|
|
||||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Department of Psychology (Dr. Quittner), University of Miami, Coral Gables, FL; Department of Statistics (Dr. Buu), University of Michigan, Ann Arbor, MI; Psychological Assessment Resources, Inc. (Ms. Messer), Tampa, FL; Division of Psychology (Dr. Modi), Cincinnati Childrens Hospital Medical Center, Cincinnati, OH; and Genentech, Inc. (Dr. Watrous), San Francisco, CA.
Correspondence to: Alexandra L. Quittner, PhD, Department of Psychology, 5665 Ponce de Leon Blvd, University of Miami, Coral Gables, FL 33146-2070; e-mail: aquittner{at}miami.edu
| Abstract |
|---|
|
|
|---|
14 years, consisting of 44 items on 12 generic and disease-specific scales. Versions of the CFQ are also available for children with CF and their parents. This study evaluated the psychometric properties of the CFQ in a national study at 18 CF centers in the United States. Participants: The CFQ-teen/adult was administered to 212 patients with CF ranging in age from 14 to 53 years. Test-retest reliability was assessed in a subset of patients over a 10- to 14-day interval.
Results: Multitrait analysis indicated a majority of items (95%) correlated more highly with their intended scale than a competing scale, supporting the conceptual model. Internal consistency coefficients indicated the CFQ scales had good reliability (Cronbach
= 0.67 to 0.94), and test-retest stability was acceptable (rs = 0.45 to 0.90). Validity was demonstrated by examining relationships between the CFQ, age, pulmonary function, and body mass index. As expected, the CFQ was inversely correlated with age, with older adults reporting lower CFQ scores than younger adults, better nutritional status was positively correlated with several weight-related scales, and the measure differentiated between individuals with varying levels of disease severity. Strong associations were also found between the CFQ and similar scales on the Short Form-36 Health Questionnaire, a well-known generic HRQOL measure.
Conclusions: The results demonstrated that the CFQ-teen/adult is a reliable and valid measure of HRQOL for individuals with CF. It may be utilized in clinical trials to assess the effects of new therapies, to document the progression of disease, and to inform clinical practice.
Key Words: adolescents and adults cystic fibrosis disease specific health-related quality of life psychometric validation
| Introduction |
|---|
|
|
|---|
Research on HRQOL has flourished as a result of advances in medical technology and treatment, the growing prevalence of chronic illnesses in adult and pediatric populations, and the need to reduce health-care costs. Although conventional measures of physical functioning are essential, they do not capture the broader impact of a disease on the patients physical, social, and emotional functioning.4 HRQOL measures are used for several purposes: (1) as primary or secondary outcomes in clinical trials, (2) to describe the impact of an illness on a patients daily functioning, (3) to evaluate new pharmaceutical and surgical interventions, (4) to aid in clinical decision making, and (5) to estimate the costs and benefits of medical interventions.567 Reliable and valid measures of HRQOL for several chronic conditions, such as asthma and cancer, are now widely used.28 The purpose of the current study was to conduct a psychometric evaluation of a new, disease-specific HRQOL measure for cystic fibrosis (CF), the Cystic Fibrosis Questionnaire (CFQ).9
Development of a well-validated HRQOL measure for CF is important for several reasons. First, discovery of the genetic defect for CF in 1989 led to dramatic advances in our understanding of the pathophysiology of this chronic illness, which in turn has led to the development of new medications and treatments.10 These treatments have shown promise for extending life span and improving quality of life. Both recombinant human deoxyribonuclease and inhaled antibiotics have been shown to positively impact FEV1 percentage of predicted and HRQOL.1112 However, in the absence of a well-validated measure of HRQOL, these studies have relied on ad hoc items that make it difficult to determine how these new treatments affect functioning in the four core HRQOL domains.
Measures of HRQOL are also important for identifying the benefits of new treatments that are not reflected in conventional health indexes, such as pulmonary functioning. For example, after controlling for changes in pulmonary functioning for patients with CF in the Tobramycin Solution for Inhalation trial,12 additional variance remained unaccounted for, suggesting the possibility of improvements in other areas of functioning, such as energy level. These patient-reported outcomes are now routinely included in US Food and Drug Administration-approved clinical trials, and this information can be added to the label if it meets certain measurement and statistical criteria.13 Thus, HRQOL data may play an important role in decision making between health-care providers and patients.
Over 15 years ago, the National Institutes of Health sponsored a conference on the behavioral and psychological aspects of CF and recommended the development of a disease-specific measure of HRQOL to be included along with other health outcomes.14 More recently, a consensus conference reviewed the use of HRQOL measures in clinical trials with CF patients and again recommended that validated measures of HRQOL be incorporated into phase 3 clinical trials for both adults and children.15 Only one other disease-specific HRQOL measure for CF, the Cystic Fibrosis Quality of Life Questionnaire, has been published for patients in the United Kingdom.16 However, this measure does not have parallel forms available for children with CF, who make up the majority of the patient population.
