| 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 Lung and Allergy Research, Division of Physiology, National Institute of Environmental Medicine (Dr. Ehrs), Centre for Allergy Research (Mr. Nokela), and Department of Medicine (Drs. Hjemdahl and Wikström Jonsson), Karolinska Institutet, Stockholm; and Department of Public Health and Caring Sciences (Dr. Ställberg), Section of Family Medicine and Clinical Epidemiology, Uppsala University, Uppsala, Sweden.
Correspondence to: Eva Wikström Jonsson, MD, PhD, Clinical Pharmacology Unit L7:05, Karolinska University Hospital (Solna), SE-171 76 Stockholm, Sweden; e-mail: Eva.Wikstrom-Jonsson{at}medks.ki.se
Abstract
Study objectives and design: Health-related quality of life (QoL) instruments are generally used for studies of asthma in specialized settings. For primary care use, there is a need for brief and simple questionnaires for structured patient-reported outcomes. We validated the Mini-Asthma Quality of Life Questionnaire (Mini-AQLQ), using the Asthma Quality of Life Questionnaire with standardized activities (AQLQ[S]) as the "gold standard." The Asthma Control Questionnaire (ACQ) was validated against the symptoms domain of the AQLQ(S). Patients were characterized by the Short Form-36 Health Survey (SF-36).
Subjects: One hundred eight patients (68 women) with asthma diagnosed by their physicians from 24 primary care centers completed two visits (2 to 3 months apart). Their mean SF-36 scores were lower than the national norm for all domains.
Results: The Mini-AQLQ and ACQ correlated well with the AQLQ(S). Reliability, determined in 57 patients with stable AQLQ(S) scores, was good. Both brief questionnaires detected improvement or deterioration of patients at the group level. Global ratings of disease severity by patients or clinicians correlated poorly with disease-specific QoL scores.
Conclusions: The Mini-AQLQ and ACQ instruments are sufficiently simple and robust to be suitable for research and quality of care monitoring in primary care at the group level. They may, after further validation, even be useful in the management of individual patients.
Key Words: asthma • Asthma Control Questionnaire • Asthma Quality of Life Questionnaire with standardized activities • health-related quality of life • Mini-Asthma Quality of Life Questionnaire • primary care
Asthma is mainly managed in primary care. Lung function data are often used to evaluate disease severity and treatment effects, but they are poorly correlated to quality of life (QoL) in asthma patients.1234 Treatment goals should therefore emphasize patient reported outcomes, such as QoL, to a greater extent in clinical practice.5 Several disease-specific QoL questionnaires have been developed and validated for use in research on asthma,6 but these questionnaires have mainly been used in specialized settings and are too extensive for routine use in primary health care. The availability of QoL questionnaires that are brief and simple to use in a nonspecialized setting would facilitate evaluations of asthma management in primary care.
To be suitable for routine use in the evaluation of disease management in primary health care, questionnaires must be simple to use and valid in that setting, which has little or no experience of research methodologies and limited time for each patient visit. Thus, the patient should be able to complete the questionnaire during a short visit and without extensive instructions. Simple questionnaires for measurements of health-related QoL in ordinary primary care would be valuable for surveys of asthma management and for research on asthma in primary care but might also be used in the management of individual patients. However, several standards need to be met for use of QoL questionnaires at the individual level.7 First, they should be easy to administer, score, and interpret. Second, the questionnaires should include a variety of health concepts. Third, floor and ceiling effects should be minimal so that substantial numbers of patients do not get the highest or lowest possible scores. Fourth, the scores should have small SEs, indicating precision for cross-sectional assessments and longitudinal monitoring. Fifth, they should be valid indicators of the constructs they represent and be sensitive to clinical change.
Our aim was to find and validate QoL instruments that are suitable for research and/or quality assessment in primary care and, possibly, also for routine use. Among the available brief questionnaires on disease-specific QoL in asthma, we chose to further examine the Mini-Asthma Quality of Life Questionnaire (Mini-AQLQ).8 As the "gold standard" for the purposes of validation, we included the Asthma Quality of Life Questionnaire with standardized activities (AQLQ[S]).9 We also examined the Asthma Control Questionnaire (ACQ),10 which is a symptom-focused questionnaire and not a QoL questionnaire, since health-related QoL is multidimensional. In this case, we used the symptoms domain of the AQLQ(S) as the "gold standard," but we also examined if asthma control correlated with health-related QoL in a broader sense by comparing ACQ ratings with scores in other domains of the AQLQ(S) scale.9
To characterize our patients further, we also administered a generic QoL instrument, the Short Form-36 Health Survey (SF-36).111213 We also included global ratings by patient and clinician in order to see how these corresponded to the more detailed information obtained by the questionnaires.
