Primary Care Utilization and Cardiovascular Screening in Adult Survivors of Childhood Cancer

Key Points Question How do adult survivors of childhood cancer use health care and cardiovascular screening with their primary care practitioners (PCPs)? Findings In this cross-sectional study including 293 survivors of childhood cancer at high risk for cardiovascular complications, participants’ PCP records had infrequent documentation referencing a cancer history (67.6%) or a need for late-effects surveillance (32.4%), and only 21.5% of participants records had a completed or planned echocardiogram in the prior 2 years. Factors associated with up-to-date cardiac screening included documentation of increased cardiovascular risks or a late-effects surveillance plan. Meaning These findings suggest that increasing participant and PCP awareness of risks and surveillance recommendations may improve adherence to screening.


Introduction
5][6][7] Cancer-related treatment exposures place survivors at up to 5-fold increased risk for cardiovascular disease and related death. 4,5,7Importantly, modifiable cardiovascular risk factors, such as hypertension, dyslipidemia, and diabetes, are more prevalent among survivors of cancer compared with the general population, and these risk factors contribute a proportionally greater impact on cardiovascular mortality compared with their impact on the general population. 6,7reening for and managing these risk factors following cancer therapy may reduce long-term morbidity and mortality from cardiovascular disease. 8Multiple guidelines exist for survivors of childhood cancer, particularly those at increased risk of heart disease due to prior treatment exposures. 9-13These guidelines recommend regular screening for hypertension, dyslipidemia, and diabetes.Early detection of anthracycline-associated cardiomyopathy using surveillance echocardiography or equivalent modalities has also been recommended.5][16][17] However, little is known about health care utilization patterns in this phase of survivorship, including the testing for and prevention of cardiovascular disease.
As part of a randomized intervention trial to improve control of cardiovascular risk factors among long-term survivors of cancer, 18 we examined cardiovascular screening and health care utilization patterns among survivors of childhood cancer at high risk for cardiovascular complications.
To achieve this, we comprehensively reviewed medical record information from participants' PCPs, covering the 2-year period leading up to trial enrollment, with the goal of identifying factors associated with cardiac screening.Furthermore, as many survivors of childhood cancer lack critical knowledge regarding their own treatment exposures and risk for late effects, [19][20][21] we compared the accuracy of PCP records and survivor self-report against research records from survivors' original cancer treatment institutions.Deficits in both PCPs' and survivors' knowledge can highlight important targets to improve communication among oncology, survivorship, and primary care teams to increase adherence to recommended screening.

Methods
This was a cross-sectional study of adult survivors of childhood cancer enrolled in a randomized trial

JAMA Network Open | Pediatrics
Primary Care Utilization and Cardiovascular Screening in Adult Survivors of Childhood Cancer

Participants
The Childhood Cancer Survivor Study (CCSS) is an ongoing multi-institutional cohort study including approximately 25 700 five-year survivors of the most common types of childhood cancer (leukemia, lymphoma, renal tumors, sarcomas, and central nervous system malignant neoplasms) diagnosed before age 21 years between 1970 and 1999.The CCSS has abstracted cancer treatment exposures within 5 years of diagnosis from treating institutions' records, including radiation therapy, chemotherapy, and surgical procedures. 22The Communicating Health Information and Improving Coordination With Primary Care study (CHIIP) is a randomized clinical trial conducted within CCSS to test the efficacy of a personalized survivorship care plan (SCP) intervention tailored to improve control (ie, reduce undertreatment) of hypertension, dyslipidemia, and diabetes. 18For CHIIP, eligible participants were members of the CCSS cohort estimated to be at elevated risk for future cardiovascular disease, defined as an approximately 10% or greater risk of ischemic heart disease or heart failure by age 50 years, based on risk models incorporating demographic and cancer treatment exposures. 23,24We excluded participants with a history of known ischemic heart disease or heart failure at enrollment.Interested participants underwent a home visit by a trained examiner recruited between September 2017 and April 2021.Individuals found to have 1 or more potentially underdiagnosed (ie, previously unknown diagnosis) or undertreated (ie, known diagnosis but measured values out of the therapeutic reference range) cardiovascular risk factors (ie, hypertension, dyslipidemia, and glucose intolerance) were then randomized 1:1 into an intervention or a control group, and their PCP medical records were requested.Details on study eligibility and measurements have been published elsewhere. 18Randomized participants who were unable or unwilling to sign a Health Insurance Portability and Accountability Act waiver granting release of medical records and individuals whose PCP's office could not or would not provide medical records were excluded from this analysis.

Data Collection and Variables
At the time of consent, we asked all participants to list their PCPs over the 2 years preceding trial enrollment.We then requested PCP outpatient clinic medical records over this time frame, including clinician notes, medication lists, laboratory results, and imaging reports for all participants.Through manual record review, we ascertained the number of PCP and specialist visits, as well as documentation of the testing for, diagnosis of, and treatment of hypertension, dyslipidemia, and diabetes.We determined whether there was documentation of participants' cancer history, cardiotoxic treatment exposures (ie, anthracycline-based chemotherapy, radiation therapy affecting the heart), need for late-effects screening, and the presence of any SCP, the latter of which we defined as a note or document providing a summary of a participant's past cancer treatment combined with an evidence-based follow-up plan.We also documented whether more in-depth cardiac testing (eg, electrocardiogram [ECG], echocardiogram, or other imaging) was performed or planned.
On trial enrollment, we asked participants to self-report their personal medical history to allow for comparisons with CCSS research records and PCP medical record documentation.Specifically, participants were asked whether they had previously received radiation or cardiotoxic chemotherapy; had undergone various medical and cardiovascular testing in the preceding 2 years (ie, echocardiogram, blood pressure measurement, diabetes screening or testing, lipid screening or testing); or were using medications for hypertension, diabetes, or dyslipidemia.Participants' historical treatment exposures were derived from CCSS research data, which were based on an abstraction of each survivor's original cancer treatment records.Participant demographics were obtained from CCSS research data.Participants self-reported race and ethnicity to the CCSS study team.Race was categorized as American Indian or Alaska Native, Asian, Black, Pacific Islander, White, or unknown.Ethnicity was categorized as Hispanic or Latino, not Hispanic or Latino, or unknown.
Other race and ethnicity categories were self-reported as other and nonspecified.Race and ethnicity were assessed to examine the study cohort and consider the generalizability of its findings to other

JAMA Network Open | Pediatrics
Primary Care Utilization and Cardiovascular Screening in Adult Survivors of Childhood Cancer populations.CCSS research data were considered the criterion standard in comparisons between different reporting methods.

Statistical Analysis
Participant demographics, diagnosis characteristics, and historical treatment information were analyzed with descriptive statistics.We calculated the proportion of participants with medical visits and other health care utilization and cardiovascular screening based on review of the PCPs' medical

Participant Characteristics
Of 347 enrolled CHIIP participants, 293 (median [range] age, 39.9 [21.5-65.0]years; 149 [50.9%] male) had evaluable data for baseline health care utilization analysis (Figure 1 and Table 1).No participants had missing data for demographic variables of interest.The median (range) age of childhood cancer diagnosis was 9.3 (0.0-20.9) years.Compared with the overall CCSS cohort who completed the most recent survey (follow-up 5, conducted between 2014 and 2016), CHIIP study participants had similar demographic characteristics, including approximately 90% with selfreported health insurance and 90% having reported receiving routine medical care within the past 2 years (Table 1).By design, CHIIP participants were selected for individuals who had a diagnosis of hypertension, dyslipidemia, or diabetes.

Baseline Utilization of General Health Care and Cardiovascular Disease Screening
In the 2 years leading up to study enrollment, 238 participants (81.2%) had a documented PCP office visit (median [IQR], 3 [2-5] visits), and 63 participants (21.5%) had a subspecialty visit noted,   records was also low for radiation (κ = 0.31) and cardiotoxic chemotherapy κ = 0.13) exposures.We found strong concordance between self-report and PCP records with regard to hypertension, lipid, and diabetes medications (κ range 0.79-0.89),but poor concordance with regard to echocardiography (κ = 0.37) and other common cardiometabolic screening tests (κ range, 0.05-0.34).

Factors Associated With Completed or Planned Cardiac Testing
In univariate and multivariable logistic regression analysis, we identified several factors associated with having completed or planned cardiac testing at initial trial enrollment (Table 3).In the univariate model, factors associated with cardiac testing included documentation of radiation therapy exposure, documentation of cardiotoxic chemotherapy exposure, presence of existing cardiovascular risk factors in the medical record, the presence of an SCP, documentation of a need for late-effects surveillance, a cardiology visit in the prior 2 years, number of existing cardiovascular disease risk factors, and high PCP utilization (Table 3).

Discussion
In this cross-sectional study of the PCP-documented screening patterns of a cohort of survivors of childhood cancer at high risk for cardiovascular disease, we found that adherence to regular cardiovascular screening was suboptimal.Few PCP records mentioned items pertinent to survivors' cancer histories or a need for late-effects surveillance.Almost one-third of evaluable participants had no documentation acknowledging a history of childhood cancer.Furthermore, concordance between participant self-report and PCP records, compared with known exposures abstracted from original oncology treatment records, varied considerably but was often poor.Only 21.5% of these individuals at elevated risk of cardiovascular disease had recommended echocardiography planned or performed in the prior 2 years.This was similar to the rate identified from an earlier CCSS intervention study focused on enhancing cardiomyopathy screening. 25National guidelines specific to long-term survivors of childhood cancer that were in place during this time period recommended consideration of echocardiograms or similar screening every 1 to 2 years for these individuals. 10,13In comparison, basic screening for hypertension, dyslipidemia, and diabetes was more prevalent, perhaps reflective of the frequency of routine screening for these conditions in the general adult population without a cancer history. 26,27We observed low rates of SCP utilization, consistent with participants who were diagnosed before SCPs were recommended and received much of their survivorship care in an era generally marked by slow adoption among PCPs. 28Factors associated with having up-to-date cardiovascular disease screening included documentation in the medical record of a participant having an increased risk for cardiovascular disease and a need for late-effects surveillance.
Although several prior studies have raised concerns for suboptimal screening for multiple conditions among survivors of childhood cancer, [29][30][31] no study to our knowledge has examined screening frequency and adherence through a direct review of PCP medical records.Therefore, we present survivors' health behaviors from a unique perspective that complements participants' selfreport, which we elicited concurrently.The discordance that we observed between participants' selfreport and medical records with regard to certain treatment exposures agrees with prior investigations among adult survivors of childhood cancer demonstrating limited recall [19][20][21] and emphasizes the importance of data collection from multiple sources.While some treatment exposures, such as radiation therapy, were reported by participants with reasonable accuracy, neither participants nor PCP medical records reliably reported anthracycline chemotherapy exposure.These discrepancies in both self-report and PCP medical records build on prior research that only assessed the accuracy of patient self-report and may highlight information to prioritize in communications from oncology and survivorship teams to patients and their PCPs.Our results cannot distinguish whether the act of documentation itself increased cardiac screening or was more reflective of a general awareness of risks for late effects.However, given that individual patients may be covered by multiple clinicians in a practice, documentation of risk by 1 clinicians could influence other clinicians' actions at follow-up, regardless of general awareness regarding surveillance needs.
Although the overall adherence to cardiovascular screening recommendations among this selected population is disconcerting, our findings provide hope that streamlined communication between oncologists, patients, and primary care could improve awareness, adherence, and overall cardiovascular outcomes.As strategies emerge to prevent or treat anthracycline-associated cardiomyopathy, either through the reduction of modifiable disease risk factors or cardiac remodeling as heart function begins to worsen, 6,32,33 multimodal strategies to address guidelineconcordant survivorship care are needed.The CHIIP intervention trial in progress focuses on providing personalized SCPs to patients and PCPs. 18This strategy aims to address knowledge and communication gaps that have previously represented barriers to uptake. 18,31Combined with other initiatives, such as those designed to promote general awareness of survivorship care, increase access to survivorship services, reduce financial barriers to care, and address critical disparities, 17,21,25,34 the CHIIP study's intervention to systematically outline survivors' individualized cardiovascular risks and other surveillance needs may improve adherence to recommended guidelines.Initial results have identified that survivors of cancer are twice as likely to have an undertreated cardiovascular risk factor compared with a matched general population sample. 35timately, a multimodal strategy consisting of interventions to promote SCPs in tandem with increased PCP engagement, targeted counseling, sustainable community partnerships, and technology-based solutions will likely be most effective. 25,34,36For example, these strategies have shown improvement in cardiomyopathy screening with echocardiograms in adult survivors of childhood cancer from approximately 20% to 50%. 25

JAMA Network Open | Pediatrics
Primary Care Utilization and Cardiovascular Screening in Adult Survivors of Childhood Cancer to promote control of cardiac risk factors.Study procedures were approved by the Fred Hutchinson Cancer Center and St Jude Children's Research Hospital institutional review boards.All participants provided written informed consent for data collection.This report follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for crosssectional studies.
records.Medical record documentation of prior treatment exposures was compared with known exposures from CCSS research data by McNemar test.To assess agreement among participant selfreport, PCP medical records, and CCSS research data, we estimated the sensitivity, specificity, and κ (concordance) among these data sources.Univariate logistic regression models identified factors associated with having completed any cardiac screening (ECG, echocardiogram, or other cardiac imaging) in the 2 years preceding trial enrollment, with odds ratios (ORs) and 95% CIs.Variables of interest included age, sex, documentation of radiation and cardiotoxic chemotherapy exposures, documentation that the participant had increased risk of cardiovascular disease, presence of an SCP, documentation of the need for late-effects monitoring, number of existing cardiovascular conditions (as determined by PCP medical records), a recent cardiology visit, and greater PCP utilization (defined as having >3 PCP visits over the preceding 2 years; 3 being the median value).Variables with P < .20 in univariate testing were then included in a multivariable logistic regression model.P values were 2-sided, and statistical significance was set at P = .05.Statistical analysis was performed with SAS version 9.4 (SAS Institute).Data were analyzed from November 2022 to July 2023.

Figure 1 .
Figure 1.Selection Flowchart of Communicating Health Information and Improving Coordination With Primary Care (CHIIP) Trial Participants for Analysis

347
Survivors of childhood cancer at increased risk for cardiovascular disease randomized into the CHIIP trial 293 Survivors of childhood cancer available for analysis, including 22 without a PCP during the study time period and 10 with a PCP but no visits during the study period 12 participants (4.1%) with records documenting a cardiology specialty visit.Of 293 participants with evaluable data, 46 participants (15.7%) did not have any documented outpatient medical visits.A total of 241 participants (82.3%) had a blood pressure measurement, 179 participants (61.1%) had undergone lipid testing, and 193 participants (65.9%) had undergone diabetes testing in the preceding 2 years.Cardiac testing was performed in the prior 2 years for 85 including

Table 1 .
Demographic and Clinical Characteristics of CHIIP Study Participants Compared With the Overall CCSS Cohort Primary Care Utilization and Cardiovascular Screening in Adult Survivors of Childhood Cancer a Members of the cohort who responded to the follow-up 5 survey, conducted between 2014 and 2016.bIncludesthosewho self-reported as other or nonspecified.cPerCCSS records.JAMA Network Open | Pediatrics

Table 2 )
. The specificity of participant report compared with CCSS data was lower than PCP records, and sensitivity was greater for a history of radiation (96.1%) than for a history of cardiotoxic chemotherapy (44.6%) (Table2).Concordance between participant report and PCP Figure 2. Frequency of Documentation of Treatment Exposures by Primary Care Provider (PCP) Records vs Data Abstracted From Childhood Cancer Survivor Study (CCSS) Research Records

Table 2 .
Sensitivity, Specificity, and Agreement Among Self-Report, PCP Records, and CCSS Research Records In the multivariable model, documentation of a participant's increased cardiovascular disease risk (OR, 11.94; 95% CI, 3.37-42.31),documentation of need for late-effects surveillance (OR, 3.92; 95% CI, 1.69-9.11),and the presence of existing cardiovascular risk factors in the medical record (OR, 2.09; 95% CI, 1.32-3.31)were independently associated with greater odds of having recent or planned cardiac screening.

Table 3 .
Factors Associated With Having Cardiac Testing Performed or Planned Among Adult Survivors of Childhood Cancer a a Cardiac testing was defined as electrocardiogram, echocardiogram, or other cardiac imaging.bParticipantswithmissingvariables were excluded from the regression model.cSexwasnot significant in univariate testing (P = .49)andtherefore was excluded from the multivariable model.dNA with nonbinary covariate.eDiagnosis of hypertension, dyslipidemia, or diabetes at baseline, modeled ordinally as 0, 1, and 2 or more conditions.fBased on median of 3 visits.