To date, most studies of HRQOL in patients with CF have utilized generic measures, such as the Quality of Well-being Scale and the Short Form-36 Health Questionnaire (SF-36),317 which include general items of physical, social, and emotional functioning that can be rated by patients with a variety of medical conditions. Results from these studies have indicated that generic measures are not sensitive to the specific concerns of patients with CF,181920 which limits their ability to quantify the benefits of new treatments (eg, lung transplantation, new medications) or the natural progression of the disease (eg, treatment of pulmonary exacerbations).421
The CFQ is a newly developed, disease-specific HRQOL measure for individuals with CF, with developmentally appropriate versions for children aged 6 to 13 years (CFQ-child), parents of children with CF aged 6 to 13 years (CFQ-parent), and adolescents and adults with CF
14 years old (CFQ-teen/adult).922 The CFQ was originally developed in France,23 and all three versions recently underwent independent forward and backward translations, followed by a two-phase cognitive testing procedure in the United States.24 The set of instruments has been developed to encompass general domains of HRQOL: physical functioning, role functioning, vitality, health perceptions, emotional functioning, and social functioning, as well as domains specific to CF: body image, eating disturbances, treatment burden, and respiratory and digestive symptoms.
This study presents data from the national psychometric validation of the English CFQ-teen/adult version at 18 CF centers across the United States. The primary objective of the study was to assess the reliability and validity of the CFQ. Specifically, item-level analyses were conducted to examine item to scale correlations, ceiling and floor effects, internal consistencies, and test-retest reliability. Convergent and discriminant validity was also evaluated by testing hypotheses related to age, disease severity, and nutritional status, and by examining associations between the CFQ and a generic measure of HRQOL (ie, the SF-36).
| Materials and Methods |
|---|
|
|
|---|
Procedures
To obtain a geographically representative sample, participants were recruited from 18 CF centers across the United States. Written informed consent and assent was obtained from all participants according to the procedures specified by the relevant institutional review boards. Patients were enrolled in this study during a routine clinic visit that was not associated with an acute illness or pulmonary exacerbation. All of the measures, including the CFQ, were administered prior to a physical examination or other laboratory procedures (eg, pulmonary function tests) in order to obtain an unbiased perception of HRQOL. Basic demographic (eg, age, gender) and medical information (ie, pulmonary functioning, body mass index) was collected, followed by completion of two HRQOL measures, the CFQ-teen/adult version and the SF-36,3 a generic HRQOL instrument. Each HRQOL instrument took approximately 15 min to complete.
Measures
CFQ-Teen/Adult:
The CFQ-teen/adult measure9 evaluated in this study consisted of 44 items across 12 scales (Table 1
). Response choices generally included ratings of frequency and difficulty on a 4-point scale (1 = always to 4 = never; 1 = a lot of difficulty to 4 = no difficulty) or true/false responses (1 = very true to 4 = very false). Scores were standardized on a 0- to 100-point scale, with higher scores representing better quality of life.
|
| Results |
|---|
|
|
|---|
0.40 with the intended scale and should correlate much lower with competing scales, after accounting for the SE of measure (item discriminant validity).2930 The results generally supported the French model, with item-scale correlations > 0.40, with the exception of one social ("feel comfortable sleeping away"), one marginalization ("people ask me annoying questions"), and one health perception item ("compared to 3 months ago, how you feel about your health?"). These poorly performing items were deleted. This left only two items on this scale that correlated highly with the social functioning scale (
= 0.45 and
= 0.41). Thus, combining the marginalization and social scales made both conceptual and psychometric sense. The final social functioning scale had five items.
Item-Level Analyses of the English CFQ
Two types of item-level analyses were conducted on the US national validation data. First, items were subjected to a multitrait analysis (revised Multitrait/Multi-item Analysis Program28) to evaluate item-to-scale relationships.30 Next, floor and ceiling effects were identified for each scale.
As can be seen in Table 2
, a large percentage of items correlated highly with their intended scale; 95% of the item-scale correlations (corrected for overlap) were
0.40. In addition, 83% of the items correlated at least two SEs greater with their hypothesized than competing scales. However, low item-scale correlations were observed for the treatment burden scale, on which the two items correlated only minimally with their intended scale (rs = 0.10). Subsequently, modifications were made to the wording of these items to reduce their retrospective nature, and an additional item was added to the treatment burden scale.
|
Scale-Level Reliability
Two scale-level analyses were conducted: (1) calculations of the internal consistency or reliability of each scale, and (2) test-retest reliability. First, scale-level reliability was calculated using Cronbach
. As can be seen in Table 3
, the reliability coefficients ranged from r = 0.18 to 0.94, with a majority of the coefficients > 0.70. Only two domains fell below that cutoff: the Cronbach
for digestion was 0.67 and for treatment burden was 0.18.