Methods and Materials
A prospective, multicenter study was performed in 24 primary health-care centers in the city and suburbs of Stockholm. The study centers were evenly distributed throughout the county and were selected to ensure inclusion of patients with different socioeconomic backgrounds. Their physicians and nurses had limited or no experience of research routines. Furthermore, the inclusion criteria were wide. The study was approved by the Ethics Committee of the Karolinska Institute, Stockholm, Sweden.
Study Population
One hundred seventeen patients 18 to 86 years of age were included. All participants were required to have asthma diagnosed by a general practitioner (GP). The study centers were instructed to consecutively invite all eligible patients who sought medical care to participate. Exclusion criteria were as follows: age < 18 years, COPD, malignant disease, severe psychiatric disease, and dementia. Patients were required to understand written Swedish to be able to complete the questionnaires and to give written informed consent. Nine patients with randomly distributed characteristics did not show up at the second visit. The final validation is thus based on 108 patients (Table 1
).
|
Study Design
Two brief questionnaires were compared to the "gold standard" we chose for purposes of validation, the AQLQ(S).9 The generic SF-36 questionnaire111213 was used to characterize patients. All questionnaires were completed on two occasions 2 to 3 months apart. If needed, the patients were advised to change their medication after the first visit.
The SF-36 was always administered first, according to instructions from the developers.1415 The remaining four questionnaires were completed in the following order: the AQLQ(S), the ACQ without spirometry,16 patient questions on the global rating of symptoms and disease severity (see below), and the short version of Asthma Quality of Life Questionnaire (AQLQ), the Mini-AQLQ.8 Thus, the two most similar questionnaires were maximally separated. All questionnaires were translated versions suitable for self-administration that were authorized and provided by the developers. They were completed in the same order on both occasions. The patients were allowed a maximum of 60 min to complete all questionnaires. The patients were informed that the evaluation concerned a questionnaires and not themselves. After having completed the questionnaire, they were not allowed to go back to change or check answers.
At the second visit, the GP or the asthma nurse, according to local routines, estimated if and how the patients’ clinical status had changed compared to visit 1. Pharmacotherapy during the week preceding each visit was recorded.
Questionnaires
SF-36:
The SF-36 provides a descriptive measure of generic health-related QoL that is suitable for comparisons of different patient populations. It consists of eight multi-item domains that may be combined into summary scores for physical and mental health.13 The SF-36 has good internal consistency and cross-sectional validity in patients with asthma.1415
AQLQ(S): The AQLQ2 is an asthma-specific health-related QoL questionnaire with strong evaluative and discriminative properties regarding data at the group level.2 The original AQLQ includes questions related to activities selected by the patients themselves, but we chose to use the version based on standardized activities, the AQLQ(S)9 to minimize the risk of data loss in the primary care setting.
The AQLQ(S) is a 32-item, disease-specific questionnaire that has strong measurement properties and validity for measurements of functional impairment in adults with asthma.9 Patients score their experiences during the last 2 weeks on a 7-point scale (1 = severe impairment to 7 = no impairment). The AQLQ(S) contains 12 items on symptoms, 11 items on activity limitations, 5 items on emotional functions, and 4 items concerning environmental stimuli. The overall AQLQ(S) score and mean responses for different domains (symptoms, activities, emotions, and environment) are calculated. A score change of 
0.5 points is considered to be clinically important and is termed the minimal important difference.17
Mini-AQLQ: A short version of the AQLQ, the Mini-AQLQ,8 consists of five items on symptoms, four items on activity limitations, three items on emotional function, and three items concerning environmental stimuli. Its measurement properties are good but not as strong as the original AQLQ.8 It is scored with the same 7-point scale as the AQLQ(S): 1 = severe impairment to 7 = no impairment. The complete user package can be obtained free of charge to all clinicians and academics after contact with Professor Elizabeth Juniper, at juniper@qoltech.co.uk.
ACQ: The ACQ is a short questionnaire for measurements of asthma control10 that also can be obtained from Professor Elizabeth Juniper. The original questionnaire consists of seven items, five of which concern symptoms and activity limitations, one of which concerns the percentage of predicted FEV1, and one of which concerns the use of ß2-agonists during the preceding week.10 All questions are scored on a 7-point scale (0 = good control to 6 = poor control), and overall control is reflected by the mean of all responses. The ACQ has strong evaluative and discriminative properties for measurements of asthma control in adults.10 The FEV1 question can be omitted without changing the validity or measurement properties of the instrument at the group level.16 Since spirometry was not readily available in all health-care centers, we used the modified questionnaire without the FEV1.
