Utilization of cardiac tests in anthracycline‐treated cancer survivors differs between young adults and children: A claims‐based analysis

Abstract Background The Children's Oncology Group Guidelines recommend a cardiacechocardiogram, or comparable functional imaging, following therapy completion in survivors of childhood/adolescent cancers exposed to anthracyclines. Methods Using the 2009–2019 Merative™ MarketScan® Commercial Database, we examined real‐world utilization of cardiac testing among 1609 anthracycline‐treated survivors of childhood/adolescent cancers. Results The cumulative incidence of receiving an initial cardiac test by 5.25 years from the index date (six months after end‐of‐therapy) was 62.3% (95% CI = 57.5%–66.7%), with median time to initial test being 2.7 years (95% CI = 2.5%–3.1%). Young adults (18–28 years) were less likely than children (≤17 years) to receive cardiac testing (hazard ratio [HR] = 0.42, 95% CI = 0.3%–0.49%). More likely to receive cardiac testing were survivors receiving hematopoietic stem cell transplantation versus chemotherapy only (HR = 2.23, 95% CI = 1.63%–3.03%), and survivors with bone or soft tissue versus hematologic cancer (HR = 1.64, 95% CI = 1.30%–2.07%). Conclusions Nearly 40% of anthracycline‐treated survivors of childhood/adolescent cancers had not received cardiac testing within 5.25 years post‐index date, with young adults least likely to receive a test.

Guidelines recommend a cardiac echocardiogram, or comparable functional imaging, following therapy completion in survivors exposed to anthracyclines. 2 There is currently a dearth of evidence on posttreatment uptake of guideline-recommended cardiac testing in exposed survivors.6][7] Using nationwide insurance claims data, we examined real-world utilization of cardiac testing following therapy completion among anthracycline-treated survivors of childhood and adolescent cancers.

| METHODS
We used the 2009-2019 Merative™ MarketScan® Commercial Database, a nationwide convenience sample of outpatient, inpatient, and pharmacy insurance claims from United States employer-sponsored health plans. 8,9e identified survivors who (a) received anthracyclines for hematologic cancers or bone and soft tissue cancers (aged ≤21 years at cancer diagnosis; Table S1); (b) completed cancer therapies; and (c) remained continuously insured for ≥1 year from their index date, defined as 6 months after the end-of-therapy date (Figure S1).In a sensitivity analysis that modified the definition of index date to the end-of-therapy date, we found similar results (available upon request).1][12] We followed the algorithm applied in previous claims data-based studies to identify hematologic cancers and bone and soft tissue cancers, [10][11][12] which represent the majority of anthracycline-treated childhood and adolescent cancers. 13][12] We used the Kaplan-Meier method to estimate the cumulative incidence of receipt of the first cardiac test-including echocardiogram, cardiac magnetic resonance imaging, and/or multiple gate acquisition scan (Table S2)-within 5.25 years post index date, allowing a leeway of 0.25 year (3 months) for potential delays in scheduling medical appointments as the 5-year cut-point for adherence approaches.Censoring occurred at the end of a year if survivors discontinued insurance enrollment in the following year, or at the end of our study period (December 31, 2019).Cumulative incidence curves were generated for all survivors, as well as among subgroups stratified by key sociodemographic characteristics that may contribute to differential utilization of cardiac tests.
Cox proportional hazards regressions were used to identify survivor characteristics that were significantly associated with the first cardiac test receipt occurring within 5.25 years post index date, applying the same censoring approach as described above.The regression models adjusted for survivors' sociodemographic characteristics (age at index, sex, region, rurality, health plan type, year of index date) and clinical factors (cancer type, treatment modality) individually (unadjusted analysis), and then combined (multivariable analysis).The proportional hazards assumption of the Cox model was tested using cumulative sums of martingale-based residuals developed by Lin et al. 14 and no violation was detected.

