Correlates of health-related quality of life among lower-income, urban adults with congestive heart failure

Correlates of health-related quality of life among lower-income, urban adults with congestive heart failure Daniel O. Clark, PhD,a,b Wanzhu Tu, PhD,a,b Michael Weiner, MD, MPH,a,b and Michael D. Murray, PharmD,b,c Indianapolis, Indiana

BACKGROUND: Improving health-related quality of life (HRQL) is a primary goal in the treatment of patients with congestive heart failure (CHF), yet few studies have explored correlates of HRQL among CHF patients. OBJECTIVES: We report on the association of demographic and pathophysiologic measures, socialcognitive measures, and environmental variables with HRQL as measured by the Kansas City Cardiomyopathy Questionnaire (KCCQ), Chronic Heart Failure (CHQ), and a single question of perceived overall health (PH). METHODS: Cross-sectional data were obtained from the baseline interview and electronic medical records of 212 patients 50 years of age and older who were enrolled during the first 7 months of a medication adherence study. RESULTS: Mean age was 63; 32% were male; 53% were black; the mean Charlson comorbidity score was 3.7; and the mean New York Heart Association class was 2.1. Correlations between KCCQ and CHQ subscale scores and PH ranged from 0.16 to 0.37. Multivariate regression analyses showed that the pathophysiologic measures ejection fraction and comorbidity were not associated with any of the HRQL measures. Overall PH was associated with greater age and more positive health beliefs. Persons of greater age, males, and black respondents had higher CHF-specific HRQL scores, as did persons reporting more positive health beliefs, greater income, social support, and communication with their physician. Variance explained ranged from 14 to 33%. CONCLUSION: These cross-sectional data highlight the potential significance of social and behavioral factors in CHF-specific HRQL. (Heart Lung® 2003;32:391-401.)

INTRODUCTION Congestive heart failure (CHF) is a disease associated with considerable suffering and enormous costs. The toll on patients and families is high and the vast majority of persons with congestive heart failure report reduced function and quality of life.1 Consequently, improving healthrelated quality of life (HRQL) has become a focus of many clinical trials and is a major goal in the From the aIndiana University Center for Aging Research, Indianapolis, Indiana, bRegenstrief Institute, Inc., Indianapolis, Indiana, cPurdue University School of Pharmacy, Purdue University School of Pharmacy, Indianapolis, Indiana. Reprint requests: Daniel O. Clark, PhD, Indiana University Center for Aging Research, 1050 Wishard Blvd., RG 6, Indianapolis, IN 46202. Copyright © 2003 by Mosby, Inc. 0147-9563/2003/$30.00 ⫹ 0 doi:10.1016/j.hrtlng.2003.07.005

HEART & LUNG

VOL. 32, NO. 6

treatment of patients with CHF. It has been estimated that CHF costs nearly $19 billion per year in direct medical costs2 and is directly or indirectly implicated in more than 250,000 deaths per year.3 In this article, we use the term HRQL to refer to patients’ perceptions of their physical, social, and psychological domains of health.4 Its purpose is to explore the correlates of HRQL within a sample of urban primary care patients who have CHF. Although there have been reports indicating that CHF-specific HRQL measures are psychometrically superior to generic HRQL measures when studying CHF,5 there is no gold standard for HRQL among patients with CHF. Consequently, we compare results based on 2 CHF-specific instruments, representing 5 domains, and a single question of patients’ perceptions of their overall health.

www.heartandlung.org

391

Correlates of health-related quality of life

Clark et al

Fig 1 Adaptation of HRQL model.

REVIEW OF LITERATURE Several CHF-specific HRQL instruments are currently available, and these instruments have received careful evaluation.6-8 A few recent studies have explored relationships between HRQL instruments and various outcomes and have shown that, independent of medical indicators, the instruments have good predictive validity with regard to mortality and hospitalization.9 Relatively few studies, however, have conceptualized and modeled correlates of HRQL among patients with CHF. Studies exploring correlates of HRQL have generally shown that medical indicators and demographic characteristics are related to HRQL, but inconsistencies have been identified.6,9,10 The relationship between HRQL and ejection fraction, for example, is not well established, with some studies finding little11 or no association.12 Gender, which could have social or biological determinants, has been shown to be associated with HRQL, but whether men or women report better HRQL has varied by study.10,13,14 Most published studies have neither specified a clear model of HRQL nor had the measures to assess such a model.