Patients’ Global Ratings
The patients were asked in writing, "How would you generally describe your health concerning your asthma?" and answered using the following alternatives: completely free from symptoms, almost free from symptoms, symptoms sometimes, symptoms often, or symptoms always. The patients were also asked in writing, "How would you generally describe the severity of your disease?" and answered as follows: very mild, mild, moderate, severe, or very severe.
Clinicians’ Global Rating
At the second visit, the GP or asthma nurse classified each patient’s change in global asthma status compared to visit 1 as follows: much worse, worse, stable, better, or much better. These ratings were made without special instructions or guidance by questionnaire results.
Statistical Analysis
We primarily used nonparametric methods, as we did not assume normality of distribution for any of the variables assessed. However, to allow comparison with previous validation studies, data in the tables are given as mean ± SD unless otherwise indicated. Statistical software was used (SPSS 12.0.1; SPSS; Chicago, IL).
Data from the Mini-AQLQ and ACQ instruments were compared to those from the AQLQ(S) to evaluate the closeness of association by calculating Spearman correlation coefficients.
Measurement properties of the instruments were evaluated by calculations of the intraclass correlation (ICC) [the ratio between subject variance and total variance] and test-retest reliability (Spearman correlation coefficients), using data from a subgroup of patients who were in stable condition according to AQLQ(S) ratings on the two occasions. A change score between visits of < 0.5 points (the minimal important difference) was categorized as stable.
Internal consistency, responsiveness, and longitudinal and cross-sectional validity were evaluated using data from the entire study population. Internal consistency was estimated by calculating the Cronbach 
coefficient.18 Responsiveness was evaluated by determining if the instruments could detect differences between patient groups that were stable, improved, or deteriorated according to the AQLQ(S) scores using the Kruskal-Wallis test and post hoc multiple comparisons. Since this estimate has guidelines for its interpretation,19 the responsiveness of each instrument was also examined by calculating the effect size according to the formula: mean score at the first visit, subtracted by the mean score at the second visit, and divided by the SD of the score at the first visit.
To examine cross-sectional validity, we postulated that if the Mini-AQLQ measures the same construct as the AQLQ(S) and the ACQ measures the same construct as the symptoms domain of AQLQ(S), these should correlate reasonably well. Longitudinal validity was defined as the ability of the change scores obtained with an instrument to correlate with the change score of the criterion/benchmark test. A correlation coefficient > 0.7 was considered to indicate good validity.
Results
According to the GINA classification,5 76 patients had severe asthma, 21 had moderate asthma, 8 had mild asthma, and 2 had intermittent asthma. Severity could not be determined in one patient due to missing data (Table 1). There was no significant correlation (r < 0.169; p = 0.169, n = 68) between AQLQ(S) scores and pulmonary function measured as FEV1 percentage of predicted.
The patients scored lower than the national norm in all domains of the SF-36, despite ongoing treatment. This impairment was most evident for the general health and role-physical domains, where the differences were 21.8 and 20.1 points, respectively, at visit 1. There were no significant differences between SF-36 scores at the two visits (not shown).
Comparison of Instruments
Correlations between Mini-AQLQ and AQLQ(S) were strong (r > 0.80), except for the environmental domain (r = 0.73) [Table 2
, Fig 1
]. ICC values were good, ranging between 0.62 and 0.94, with the lowest value for the environmental domain (not shown).
|
|
values, expressing internal consistency, were generally good (Table 2). The ICC values were similarly good (data not shown).
Measurement Properties
Reliability:
The reliabilities of the Mini-AQLQ and the ACQ were determined in the 57 patients who were in stable condition according to AQLQ(S) ratings. Within-subject and between-subject SDs as well as the ICC and test-retest reliability coefficients were good for the "gold standard" instrument, the AQLQ(S), also under the present conditions (Table 3
).
|
The within-subject SDs were somewhat higher and between-subject SDs somewhat lower for the ACQ than for the AQLQ(S) overall and symptom domains. The ICC was slightly lower for the ACQ than for the overall and symptoms domain scores of the AQLQ(S) [Table 3]. The correlation coefficient of the ACQ for stable patients was 0.89 (Table 3), indicating good reliability.