| RESULTS
We identified 1609 survivors treated with anthracyclines during childhood and adolescence; 48.4% had attained young adult age (18-28 years) at index date, 47.4% were female, and 13.1% were residing in rural areas (Table 1).The majority of our sample were diagnosed with hematologic cancers (90.3%), followed by bone and soft issue cancers (9.7%).Hematopoietic stem cell transplantation was part of cancer therapies for 5.5% survivors, and 42.4% received chemotherapy alone.The most commonly used anthracycline agent was doxorubicin (93.1%).
Overall, the cumulative incidence of receiving an initial cardiac test by 5.25 years post index date was 62.3% (95% confidence interval [CI]: 57.5% to 66.7%), with median time to the initial test receipt being 2.7 years (95% CI: 2.5 to 3.1 years).Among cardiac test recipients (n = 694 survivors), the vast majority underwent an echocardiogram (n = 691 survivors).When stratified by age at index, the cumulative incidence of receiving an initial cardiac test within 5.25 years post index date remained lower in young adults than children ≤17 years (40.3% [95% CI: 34.2% to 46.4%] vs. 83.7%[95% CI: 76.6% to 88.8%]; Gray's test p < 0.001; Figure 1).The cumulative incidence of receiving an initial cardiac test within 5.25 years post index date did not differ significantly by other sociodemographic characteristics, except for geographic region (Figure S2).
In the multivariable regression analysis, young adults were less likely to receive cardiac testing than children (hazard ratio [HR] = 0.42; 95% CI: 0.35 to 0.49; Table 1).More likely to receive cardiac testing were survivors who underwent hematopoietic stem cell transplantation versus chemotherapy only (HR = 2.23; 95% CI: 1.63 to 3.03), and survivors with bone and soft tissue cancers versus hematologic cancers (HR = 1.64; 95% CI: 1.30 to 2.07).In addition, survivors living in the Western region were less This nationwide, claims data-based study of a privatelyinsured cohort showed that nearly 40% of anthracyclinetreated survivors of childhood and adolescent cancer at risk of cardiac dysfunction had not received any cardiac test within 5.25 years following index date, despite national guidelines during the study period detailing the importance of timely screening for early detection of cardiac dysfunction. 15,16Our claims data provide precise measurement of realized cardiac testing and new evidence of age differences, with young adult survivors less likely than children to receive cardiac tests despite having insurance.There are potential explanations for the low uptake of guideline-recommended cardiac testing.First, the transition into survivorship often involves primary care providers, who may not feel well-equipped to provide survivorship care or lack familiarity with the guidelines or survivors' therapy risk. 17,18Inadequate care coordination between primary care providers and oncology specialists may exacerbate care fragmentation.Second, existing guidelines may be too complex to interpret and lack effective dissemination among providers who are not survivorship focused.Third, survivor-level challenges-including limited awareness of future health complications, limited knowledge and self-efficacy to pursue recommended testing and therapies, and logistical barriers including work limitations and travel time-may contribute to low uptake.The psychosocial effects of cancer-such as anxiety, depression, and financial hardship-may further deter survivorship care visits, particularly among young adults.Additionally, young adults face competing responsibilities (e.g., childcare, work schedule) that may limit attention to survivorship care.More research is needed to understand the multi-level causes of low uptake of cardiac testing, including challenges during young adults' transition from pediatric to adult care.
While our approach delineates feasibility of an algorithm for assessing specific screening service utilization in survivors of a childhood or adolescent cancer, 19 several study limitations are notable.First, the MarketScan® database contains a convenience sample derived largely from medium-to-large employers 8,9 ; our findings may not generalize to other childhood and adolescent cancer-survivor populations in the United States.1][22] Second, these claims data do not provide the cumulative dose of anthracycline or capture tests self-paid or paid by public insurance.Similarly, our data lacked information on race and ethnicity.
Third, our claims-based analysis relied on ICD codes that were unable to clearly distinguish hematologic cancer subtypes.Fourth, to minimize sample attrition due to discontinued insurance enrollment over time, we did not assess long-term use of cardiac testing, an area meriting future investigation.In addition, we lacked information on whether survivors' discontinued enrollment was due  Age Groups to death.Since our model does not account for death as a competing risk, we might have overestimated test receipt rate. 23However, sensitivity analysis using a subset of survivors with 5-year continuous enrollment showed results similar to our main analyses (results available upon request).Finally, as adult guidelines from the American Society of Clinical Oncology are discordant with the COG guidelines, 24 the recommended time to initiation of young adult cardiac testing remains ambiguous in the absence of symptoms or abnormal echocardiogram.
Our findings underscore the importance of future interventions to improve utilization of guideline-recommend cardiac tests among survivors of childhood and adolescent cancers.We demonstrate the feasibility of using insurance claims data to identify gaps in real-world cancer survivorship care.This approach informs future work to assess and implement guideline-recommended screening for late effects associated with therapies other than anthracycline.
Unadjusted and adjusted Cox proportional hazards models estimating time to initial cardiac test.
T A B L E 1Abbreviations: CI, confidence interval; HSCT, hematopoietic stem cell transplantation; HR, hazard ratio; Ref, reference.Note: N = 1609.a The index date was defined as 6 months after the end-of-therapy date.b Hematologic cancer includes leukemia and lymphoma.Using insurance claims data, our analysis relied on the International Classification of Diseases (ICD) diagnosis codes, which make it difficult to distinguish subtypes of hematological cancer.c Other types of anthracyclines include mitoxantrone, idarubicin, epirubicin, and alemtuzumab.d A multivariable Cox model was estimated that controlled for all covariates listed in this table.