392

www.heartandlung.org

In our attempt to assess correlates of HRQL among patients with CHF, we draw from a model of HRQL presented by Wilson and Cleary.15 This is a heuristic, theoretical model that identifies demographic and pathophysiologic antecedents to HRQL. The model also specifies indirect pathways via individual characteristics, such as beliefs and perceptions about health, and environmental factors, such as social support. Fig 1 presents an adaptation of this HRQL model and shows the variables we have been able to measure. Demographic and pathophysiologic factors have direct and indirect associations with HRQL. Patients’ characteristics and environmental resources are hypothesized to have direct associations with HRQL and may serve as pathways in the indirect associations of demographic and pathophysiologic measures with HRQL. In determining what measures to include in the models estimated here, we were guided by several factors. First, we were limited to those available in the study that is described below. Second, we were limited by our sample size of about 200. Finally, we selected variables that were consistent with the the-

NOVEMBER/DECEMBER 2003

HEART & LUNG

Clark et al

Correlates of health-related quality of life

oretical components (ie, boxes) of Fig 1. There were few environmental resource variables. All participants have a common exposure to 1 health care system, and most live in urban, low-income neighborhoods of 1 Midwestern city. The study did not have resources for home visits. Variables reflecting participants’ personal characteristics were more abundant, but some are not well established in the literature. We included variables that have apparent theoretical value and have some basis in the literature on chronic disease and health. Although research in the area of CHF and HRQL has been insufficient to provide a priori hypotheses regarding each of the variables in our model, the model presented serves as a simple guide in our exploratory analyses of the associations of demographic and pathophysiologic factors with HRQL, particularly the indirect associations via cognitive characteristics and environmental resources. We use baseline data from an ongoing randomized controlled trial. The population from which the sample is drawn consists of patients of an urban community health system who are characterized by low educational attainment, limited resources, and ethnic diversity. We have both medical records and questionnaire data on 209 patients with confirmed CHF.

METHODS This report is based on a multivariate analysis of cross-sectional data obtained through patient interviews and electronic medical records data.

Study population Participants for this study were recruited from Wishard Health Services (WHS) in Indianapolis, Indiana. WHS is a large, urban public health system. English-speaking patients 50 years of age or older who had CHF diagnosed by a physician were eligible. Participants planned to continue to receive their medical care and prescriptions through WHS including angiotensin-converting enzyme (ACE) inhibitors, beta-adrenergic receptor antagonist medications, diuretics, digoxin, or spironolactone. Of 40,334 outpatients 18 years of age or older seen at WHS in 1999, 2,165 (5.4%) had a clinician’s diagnosis of CHF; 1,720 (79%) of the patients with CHF were 50 years of age or older, and, among these older patients with CHF, 61% were women and 42% were black. Common comorbid conditions included hypertension (68%) and coronary artery disease (29%). Potential participants were identified through clinicians’ diagnostic codes stored in a comprehensive electronic medical records system.

HEART & LUNG

VOL. 32, NO. 6

After identification, the candidates were invited to participate during office visits at the general medicine clinics of WHS. Between March 19, 2001 and October 7, 2002, a total of 1290 patients were approached. Of the 1290 patients approached, 754 (58%) did not meet the inclusion criteria of the study; 332 (26%) consented to participate in the study. Of the 332 consented subjects, 257 were actually recruited and 228 of them completed the baseline phase of the study. Sixty-eight percent of the 228 patients were female and 47% were nonHispanic white with the all but a few of the remainder self-identified as black.

Measures After enrollment, an assistant interviewed the patient by telephone or in person. All interviews were conducted by any of 3 trained interviewers. On average, interviews took 35 to 45 minutes. Data from medical records supplemented the interview data.

Demographic and pathophysiologic measures Age, gender, ejection fraction and data to construct the Charlson Comorbidity Score16 were extracted from the Regenstrief Medical Records System (RMRS). The RMRS is a comprehensive electronic medical record system containing outpatient, inpatient, and pharmacy visit data.17 Workstations are located in every WHS care site to record all orders, prescriptions, and diagnoses electronically.18 CHF may be characterized by systolic or diastolic dysfunction, and an ejection fraction of 40% or less is considered an indication of left ventricular systolic dysfunction. The Charlson is an established and widely used measure of comorbidity19 and provides a single score reflecting the number and severity of selected conditions, including history of myocardial infarction and chronic heart failure. The RMRS data were used to identify diagnoses on record as many as 5 years before the participant’s enrollment date and to construct Deyo et al’s16 ICD-based Charlson score. Scores range from 0 to 18, with a higher score indicating more comorbidity. Race was determined by self-report on the baseline survey using the question, “What race do you consider yourself to be?” There was a large response set including white, black, and Hispanic. Age, race, and gender are difficult to conceptualize in a model such as in Fig 1, because they have both social and biological components. For the purposes of this report, we included these measures with ejection

www.heartandlung.org

393

Correlates of health-related quality of life

fraction and comorbidity but recognize that this is a choice of convenience.

