Quality of Life Research

, Volume 22, Issue 6, pp 1295–1304

Perceived health is partially associated with the symptomatological profile in patients with benign and severe conditions: the case of congenital heart disease

Authors

    • Department of CardiologyAcademic Medical Center
    • Department of Medical PsychologyAcademic Medical Center
    • Interuniversity Cardiology Institute of the Netherlands
  • M. A. G. Sprangers
    • Department of Medical PsychologyAcademic Medical Center
  • W. Budts
    • Division of Congenital and Structural CardiologyUniversity Hospitals Leuven
  • B. J. M. Mulder
    • Department of CardiologyAcademic Medical Center
    • Interuniversity Cardiology Institute of the Netherlands
  • S. Apers
    • Center for Health Services and Nursing ResearchCatholic University of Leuven
  • P. Moons
    • Division of Congenital and Structural CardiologyUniversity Hospitals Leuven
    • Center for Health Services and Nursing ResearchCatholic University of Leuven
    • The Heart CentreCopenhagen University Hospitals
Article

DOI: 10.1007/s11136-012-0241-4

Cite this article as:
Schoormans, D., Sprangers, M.A.G., Budts, W. et al. Qual Life Res (2013) 22: 1295. doi:10.1007/s11136-012-0241-4

Abstract

Purpose

Individuals with serious medical conditions can perceive their health status as good. This might be explained by the symptomatology inherent to the condition. Research in this respect is scarce. Congenital heart disease (CHD) is a spectrum of mild, moderate, and complex heart defects, representing more benign and severe chronic conditions. We investigated (1) symptomatology (i.e., symptom frequency and symptom distress) of CHD patients; (2) the extent to which symptomatology was independently related to perceived health; and (3) the relative importance of individual symptoms for perceived health.

Methods

A secondary data analysis on two separate patient samples (629 Belgian and 1,109 Dutch patients) was conducted. Patients’ symptomatology was measured with the TAAQOL–CHD. Perceived health was measured by the EQ-5Dvas in Belgian patients, and by a single item (EVGFP rating) of the SF-36 in Dutch patients. Linear regression analyses were performed to investigate the relationship between symptoms and perceived health, while controlling for sex, age, disease complexity, and functional status.

Results

The most frequently occurring symptoms were dizziness, palpitations, and nycturia. Symptom distress was associated with perceived health, independent of confounders. Symptom distress with respect to shortness of breath while walking; palpitations; and dizziness were independently related to perceived health.

Conclusions

Perceived health in CHD patients is partially associated with their symptomatology. This finding underscores the possibility that differences in perceived health across patient groups with more benign and severe conditions may be caused by the different impact conditions have—in terms of symptoms—on the day-to-day life.

Keywords

Perceived health statusSymptom frequencySymptom distressCongenital heart disease

Introduction

Perceived health is the subjective rating of one’s general health status [1]. This rating is based on various aspects including patients’ knowledge about their medical condition and symptoms [2]. Patients’ perceived health is frequently measured in research and clinical practice, because it can be easily assessed and gives a general idea about how individuals appraise their condition. In addition, it is a very powerful variable, because studies have found that it is predictive for mortality, and poor health outcomes such as morbidity, self-management, and health-care use [35]. Hence, it can guide health-care providers’ decision-making.

Research findings on perceived health show sometimes paradoxal and counterintuitive results. For instance, patients with a transposition of the great arteries or single ventricle physiology, which are obviously complex congenital heart defects (CHD), had a median score of 75–90 on the EQ-5Dvas that ranged from 0 to 100 (higher scores represent a better perceived health) [68]. Patients with an unruptured cerebral or an abdominal aortic aneurysm also perceived their health to be good (EQ-5Dvas = 80) [9, 10]. In contrast, patients with irritable bowel syndrome (EQ-5Dvas = 66) [11], psoriasis (EQ-5Dvas = 64) [12], or asthma (EQ-5Dvas = 61) [13] reported substantially lowers scores. In sum, patients with a severe condition such as a transposition of the great arteries, single ventricle physiology, or cerebral/aortic aneurysm perceive their health as good, whereas the other patients with a more benign condition perceive their health as poorer.

It is well-known that within-group variations in perceived health are partially due to differences in experienced symptoms. The question then emerges whether the symptomatology is also responsible for differences in perceived health of individuals with different chronic conditions. It can be hypothesized that differences in the symptomatological profile inherent to a condition are the reason why patients with more benign conditions perceive their health as poorer than individuals with obviously serious medical conditions.

A study on the direct relationship of perceived health and symptoms is not self-evident. Symptoms are often disease-specific, hampering comparisons of the symptoms across patients with different disorders. Furthermore, prior studies were mainly undertaken in patient populations with a prominent symptomatological profile, such as asthma or eczema. Therefore, there is a need for studies conducted in patients with a more subtle symptom experience, and in disorders that represent the spectrum of benign to severe conditions.

In the present study, we assessed the relation between symptomatology and perceived health in CHD patients. Indeed, CHD is an appropriate patient population to test this relationship because CHD comprises a broad spectrum of defects (ranging from mild to complex defects) representing more benign and serious chronic conditions. Moreover, in contrast to other conditions where severity is often based on the symptoms displayed (e.g., eczema), the complexity of a CHD is objectively assessed based on the type of defect. Furthermore, the complexity of a CHD is found to be (largely) unrelated to perceived health [1416]. Thus, variation in perceived health across CHD patients is automatically the result of other factors than disease severity, such as symptom experience. It may be the case that a patient with a complex defect who experiences few symptoms consequently reports a better perceived health, than a patient with a simple defect who experiences many symptoms. Hence, our specific aims were (1) to examine which symptoms are most frequent and distressing in patients with CHD; (2) to examine the extent to which symptomatology (i.e., symptom frequency and symptom distress) was independently related to perceived health; and (3) to examine the relative importance for individual symptoms on patients’ perceived health.

Methods

We undertook a secondary analysis on data from two prior studies (i.e., a Belgian and a Dutch patient sample) that used similar methods [15, 16].

Study population and procedure

Belgian patient sample

Seven hundred and sixteen adult patients with CHD attending the outpatient clinic of the Adult Congenital Heart Disease Program of the University Hospitals Leuven were eligible for inclusion in this study [15]. Patients were excluded if they were mentally impaired, illiterate in Dutch, visited the outpatient clinic for the first time, referred for or in follow-up after percutaneous closure of an atrial septal defect (ASD), or a patent foramen ovale [15]. Institutional Review Board approval was provided by the medical ethics committee of the University Hospital Leuven. Verbal informed consent was provided by all included patients.

Dutch patient sample

Sixteen hundred and seventy adult patients with CHD from five tertiary referral and three regional Dutch centers were randomly selected from the CONCOR database, a Dutch registry for patients with CHD [17]. Exclusion criteria were being diagnosed with Marfan syndrome, being mentally impaired, or illiterate in Dutch [16]. Under Dutch law, ethical approval is not required for questionnaire studies. Therefore, the study was exempted from approval of the Medical Ethics Committee of the Academic Medical Center in Amsterdam.

Measurements

Demographics and clinical variables

Belgian patient sample

Sex, age, and type of CHD were obtained from patients’ medical records. Based on the type of CHD, we have categorized patients into three classes of disease complexity (i.e., mild, moderate, and complex) according to the Task Force 1 of the 32nd Bethesda conference of the American College of Cardiology (ACC) [18]. Functional status was assessed by the New York Heart Association (NYHA) classification [19] and was scored by one cardiologist [14].

Dutch patient sample

Sex and age were both measured through self-report. Type of CHD was subtracted from the CONCOR database and categorized into classes of disease complexity, according to the Task Force 1 of the 32nd Bethesda conference of the ACC [18]. Functional status was measured by one patient-based question assessing the NYHA classification [19].

Symptomatology

Belgian and Dutch patient sample

Somatic symptoms as experienced by the patients were measured by the 9-item Symptoms subscale of the TNO/AZL Adult Quality Of Life–CHD (TAAQOL–CHD) in both patient samples [20]. For each item, both the frequency and perceived level of distress were scored. First, the frequency of each symptom during the past month is scored on a 3-point scale (0 = never; 1 = occasionally; 2 = often). If a symptom occurred (i.e., score 1 or 2 on frequency), the level of the associated distress is assessed on a 4-point scale, ranging from 1 (not at all) to 4 (very much). Consistent with the TAAQOL–CHD algorithm [20], the scores on both frequency and distress were combined into one score ranging from 0 to 4, with a zero if the symptom did not occur, and the one to four distress-scores if the symptom did occur. For the purpose of this study, we calculated two sum scores. First, after dichotomizing the frequency score according to the presence of the symptom (no = 0; yes = 1), a total symptom frequency score was computed by summing the scores over all 9 items, yielding a total score ranging from 0 to 9. Second, we calculated symptom distress, based on the combined item scores. The total symptom distress score could range from 0 to 36, with higher scores reflecting a higher level of symptom distress.

Perceived health

Belgian patient sample

Perceived health was measured by the EQ-5Dvas, a vertically oriented 10-cm line. The line was graded with indicators that ranged from 0 (worst perceived health) to 100 (best perceived health). Patients were asked to mark the place on the line indicating their perceived health. The EQ-5Dvas has good validity, reliability, and responsiveness when used in patients with CHD [15].

Dutch patient sample

Perceived health was assessed by the first item of the Short Form Survey-36 (SF-36): “In general would you say your health is?” also known as the EVGFP rating [21]. Answers were given on a 5-point Likert scale with the options; “excellent,” “very good,” “good,” “fair,” and “poor.” To facilitate easy interpretation, the response options were recoded, so that a higher score indicated better perceived health.

Statistical analyses

Exploratory analyses showed that there were missing values for both symptomatology variables (symptom frequency and symptoms distress). For both variables, missing values were imputed with a mean-value if more than 75 % of the data was available for that person. Given this imputation of missings, we ran our analyses with missings imputed and by excluding all missings. Results were compared via a sensitivity analysis and showed stable results. We therefore chose the options were missings were imputed.

First, to establish the rank order of the most frequently occurring symptoms (symptom frequency) and the most distressing ones (symptom distress), we calculated frequencies and means, respectively.

Both samples were tested in a different way given the measurement differences in assessing perceived health. The EQ-5Dvas is a continuous variable and was used in the Belgian sample, whereas the categorical EVGFP rating was employed in the Dutch patient sample. Therefore, linear regression analyses were used for the Belgian patient sample, and ordinal regression analyses for the Dutch patient sample. As part of the ordinal regression analyses, an indicator for goodness-of-fit for each model was provided as the Akaike information criterion (AIC). A lower score on the AIC indicates a better fitting model.

The assumption that disease complexity is not related to perceived health was tested by means of a regression analysis. Subsequently, we examined the extent to which symptomatology (i.e., symptom frequency and symptom distress) was independently related to perceived health, while adjusting for the four confounders: sex, age, disease complexity, and functional status. As there was multicollinearity between the two symptomatology variables, we employed two separate regression analyses for each symptomatology variable. Thus, the confounders and (a) symptom frequency and (b) symptom distress were associated with perceived health. Given that data from the Dutch sample came from multiple centers, we correct for clustering by adding a categorical variable representing these centers.

Third, to explore the relevant importance of each symptom for perceived health, an additional regression analysis was employed. Symptom distress of all nine symptoms was simultaneously entered as independent variables together with the confounders (i.e., sex, age, disease complexity, and functional status). In the Belgian study, over 50 % of patients had missing values on the item “shortness of breath walking 1–5 km” due to misinterpretation by patients. Therefore, this item was excluded from the analysis in the Belgian study. In analyzing the data from the Dutch sample, we again corrected for clustering within the centers.

All analyses were done with SPSS 16.0. We used two-sided tests, and a significance level of 5 % was employed.

Results

Patient characteristics

Belgian patient sample

Of the 716 eligible Belgian patients, 629 participated in this study (response rate 87.8 %). Approximately 60 % of patients were male, with a mean age of 40 years (Table 1). Twenty-six percent of patients had a mild congenital heart malformation, whereas 58.0 % had a moderate malformation. Most patients (81.5 %) were categorized in NYHA-I. The most common types of congenital heart malformations were tetralogy of fallot (TOF) and pulmonary stenosis (PS) (n = 160, 25.4 %).
Table 1

Patient characteristics

 

Belgian patient sample (n = 629)

Dutch patient sample (n = 1,109)

Sex (male)

378 (60.1)

502 (45.8)

Age in mean years (SD)

39.5 (13.4)

40.7 (14.42)

Disease complexity

 Mild

164 (26.1)

644 (58.1)

 Moderate

365 (58.0)

366 (33.0)

 Complex

100 (15.9)

99 (8.9)

Functional status

 NYHA-I

511 (81.5)

754 (69.9)

 NYHA-II

85 (13.6)

232 (21.5)

 NYHA-III

26 (4.1)

62 (5.8)

 NYHA-IV

5 (0.8)

30 (2.8)

Congenital heart defect

 VSD/ASD/ASD-I/PDA

140 (22.3)

397 (35.8)

 TOF/PS

160 (25.4)

218 (19.7)

 BAV/AS/SubvAS

93 (14.8)

151 (13.6)

 CoA

89 (14.1)

137 (12.4)

 TGA/ccTGA

46 (7.3)

42 (3.8)

 Other CHD diagnoses

101 (16.1)

164 (14.8)

Numbers are given in frequencies (percentages), age in mean years (SD)

n numbers, SD standard deviation, NYHA New York Heart Association functional class, VSD ventricular septum defect, ASD atrial septum defect, ASD-I atrial septum defect premium, PDA patent ductus arteriosus, TOF tetralogy of fallot, PS pulmonary stenosis, BAV bicuspid aortic valve, AS aortic stenosis, SubvAS subvalvar aorta stenosis, CoA coarctation of the aorta, TGA transposition of the great arteries, ccTGA congenital corrected transposition of the great arteries, CHD congenital heart defect

Dutch patient sample

Of the 1,670 eligible Dutch patients, 1,109 participated in this study (response rate 66.4 %). Almost half of the patients (45.8 %) were male. The average age for the total group was 41 years (Table 1). More than half of the patients (58.1 %) were diagnosed with a mild heart malformation, whereas 33.0 % had a moderate malformation. Almost 70 % was classified in NYHA-I. The most common types of congenital heart malformations were Ventricular Septum Defect (VSD), (ASD) or Patent Ductus Arteriosus (PDA) (n = 397, 35.8 %).

Symptomatology and perceived health

Belgian patient sample

Table 2 shows the rank order of the most frequently occurring (symptom frequency) and distressing symptoms. The three most frequently occurring symptoms in Belgian patients were dizziness, palpitations, and nycturia. The three most distressing symptoms were orthopnea, edema, and shortness of breath walking <100 m. Patients experienced on average 2.57 symptoms (SD = 1.93) and reported an average symptom distress score of 4.76 (SD = 4.64). Patients perceived their health to be good (median EQ-5Dvas = 80, SD = 14.38).
Table 2

Rank order of most frequently occurring and distressing symptoms

Symptom frequency (yes)

Symptom distress*

Belgian patient sample

N (%)

Dutch patient sample

N (%)

Belgian patient sample

Mean (SD)

Dutch patient sample

Mean (SD)

1. Dizziness

265 (42.4)

1. Nycturia

560 (51.1)

1. Orthopnea

2.29 (0.83)

1. Orthopnea

2.45 (0.87)

2. Palpitations

251 (40.0)

2. Palpitations

527 (48.0)

2. Edema

2.12 (0.88)

2. Short of breath walking <100 m

2.42 (0.87)

3. Nycturia

235 (37.5)

3. Dizziness

495 (45.3)

3. Short of breath walking <100 m

2.11 (0.88)

3. Short of breath walking 1–5 km

2.29 (0.89)

4. Pale

214 (34.3)

4. Short of breath walking 1-5 km

326 (30.4)

4. Palpitations

2.06 (0.88)

4. Palpitations

2.27 (0.89)

5. Short of breath walking <100 m

168 (31.8)

5. Pale

286 (26.2)

5. Cyanosis during exercise

1.90 (1.04)

5. Edema

2.23 (0.87)

6. Orthopnea

108 (17.3)

6. Edema

250 (22.9)

6. Dizziness

1.86 (0.82)

6. Dizziness

1.85 (0.79)

7. Cyanosis during exercise

51 (8.2)

7. Orthopnea

184 (16.8)

7. Pale

1.51 (0.72)

7. Cyanosis during exercise

1.81 (0.79)

8. Edema

43 (7.0)

8. Short of breath walking <100 m

179 (16.4)

8. Nycturia

1.50 (0.79)

8. Pale

1.81 (0.80)

  

9. Cyanosis during exercise

88 (8.0)

  

9. Nycturia

1.75 (0.87)

Dutch patient sample

In the Dutch patient sample, nycturia, palpitations, and dizziness were the most frequently occurring symptoms (Table 2). Patients were mostly distressed about orthopnea and shortness of breath when walking (Table 2). Patients had a mean symptom frequency score of 2.66 (SD = 2.22) and a mean symptom distress score of 5.50 (SD = 5.84). Around 80 % of patients rated their perceived health either as “good” (9.9 %), “very good” (18.4), or “excellent” (52.9 %).

Symptom perception as an associate of perceived health

The assumption that disease complexity was not related to perceived health was confirmed in both patient samples (data not shown).

Belgian patient sample

Results of the multivariable linear regression analysis revealed that symptom distress was significantly associated with perceived health (B = −1.301; p < 0.001), even when adjusted for sex, age, disease complexity, and functional status (Table 3). More specifically, patients with a high level of symptom distress rated their health as poor. Other factors that contributed to a worse perceived health were male gender or poor functional status (i.e., higher NYHA score). Nearly 24 % of the variance in perceived health was explained by this model. Symptom distress was the strongest explanatory factor in this model, as illustrated by the highest absolute β value: −0.420 (Table 3). Result of the model including symptom frequency instead of symptom distress was similar with symptom frequency as the strongest explanatory factor (β value was −0.326).
Table 3

Relationship between symptom distress and perceived health, adjusted for potentially confounding factors for the Belgian patient sample

 

B

Standard error

β

p value

Belgian patient sample

    

 Males

−2.508

1.074

−0.085

0.020

 Age

−0.073

0.064

−0.041

0.259

 Disease complexitya

    

  Moderate

0.313

1.214

0.014

0.796

  Severe

2.687

1.727

0.012

0.120

 Functional statusb

    

  NYHA-II

−6.871

1.583

−0.268

<0.001

  NYHA-III

−8.833

2.744

−0.345

0.001

  NYHA-IV

−1.296

5.724

−0.051

0.821

 Symptom distress

−1.301

0.119

−0.420

<0.001

NYHA New York Heart Association functional class

aReference category is mild complexity

bReference category for functional status is NYHA-I

Dutch patient sample

Results of the ordinal regression show that also in the sample of Dutch patients, symptom distress was related to perceived health (OR = 0.906, 95 % CI 0.890–0.922; p < 0.001), while adjusting for the confounders sex, age, disease complexity, and functional status. In addition to a higher level of symptom distress, being male, older age, and a poorer functional status were significantly associated with worse perceived health. Symptom distress was the variable that contributed the most (highest Wald statistic, Wald χ2 = 118.143) (Table 4). The AIC was 2,035.989. Similar results were found in the model including symptom frequency instead of symptom distress. Here, symptom frequency (Wald χ2 was 96.318) was the strongest explanatory factors. The AIC was 2,071.508.
Table 4

Relationship between symptom distress and perceived health, adjusted for potentially confounding factors for the Dutch patient sample

 

Wald χ2

Adjusted OR (95 % CI)

p value

Dutch patient sample

   

 Males

5.441

0.841 (0.727–0.973)

0.020

 Age

12.052

0.990 (0.984–0.996)

0.001

 Disease complexitya

   

  Moderate

0.437

1.055 (0.901–1.234)

0.509

  Severe

0.825

0.887 (0.685–1.149)

0.364

 Functional statusb

   

  NYHA-II

33.413

0.544 (0.443–0.669)

<0.001

  NYHA-III

38.379

0.323 (0.226–0.462)

<0.001

  NYHA-IV

20.024

0.293 (0.171–0.502)

<0.001

 Symptom distress

118.143

0.906 (0.890–0.922)

<0.001

 Centerc (overall effect)

9.721

Not available

0.205

NYHA New York Heart Association functional class

aReference category is mild complexity

bReference category for functional status is NYHA-I

cThe overall effect for the categorical variable center and is provided by the Wald χ2 test and its p value

The relative impact of each symptom on perceived health

Belgian patient sample

Adjusted for potentially confounding factors, we found six symptoms to be associated with worse perceived health: shortness of breath walking <100 m, orthopnea, edema, palpitations, dizziness, and cyanosis during exercise (Table 5). Moreover, male sex and a poor functional status were independently related to a poor perceived health. In total, 27.5 % of the variance in perceived health was explained.
Table 5

Relative importance of individual symptom to perceived health, while controlling for confounders for the Belgian patient sample

 

B

Standard error

β

p value

Belgian patient sample

    

 Males

−1.865

1.111

0.063

0.094

 Age

−0.092

0.067

−0.051

0.169

 Disease complexitya

    

  Moderate

−0.019

1.233

0.001

0.988

  Severe

2.239

1.758

0.100

0.203

 Functional statusb

    

  NYHA-II

−6.675

1.631

−0.261

<0.001

  NYHA-III

−6.690

2.905

−0.261

0.022

  NYHA-IV

1.639

5.817

0.064

0.778

 Shortness of breath walking <100 m

−2.295

0.520

−0.176

<0.001

 Orthopnea

−1.606

0.677

−0.104

0.018

 Nycturia

−0.679

0.616

−0.041

0.271

 Edema

2.174

0.933

0.089

0.020

 Palpitations

−1.620

0.564

−0.130

0.004

 Dizziness

−2.835

0.570

−0.210

<0.001

 Cyanosis during exercise

−2.301

0.969

−0.096

0.018

 Looking pale

0.038

0.710

0.002

0.957

NYHA New York Heart Association functional class

aReference category is mild complexity

bReference category for functional status is NYHA-I

Dutch patient sample

In the Dutch sample, the following symptoms were related to a lower level of perceived health when adjusted for confounders: shortness of breath walking 1–5 km, orthopnea, nycturia, palpitations, and dizziness (Table 6). Additionally, being male, older age, and a poor functional status were independently related to a poor perceived health. The AIC was 1,923.721.
Table 6

Relative importance of individual symptom to perceived health, while controlling for confounders for the Dutch patient sample

 

Wald χ2

Adjusted OR (95 % CI)

p value

Dutch patient sample

   

 Males

4.677

0.845 (0.726–0.984)

0.031

 Age

8.899

0.990 (0.984–0.997)

0.003

 Disease complexitya

   

  Moderate

0.385

1.053 (0.894–1.240)

0.535

  Severe

0.794

0.884 (0.674–1.159)

0.373

 Functional statusb

   

  NYHA-II

30.402

0.547 (0.441–0.678)

<0.001

  NYHA-III

35.145

0.302 (0.203–0.449)

<0.001

  NYHA-IV

21.550

0.246 (0.136–0.445)

<0.001

 Shortness of breath walking <100 m

0.007

1.006 (0.884–1.143)

0.933

 Shortness of breath walking 1–5 km

13.147

0.829 (0.749–0.917)

<0.001

 Orthopnea

4.202

0.905 (0.822–0.996)

0.040

 Nycturia

4.349

0.918 (0.847–0.995)

0.037

 Edema

2.877

0.926 (0.846–1.012)

0.090

 Palpitations

12.312

0.881 (0.820–0.946)

<0.001

 Dizziness

8.866

0.893 (0.829–0.962)

0.003

 Cyanosis during exercise

0.542

1.064 (0.902–1.254)

0.462

 Looking pale

0.711

0.958 (0.686–1.058)

0.399

 Centerc (overall effect)

10.647

Not available

0.155

NYHA New York Heart Association functional class

aReference category is mild complexity

bReference category for functional status is NYHA-I

cThe overall effect for the categorical variable center is provided by the Wald χ2 test and its p value

Discussion

To our knowledge, this is the first study examining the direct relation between symptomatology and perceived health in a population with a subtle symptom experience profile that represents a broad spectrum of benign to severe conditions. We tested this relation in two large samples of patients with CHD, an appropriate population, because CHD comprises a broad range of defects (mild to complex) representing benign to severe chronic conditions. We found that perceived health in patients with CHD is associated with their symptomatology, in particular symptom distress, rather than disease severity. Indeed, in both samples, symptom distress was the strongest predictor (highest absolute beta coefficient in the Belgian patient sample and the highest Wald χ2 in the Dutch patient sample).

Comparison of the analyses including either symptom frequency or symptom distress showed that the model including symptom distress explained the highest proportion of variance in the Belgian sample and the lowest AIC score in the Dutch sample. This indicates that symptom distress is a stronger correlate of perceived health than symptom frequency.

The most frequently experienced symptoms in patients with CHD were dizziness, palpitations, and nycturia. The most distressing symptoms, however, were orthopnea and shortness of breath walking <100 m. This finding is in line with prior studies in various patient groups, showing that the most frequently occurring symptoms are not necessarily the most distressing ones [22, 23]. In relation to patients’ perceived health, distress about shortness of breath when walking, palpitations, and dizziness were important in both patient samples. In addition, for the Belgian patients, distress about edema and cyanosis during exercise were independently related to their perceived health. A potential explanation for the difference in the relation between specific symptoms and perceived health across samples may be the different measurement instruments for perceived health.

The findings of our study seem to be robust. The relationship between symptoms and perceived health was confirmed in the two independent studies that we report in this article. Furthermore, it was corroborated in a recent article by Berghammer et al. [7] that was published online at the time of this writing. Berghammer et al. concluded that symptomatic patients with CHD rate their health as poorer than asymptomatic patients. This relation remained significant after controlling for age, sex, functional status, and diagnosis. An important difference between our studies and the Berghammer study is that in the latter, symptoms were rated by cardiologists, whereas in our studies, symptoms were measured using a structured self-report questionnaire. Patients’ symptomatology is based on patients’ subjective perceptions. The patient is therefore by definition the expert on these subjective perceptions and can thus be considered the gold standard. It is known that physicians’ report of symptoms is underestimating the symptomatological profile of patients, when compared with standardized assessment through self-report questionnaires (the gold standard) [24]. Hence, it is likely that symptoms are underrated in the Berghammer study. Indeed, only 30 % of their patients were considered to be symptomatic, whereas this proportion was 82.7 and 80.6 % in our studies. Furthermore, the Berghammer study only included the presence of symptoms, whereas we have assessed both symptom frequency and symptom distress. Irrespective of these differences, their findings are in keeping with our results.

Clinical implications

In general, it is recommended that physicians assess patients’ perceived health, since it gives the physician a general sense of how patients rate their health. In addition, we advise that next to clinical objective data, patients’ perceived health is taken into consideration when physicians discuss treatment planning. Information on why patients rate their perceived health the way they do is inevitable, particularly in the case where there is a discrepancy between patients’ perceived health and physicians’ evaluation. This knowledge may help clinicians in deciding which perspective preponderates in the decision-making. The results of our study, moreover, suggest that a systematic appraisal of symptoms, with in particular symptom distress, is essential in the assessment of perceived health, irrespective of disease complexity. Indeed, one of the included Dutch patients was a 39-year-old male with a complex defect (i.e., transposition of the great arteries) who reported no symptoms or symptom distress and perceived his health to be excellent. In contrast, a Dutch female 25-year-old patient with a mild defect (i.e., atrial septum defect) who reported seven symptoms and experienced severe symptom distress perceived her health to be only fair. It is advocated to use standardized assessments to appraise symptoms, because spontaneous reports of symptoms by patients during clinical interview only give a partial view on the symptomatological profile of patients [24]. This would give clinicians better insight in the illness perception of patients and allow them to better understand why patients consider their health status as good or bad, irrespective of the objective condition.

Methodological issues

The strengths of this study are worth mentioning. First, data of two large patient samples (total number of patients is 1,738) recruited at nine different centers and across two countries were included. Second, not only symptom frequency but also perceived distress regarding symptoms was measured.

However, our study also has some limitations. First, although the study designs were quite similar across both patient samples, data could not be pooled due to differences in measurements. Primarily, perceived health was measured by the EQ-5Dvas in Belgian patients and by the EVGFP in Dutch patients. Moreover, functional status was appraised by a cardiologist in the Belgian sample but was assessed by patients themselves in the Dutch sample. Furthermore, one item of the TAAQOL–CHD symptoms subscale was missing in the Belgian sample. Second, the Belgian and Dutch sample differed with respect to disease complexity; the number of patients with mild congenital malformations was larger in the Dutch study. This might be a result of the selection procedure. Belgian patients were recruited at the outpatient clinic, whereas Dutch patients were randomly selected from a national database. This difference in distribution of disease complexity was controlled for in the multivariable statistical analyses. Although some further baseline characteristics of the patient groups were slightly different, the relationship between symptomatology and perceived health was confirmed in both patient samples. Third, it is important to note that both studies were conducted in Western Europe. It is well known that studies in other countries or continents conclude that patients with CHD are more negatively affected by their condition than Belgian or Dutch patients [25, 26]. Therefore, results from this study cannot be automatically generalized to populations of persons with CHD. Fourth, other factors that are not measured in our study are likely to be involved as well. For example, illness perceptions, depression, anxiety, and Type D personality are also found to be related to patients’ perceived health [16, 27] and symptom perception [2830]. Indeed, our multivariable model explained only 23.6 % of the variance in perceived health. Fifth, although the SF-36 is a well-validated questionnaire, we were faced with the asymmetric nature of the EVGFP rating. More specifically, the 5 response categories for the question “In general, would you say your health is” are as follows: excellent, very good, good, fair, and poor. The middle category represents a good health status, which has a positive instead of a neutral connotation. Therefore, direct comparison with the EQ-5Dvas is not possible, even if the EVGFP rating is transformed into a 100-point scale. However, we believe that the inclination of our regression analyses is not affected by these differences in measurements of perceived health, although it may have affected the strength of our relations and model fit.

Conclusion

In a heterogeneous group of adult patients with CHD, symptomatology, and in particular symptom distress, is related to perceived health. Confirmation that symptoms are related to perceived health in a patient population comprising more benign and severe conditions and experiencing a low symptom burden underscore the possibility that the differences in perceived health across patient groups are caused by the different impact that conditions have—in terms of symptoms—on the day-to-day life. Studies examining perceived health should be interpreted in light of our conclusions, as paradoxal results (e.g., good perceived health in seriously ill patients) could be explained by the symptomatological profile.

Acknowledgments

This study was partially funded by the Interuniversity Cardiology Institute of the Netherlands.

Copyright information

© Springer Science+Business Media B.V. 2012