Author + information
- Received November 22, 2005
- Revision received March 21, 2006
- Accepted March 28, 2006
- Published online July 18, 2006.
- Graham J. Reid, PhD⁎,†,‡,§∥,¶,1,⁎ (, )
- Gary D. Webb, MD§∥,#,
- Mor Barzel, MA⁎,
- Brian W. McCrindle, MD⁎⁎,††,
- M. Jane Irvine, PhD‡∥,‡‡ and
- Samuel C. Siu, MD§∥,¶
- ↵⁎Reprint requests and correspondence:
Dr. Graham J. Reid, The University of Western Ontario, SSC 7310, London, Ontario N6A 5C2, Canada.
Objectives This study sought to determine what adolescents and young adults with moderate to complex congenital heart disease (CHD) believe their life span to be and to examine correlates of their beliefs.
Background Patients with moderate to complex CHD have a shortened life expectancy. Patients’ perceptions of their life expectancy have not been examined.
Methods Young adults and older adolescents with moderate or complex CHD (n = 296) estimated their own life expectancy and that of healthy peers, and rated their health status and risk of CHD complications. Adults with CHD discussed reasons for life expectancy ratings in an interview.
Results Patients with CHD expected to live to age 75 ± 11 years, only 4 years less than their healthy peers. Over 85% of patients expected to live longer than our estimates of their life expectancy. Poorer health status and higher perceived risk of CHD complications related to shorter perceived life expectancy. Young adults lacked awareness and understanding of CHD-specific risks, and their life expectancy perceptions often related to risk factors for coronary artery disease.
Conclusions Patients with moderate to complex CHD expect to live almost as long as their healthy peers. For most patients, this is unlikely. The implications of these beliefs on health behaviors and life choices are unknown, but should be examined. Nevertheless, patients need accurate information delivered in a sensitive manner to make informed life choices regarding education, careers, and family.
In the 1950s, about 20% of infants born with moderate or complex congenital heart disease (CHD) survived their first year, but now up to 90% of these children live to adulthood (1). As adolescents and young adults, they face a variety of health and psychosocial issues (2). The emphasis has been on increasing longevity for patients with CHD through advances in diagnostic methods and cardiac surgery (1), but one issue that has received relatively little attention is the prospect of earlier mortality despite these advances (3,4). A clinic-based study of patients with moderate to complex CHD found that those who died did so at a mean age of 37 ± 15 years (range 18 to 80 years) (3). We examined adolescents and young adults’ expectations for their life expectancy.
Expectations affect current behavior and decision making regarding future goals and aspirations. However, people tend to be unrealistically optimistic and consistently underestimate their own likelihood of experiencing adversity (5). For example, people believe they will live longer than actuarial estimates of life expectancy (6). When individuals with a chronic illness expect to live longer than their peers, it might be viewed as denial or lack of information, rather than unrealistic optimism. For individuals with moderate or complex CHD, expecting to live as long as their healthy peers is unlikely given current data for this population. Perceived life expectancy could influence decisions such as pursuing higher education, choice of vocation, marriage, and having children.
Individuals with CHD may base estimates of their life expectancy on what they have been told and past medical experiences. Similar to healthy individuals, their estimates may be influenced by their current health status and behaviors (e.g., exercising, avoiding drug use ). Patients’ perceived risk of complications is also likely important, because risk perceptions are consistently related to engagement in positive and negative health behaviors (5).
Self-reported beliefs regarding life expectancy have not been examined among patients with CHD, or any population of adolescents or young adults with chronic illness. The present study sought to determine what adolescents and young adults with moderate to complex CHD believed their life span would be, and examined correlates of their beliefs.