|
Construct Validity
Several hypotheses were tested to establish the validity of the CFQ. First, because CF is a deteriorating medical condition, a strong correlation was expected between age and CFQ scores, with higher HRQOL reported by adolescents vs adults. This hypothesis was supported with inverse relationships found between age and the physical, role, vitality, emotion, social, eating, health perceptions, and respiratory scales (rs = 0.17 to 0.36, p < 0.05). Next, respondents were classified into three age groups (adolescents, age 14 to 17 years, n = 68; young adult, age 18 to 25 years, n = 66; and adult, age > 25 years, n = 74), and their scores were compared using a multivariate analysis of variance. As expected, younger individuals with CF reported higher quality of life than older individuals on most dimensions (Hotelling T2 = 0.30, F [22,388] = 2.60, p < 0.001). No age-related differences were found on the eating disturbances, treatment burden, and digestion scales. Significant associations were also found between the CFQ weight-related domains and nutritional status, with positive relationships obtained between the body image, eating disturbances, and weight domains and body mass index scores (Table 4
).
|
70%, n = 91; moderate = FEV1 percentage of predicted
40%, n = 71; and severe = FEV1 percentage of predicted < 40%, n = 44) based on pulmonary functioning scores measured at the time of the CFQ assessment. A multivariate analysis of variance indicated significant differences in CFQ scores by disease severity (Hotelling T2 = 0.44, F[22,384] = 3.84, p < 001). Individuals with less severe disease reported higher scores on all of the CFQ scales except for digestion, compared to individuals with more severe disease (Fig 1
). Thus, the CFQ successfully differentiated between those with mild, moderate, and severe disease.
|
| Discussion |
|---|
|
|
|---|
In an effort to improve the internal consistency of the treatment burden and digestion scales, both of which had only two items, new items have been added to these scales (treatment burden: "How difficult is it for you to do your treatments [including medications] each day?; digestion: "Have you had problems with gas?"). An analysis of these additional items in two studies3435 indicates that the scale-level reliability has been improved (average
> 0.70). Ceiling effects were also noted on the eating disturbances, role functioning, and weight scales. This is a common problem for HRQOL measures1820 that can be addressed by using item-response theory techniques to generate more difficult items.3637 We have recently added more items to the role functioning scale, which may remedy this problem.
Increases in the life span of individuals with CF, as well as the development of new medications, have highlighted the importance of measuring HRQOL using a disease-specific instrument. Thus, the CFQ has a number of potential applications and is currently ready to be used for research purposes. First, it can be used in clinical trials as a secondary outcome to assess the benefits of new medications and treatments from the patients perspective. Second, it can be used to understand the natural course and progression of the disease in terms of its effects on several domains, including role, social, and emotional functioning. It may also illuminate the mechanisms associated with differential survival. For example, there are well-documented gender differences in morbidity and mortality for male and female CF patients.3839 Studies404142 using the CFQ have found significant gender differences on the body image and weight scales, suggesting that female CF patients are more satisfied with their "thinness" and weight than male CF patients, although this is clearly detrimental to their health. Finally, the CFQ can be used as a clinical tool in annual visits to provide a broader assessment of the individuals functioning and to identify problem areas that require intervention. Computerized versions of the CFQ are now available that permit patients to complete the measure in clinic, with real-time scoring and interpretation. The CFQ is currently being used in a national, Web-based data entry system for the Epidemiologic Study of Cystic Fibrosis II. This database includes > 20,000 patients at 203 CF centers.42 Patients complete the CFQ at their annual visit with comparisons to normative information based on age and gender. These data can then be used to generate a patient profile for that clinic visit or to examine changes in quality of life over the course of a year.
There are several important directions for future research in quality-of-life measurement for individuals with CF. As the CFQ is used more commonly in research and clinical contexts, it will be important to determine the minimal clinically important difference score.43 This will facilitate interpretation of the clinical significance of observed changes in CFQ scores. In addition, because adults with CF are now living much longer, we are developing an adult-focused module that includes items related to job satisfaction, intimate relationships, and having children.44 There are an increasing number of clinical trials being conducted internationally, and this requires conceptually and linguistically equivalent translations of the CFQ. In addition to the French version, the CFQ has now been translated into German, Dutch, Italian, and Portuguese, and we are currently completing the cognitive testing phase of the Spanish translation.4546 The CFQ, scoring information, software program, and manual are available by request from the authors (e-mail: aquittner{at}miami.edu).
| Acknowledgements |
|---|
| Footnotes |
|---|
This study was funded in part by the Cystic Fibrosis Foundation and Genentech, Inc.
Received for publication January 10, 2005. Accepted for publication March 14, 2005.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A. Jaques, E. Daviskas, J. A. Turton, K. McKay, P. Cooper, R. G. Stirling, C. F. Robertson, P. T. P. Bye, P. N. LeSouef, B. Shadbolt, et al. Inhaled Mannitol Improves Lung Function in Cystic Fibrosis Chest, June 1, 2008; 133(6): 1388 - 1396. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. A. Riekert, S. J. Bartlett, M. P. Boyle, J. A. Krishnan, and C. S. Rand The Association Between Depression, Lung Function, and Health-Related Quality of Life Among Adults With Cystic Fibrosis Chest, July 1, 2007; 132(1): 231 - 237. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |