cell transplant 0.04 Yes (referent) 64 (42.2) No 378 (56.9) 2 ( ) Global overall effect P-value is calculated by logistic regression. CI indicates confidence interval; OR, odds ratio. e146

5 J Pediatr Hematol Oncol Volume 39, Number 3, April 2017 Long-Term Follow-up in Childhood Cancer Patients between the findings of these studies suggest ways to facilitate follow-up care by reducing financial and practical barriers to attendance. In our study, treatment modality was a strong predictor of long-term follow-up. As treatment is closely associated with different risks and different recommendations for follow-up, this finding can give insight into factors that can be modified to better facilitate follow-up for care. Continued follow-up was highest among patients treated with chemotherapy alone (58.6%) or in combination with other therapies, consistent with other research. 24,25 Patients who receive chemotherapy are at higher risk for late effects and tend to have better long-term follow-up, perhaps as the result of better counseling about long-term risks. 20 In addition, chemotherapy patients generally see their providers more frequently and for a longer time than other cancer groups, so they may develop greater familiarity with and attachment to their treatment team. A growing body of research suggests the powerful impact of provider-patient relationships on adherence to medical care. 26 Developing patients ties to the care team may be an especially important aspect of improving long-term follow-up. Patients treated with surgery alone were more likely to be lost to follow-up by our criteria, and this was equally true of both solid and CNS tumors. Interestingly, a similar finding was reported by Barakat et al, 24 who also found that brain tumor and solid tumor patients were less likely to attend survivorship follow-up visits compared to other patient groups and others have also reported reduced attendance rates for patients who had surgery alone. 21,25 Solid tumor patients receive less instruction regarding the need to screen for late effects compared to those receiving chemotherapy and radiation; however, this is not true of brain tumor patients, who had equally poor follow-up. Another possibility is that surgery patients are more likely to believe themselves cured, 27 which may lead them to conclude that follow-up is unnecessary. 20 Our findings suggest that providers may need to find more effective ways to educate surgery patients about the importance of receiving long-term follow-up care, especially given the lack of awareness and understanding of long-term risks that has been documented in pediatric cancer patients. 11 Education may be particularly important for surgery patients with CNS tumors, as they have a 2 to 3-fold higher incidence of cancer recurrence compared to other diagnostic groups. 28 One other possible way to improve follow-up rates in this group is to educate our surgical oncology colleagues about the need and risks for late effects. There are very little published data evaluating the impact of surgery on late effects independent of other modalities in cancer survivors. Pediatric low grade glioma survivors who were treated with surgical resection alone had increased rates of multiple adverse effects including neurocognitive impairment, seizure disorders, vision loss, endocrinopathies and psychological disorders. 29 In a recent review of the literature in adults with sarcomas half of the patients with extremity tumors had an impairment in their muskuloskeletal systems primarily attributed to their surgical intervention. 30 Consistent with other studies, socioeconomic and demographic factors were also important predictors of follow-up. The rate of being lost to follow-up was especially high among patients who lacked insurance (77.8%), children aged 10 or older at diagnosis (74.2%), and black patients (73.5%). Being uninsured is an obvious risk factor for suboptimal follow-up care 19,25 and poor attendance to medical care is a well-known phenomenon among adolescents and black Americans. 6,12,20,21,23,25,31 Efforts are needed to reduce these disparities in service utilization. Interestingly, Hispanic patients in our sample were less likely to be lost to follow-up (46.7% vs. 60.3% for non- Hispanic groups). A similar trend was evident in the Yale study, though the finding did not quite reach statistical significance. 25 Hispanic patients in the CCSS had follow-up rates comparable to non-hispanic whites, 6,19,32 despite the fact that Hispanics in the United States are more than twice as likely to be uninsured than non-hispanic whites 33 and tend to have lower income. 32 In contrast to these studies, a study from California using statewide registry data reported an increased risk for Hispanic pediatric cancer survivors for being lost to follow-up. 34 Taken together, these findings suggest that cultural differences may be important factors in whether patients seek follow-up care, though the relationship is likely to be complex, given the heterogeneity within the Hispanic population. 35 Issues to consider and that may contribute to study differences include language, level of acculturation, and country of origin. Further research is needed to better understand these trends. Interpretation of our findings must be tempered by a number of limitations. First, this was a retrospective chart review, and we do not have data on the current care of the patients considered lost to follow-up. It is possible that some of these patients are receiving cancer-focused care elsewhere, but we believe that this is likely a small number, given the limited options for cancer-focused care in our catchment area. It seems more likely that these patients were either receiving general medical care in the community or no follow-up care at all, as was true of patients in the CCSS. Some patients may also have moved. Although our sample size was moderate for a single-center pediatric survivorship study we acknowledge that a larger sample size is needed to obtain more precise ORs and 95% CIs in the multivariable analyses. In contrast, generalizability was enhanced by the diverse sample, which was 40% Hispanic and 11% black, and similar to most pediatric cancer centers in the United States in terms of available support to facilitate long-term follow-up. We also lacked data on other potentially important contributors to suboptimal followup, including patients socioeconomic status, and the distance they would have to travel to our medical center. However, the effects of socioeconomic status and distance are well known, and inclusion of these variables is not likely to have changed our main findings. Finally, it remains possible that individual treatment providers gave different follow-up recommendations to their patients, despite the policy in our center to refer all off-therapy cancer patients for long-term survivorship care 2 to 5 years after completion of therapy. Investigating providers actual practice in educating patients and recommending long-term care is a critical area for future research. More work is also needed to examine interventions focused on improving follow-up for care. In one recent study from a center in the United Kingdom, investigators decreased the proportion of childhood cancer survivors who were lost to follow-up from 46% to 36% with a simple intervention, a postal questionnaire. 36 Internet-based software programs such as Passport for Care, developed to enhance survivorship care, are also promising interventions that have the added potential of keeping survivors linked into a system capable of providing appropriate risk-based follow-up e147

6 May et al J Pediatr Hematol Oncol Volume 39, Number 3, April 2017 Currently, there is no consensus on the best model in which to provide follow-up care to survivors of childhood cancers. Patients may continue to receive care at the same pediatric cancer center in which they were initially treated, transition into a survivorship program or to an adult cancer center, or they may transition out of cancer care altogether into the community primary care system. 38,39 The findings from this study suggest that follow-up appears to be enhanced at institutions with long-term survivorship programs such as ours, at least compared to attendance rates reported in multi-institution studies. Nonetheless, studies are needed to directly compare the different models of providing follow-up care to determine which is most effective in the diagnosis, prevention, and treatment of childhood cancer survivors based on their relative risk for therapy-related late effects. Our observations extend previous research into attendance to long-term follow-up care for survivors of pediatric cancer by identifying predictors applicable to the majority of the institutions providing service to this population in the United States. Given the documented risks facing this growing population, this is a vital area of research and concern. Focus group studies are underway at our center to inform the development of an assessment tool to identify patients who at increased risk for being lost to follow-up and who might benefit most from targeted interventions to promote attendance. REFERENCES 1. Horner MJ, Ries LAG, Krapcho M, et al. SEER Cancer Statistics Review, Bethesda, MD: National Cancer Institute; Based on November 2008 SEER data submission, posted to the SEER Web site. 2. Robison LL, Hudson MM. Survivors of childhood and adolescent cancer: life-long risks and responsibilities. Nat Rev Cancer. 2014;14: Oeffinger KC, Mertens AC, Sklar CA, et al. Chronic health conditions in adult survivors of childhood cancer. N Engl J Med. 2006;355: Hudson MM, Ness KK, Gurney JG, et al. Clinical ascertainment of health outcomes among adults treated for childhood cancer. JAMA. 2013;309: Geenen MM, Cardous-Ubbink MC, Kremer LC, et al. Medical assessment of adverse health outcomes in long-term survivors of childhood cancer. JAMA. 2007;297: Nathan PC, Greenberg ML, Ness KK, et al. Medical care in long-term survivors of childhood cancer: a report from the childhood cancer survivor study. J Clin Oncol. 2008;26: Byrne J, Lewis S, Halamek L, et al. Childhood cancer survivors knowledge of their diagnosis and treatment. Ann Intern Med. 1989;110: Kadan-Lottick NS, Robison LL, Gurney JG, et al. Childhood cancer survivors knowledge about their past diagnosis and treatment: childhood Cancer Survivor Study. JAMA. 2002;287: Oeffinger KC, Hudson MM. Long-term complications following childhood and adolescent cancer: foundations for providing risk-based health care for survivors. CA Cancer J Clin. 2004;54: Oeffinger KC. Longitudinal risk-based health care for adult survivors of childhood cancer. Curr Probl Cancer. 2003;27: Landier W, Bhatia S, Eshelman DA, et al. Development of risk-based guidelines for pediatric cancer survivors: the Children s Oncology Group Long-Term Follow-Up Guidelines from the Children s Oncology Group Late Effects Committee and Nursing Discipline. J Clin Oncol. 2004;22: Hewitt M, Simone JV. Ensuring Quality Cancer Care. Washington, DC: National Academies Press; Rebholz CE, Reulen RC, Toogood AA, et al. Health care use of long-term survivors of childhood cancer: the British Childhood Cancer Survivor Study. J Clin Oncol. 2011;29: Krull KR, Annett RD, Pan Z, et al. Neurocognitive functioning and health-related behaviours in adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. Eur J Cancer. 2011;47: Nathan PC, Ness KK, Mahoney MC, et al. Screening and surveillance for second malignant neoplasms in adult survivors of childhood cancer: a report from the childhood cancer survivor study. Ann Intern Med. 2010;153: National Cancer Policy Board. Childhood Cancer Survivorship: Improving Care and Quality of Life. Washington, DC: The National Academies Press; Kimberg CI, Klosky JL, Zhang N, et al. Predictors of health care utilization in adult survivors of childhood cancer exposed to central nervous system-directed therapy. Cancer. 2015;121: Henderson TO, Hlubocky FJ, Wroblewski KE, et al. Physician preferences and knowledge gaps regarding the care of childhood cancer survivors: a mailed survey of pediatric oncologists. J Clin Oncol. 2010;28: Casillas J, Oeffinger KC, Hudson MM, et al. Identifying predictors of longitudinal decline in the level of medical care received by adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. Health Serv Res. 2015;50: Michel G, Kuehni CE, Rebholz CE, et al. Can health beliefs help in explaining attendance to follow-up care? The Swiss childhood cancer survivor study. Psychooncology. 2011;20: Nathan PC, Agha M, Pole JD, et al. Predictors of attendance at specialized survivor clinics in a population based cohort of adult survivors of childhood cancer. J Cancer Surviv. 2016;10: Klosky JL, Cash DK, Buscemi J, et al. Factors influencing long-term follow-up clinic attendance among survivors of childhood cancer. J Cancer Surviv. 2008;2: Johnson R, Horne B, Feltbower RG, et al. Hospital attendance patterns in long term survivors of cancer. Arch Dis Child. 2004;89: Barakat LP, Schwartz LA, Szabo MM, et al. Factors that contribute to post-treatment follow-up care for survivors of childhood cancer. J Cancer Surviv. 2012;6: Zheng DJ, Sint K, Mitchell H, et al. Patterns and predictors of survivorship clinic attendance in a population based sample of pediatric and young adult childhood cancer survivors. J Cancer Surviv. 2016;10: Zolnierek KB, DiMatteo MR. Physician communication and patient adherence to treatment: a meta-analysis. Med Care. 2009;47: American Cancer Society. A Guide to Cancer Surgery treatmenttypes/surgery/surgery-learning-about-cancer-surgery. Accessed October 11, Wasilewski-Masker K, Liu Q, Yasui Y, et al. Late recurrence in pediatric cancer: a report from the Childhood Cancer Survivor Study. J Natl Cancer Inst. 2009;101: Turner CD, Chordas CA, Liptak CC, et al. Medical, psychological, cognitive and educational late-effects in pediatric low-grade glioma survivors treated with surgery only. Pediatr Blood Cancer. 2009;53: Parsons JA, Davis AM. Rehabilitation and quality-of-life issues in patients with extremity soft tissue sarcoma. Curr Treat Options Oncol. 2004;5: Butow P, Palmer S, Pai A, et al. Review of adherence-related issues in adolescents and young adults with cancer. J Clin Oncol. 2010;28: Bhatia S, Gibson TM, Ness KK, et al. Childhood cancer survivorship research in minority populations: a position paper e148

7 J Pediatr Hematol Oncol Volume 39, Number 3, April 2017 Long-Term Follow-up in Childhood Cancer Patients from the Childhood Cancer Survivor Study. Cancer. 2016;122: CDC. Summary Health Statistics: National Health Interview Survey, Accessed August 10, Milam JE, Meeske K, Slaughter RI, et al. Cancer-related followup care among Hispanic and non-hispanic childhood cancer survivors: the Project Forward study. Cancer. 2015;121: Weinick RM, Jacobs EA, Stone LC, et al. Hispanic healthcare disparities: challenging the myth of a monolithic Hispanic population. Med Care. 2004;42: Edgar AB, Borthwick S, Duffin K, et al. Survivors of childhood cancer lost to follow-up can be re-engaged into active long-term follow-up by a postal health questionnaire intervention. Eur J Cancer. 2012;48: Horowitz ME, Fordis M, Krause S, et al. Passport for care: implementing the survivorship care plan. J Oncol Pract. 2009;5: Aziz NM, Oeffinger KC, Brooks S, et al. Comprehensive longterm follow-up programs for pediatric cancer survivors. Cancer. 2006;107: Kenney LB, Bradeen H, Kadan-Lottick NS, et al. The current status of follow-up services for childhood cancer survivors, are we meeting goals and expectations: a report from the Consortium for New England Childhood Cancer Survivors. Pediatr Blood Cancer. 2011;57: e149