Patients with CHD
The cardiac clinic database at the Hospital for Sick Children (HSC) in Toronto, the largest pediatric cardiac center in Canada, was used to identify patients. We identified patients with CHD recommended to receive annual follow-up at a specialized adult CHD clinic using the HSC diagnostic codes meeting consensus criteria (1,8). Two cohorts were selected: 1) young adults born between June 1, 1978, and May 31, 1980, and 2) older adolescents born between April 1, 1982, and March 31, 1984. These cohorts were identified as part of a larger project on psychosocial issues and transfer to adult care. The older cohort was of age to have transferred to adult care (i.e., age 19 years or older), whereas the younger cohort should still be receiving pediatric care. Because pediatric cardiologists may be reluctant to discuss life expectancy with their patients, it would be informative to compare the self-reported life expectancy of these two cohorts.
For the young adult cohort, a computer algorithm identified 891 potential patients, whose charts were reviewed. Patients were excluded (n = 490) at chart review or initial contact if their CHD diagnosis did not warrant specialized CHD follow-up or had a low risk for a shortened life expectancy (e.g., unrepaired atrial septal defect, closure of ventricular septal defect), or they had Down syndrome or another developmental disability, lived outside of Canada, had died, or did not speak English; a further 63 could not be contacted. Patients were mailed a letter of information, consent form, questionnaires, and a stamped addressed return envelope. The letter did not mention that we would ask about their life expectancy. Two weeks later, patients were telephoned to address any difficulties with the questionnaires and to schedule an interview. Interviews were audiotaped and transcribed. A total of 212 patients consented and participated; 87 refused participation and 39 did not return questionnaires. Most patients completed questionnaires and interviews (n = 184); 15 completed only interviews, and 13 completed only questionnaires. Seven audiotapes were corrupted, 7 patients were not asked the question regarding life expectancy, and 5 stated they did not know in response to one or both of the life expectancy questions, leaving a final sample of 179 patients.
For the adolescent cohort, 533 potential patients were identified using a modified computer algorithm refined by removing those diagnoses in the algorithm from the adult cohort that were in fact ineligible after chart review. Patients were excluded (n = 223) after chart review or initial contact if they did not meet study criteria; 58 patients could not be contacted. Patients and their parents were mailed a consent form, questionnaires, and a stamped addressed return envelope. In total, 133 adolescents consented and returned questionnaires (54 refused participation; 65 did not return questionnaires despite follow-up telephone calls) and 16 did not answer the questions regarding their own and/or their peers’ life expectancy, leaving a final sample of 117.
Undergraduate university students completed the same questions as the CHD patients and received credit in an introductory psychology course. Students were age 18 (n = 6, 10%), 19 (n = 45, 70%), or 20 years old (n = 13, 20%), 70% were women, and they had the following medical conditions: arthritis (n = 3), diabetes (n = 1), kidney trouble (n = 2). Life expectancy questions from this sample were compared with those of patients with CHD. Potential differences in ratings of life expectancy by interview versus questionnaire were examined in this sample by randomly assigning participants to either questionnaire or interview format at the time of consent. Some students did not keep their appointments, and students who were age 21 or older (n = 8) or who did not answer the life expectancy questions (n = 2) were excluded; thus, 67% of the students answered the questions in an interview, and 33% completed questionnaires.
Participants were asked two questions: 1) What do you think is the average life expectancy for your peers who do not have a congenital heart problem? and 2) What do you think your life expectancy is? The exact wording of these questions varied slightly for the CHD or undergraduate samples, and in response to the general flow of the interview. For example, the undergraduate sample was asked, “What do you think is the average life expectancy for your peers?” When participants reported a range for life expectancy, the lowest number was recorded to capture perceptions of shortest life expectancy. Participants who gave a numeric rating for their peers’ life expectancy and stated they expected to live as long as their peers were given the identical age for their own life expectancy.
For the patient sample, the primary outcome variable was a shortened life expectancy score computed from the difference between patients’ self-reported life expectancy and ratings of their peers’ life expectancy. The difference score removes the general effect of patients expecting that everyone, including themselves, will have a long or short life.
Pediatric health care at the HSC since age 10 years was abstracted from the medical records. Diagnoses for each patient were reviewed (by G.D.W.), and a primary diagnosis was determined. Patients were categorized as being at medium or high risk of earlier death based on their diagnoses and surgical history; this categorization was based on the Bethesda conference division of lesion complexity according to expert consensus (i.e., complex/severe vs. moderately severe) (1), and took into consideration multiple diagnoses when categorization by the primary defect was unclear.
