| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Division of Adolescent Medicine (Dr. Britto), Health Policy and Clinical Effectiveness (Dr. Kotagal and Mr. Atherton), and Pulmonary Medicine, Allergy & Clinical Immunology (Dr. Wilmott), Children’s Hospital Medical Center, Cincinnati; Institute For Health Policy and Health Services Research (Dr. Hornung), and Division of General Internal Medicine (Dr. Tsevat), University of Cincinnati, Cincinnati, OH.
Correspondence to: Maria T. Britto, MD, MPH, Children’s Hospital Medical Center, Division of Adolescent Medicine, 3333 Burnet Ave, Cincinnati, OH; e-mail: maria.britto{at}chmcc.org
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
Design: Cross-sectional analysis of observational cohort.
Setting: Outpatient clinics of a Midwestern CF center.
Subjects: One hundred sixty-two subjects with CF aged 5 to 45 years.
Main outcome measures: Physical and psychosocial summary scores and individual scale scores for the Child Health Questionnaire and Short Form-36.
Results: Compared with the general population, people with CF reported similar scores for most psychosocial measures, but lower scores for most physical measures, with the lowest scores on the general health perceptions scale. In multivariable analyses, pulmonary exacerbations in the past 6 months were strongly associated with the physical (p = 0.001) and psychosocial (p = 0.0003) scores. The physical score fell, on average, 6 points per exacerbation and the psychosocial score fell 3 points. Lung function, nutrition, 6-min walk distance, age, gender, and insurance status were not significantly associated with HRQOL in this study population. Those who declined to participate had significantly lower FEV1 percent predicted and nutritional indexes. Our findings may not be generalizable to the entire CF population.
Conclusion: Recent pulmonary exacerbations have a profound negative impact on HRQOL that is not explained by differences in lung function, nutritional status, or demographic factors.
Key Words: Child Health Questionnaire • health-related quality of life • health status • questionnaire • Short Form-36
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
Measures of HRQOL that are valid and suited to the CF population are needed for research, for policy making, and for improving clinical care. HRQOL measures can serve as outcomes in clinical trials and studies of health-care effectiveness. They can also be used to determine the cost-effectiveness of new treatments. These latter analyses, which are based partly on patients’ assessment of HRQOL, may impact policy decisions regarding allocation of resources for different CF treatments, or for treatments for CF vs other conditions. HRQOL measures are especially important for interventions that may improve the patient’s well-being or ability to function, but that do not result in changes in FEV1 percent predicted or other traditional clinical outcomes. They also may be useful in assessing disease progression and monitoring the clinical course of individual patients. Finally, HRQOL measures can describe health outcomes in ways that are meaningful to patients and families as well as to health professionals.3
Among adults with a variety of chronic conditions, self-reported HRQOL has been found to be a powerful predictor of future health and social outcomes, such as hospital days, mortality, total health costs, and employment status.4 In older adolescents and adults with CF, scores on the Nottingham Health Profile correlated with FEV1 percent predicted, breathlessness, and time receiving home antibiotic therapy.5 The RAND functional status index (RAND Health Insurance Study) by itself significantly predicted mortality in adults with CF over the following 5 years.6 The quality of well-being scale correlated well with measures of pulmonary function and exercise tolerance in cross-sectional7 and longitudinal8 studies including both children and adults. A larger study of 199 pediatric patients with CF, however, found only low-to-moderate correlations between the quality of well-being scale and other measures of physical and psychosocial health.9 One consistent finding in all these earlier studies is that the correlation between FEV1 percent predicted and HRQOL, while usually significant, is typically weak.3 These previous studies may have limited applicability to children and adolescents, since none of the measures used were developed for a pediatric population. In addition, most earlier studies used convenience samples that may not be representative of the underlying CF population. New HRQOL measures, specifically designed for patients with CF, have recently been developed for adults and adolescents10 and for children.11 These measures may correlate more highly with physiologic outcomes and may be particularly suited to CF clinical trials. Generic HRQOL measures, however, will be important for comparing outcomes across diseases and for determining the relative cost-effectiveness of different treatments.
This study sought to build on previous work by using recently developed age-specific HRQOL measures and by attempting to include the entire population of a large center, rather than a convenience sample. The specific aims of this study were to determine the impact of CF on HRQOL, to compare HRQOL of a population with CF to the general population, and to determine the relationship between HRQOL and clinical and demographic factors. We hypothesized that CF would have a greater impact on physical than psychosocial functioning, and that physical functioning would be significantly associated with FEV1 percent predicted and with exercise capacity as measured by 6-min walk distance.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
5 years old
and who had not undergone lung transplantation were invited to
participate. Patients were enrolled during routine quarterly clinic
visits. For those who had no clinic visits during the 27-month
enrollment period or who were missed by study personnel, attempts were
made to enroll by mail or telephone. Overall, 162 of 191 eligible
patients (83.9%) participated. Three of these patients were enrolled
by telephone. The remainder enrolled in the clinic. The participants,
compared with those who declined, were significantly better nourished
(body mass index [BMI] percentage of ideal, 98.5 vs 92.3; weight
z score, - 0.5 vs - 1.2; height z score,
- 0.6 vs - 1.2), had a higher FEV1 percent
predicted (72.8 vs 58.6), and were more likely to have commercial
insurance (81.6% vs 61.5%). They did not differ significantly with
respect to age, gender, rates of colonization with Pseudomonas
aeruginosa, or presence of diabetes mellitus.
Quality-of-Life Measures
The HRQOL measures used in this study were based on the World
Health Organization conceptual model of health, which incorporates
physical, mental, and social well-being. In addition to being derived
from the same conceptual framework, the measures share similar
construction and identical scaling of summary scores. Subjects < 18
years old completed the Child Health Questionnaire
(CHQ).12
The Parent Form-50 (PF-50) of the CHQ consists of
50 items comprising 11 multi-item scales: physical function, role
function as limited by physical problems, general health perceptions,
bodily pain, role function as limited by social/emotional/behavioral
problems, self esteem, mental health, behavior, family activities,
parental impact-emotional, and parental impact-time. The
parental-impact scales capture the amount of emotional distress and
time limitation experienced by the parent due to the child’s physical
health, emotional well-being, attention/learning abilities, ability to
get along with others, and general behavior.12
)
In addition, there are single-item measures of family cohesion, global
health, global behavior, and change in health perceptions. The Child
Form-87 (CF-87) of the CHQ, designed for those aged 10 to 17 years,
consists of 87 questions and contains the same scales as the PF-50,
except that it omits the parental impact scales. For patients 18
years old, we administered the Short Form-36 (SF-36), which contains
eight scales: physical functioning, role-physical, bodily pain, general
health, vitality, social functioning, role-emotional, and mental
health. For all three measures, the reporting frame for all items is
the past month, except for the change in health item and general health
perceptions, for which the reporting frame is the past year. Individual
scale scores range from 0 to 100, with higher scores representing
better HRQOL. For the PF-50 and the SF-36, physical and psychosocial
summary scores have been created using factor analytic
techniques.12
13
All three measures have been found to be
valid and reliable.12
Clinical and Demographic Measures
Height and weight were obtained by clinic personnel using a
stadiometer and digital scale, and were recorded on clinic intake
sheets. Height and weight z scores and BMI were calculated
using Epi-Info 2000 from the Centers for Disease Control and
Prevention.15
The BMI percentage of median expected for
age was derived from the 1998 standards proposed by Rosner and
colleagues.16
As has been commonly done in CF research,
for those aged 18 years, BMI percentage was calculated using
standards for 18 year olds. Pulmonary function was determined using
spirometry, and total body plethysmography according to American
Thoracic Society Standards.17
Pulmonary function test data
were excluded if they did not meet these standards (n = 30); these
were from mostly young children. Demographic data and other clinical
indicators (such as pulmonary exacerbations) were obtained from medical
record review and hospital databases. Pulmonary exacerbations
were defined using the criteria suggested by Fuchs et
al.18
Patients with four or more of the following signs or
symptoms were considered to have an exacerbation and were treated with
antibiotics: (1) increased cough; (2) increased dyspnea (without
reactive airways component); (3) increased sputum production/change in
sputum appearance; (4) new/increased hemoptysis; (5) temperature
> 38°F; (6) loss of appetite or weight loss (unexplained by stool
pattern); (7) malaise/fatigue/lethargy; (8) new finding on chest
examination; (9) new infiltrate on chest radiograph; (10) decline in
pulmonary function FVC or FEV1 percent predicted
> 10% from previous measurement; (11) sinus pain/tenderness. No
patients at this center received routine tri-monthly antibiotics.
Exercise tolerance was determined by timed walking distance. In
children with CF and other lung diseases, walking distance has
correlated well with more complicated measures of functional capacity,
such as peak oxygen uptake and physical work
capacity.19
20
Baseline oxygen saturation and vital signs
were obtained. A portable pulse oximeter was attached to the patient
and carried by study personnel. The patient was instructed to walk as
quickly as comfortably possible on a 41-m long, level hospital corridor
for 6 min. Each patient completed the protocol one time. Supplemental
oxygen was administered as needed to keep the patient’s oxygen
saturation 88%. Owing to infection control concerns, patients
known to have Burkholderia cepacia or methicillin-resistant
Staphylococcus aureus colonization did not complete the walk
test. A few additional patients declined the walk test, leaving a total
of 121 patients with walk test data.
Data Collection Procedures
The study was approved by the institutional review boards at the
children’s and university hospitals. Families were informed of the
study by mail. At the time of registration for a routine quarterly
visit, families were invited formally to participate. Informed assent
or consent was obtained from the patient, and from the parent or
guardian for those patients < 18 years of age. A parent completed the
PF-50 for children < 18 years old, and children 10 to 17 years old
completed the CF-87. Those 18 years old completed the SF-36. The
SF-36 could be completed in 5 to 7 min, and the CHQ could be completed
in 15 to 20 min. HRQOL measures were completed prior to obtaining any
clinical information (such as weight) from the patient. Pulmonary
function testing was done as part of routine clinical care. The 6-min
walk test was completed during the visit. For patients who missed more
than two quarterly visits, the study team attempted to reschedule a
visit. If a visit could not be scheduled, the HRQOL measure was
administered by telephone using standardized interview protocols
developed by the CHQ and SF-36 developers.12
13
Data Management
Data were double entered into Microsoft Access (Microsoft;
Redmond, WA). Standardized software provided by the CHQ and SF-36
developers were used to compute scale and summary scores. For missing
scale items, the developers recommend that scale scores be calculated
if more than half of the items within the scale have been answered,
substituting the mean score for that scale for the missing value. If
fewer than half of the items in a scale were completed, then the scale
was coded as missing.
Power Calculation
Although there is no "gold standard," a 5-point
difference in scale scores is a conservative estimate of clinically and
socially important differences. For example, a difference of 5.5 points
on the physical function scale of the CHQ PF-50 discriminated children
with asthma in the 75th percentile of function from those in the 50th
percentile.12
Power was calculated in two ways: (1)
comparing the study population with a national population norm, and (2)
comparing patients with FEV1 percent predicted of
> 70% predicted or < 70% predicted. SDs for CHQ scores were
obtained from children with asthma.12
)
FEV1 percent predicted means and SDs were
obtained from the clinical database for the prospective study
population. With a projected sample of 165, we had 80% power to detect
2-point differences in the physical or psychosocial summary scores
between the study population and published norms. For the individual
scale scores, we had 80% power to detect a 3- to 4- point difference
between the study population and published norms. For the high vs low
FEV1 percent predicted groups, we had 80% power
to detect a difference of 3 to 4 points on the summary scores and 5
points on an individual scale score.
Analysis
Pairwise comparisons between the means of each scale and summary
score for the SF-36 and PF-50 and the published norms for the general
US population12
) were conducted using
two-tailed t tests. Published normative data are summary in
nature. Hence, the study patients could not be individually matched but
only compared at the group level. Detailed sociodemographic data for
the normative samples were also not available, so comparison could not
be made on these variables. Since the mean age of our SF-36 completers
was 27 years, we compared their scores with both the 18- to 24-year-old
and 25- to 34-year-old norms. The results were the same, so only the
comparisons with the 18- to 24-year-old norm are presented. National
norms and summary scores are not yet available for the CF-87. The
physical and psychosocial summary scores for the SF-36 and PF-50 are
scaled to have a mean of 50 and a SD of 10 in the general population.
In order to determine whether the SF-36 and PF-50 scores could be
combined in subsequent analyses, we first tabulated them separately and
tested the homogeneity of the SF-36 summary scales vs the PF-50 summary
scales by stratifying on other suspected predictors of difference in
scores, such as FEV1 percent predicted and number
of exacerbations. The means and the variances of the summary scores
were not significantly different after stratifying by these factors.
They were therefore combined in subsequent analyses.
In preliminary analyses, bivariate associations between physical and
psychosocial summary scores and the independent variables were
determined using t tests for continuous variables and
2 tests for categorical variables. Multiple
linear regression was then used to identify variables independently
associated with summary scores. The two a priori predictors
of differences in scores were FEV1 percent
predicted and 6-min walk distance. The other independent variables
(age, gender, height, and weight z scores, percent predicted
BMI, number of pulmonary exacerbations, presence of P
aeruginosa in the sputum, presence of diabetes mellitus, and
insurance status) were added to the model and retained if significant
or if their inclusion caused a significant change in the coefficients
of the model (confounding). These variables were also examined for
significant interactions with the measures of disease severity (effect
modification). The number of pulmonary exacerbations was not initially
included as a key predictor, but preliminary analyses demonstrated a
significant association between HRQOL and exacerbations. Thus
additional post hoc analyses were conducted to determine if
there were any significant interaction terms involving this variable.
Those with no exacerbations in the preceding 6 months were coded as
having 183 days (6 months) since last exacerbation.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
|
|
|
|
|
|
Due to the unexpected lack of significant association between the summary scores and FEV1 percent predicted, walk distance, or the nutritional indexes, we performed further analyses to examine the associations between these variables and the individual scale scores. We thought that the overall nature of the summary scores may obscure real differences between groups that could be apparent comparing the more specific scale scores. There were a number of significant bivariate associations, especially for the global health scale. When multivariable modeling was conducted, however, the only significant independent predictors were those related to exacerbations (data not shown).
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
70 had physical
scores 3 points lower than those with higher FEV1
percent predicted. We had 80% statistical power to detect a 3- to
4-point difference between groups, so it is possible that we failed to
detect a significant difference by chance. Nonetheless, the modest
magnitude of the difference (3 points) suggests that the impact of
FEV1 percent predicted on HRQOL is relatively
small, especially compared to the impact of exacerbations. Our findings
are similar to some other studies of patients with CF3
and
other pulmonary disease21
that demonstrated only moderate
correlations between physiologic parameters and HRQOL. The powerful
association of HRQOL with exacerbations, and the weaker association
with FEV1 percent predicted may imply that for
patients and families, HRQOL has less to do with how severe one’s
underlying disease is, and more to do with the disruptive effect of
exacerbations. In this analysis, the available data on exacerbations
was limited to the number and timing of protocol-defined exacerbations.
Thus, we are unable to fully explore how exacerbations cause their
detrimental effect on HRQOL. For example, do different treatment
strategies have the same impact? Prospective longitudinal information
regarding the association between FEV1 percent
predicted, HRQOL, and exacerbations will help to clarify these issues. The lack of association between HRQOL and 6-min walk distance was surprising. We had expected this measure, which is related to functional exercise capacity,19 20 to be reflected in the physical HRQOL scores. One possible explanation for the lack of association may be the relatively long distances walked by our patients: 507 m or 84 m/min on average. In the study of severely ill patients by Nixon et al,19 the mean distance was only 407 m. Some of our most physically limited patients declined to undertake the walk test. It may have been that in order to reach submaximal exercise capacity, our participating patients would have needed to run rather than walk quickly. This explanation seems less likely, though, because the HRQOL scores were similar for even the lowest tertile of walkers whose maximum distance was only 440 m. Thus, it may be that exercise tolerance is not closely related to HRQOL in this relatively healthy population. Other studies of patients with CF have not assessed the relationship of HRQOL to exercise tolerance.
The CHQ and SF-36 overall demonstrated wide variability in this population except for the role function scales, where most participants scored at the highest possible level. Ceiling problems with the role function scales have been a common problem with generic HRQOL measures. Studies for which role function is of particular interest should use more detailed role function measures that explore a broader range of function.
In the psychosocial domains, the participants’ scores were similar to the overall population. Our findings are in agreement with other studies22 23 demonstrating the reasonably good psychosocial function of patients with CF. One important caveat here is that our participants were healthier than our nonparticipants. They were also more likely to have commercial, rather than public, health insurance, which likely reflects higher socioeconomic status. Thus, we cannot conclude from this work that CF patients, in general, experience psychosocial HRQOL similar to the overall population.
As expected, the majority of the physical functioning scores were worse than the general population, although the magnitude of the difference was modest. Differences of the largest magnitude were seen in the general health perceptions scale. Items in that scale require respondents to compare explicitly their state of health to that of their peers. The relatively higher scores on the other scales likely reflect CF patients’ and families’ long-term acclimatization to living with CF.
The issue of systematic differences between those who participate and those who decline is a significant one in behavioral and HRQOL studies in CF and other pediatric chronic illnesses. Relatively small population sizes preclude random sample techniques used in general population studies. Thus, most studies of psychosocial adaptation or HRQOL in patients with CF have relied on convenience samples. It is likely that these samples overestimate the true health of the population. With an extended 27-month recruitment period and outreach efforts, we were able to recruit 83% of the population of our center. Nonetheless, those who declined were sicker and may well have had a substantially lower HRQOL than participants. Populations in other geographic areas or at other centers may have different characteristics and subsequently different HRQOL than our population.
|
Conclusion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
|
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by grant BRITTO98AO from the Cystic Fibrosis Foundation.
Presented in part at the 1999 North American CF Conference, Seattle, WA, October 9, 1999; and Pediatric Academic Societies and American Academy of Pediatrics Joint Meeting, Boston, MA, May 12–15, 2000.
Received for publication February 2, 2001. Accepted for publication June 29, 2001.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  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]
|
|
|
|
 |
|
 J K Block, K L Vandemheen, E Tullis, D Fergusson, S Doucette, D Haase, Y Berthiaume, N Brown, P Wilcox, P Bye, et al. Predictors of pulmonary exacerbations in patients with cystic fibrosis infected with multi-resistant bacteria Thorax, November 1, 2006; 61(11): 969 - 974. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. C. Marshall Pulmonary Exacerbations in Cystic Fibrosis: It's Time to Be Explicit! Am. J. Respir. Crit. Care Med., April 1, 2004; 169(7): 781 - 782. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