Health-Related Quality of Life Measures As noted above, HRQL represents the health component of patients’ ratings of their quality of life. This generally involves several domains, including physical and emotional symptoms, function, and perceived overall health. These domains can be modeled separately as specified by Wilson and Cleary15 or as a group. We chose to model them separately as a summary score of functional and clinical status from the Kansas City Cardiomyopathy Questionnaire (KCCQ)6 and assessments of dyspnea, emotion, and fatigue from the Chronic Heart Failure Questionnaire (CHQ).8 In addition, we model a single, common question regarding perceived overall health. The functional status and clinical summary scores from KCCQ represent physical limitations, physical and psychological symptoms, and their impact on quality of life. The KCCQ questionnaire was recently shown to have good reliability and validity relative to the Minnesota Living with Heart Failure Questionnaire, the New York Heart Association (NYHA) classification, and the ShortForm 36.6 The CHQ has shown moderate correlations with global ratings of health and walk tests and clinical assessments.8 Perceived health has been widely used as a measure of general wellbeing and has been consistently shown to be a good predictor of disability and mortality.20 This measure consists of a single question asking a participant to rate his or her own health, with a response of excellent, good, fair, or poor. Each of these 3 measures, KCCQ, CHQ, and perceived health, is coded such that a higher score represents better HRQL. We also report the sample’s distribution on the NYHA classification tool. The NYHA classifications represent a clinician’s judgment of the patient’s disease severity, based on symptoms and physical function. Scores range from 1 to 4, with 4 being the most severe. Because symptoms and function make up central components of HRQL, and we have more detailed HRQL scores through the KCCQ and CHQ, we do not use NYHA scores in our HRQL models. We do report distributions on the NYHA to allow comparison of our sample to that of other studies.

Participants’ Characteristics As shown in the model of Fig 1, the measures in this group represent factors that, theoretically, affect a person’s physical and emotional expression of

394

www.heartandlung.org

Clark et al

demographic and biological factors. With the exception of education, the instruments and variables representing participants’ characteristics are not well established, so we have presented psychometric properties of these variables using our data. The Health Belief Scale (HBS) is designed after the Health Belief model and has been explored by Jette et al21 and Bennett et al.22 Consistent with the Health Belief model, the HBS focuses on perceived threat and severity of disease. For 7 items, responses of very likely to very unlikely and very serious to not serious at all are available. An example of a question is, “If you were to do nothing in particular to protect yourself, how likely is it that you will get the flu during the next 12 months?” The sum of responses across items represents the HBS score for the participant, with a higher score indicating better health beliefs (ie, lower perceived threat). In addition to the HBS, we use 4 questions about the participants’ communication with their health care provider. Responses can range from never to always, and a higher score indicates more frequent communication. An example of a question is, “In the past 6 months, how often did you tell your doctor about side effects or other problems that you were having with your heart medications?” Health literacy has been shown to be very low among inner-city patients23 and may be important for management of chronic disease. It has recently been shown to be associated with glycemic control among patients with type 2 diabetes mellitus.24 To assess this, we have used the Short Test of Functional Health Literacy in Adults.25 Scores range from 0 to 36, with a higher score indicating more errors on the reading comprehension test. Scores of 16 or lower are considered adequate for health literacy.

Environmental Resources Social support and satisfaction with income are available as measures of environmental resources. Social support, in particular, has been shown to be an important component of CHF-related treatment and outcomes.26 The social support scale comes from the Medical Outcomes Study,27 which covers both instrumental and emotional support and is a validated and widely used scale. The first question asks how many close family and friends one has around, and the next 19 items ask about the frequency of instrumental and emotional support. Asking about satisfaction with income—the participant’s rating of adequacy of his or her household’s income—is less intrusive than asking more specifi-

NOVEMBER/DECEMBER 2003

HEART & LUNG

Clark et al

Correlates of health-related quality of life

cally about income and results in greater response rates.

Analyses Missing data existed in most of the baseline measurements. Among the single item measurements (age, gender, race, Charlson Index, ejection fraction, NYHA, education, income satisfaction, health literacy, and perceived health), percentages of missing ranged from 0 to 12%, with the Charlson index having the highest missing percentage. In the multi-item composite measures (KCCQ functional and clinical summaries, CHQ dyspnea, fatique, and emotional scores, MOS social support; health belief scale; patient communication score), many subjects had missed 1 or 2 item responses. To avoid losing the entire composite scores for subjects with few missing items, we used a mean imputation procedure to replace the missing items by the means of the non-missing values of the same items across the subjects. In doing so, item response distributions are not biased to favor a particular direction.28 We have first provided basic descriptive data for the sample on the measures used in this study. Second, we provide psychometric data for the measures about patients’ characteristics because of their limited prior investigation and their importance to our assessment. Third, we estimate bivariate associations between each of the demographic and biological measures and the KCCQ, CHQ, and perceived health. Finally, using Fig 1 as a guide, we estimate a multivariate model with demographic and biological measures as the independent variables and the KCCQ and perceived health as the dependent variables. The cognitive characteristic and environmental resource measures are then incorporated into the model to estimate indirect effects of demographic and biological measures via social-cognitive factors and environmental resources, and the direct effects of social-cognitive factors and environmental resources.

RESULTS The descriptive data for the measures are shown in Table I. As noted, nearly two-thirds of the sample is female, and the mean age is 63. The Charlson score indicates a relatively high level of comorbidity with an average of 3.7. Mean ejection fraction was 0.50, which is debatably and only marginally normal. Health belief and patient communication scores were 33 and 49, which is approximately at the centers of the scales. Although not shown in Table 1, the Cronbach alpha scores for the health belief

HEART & LUNG

VOL. 32, NO. 6

and patient communication scales were 0.71 and 0.79, respectively. Health literacy scores indicate that 72% of the sample has adequate health literacy, and 16% does not. Similarly, income is satisfactory for about two-thirds, and the MOS social support mean score is 70 of 100 possible points. The MOS social support scale had a Cronbach Alpha of 0.96 (not shown). The HRQL mean scores of 58 and 56 for functional and clinical domains indicate a moderate level of functioning, as do CHQ scores. Perceived health has a mean of 1.9, which is close to a report of fair. The mean score for the NYHA classification is 2.1. The correlations between the HRQL measures and each of the demographic and biological variables, including NYHA classification, is shown in Table 2. Age has a modest positive association with emotional and fatigue scores (CHQ) and greater perceived health. Men report higher KCCQ clinical scores and slightly higher CHQ fatigue and emotional scores. White respondents report modestly lower HRQL, but not lower perceived health, on all 5 dimensions of the KCCQ and the CHQ. The Charlson comorbidity score has a weak negative association with the KCCQ scores (sicker patients have lower KCCQ scores), whereas ejection fraction is not associated with any of the HRQL measures. The NYHA classification score has moderate associations with the KCCQ scores and CHQ dyspnea and fatigue scores. The HRQL measures are all correlated from 0.3 to 0.7, with the exception of CHQ dyspnea and perceived health, which have a correlation of 0.16. The results of the multivariate regression analyses are shown in Table 3. Model 1 includes the demographic and biological measures as correlates, and Model 2 includes these and social-cognitive factors and environmental resources. The variance explained in Model 1 is low for each dependent variable, ranging from 1 to 14%. The importance of variables in bivariate analysis (Table 2) change little in multivariate analysis. The only consistent association across HRQL measures is race. From Model 1, white respondents report poorer HRQL on all dimensions of the CHQ and KCCQ but not perceived health. Being white, for example, is associated with an 11-point lower KCCQ functional score (100-point scale) and an 8-point lower KCCQ clinical score (100-point scale). To a lesser degree, age is associated with CHQ emotional and fatigue scores, KCCQ function scores, and perceived health. Assuming linearity, relative to a 50-year-old respondent, a 60year-old respondent had a 2.5-point greater score on the CHQ emotional domain (42-point scale). The

www.heartandlung.org

395

Correlates of health-related quality of life

Clark et al

Table I Means and frequencies for all variables Percent

Demography/Biology: Age (47-89) Male White Charlson Score (0-12) Ejection Fraction NYHA Class (1-4) Individual Characteristics: Education (0-18) Health Belief Scale (0-85.7) Patient Communication (0-100) Adequate Health Literacy Environmental Resources: Income Satisfactory Social Support (0-100) HRQL: KCCQ Functional Status (0100) Clinical Summary (0-100) CHQ Dyspnea (5-35) Fatigue (4-28) Emotional (7-49) Perceived Health (1-4) NYHA Class (1-4) I II III IV

Mean

62.9

SD

n

8.6

228 228 228 203 206 208

32 47 3.68 0.50 2.08

2.1 0.16 0.78

10.6 33.8 49.3

2.86 19.9 28.3

70.2

22.7

228 228

58.3 55.9

21.1 20.8

228 228

72 65

19.6 13.6 30.1 1.9 2.08

6.9 5.5 9.4 0.70 0.78

23 51 23 4

224 228 228 213

228 228 228 228 208 47 105 48 8

NYHA ⫽ New York Heart Association; KCCQ ⫽ Kansas City Cardiomyopathy Questionnaire; CHQ ⫽ Chronic Heart Failure Questionnaire.

Charlson comorbidity score and ejection fraction are not associated with any of the HRQL measures. Model 2 of Table 3 shows for each HRQL measure the associations of social-cognitive factors and environmental resources. We can estimate the indirect associations of the demographic and biological measures by comparing the parameter estimates of Model 1 to those of Model 2. The incorporation of the social-cognitive factors and environmental variables (Model 2) at least doubles the variance explained, except when perceived health is the dependent variable. Nevertheless, the variance explained remains relatively low for each outcome, with the exception of the emotional dimension of the CHQ, where the 11 independent variables account for 33% of the variance.

396

www.heartandlung.org

Comparing the parameter estimates for the demographic and biological variables of Models 1 and 2 indicates only relatively small changes in them. Although the changes do not reach statistical significance, the association of age appears to lessen across models, whereas the association of male gender appears to grow. Overall, however, there is little apparent indirect effect of the demographic and biological measures via social-cognitive factors and environmental variables. There are some statistically significant, independent effects of social-cognitive factors and environmental variables. Satisfaction with income is associated with higher KCCQ functional status, CHQ fatigue, and CHQ emotional scores. Health literacy is associated with higher KCCQ clinical and CHQ emotional scores. In fact,

NOVEMBER/DECEMBER 2003

HEART & LUNG

Clark et al

Correlates of health-related quality of life

Table II Bivariate associations of demographic and biological measures, social-cognitive factors, environmental resources, with HRQL measures, including the NYHA

Age Male White Charlson EF NYHA KCCQ function KCCQ clinical CHQ dyspnea CHQ fatigue CHQ emotional

KCCQF

KCCQC

0.15* 0.02 ⫺0.23*** ⫺0.16* 0.06 ⫺0.20**

0.05 0.22** ⫺0.15* ⫺0.16* ⫺0.01 ⫺0.33*** 0.65***

CHQD

0.02 0.03 ⫺0.17* ⫺0.10 0.02 ⫺0.22** 0.51*** 0.68***

CHQF

0.22*** 0.16* ⫺0.23*** ⫺0.12 ⫺0.04 ⫺0.19** 0.52*** 0.61*** 0.52***

CHQE

0.26*** 0.15* ⫺0.24*** ⫺0.02 ⫺0.01 ⫺0.12 0.59*** 0.53*** 0.46*** 0.71***

PH

0.35*** ⫺0.02 ⫺0.11 ⫺0.07 ⫺0.03 ⫺0.06 0.37*** 0.28*** 0.16* 0.39*** 0.31***

EF ⫽ Ejection Fraction; NYHA ⫽ New York Heart Association; KCCQ ⫽ Kansas City Cardiomyopathy Questionnaire; CHQ ⫽ Chronic Heart Failure Questionnaire. *P ⬍ .05 **P ⬍ .01 ***P ⬍ .001

adequate literacy is associated with a 7-point greater KCCQ clinical score (100-point scale) and a 3.7-point greater CHQ emotional score (42-point scale). Patient communication, social support, or health belief is significant in every model, and patient communication is significant in all but perceived health and CHQ fatigue. A higher belief score (ie, lower perceived illness threat) and greater social support are associated with better HRQL, and more communication with provider is associated with lower HRQL. The magnitude of the associations, however, is relatively low. For example, a 10-point higher health belief score (85-point scale) is associated with a 2.2-point greater KCCQ functional status score (100-point scale) and a 10-point higher patient communication score is associated with a 1.6-point lower KCCQ functional status score.

DISCUSSION Our analyses of CHF-specific HRQL measures found associations of demographic variables, social-cognitive factors, and environmental factors. Age and gender were associated with some of the HRQL measures, but an indicator variable for white had the most consistent association across CHFHRQL measures. Virtually all of the respondents in this study are either non-Hispanic white or black, and most are of low income and education and live in inner-city neighborhoods of a Midwestern city. In analyses not shown, we explored differences between white and black respondents following the

HEART & LUNG

VOL. 32, NO. 6

format of Table 1. The only differences we were able to identify were that the black respondents were more likely to be female and to have a somewhat lower level of health literacy. Interestingly, it is the black respondents who report higher CHF-HRQL. Most studies of CHF have not included many black respondents, if any, and further study with samples that represent white and black patients will be necessary to determine if there are HRQL differences or if our finding is unique to our sample or patient population. Despite associations between demographic variables and HRQL, we identified no clear indirect pathways via patient characteristic and environmental variables. In other words, the demographic variation in social-cognitive factors and environmental variables did not account for associations between demographics and HRQL. Social-cognitive factors and environmental variables, however, did have significant associations with the CHF-HRQL measures. Associations of health beliefs, patient communication, and social support with HRQL measures were apparent but modest. Entire theoretical frameworks have been developed around each of these constructs,29 and our measures are not comprehensive, but each of the 3 measures has been shown in other studies to have acceptable psychometric properties. Continued research into the association between health beliefs, patient communication, and social support may indicate opportunities for improving these with the possibility of improving CHF-HRQL.

www.heartandlung.org

397

Correlates of health-related quality of life

Clark et al

Table III Ordinary least squares regression models of perceived health, KCCQ, and CHQ. Perceived Health (n ⴝ 186) Model 1 (SD)

Age Male White Comorbidity Ejection Fracture Education Income Health Beliefs Patient Communication Social Support Adequate Literacy Adjusted R-Square

Model 2 (SD)

0.03*** (0.01) 0.00 (0.11) ⫺0.14 (0.10) ⫺0.01 (0.01) 0.22 (0.31)

0.12

0.03*** (0.01) 0.09 (0.11) ⫺0.15 (0.11) ⫺0.01 (0.02) ⫺0.23 (0.31) 0.01 (0.01) 0.01 (0.10) 0.01*** (0.00) ⫺0.00 (0.00) ⫺0.00 (0.00) 0.14 (0.12) 0.14

CHQ Fatigue (N ⴝ 186) Model 1

Age Male White Comorbidity Ejection Fracture Education Income Health Beliefs Patient Communication Social Support Adequate Literacy Adjusted R-Square

0.14 (0.05)** 1.79 (0.84)* ⫺2.53** (0.76) ⫺0.13 (0.18) 0.50 (2.41)

0.11

0.10 (0.05)* 2.32** (0.83) ⫺2.35** (0.78) ⫺0.15 (0.17) ⫺0.11 (2.28) ⫺0.04 (0.14) 2.49** (0.76) 0.03 (0.02) ⫺0.02 (0.01) 0.04** (0.02) 1.40 (0.89) 0.21

KCCQ ⫽ Kansas City Cardiomyopathy Questionnaire; CHQ ⫽ Chronic Heart Failure Questionnaire. *P ⬍ .05, **P ⬍ .01, ***P ⬍ .001

Current efforts include the development of assessment tools for CHF self-management that include health beliefs, social support, and provider-patient communication.30 Interventions to manipulate these factors have occurred within multi-component self-management interventions targeting chronically ill older adults and have shown improved outcomes.31-33 Ejection fraction and comorbidity were not associated with any of the HRQL measures. This is not entirely surprising given that prior studies have had mixed results regarding these associations. Nonetheless, the consistency of our finding across every model is impressive. Whether comorbidity as measured by the Charlson Comorbidity Score should be associated with CHF HRQL is not clear, but we

398

www.heartandlung.org

would expect the Charlson score to be associated with overall perceived health, as has been shown in the general population.34 Similarly, it is possible that ejection fraction is not a good measure of the severity of CHF, and in our sample, most respondents had ejection fractions of greater than 0.40. The prevalence of concurrent diastolic heart failure and normal (systolic) ejection fraction among older adults may also explain part of this finding.36 Recognizing this, we split ejection fraction into 2 groups (0.35 or less and 0.35 or greater) and estimated ejection fraction and HRQL correlations by group. Again, we found no association between ejection fraction and HRQL. The severity of CHF may be better represented through the NYHA classification. We did not include NYHA class in our analyses

NOVEMBER/DECEMBER 2003

HEART & LUNG

Clark et al

Correlates of health-related quality of life

KCCQ Functional Status (n ⴝ 186) Model 1 (SD)

Model 2 (SD)

0.36* (0.18) 0.18 (0.18) 1.54 (3.40) 5.11 (3.30) ⫺11.35*** (3.08) ⫺10.06** (3.1) ⫺1.25 (0.74) ⫺1.17 (0.68) 7.71 (9.8) 7.02 (9.05) 0.55 (0.55) 7.00* (3.01) 0.22** (0.08) ⫺0.16** (0.53) 0.10 (0.07) 5.58 (3.60) 0.09 0.25 CHQ Dyspnea (N ⴝ 186) Model 1

Model 2

0.01 (0.06) 0.11 (1.15) ⫺2.11* (1.03) ⫺0.25 (0.25) 0.32 (3.31)

0.01

⫺0.04 (0.06) 0.58 (1.16) ⫺1.72 (1.08) ⫺0.20 (0.24) 0.06 (3.18) ⫺0.00 (0.19) 0.33 (0.24) 0.05 (0.03) ⫺0.06** (0.02) 0.03 (0.02) 0.85 (1.24) 0.10

because it is an assessment based on participants’ reports of their functioning and similar to HRQL measures. In analyses not shown, however, we ran the models of Table 3 with NYHA classification included as an independent variable. The parameter estimates for NYHA classification were moderate and significant in 3 of the 6 models. Also, because the factors associated with HRQL may differ for those with more severe heart failure, we ran models for those with NYHA class 1 or 2 separately from those with NYHA class 3 or 4. Sample size becomes a concern in these stratified models, but our findings did not change. Although we have a model of HRQL and independent measures consistent with that model, the overall variance explained for the HRQL measures is

HEART & LUNG

VOL. 32, NO. 6

KCCQ Clinical (n ⴝ 186) Model 1 (SD)

0.10 (0.18) 8.15* (3.36) ⫺8.26** (3.02) ⫺1.02 (0.73) 6.48 (9.64)

0.05

Model 2 (SD)

0.02 (0.18) 11.57*** (3.34) ⫺7.49* (3.11) ⫺0.95 (0.69) 4.42 (9.16) 0.51 (0.56) 3.57 (3.05) 0.12 (0.08) ⫺0.17** (0.05) 0.12 (0.07) 7.40* (3.59) 0.16

CHQ Emotional (N ⴝ 186) Model 1

0.32*** (0.08) 2.72 (1.43) ⫺4.43*** (1.29) 0.25 (0.31) ⫺0.05 (4.10)

0.14

Model 2

0.25** (0.07) 4.12** (1.33) ⫺4.03*** (1.24) 0.19 (0.27) 0.45 (3.65) ⫺0.03 (0.22) 4.96*** (1.22) 0.05 (0.03) ⫺0.05* (0.02) 0.11*** (0.03) 3.70* (1.44) 0.33

low, ranging from 10 to 33%. We are not aware of any similar analyses specific to CHF, but explained variance in perceived health has tended to be higher in samples with greater variance in socioeconomic and health status.36 Compared with other models, our sample is relatively small and specific to a particular clinic population. Moreover, our analyses are based on cross-sectional data. Nonetheless, our findings suggest that further research to identify the correlates and potentially modifiable determinants of HRQL among persons with CHF will require greater attention to behavioral and social factors and greater exploration of relationships among clinical indicators. Recognizing the limits of traditional medicine, the Institute of Medicine has recommended

www.heartandlung.org

399

Correlates of health-related quality of life

greater focus on social and behavioral components of health.37 Holman and colleagues recognized several decades ago that patients’ reports often predict outcomes better than traditional medical indicators.38 Consequently, a system of care was developed whereby patients’ confidence and support received priority, leading to a long line of research in self-management of chronic disease. With a high probability of increased rates of chronic disease (including CHF) in the years to come and the mounting evidence of the significant influence of environmental and behavioral factors on HRQL, medicine will be under evergreater pressure to evolve to a system of care for chronic disease.39 This will include greater consideration of behavioral and environmental factors and recognition of the centrality of patients and their resources in the care of chronic disease.40,41 The centrality of patients to successful chronic disease management suggests that special education and support for patients with limited personal and social resources may be required.

REFERENCES 1. Van Jaarsveld CHM, Sanderman R, Miedema I, Ranchor AV, Kempen GIJM. Changes in health-related quality of life in older patients with acute myocardial infarction or congestive heart failure. A prospective study. J Am Geriatr Soc 2001;49: 1052-8. 2. American Heart Association. 1998 Heart and Stroke Statistical Update. American Heart Association: Dallas, 1998. 3. American Heart Association. Older Americans and Cardiovascular Diseases. Biostatistical Fact Sheets, 1-5. American Heart Association: Dallas, 1999. 4. Testa MA, Simonson DC. Assessment of quality of life outcomes. N Engl J Med 1996;334:835-40. 5. Wolinsky FD, Wyrwich KW, Nienaber NA, Tierney WM. Generic versus disease-specific health status measures. Eval Health Prof 1998;21:216-43. 6. Green CP, Porter CB, Bresnahan DR, Spertus JA. Development and evaluation of the Kansas City Cardiomyopathy Questionnaire. A new health status measure for heart failure. J Am Coll Cardiol 2000;35(5):1245-55. 7. Rector TS, Cohn JN. Assessment of patient outcome with the Minnesota Living with Heart Failure Questionnaire. Reliability and validity during a randomized, double-blind, placebocontrolled trial of pimobendan. Am Heart J 1992;124:1017-25. 8. Guyatt GH, Nogradi S, Halgrow S, Singer J, Sullivan MJ, Fallen EL. Development and testing of a new measure of health status for clinical trials in heart failure. J Gen Intern Med 1989;4:101-7. 9. Konstam V, Salem D, Pouleur H, Kostis J, Gorkin L, Shumaker S, et al. Baseline quality of life as a predictor of mortality and hospitalization in 5,025 patients with congestive heart failure. SOLVD Investigations. Studies of Left Ventricular Dysfunction Investigators. Am J Cardiol 1996;78(8):890-5. 10. Cline CM, Willenheimer RB, Erhardt LR, Wiklund I, Israelsson B. Health-related quality of life in elderly patients with heart failure. Scand Cardiovasc J 1999;33(5):278-85. 11. Riegel B, Moser DK, Glaser D, Carlson B, Deaton C, Armola R, et al. The Minnesota Living With Heart Failure Question-

400

www.heartandlung.org

Clark et al

12.

13. 14. 15. 16. 17. 18. 19.

20. 21. 22.

23.

24. 25. 26. 27. 28. 29. 30. 31.

32.

naire. Sensitivity to differences and responsiveness to intervention intensity in a clinical population. Nurs Res 2002; 51(4):209-18. Juenger J, Schellberg D, Kraemer S, Haunstetter A, Zugck C, Herzog W, et al. Health related quality of life in patients with congestive heart failure. Comparison with other chronic diseases and relation to functional variables. Heart 2002;87(3): 235-41. Evangelista LS, Kagawa-Singer M, Dracup K. Gender differences in health perceptions and meaning in persons living with heart failure. Heart Lung 2001;30(3):167-76. Chin MH, Goldman L. Gender differences in 1-year survival and quality of life among patients admitted with congestive heart failure. Med Care 1998;36(7):1033-46. Wilson IB, Cleary PD. Linking clinical variables with healthrelated quality of life. A conceptual model of patient outcomes. JAMA 1995;273(1):59-65. Deyo R, Cherkin D, Ciol M. Adapting a clinical comorbidity index for use with ICD-9 administrative databases. J Clin Epidemiol 1992;45(6):613-19. McDonald CJ, Overhage JM, Tierney WM, Dexter PR, Martin DK, Suico JG, et al. The Regenstrief Medical Record System. A quarter century experience. Int J Med Inf 1999;54(3):225-53. McDonald CJ, Tierney WM. The Medical Gopher—a microcomputer system to help find, organize and decide about patient data. West J Med 1986;145:823-9. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies. Development and validation. J Chronic Dis 1987; 40(5):373-83. Idler EL, Benyamini Y. Self-rated health and mortality. A review of twenty-seven community studies. J Health Soc Behav 1997;38(1):21-37. Jette AM, Cummings KM, Brock BM, Phelps MC, Naessens J. The structure and reliability of health belief indices. Health Serv Res 1981;16:81-97. Bennett SJ, Perkins SM, Lane KA, Forthofer MA, Brater DC, Murray MD. Reliability and validity of the compliance belief scales among patients with heart failure. Heart Lung 2001a; 30:177-85. William MV, Parker RM, Baker DW, Parikh NS, Pitkin K, Coates WC, et al. Inadequate functional health literacy among patients at 2 public hospitals. JAMA 1995;274:167782. Schillinger D, Grumbach K, Piette J, Wang F, Osmond D, Daher C, et al. Association of health literacy with diabetes outcomes. JAMA 2002;288(4):475-82. Baker DW, Williams MV, Parker RM, et al. Development of a brief test to measure functional health literacy. Patient Educ Couns 1999;38:33-42. Bennett SJ, Perkins SM, Lane KA, et al. Social support and health-related quality of life in chronic heart failure patients. Qual Life Res 2001b;10(8):671-82. Measuring Functioning and Well-Being. In: Stewart AL, Ware JE (eds). The Medical Outcomes Study Approach. Durham (NC): Duke University Press; 1992. Little RJA, Rubin DB. Statistical Analysis with missing data, 2nd Ed. Wiley Interscience: New York, 2002. Health Behavior and Health Education. In: Glanz K, Rimer B, Lewis F (eds). Theory, Research and Practice. 3rd Ed. JoseyBass: San Francisco; 2002. Riegel B, Carlson B, Glaser D. Development and testing of a clinical tool measuring self-management of heart failure. Heart Lung 2000;29:4-12. Leveille S, Wagner EH, Davis C, et al. Preventing disability and managing chronic illness in frail older adults. A randomized trial of a community-based partnership with primary care. J Am Geriatr Soc 1998;10:1191-98. Lorig KR, Ritter P, Stewart AL, et al. Chronic disease selfmanagement program. 2-year health status and health care utilization outcomes. Med Care 2001;11:1217-23.

NOVEMBER/DECEMBER 2003

HEART & LUNG

Clark et al

Correlates of health-related quality of life

33. Reuben D, Frank J, Hirsch H, et al. A randomized clinical trial of outpatient comprehensive geriatric assessment coupled with an intervention to increase adherence to recommendations. J Am Geriatr Soc 1999;3:269-76. 34. Clark D, Von Korff M, Saunders K, et al. A chronic disease score with empirically derived weights. Med Care 1995;33(8): 783-95. 35. Kitzman DW, Little WC, Brubaker PH, Anderson RT, Hundley WG, Marburger CT, et al. Pathophysiological characterization of isolated diastolic heart failure in comparison to systolic heart failure. JAMA 2002;288(17):2144-50. 36. Bryant LL, Beck A, Fairclough D. Factors that contribute to positive perceived health in an older population. J Aging Health 2000;12(2):169-92. 37. Pellmar TC, Brandt E, Baird M. Health and behavior. The interplay of biological, behavioral, and social influences:

HEART & LUNG

VOL. 32, NO. 6

38. 39.

40.

41.

summary of an Institute of Medicine report. Am J Health Promot 2002;16:206-19. Schoor S, Lorig KR. Self-care and the doctor-patient relationship. Med Care 2002;40(4 Suppl):II40-44. Wagner EH, Austin BT, Von Korff M. Organizing care for patients with chronic illness. Milbank Q 1996;74(4):51145. Von Korff M, Gruman J, Schaefer J, Curry SJ, Wagner EH. Collaborative management of chronic illness. Ann Intern Med 1997;127:1097-1102. Glasgow RE, Funnell M, Bonomi A, Davis C, Beckham V, Wagner EH. Self-management aspects of the improving chronic illness care breakthrough series. Implementation with diabetes and heart failure teams. Ann Behav Med 2002; 24:80-7.

www.heartandlung.org

401