The American College of Cardiology 2008 guidelines for the management of adults with congenital heart disease recommend that care of adults with moderate and complex congenital heart disease (CHD) be guided in collaboration with clinicians trained in adults with congenital heart disease (ACHD) (1). A review of data from approximately 90 self-described ACHD programs in the United States indicates that the number of ACHD patients seen in these specialized clinics is far below targeted estimates ((2),(3),(4),(5),6); it is likely that a substantial number of ACHD patients continue to be cared for by pediatric cardiologists. Pediatric cardiologists might choose or feel obligated to continue caring for or to assume care for ACHD patients for a variety of valid or perceived reasons. To develop effective strategies to improve access to and quality of ACHD care for all adult survivors, it is imperative to better understand referral patterns and perceived barriers to specialized care. Therefore, we sought to characterize patterns of referral to ACHD care and identify barriers to such care, as perceived by pediatric cardiologists.

We conducted a cross-sectional mail survey of clinicians in the United States certified by the American Board of Pediatrics (ABP) to practice pediatric cardiology. Access to our study population was requested from the ABP in May 2010 and approved in August 2010. From a total of 1,824 clinicians certified in pediatric cardiology by the ABP at the time of our request, 1,766 had registered addresses in the United States and therefore were retained for sampling and randomly assigned to sampling replicates stratifying across 4 geographic regions (Northeast, South, Midwest, and West). A power analysis was run to determine minimum sample size needed/group to detect a difference of 20 percentage points across the regions in terms of the prevalence of pediatric cardiologists taking care of at least 20 patients >40 years of age. A minimum of 60 surveys/region was determined to be sufficient to detect differences with a power of 0.8 and an alpha of 0.05. Surveys were sent until a minimum of 60 pediatric cardiologists from each region met inclusion criteria. Selection criteria were verified at the point of survey entry. Clinicians not actively seeing patients over the age of 11 years in the outpatient setting were excluded. The survey was distributed by the U.S. Postal Service to addresses provided by the ABP between November 2010 and July 2011. No less than 1 month after the initial survey was mailed, a reminder notice was sent to subjects who had not previously returned their survey. The database was closed to further entries on November 1, 2011. The project was approved by the local institutional review board at Boston Children's Hospital.

The survey tool was developed over a 12-month time frame, through a working group of 10 clinicians and researchers with an interest in healthcare transition. The working group included physician, nursing, physician assistant, and social worker representation. Questions were developed on the basis of a review of the healthcare transition published data including a recent national survey (7) and guided by a survey methodologist (S.Z.). The final survey included a total of 20 questions within the following 6 categories: 1) inclusion criteria; 2) demography of ACHD care; 3) transitioning assessment and education; 4) referral patterns to ACHD care; 5) perceived barriers to ACHD care; 6) resources; 7) demographic data; and 8) an open-ended question inviting further input on the transition and transfer process.

Transitioning education and assessment was defined for survey participants as encompassing disease knowledge; understanding of medication use and side effects; symptoms requiring urgent care; congenital/genetic anomalies in offspring; impact of high-risk behaviors; and disease impact on education, vocation, and insurability.

The section regarding demography of ACHD care asked clinicians if they provided care to patients in various age groups (i.e., 11 to 14, 15 to 16, 17 to 18, 19 to 21, 22 to 24, 25 to 30, 31 to 40, 41 to 50, and >50 years) and whether they felt comfortable providing such age-appropriate care. Response options for each age group were “yes” “no” or “don't know”. The section on transitioning inquired whether transitioning education/assessment was provided to patients and, if so, in what manner (formal vs. informal) and at what age of initiation. Sections regarding referral patterns and barriers to referral included questions regarding characteristics that prompt referral to an ACHD program and barriers to transfer. For each characteristic prompting transfer and for each perceived barrier, respondents were provided with the options: “yes” “no” or “don't know”. The resource section addressed the desire for resources to improve the delivery of transition education and to streamline the transfer process.

Descriptive analyses of clinician demographic data were summarized with mean ± SD for continuous variables and proportions for categorical variables. Demographic and regional differences were analyzed with Pearson's chi-square or Fisher exact test for categorical variables and analysis of variance for continuous variables. Barriers to and impetus for the transitioning of patients into adult care were ranked, and differences in clinician beliefs were assessed with chi square analysis. In the multivariate analysis, a multiple logistic regression model was used to identify provider factors that might be associated with provision of ACHD care to patients older than 18 years of age. Factors identified in the univariate analysis with p value <0.2 were initially added to the logistic regression model. Final model results are presented in the tables. Multiple logistic regression models were also run for the top 3 identified incentives and barriers to transitioning of care to identify provider characteristics. A 2-sided p value of <0.05 was considered indicative of statistical significance. Data analysis was performed with SAS software (version 9.2, SAS Institute, Cary, North Carolina). Survey results were reported in aggregate so that clinician confidentiality was ensured.

A content analysis approach to the qualitative data interpretation was employed ((8),9). Comments about the perceptions of participants with regard to their experiences constituted the units of analysis. The text was separated into meaning units that were condensed. The condensed meaning units were summarized and labeled with a code. The codes were reviewed for similarities and differences and sorted into categories and then studied for underlying meanings threaded through the whole as themes.

The overall response rate was 47.7% (291 of 610), and 88.3% (257 of 291) of respondents met inclusion criteria (i.e., provision of outpatient care to patients >11 years of age). Qualifying participants geographically represented 39 of 50 U.S. states, with 62 from the Midwest, 69 from the Northeast, 62 from the South, and 64 from the West. The response rate/region was 54.5% for the Midwest, 52.9% for the Northeast, 37.4% for the South, and 49.7% for the West (p < 0.001). Clinicians were in practice for an average of 18.2 ± 10.7 years, and 70.2% were male. An affiliation with an academic institution was reported by 72% of respondents. Additional clinician characteristics are summarized in (Table 1).

Most (79.1%) pediatric cardiologists reported caring for at least 20 adult patients with CHD (≥18 years of age); 53.5%, 21.7%, 13.4%, and 9.8% reported caring for at least 20 patients ≥21, >30, >40, and >50 years of age, respectively. Factors associated with the provision of adult care are summarized in (Table 2). The mean years in practice was longer for those clinicians who reported caring for adult patients compared with those that did not (20.0 ± 10.5 years vs. 11.21 ± 8.62 years, p < 0.001), and male clinicians were more likely to provide this care than female clinicians (84.8% vs. 66.2%, p = 0.002). Clinicians practicing in the West were less likely to care for adult patients than those from the other U.S. regions (67.7% vs. 82.8%, p = 0.014). The mean years in practice was longer for clinicians caring for patients >30 years of age compared with those who did not report caring for this age group (22.1 ± 12.4 years vs. 17.1 ± 10.0 years, p = 0.003), and respondents from the Midwest were more likely to care for patients >50 years of age compared with other U.S. regions (17.7% vs. 7.3%, p = 0.016). The proportion of respondents with and without internal medicine certification who felt competent providing age-appropriate ACHD care is provided in (Table 3). The mean years in practice was longer for clinicians who reported feeling capable of providing age appropriate care to ACHD patients >18 years of age (18.4 ± 10.6 years vs. 11.4 ± 6.8 years, p = 0.016). However, mean years in practice did not differ according to whether or not respondents felt capable of providing age appropriate care to patients >30 years of age (19.3 ± 10.2 vs. 17.5 ± 11.2 years, p = 0.216). Among clinicians caring for patients >30 years of age, 26.9% doubted their capacity to provide age-appropriate care.

Most clinicians (91.8%) stated that they prepare their patients for self-care management via transitioning education as defined in our methodology. Only 6.9% of respondents employed a formal assessment tool such as a questionnaire, checklist, or survey. Most clinicians reported that they begin transition education/assessment before patient adolescence: 5.3%, 49.4%, and 29.4% before patient age 11 years, between ages 12 and 16 years, and between 17 and 18 years, respectively. A minority (10.2%) replied that they wait until after patient age 18 years to begin preparation for self-care management; 5.7% answered that they were not sure when this education/assessment began.

The most commonly perceived patient characteristics prompting referral to an ACHD program or adult-oriented health care system are presented in (Table 4). Predictors of the most common characteristic prompting referral are provided in (Table 5). The mean years in practice for clinicians who did not identify adult comorbidities, age, and pregnancy as characteristics that should trigger referral to adult-oriented care was longer than those who did (i.e., 20.2 ± 10.2 years vs. 17.6 ± 10.7 years, p = 0.021; 20.0 ± 10.7 years vs. 17.2 ± 10.6 years, p = 0.047; 20.5 ± 11.5 years vs. 16.9 ± 10.0 years, p = 0.011, respectively). Clinicians who reported an affiliation with an academic institution were more likely to identify age (68.9% vs. 52.1%, p = 0.013), graduation from college (56.3% vs. 33.8%, p = 0.001), marriage (46.4% vs. 28.2%, p = 0.008), and pregnancy (68.3% vs. 52.1%, p = 0.011) as characteristics that should prompt referral.

Perceived barriers to transfer and their rankings as well as predictors of the most common barriers are summarized in Tables (Table 6) and (Table 7), respectively. Overall, 93.4% of respondents reported that patient, parent, and/or clinician attachment was a barrier to ACHD care. A lack of qualified care providers in ACHD was less likely to be identified by those who stated affiliation with an academic institution (71.0% vs. 88.7%, p = 0.003) and the mean years in practice was longer for clinicians who identified this barrier than those who did not (19.5 ± 10.8 years vs. 14.1 ± 9.3 years, p = 0.002). The attachment of the provider to the ACHD patient and/or family was more likely to be noted as a barrier to transfer by those with an affiliation to an academic institution (76.0% vs. 53.5%, p = 0.001). There was no significant difference in the mean years in practice among those who did and did not identify attachment to the ACHD patient and/or family as a barrier to transfer (17.6 ± 10.7 years vs. 19.8 ± 10.7 years, p = 0.121).

Most respondents perceived a need for additional institutional resources to improve the delivery of transitioning (self-care management) education and to streamline the transfer process to adult care (82.3% and 79.2%, respectively).

A total of 85 of 257 (33%) respondents replied to the open-ended question inviting further input on the transition and transfer process. Four major themes emerged: 1) safe care; 2) help; 3) teamwork; and 4) no rules need apply.

Many respondents expressed concern about providing adult care in their pediatric hospital, especially in the context of adult comorbidities. An equally common concern, however, was patient safety due to lack of experience and expertise for the provision of quality congenital care by adult institutions and providers.

Several respondents indicated that additional staff (physicians, nurses, and social workers) were required to address the often labor-intensive needs of the ACHD population, including management of adult comorbidities and psychosocial issues. These needs were felt to be under-appreciated by their practice or institution.

Most respondents favored a multidisciplinary team approach to caring for ACHD patients, including adult internists, adult cardiologists, and ACHD specialists where available.

Most respondents do not favor strict policies regarding where a patient should be cared for or by whom, because—as many point out—no 2 patients are alike clinically, emotionally, and developmentally. Some respondents from institutions with set age policies for transfer were strongly opposed to such.

The overwhelming majority of patients with CHD are now expected to live well into and through adulthood (10), although they remain at risk for long-term and life-defining sequelae related to their CHD ((1),11). In addition, these patients are at equal or sometimes greater risk of and from adult (internal medicine) comorbidities (including atherosclerosis, systemic hypertension, hyperlipidemia, diabetes, obesity, thyroid disease, psychiatric illness, and cancer), activities, and life milestones (such as pregnancy) ((11),(12),(13),(14),15). Although it might be tempting to consider growing adults with a chronic medical condition such as CHD to have a physiology with constant ramifications that can be cared for by a singular team, it is increasingly clear that adults with CHD have unique healthcare needs that are different from their pediatric counterparts. The complex inter-relationships between congenital heart lesions, changing physiologies, demands of aging, and comorbidities acquired in adulthood require expertise in both CHD and adult internal medicine and cardiovascular care. To serve this population of patients, an insufficient but growing number of ACHD clinical and training programs are evolving; most combine resources from partnering pediatric and adult internal medicine programs and facilities. Standardization of such training via aligned ABP and American Board of Internal Medicine collaboration toward ACHD subspecialty board certification is underway. In addition, numerous task forces have been established to develop strategies to improve access to specialized ACHD care in collaboration with primary care teams. A better understanding of patterns of referral and of perceived barriers to ACHD care seems important in optimizing the essential partnership between pediatric cardiologists and ACHD care practitioners in such initiatives.

Our current study indicates that most pediatric cardiologists care for adult patients. The longer they have been in practice, the more likely they are to care for adult patients. A substantial proportion continue to care for patients in their third, fourth, and even fifth decades of life, despite many respondents stating that they had mandatory institutional patient transfer policies, often at 18 or 21 years of age. Reasons as to why younger pediatric cardiologists are less involved in ACHD care remain speculative and might include an overall younger patient population that has yet to mature to adulthood, less time to form deep emotional ties that might impede transfer, and/or increased awareness and acceptance of ACHD specialists.

Although the published data are scarce with regard to the practice patterns of pediatric subspecialists, evidence suggests that caring for adults in pediatric hospitals is not uncommon for patients with CHD or other lifelong ailments of pediatric onset ((16),(17),18). It has been estimated that 6.3% of pediatric hospital admissions, accounting for at least $1 billion a year in hospital charges, relate to adult care ((19),20). Debate continues with regard to optimal location of adult-aged care for patients with congenital or pediatric-onset disease. Caring for such adult survivors has generally seemed to be more expensive in pediatric hospitals (20). However, recent disease-specific evidence has suggested that ACHD surgery in pediatric as opposed to adult hospitals might not be associated with an increase in resource use (21). Morbidity and mortality for such surgical interventions consistently seems lessened when provided in a setting that maximizes both congenital and adult internal medicine experience and expertise, regardless of specific pediatric versus adult internal medicine location of care ((17),22).

Most responding pediatric cardiologists reported preparing their adolescent patients for self-care management through transitioning assessment and education, albeit typically in informal fashion. Of import, despite the recognition by pediatric congenital cardiology care providers that their adolescent and young adult patients would need collaborative care with specialists trained in adult internal medicine, our study identified that the presence of adult internal medicine comorbidities seemed to be the most common characteristic prompting patient referral from pediatric to ACHD-specialist care. The uncertainties and insecurities present in understanding and facing ramifications of new physiology and disease might place patient outcomes at risk and might strain or threaten the goal of a therapeutic patient-physician relationship ((23),24); such outcomes might be improved if both the transition and transfer processes are successfully underway, before clinical instability ((25),26).

Most respondents (93.4%) identified attachment issues (between patient or family to or from the clinician) as barriers to transfer to adult-oriented care; attachment concerns were pronounced in academic institutions. These results are consistent with but more pronounced than previously reported findings (27). Studies suggest that a coordinated transfer process, which can be flexible on the basis of factors such as target age, clinical stability, emotional maturity, and cognitive ability, is more likely to result in successful transfer (28). Implementation of a formal transitioning program including educational programming for patients, families, and clinicians might assist in providing the necessary skills, confidence, support, and guidance for all participants to overcome emotional hurdles, inherent in transition and termination of direct care relationships.

A lack of qualified ACHD providers was identified as an important perceived barrier to transfer to ACHD care. The recently published American Heart Association scientific statement regarding “best practices in managing transition to adulthood for adolescents with congenital heart disease” supports a “policy on timing” of transfer, ideally between 18 and 21 years of age, depending on the developmental ability of the patient (28); but given this widely perceived lack of qualified ACHD providers, such a recommendation might be premature. Of interest, many of the pediatric cardiologists in this study were strongly opposed to policies mandating transfer of ACHD patients. The perceived lack of qualified ACHD provider was noted even by clinicians from institutions affiliated with established ACHD care programs. Currently, the terms “ACHD specialist” and “ACHD program” are self-described, and specialist training and experience are highly variable. Standardization of eligibility for training criteria, ACHD training, subspecialty certification, and program accreditation might improve confidence in clinicians and programs providing ACHD care. Such a model of regulation and transparency in credentialing has been successful for the care of adults with other child-onset illnesses, such as cystic fibrosis (29). In the interim a team approach to care—including (when applicable to given regional care models) the pediatric cardiologist, ACHD specialist, general internist or primary care provider, and general adult cardiologist—seems necessary and prudent to ensure continuity of care for all ACHD survivors. Further study is required to understand the impact of the recommended and evolving ACHD care models on continuity of care and quality of life as well as on morbidity and long-term survival of ACHD patients.

Given that we targeted pediatric cardiologists practicing in the United States, results might not be generalizable to other countries. Although the response rate of <50% might result in a sampling bias that undermines external validity, it should not impact internal validity. Moreover, the population of respondents seems demographically similar (on the basis of sex and age) to the overall target population of pediatric cardiologists practicing in the United States (30). Our results reflect self-reporting, and social desirability response bias might have over- or underestimated the percentage of pediatric cardiologists who care for adult patients. Having only 1 open-ended survey question might result in leading question bias. This study did not examine or analyze the complexity of ACHD patients that pediatric cardiologists care for, and we did not explore which proportion of pediatric cardiologists care for ACHD patients in collaboration with ACHD specialists or general adult cardiologists. Finally, the study is cross-sectional in nature such that cause-and-effect relationships (e.g., emotional attachment and barrier to ACHD care) cannot be inferred.

Most U.S. pediatric cardiologists provide care to adults and are hesitant to transfer age-appropriate patients to internal medicine ACHD specialists. Reasons seem to include widely perceived lack of expert ACHD caregivers as well as sensed emotional bonding of patients or families to and from caregivers. Transfer of care seems to be commonly triggered by comorbidities acquired in adulthood. Knowledge of referral patterns and perceived barriers to care can inform health care strategists as they face such in the development of educational, transition, and transfer programming aimed at retaining patients within lifelong, coordinated CHD care and achieving optimal outcomes for patients, families, and clinicians.