Patients reported basic demographics and any comorbid medical conditions or activity restrictions, and completed the Short Form-36 (SF-36). Four dichotomous scores were computed from the SF-36 using age and gender norms and clinical cutoffs: 1) poor health status (i.e., poor or fair vs. good, very good, excellent); 2) physical health limitations (i.e., difficulties performing physical activities); 3) mental health problems (i.e., symptoms of depression and anxiety); and 4) limitations in work and daily activities caused by physical or mental health problems (9). Health behaviors included physical activity level (10) and substance use (e.g., binge drinking) (11,12). Dental health behaviors included frequency of dental appointments, brushing and flossing teeth, and use of prophylactic antibiotics for procedures that may cause bleeding.
Patients rated their beliefs regarding CHD risks, on a 6-point scale ranging from highly likely to not at all likely, for: 1) overall likelihood of CHD complications; 2) risks attributable to not always attending cardiac appointments; 3) risks of lack of exercise; and 4) risks of not using antibiotic prophylaxis for dental procedures. To examine how optimism affected patient perceptions, general optimism was measured using two items (e.g., “I’m always optimistic about my future”) from an established questionnaire (13) with 5 response options (strongly disagree, disagree, neutral, agree, strongly agree), and health optimism was based on 2 items using a similar format. Items were averaged and dichotomized to reflect greater pessimism; 1 = neutral or more pessimistic and 0 = more optimistic views.
Patients reported medical visits during the previous 2 years for adolescents or since age 18 years for the young adults; visits were categorized as no medical appointments, visits only to a family physician, or visits to a cardiologist. Frequency of pediatric cardiac appointments was obtained by interview for the adults and by questionnaire for the adolescents. Independence in attending cardiology appointments (i.e., alone or with friend vs. with parents) was obtained by interview for the adults and by questionnaire for the adolescents, because parental presence might affect the information discussed during appointments.
Interviews and qualitative data analyses
The young adults with CHD completed a semistructured interview in which the following issues were covered in the order listed: 1) their biggest fear or worry regarding future health; 2) general expectations for experiencing negative health effects because of CHD; 3) specific expectations regarding: a) future surgery, b) heart rhythm problems, c) decreased ability to work due, or d) to perform recreational or household activities due to poor health; and 4) life expectancies and reasons for life expectancy estimates.
Thematic analyses of patients’ responses were conducted independently by two investigators (G.J.R. and M.B.) (14). The investigators discussed key themes and concepts that emerged from the data related to reasons for patients’ perceptions of life expectancy and other notable issues. The reviewers agreed on all major themes, and that theme saturation (i.e., no new ideas or concepts were emerging from the data) occurred. For themes not identified by both reviewers, transcripts were jointly reviewed and discussed to reach a consensus.
This study was approved by the Research Ethics Boards at the HSC, University Health Network, and The University of Western Ontario.
Patients who did not provide life expectancy data were compared with those who did using chi-square tests. For the patient sample, a dichotomous shortened life expectancy score was created: 1 = more than 5 years shorter life expectancy than peers, and 0 = life expectancy within 5 years of peers or longer than their peers. The average difference between self and peer life expectancy was +3.1 years for undergraduates and −4.0 years for patients. Given this, we categorized a difference of 5 years as “equal” to peers. A dichotomous score was used because the absolute difference score was not normally distributed, and we were interested in correlates of whether or not patients expected to live as long as their peers. Variables significantly related with shortened life expectancy in univariate analyses were used in a multivariate logistic regression with forward, stepwise variable entry (p < 0.05). The 13 patients with missing data for more than 2 of the predictor variables were dropped from this analysis. For other patients with missing data, the mode, for categorical variables, and the mean, for continuous variables, were substituted (15).
Patients were 16 (n = 29; 10%), 17 (n = 59; 20%), 18 (n = 28; 10%), 19 (n = 93; 31%), or 20 years of age (n = 87; 29%) when they completed the questionnaires; 49% were men and 80% lived with their parents. Patients had undergone an average of 1.7 ± 1.5 surgeries. The most common diagnoses for patients with a high risk of early death (complex/severe CHD, n = 93) were as follows: Fontan procedure, 31%; Mustard repair of transposition of the great arteries, 23%; conduit and tunnel repairs of complex conditions, 18%; and palliative shunts in cyanotic patients, 16%; the remaining 12% of patients had a variety of other defects. For patients with medium risk of shortened life expectancy (moderate CHD, n = 203), the most common diagnoses were: tetralogy of Fallot, 24%; coarctation of aorta, 17%; atrial septal defect with complicating factors, 14%; atrioventricular septal defect, 14%; left ventricular outflow tract obstruction, 11%; right ventricular outflow tract obstruction, 5%; repaired total anomalous pulmonary venous connection, 5%; the remaining 10% of patients had a variety of other defects.
Patients with CHD who did not provide life expectancy data (16% of adults; 12% of adolescents) were compared with those who did. The only significant difference was that patients with missing data were less likely to have had their teeth cleaned at least annually (61% vs. 76%). The adolescents and adults did not differ on self or peer life expectancy, or the difference between self and peer life expectancy. There were no differences in self or peer life expectancy ratings between undergraduates who completed interviews versus questionnaires.
Patients with CHD expected to live to be 75 years old on average, which was significantly younger than their estimates for their peers without CHD (Table 1).Patients expected to live 4.0 years less than their peers. Figure 1shows the distribution of life expectancy ratings in 10-year groupings. Over 87% of patients with moderate CHD and 68% with complex CHD expected to live past age 55 years.
By comparison, the undergraduate sample expected to live to be 81 years old on average, which was significantly older than their estimates for their peers (Table 1). Patients with CHD did not differ from the undergraduates in their estimates of peers’ life expectancy. Patients with CHD had significantly shorter self-reported life expectancies than the undergraduates.
Correlates of shorter life expectancies among CHD patients
About one-fourth of the patients (27%) were categorized as having a shortened self-reported life expectancy (i.e., at least 5 years shorter than peers). Type of medical care (i.e., cardiac, family physician, none) and gender were not related to life expectancy; results for other nonsignificant predictor variables are not reported. Significant univariate correlates of a shorter life expectancy (Table 2)were entered in a logistic regression. In the multivariate analyses, having a diagnosis rated at a higher risk for early death, poorer self-rated health status, and higher perceived risk of CHD complications were independent predictors of perceptions of a shortened life expectancy (Table 3).Patients with complex CHD expected to live to be 70 ± 12.7 years of age (median 70 years, range 30 to 99 years), which was significantly younger than patients with moderate CHD (77 ± 9.7 years, median = 78 years, range = 50 to 120 years; t = 5.1, p < 0.001).
Qualitative analyses of interviews with young adults with CHD
Patients’ interviews were characterized by a lack of awareness and understanding of future health risks related to their CHD. An overriding theme in patients’ rationales for their life expectancy related to risk factors for coronary artery disease. Patients thought that lack of exercise, poor dietary habits, smoking, stress, and a family history of heart problems meant they were more likely to die younger than their peers. Specific sources of information about CHD risks were rarely mentioned. Parents were mentioned as sources of information more often than physicians. (See the Appendixfor quotations related to the key themes.) Patients often believed that advances in health care would extend their life. Although many patients were probably overly optimistic about their life expectancy, for some patients who held a more realistic view, their overall outlook on life seemed fairly bleak.
Older adolescents and young adults with CHD expected to live well past middle age. They expected to live 75 years, just 4 years less than their expectations for their peers without CHD. A sample of undergraduates expected to live to 81 years, about 3 years longer than their peers. Estimated life expectancy at birth for Canadians born in 1990 is 77.8 years (16), and 79.7 years for births in 2002 (17). Thus, both the patients and undergraduates were quite accurate in estimating life expectancy for their peers.
Patients with CHD had optimistic expectations for their own life spans. The majority expected to live beyond age 65 years. Survival of individuals with CHD has increased markedly (18). Most data on late mortality are not reported separately for individuals surviving to adulthood (e.g., Nieminen et al. ). We estimate that the average life expectancy of these adult patients will be 35 to 40 years for those with complex CHD, and 55 years for those with moderate CHD. Thus, patients with complex CHD believed they would live about 35 years longer than expected, and those with moderate CHD about 20 years longer than expected. About 9% of the young adults and adolescents with complex and 3% of those with moderate CHD might be considered to have “realistic” views of their longevity given these projections.
Correlates of self-reported life expectancy
We examined correlates of relatively shorter, rather than absolutely shorter, life expectations. That is, patients were considered to have a shortened life expectancy if they believed they would die 5 years or more before when they thought their healthy peers would die. This controls for a patient’s general bias toward shorter or longer life expectancies. Future studies could compare patient perceptions with diagnosis-specific life expectancies or patient-specific estimates as rated by the patient’s cardiologist.
Individuals with poorer self-rated overall health status predicted shorter life expectancies for themselves. Patients may compare themselves with healthy peers, and conclude that if they are in poor physical health at a young age, it is unlikely that they will live a long life. Self-rated health is a consistent predictor of mortality in the general population (19) and among adults with chronic illness (20), and predicts mortality when assessed as early as age 18 to 20 years (21). Self-rated health and self-rated life expectancy were independently related to mortality in a community sample of elderly individuals, but only self-rated health predicted mortality among middle-age individuals (22). Among adolescents or young adults with chronic illness, self-reported life expectancy has not previously been examined as a predictor of mortality, nor has the relationship between self-rated health and self-reported life expectancy been examined.
The role of risk perception in patients’ self-reported life expectancies
Individuals who perceived themselves to be at higher risk of CHD complications expected to have a shorter life than their peers. We recently showed that among the adult cohort of patients in this study, patients’ beliefs were significantly related to whether or not they had successfully transferred to adult cardiac care (23). Given that risk perception was an important variable in patients’ estimates of their life expectancies, we sought to understand how patients arrive at their perceptions of risk. The interviews conducted with the young adults suggested that patients do not seem to understand the specific risks related to CHD. Many patients’ beliefs were based on risks for coronary artery disease, rather than CHD. In public health and educational initiatives, the role of diet, exercise, and smoking are often discussed in relationship to “heart disease” or stroke. Many patients with CHD seem to identify with the term heart disease and examine their own health behaviors and risks in terms of coronary artery disease. Even among patients who held more realistic views of their life expectancy, many were unable to articulate specific reasons. Accurate knowledge about CHD-related risks may have more impact on patients’ behaviors, including attending regular cardiac care, than their knowledge of their cardiac anatomy or surgical history (2).
A fairly high percentage of eligible patients declined participation, and life expectancy questions were not obtained from a number of patients who did participate. If patients who had more negative views about their CHD or who were more ill were less likely to participate, this could result in overestimates of self-reported life expectancy. The letter of information did not mention that patients would be asked about their life expectancy; thus, awareness of this question could not have affected the decision to participate. There were virtually no differences between patients with versus without missing data, suggesting a lack of bias attributable to patients who participated but did not provide life expectancy data.
Different assessment methods were used for the two cohorts of patients. However, there were no differences between the cohorts in life expectancy estimates. Further, there were no differences between interview and questionnaire ratings from the undergraduates, suggesting that this difference did not have a substantive impact on the results.
Our comparison sample was not matched to the patient sample in terms of educational attainment, location, or other potentially relevant variables. The fact that there were no differences between the patients’ and the undergraduates’ ratings of average life expectancy for their peers suggests that the lack of a matched control group was not a substantive issue.
Current data on longevity among patients with moderate or complex lesions are based on outcomes from tertiary care centers. Determination of population-based survival for contemporary patients with moderate or severe lesions is needed. In particular, estimates of life expectancy for patients surviving to adulthood are needed. Given constantly improving standards of care for this population, which will hopefully enhance life expectancy, obtaining accurate data or estimates of life expectancy for specific birth cohorts and defects is difficult (18). Future research might examine ratings of expected longevity by experts for specific CHD groups; use of vignettes presenting prototypical cases including age, surgeries and symptoms might also be of interest.
The implications of unrealistic or optimistic life expectations are unknown. To our knowledge, this is the first time adolescents or young adults with a chronic illness have been asked this question. Being overly optimistic may be an adaptive coping strategy in the face of acute health problems, and denial can be effective immediately after serious acute life events, but may become problematic with ongoing issues such as a chronic illness (e.g., Miller et al. ). For individuals with CHD, denial or unrealistic expectations could lead to avoidance of appropriate health care, and if regular adult congenital cardiac follow-up is lacking, could lead to the acute and sudden onset of adverse events ranging from arrhythmia to heart failure or sudden cardiac death. However, interviews of some patients with fairly realistic views had an element of hopelessness about them, which could be as problematic as being too optimistic. Interventions to improve patients’ adherence to medical follow-up and other health behaviors (e.g., good dental hygiene minimizes the risk of infective endocarditis ) that could reduce premature mortality are needed. Further, we do not know whether patients are in fact denying or being unrealistically optimistic about their life expectancies or if their ratings reflect missed educational opportunities. As a number of the interviews suggested, rarely were patients’ views based on discussions with their cardiologist or another health professional.
Discussion of CHD risks, and particularly life expectancy, with adolescents and young adults is a sensitive topic. If generally uninformed patients are suddenly told of all their health risks, they may become overwhelmed, which could lead to avoidance of cardiac care and failure to live up to their potential in social, educational, and occupational endeavors. Discussions need to account for the patient’s diagnoses, surgical history, current complications, ongoing improvements in medical and surgical care, and other factors that may contribute to a specific patient’s life expectancy. Discussions may also need to reflect the variability in outcomes, even among patients with similar CHDs, and the uncertainty in our current estimates of life expectancy in the CHD population. Patients need ongoing discussions, beginning in adolescence, that are also sensitive to their psychosocial development and life circumstances (2,23). As many adolescents and young adults turn to parents for advice and support, separate discussions with parents would also be prudent. With accurate knowledge, delivered in sensitive manner, patients will be in the best position to make informed decisions about life choices such as higher education and career choice, contraception and family planning, etc. Ideally patients need to hold accurate, but positive, views of their lives, and be able to tolerate the uncertainty associated with having a complex chronic illness.
For the qualitative analyses of interviews with young adults with congenital heart disease, please see the online version of this article.
↵1 Dr. Reid was supported as the Bill and Anne Brock Professor in Child Health at the time this paper was written.
This study was support by a grant from the Hospital for Sick Children Foundation (No. XG95-017).
- Abbreviations and Acronyms
- congenital heart disease
- Hospital for Sick Children
- Received November 22, 2005.
- Revision received March 21, 2006.
- Accepted March 28, 2006.
- American College of Cardiology Foundation
- Warnes C.A.,
- Liberthson R.,
- Danielson G.K.,
- et al.
- Foster E.,
- Graham T.P. Jr..,
- Driscoll D.J.,
- et al.
- Nieminen H.P.,
- Jokinen E.V.,
- Sairanen H.I.
- Weinstein N.D.,
- Klein W.M.
- Ware J.E.,
- Snow K.K.,
- Kosinski M.,
- Gandek B.
- Health Canada
- Thacker S.B.,
- Goodman R.A.,
- Koo D.
- Kolbe L.J.
- Patton M.Q.
- Tabachnick B.G.,
- Fidell L.S.
- Statistics Canada
- Statistics Canada
- Idler E.,
- Leventhal H.,
- McLaughlin J.,
- Leventhal E.
- Larsson D.,
- Hemmingsson T.,
- Allebeck P.,
- Lundberg I.
- Reid G.J.,
- Irvine M.J.,
- McCrindle B.W.,
- et al.