Responsiveness: All three questionnaires yielded minimal changes in the stable group (Table 4 ). Effect sizes were similar for the AQLQ(S) [0.19], the Mini-AQLQ (0.21), and the ACQ questionnaires (0.21), indicating similar responsiveness of the questionnaires.
|
Validity: Good cross-sectional correlations between the Mini-AQLQ and AQLQ(S) scores indicate that the former measures asthma-specific QoL acceptably (Table 5 ). The correlation between the Mini-AQLQ and the physical index of SF-36 was moderately strong (r = 0.71). The correlation between the ACQ and the AQLQ(S) was weaker than that between the Mini-AQLQ and AQLQ(S), with higher values for the overall and symptoms domain scores of AQLQ(S). The ACQ and SF-36 physical domain ratings were moderately correlated (r = – 0.60). The longitudinal validity was good. The correlation between Mini-AQLQ and AQLQ(S) overall change scores was 0.85. The ACQ performed equally well (r = – 0.78). The longitudinal validity was weakest for the environment and emotions domains of Mini-AQLQ. Omission of the ß2-agonist question from the ACQ did not significantly affect its validity.
|
Floor and Ceiling Effects: The ACQ had floor effects (0 points in the overall score for several patients, leaving no room for improvement) but no ceiling effects. The AQLQ(S) and the Mini-AQLQ did not suffer from any floor effects and negligible ceiling effects (< 5%) in the overall and symptom domain scores. The three remaining domains had moderate ceiling effects (indicating small room for improvement) in both questionnaires (Table 6 ).
|
The brief questionnaires for patient-reported outcomes evaluated in this study had good measurement properties in an ordinary Swedish primary care setting. Results obtained with the Mini-AQLQ and our "gold standard," the AQLQ(S), correlated well, and the ACQ correlated well with the overall and symptoms domain scores of the AQLQ(S). The questionnaires appeared to be equally responsive, but this was not tested with a structured intervention. Our results show that the brief instruments can discriminate between stable and improved or deteriorated groups, and that they may be very useful in primary care research, as well as for quality assessments of the management of asthma in primary health care. Correlations between changes in the Mini-AQLQ or ACQ and AQLQ(S) scores agreed well with previous results obtained in other settings.810
The Mini-AQLQ exhibited good measurement properties and should be useful for both cross-sectional and longitudinal studies. The weakest domain was the environmental domain, which is composed of few items.
The ACQ is constructed to evaluate symptom control and consequently correlated best with the symptoms domain of AQLQ(S); there was, however, also a good correlation with the overall score of the AQLQ(S). Leaving out the question on ß-agonist use did not destroy its properties. This question has lost its discriminative value in modern asthma therapy, which commonly includes long-acting ß-agonists. Thus the ACQ is suitable for evaluations of asthma control based on symptoms alone, without FEV1 and ß-agonist use, in agreement with previous studies in a smaller patient group.16 We noted limited technical skills in obtaining good lung function data, as well as problems with their interpretation at several centers. This finding also favors use of the ACQ version without FEV1. The modified ACQ is obviously useful for symptom evaluation in adult asthmatics at the group level. We also agree with the suggestion that the ACQ may be useful in the management of individual patients in different clinical settings, including general practice.16 However, several standards need to be met for questionnaire use at the individual level,7 and such use will require further validation.
The present results further strengthen the view that asthma patients in primary care should be monitored on the basis of patient-reported outcomes, rather than on lung function data. However, our results apply to patients at the group level. We found that the ACQ and Mini-AQLQ instruments were easy to administer, but the primary care centers did not have to deal with scoring and interpretation. The Mini-AQLQ includes several domains and health concepts, whereas the ACQ focuses on symptoms. The ACQ had limited room for improvement, indicating that many patients had good asthma control and few symptoms even though they had severe asthma according to the GINA classification. However, patients adapt to their condition and may also deny or neglect symptoms. There was no relationship between asthma severity according to GINA and the risk of hitting the floor or the ceiling with the questionnaires. The scores obtained by the Mini-AQLQ had small SEs, and it appears to be valid also in the routine primary care setting. Furthermore, it was sensitive to clinical change. Commonly accepted requirements for reliability coefficients are 0.70 for group comparisons, whereas individual monitoring requires coefficients > 0.90.20 The reliability coefficient of the overall score of Mini-AQLQ was 0.91 in our study. Altogether, the present data indicate that the Mini-AQLQ may be useful for the monitoring of individual asthma patients in primary care.
Due to the consecutive inclusion of patients seeking medical care, 76 of our 108 patients had severe persistent asthma according to the GINA criteria.5 Since patients with more severe disease seek medical care to a greater extent than patients with mild disease, they constitute a major part of the workload in primary care. However, the subjective ratings of disease severity of our patients correlated poorly with their GINA classifications. This is in line with previous findings.21 Inhaled corticosteroid and long-acting ß2-agonist treatment are instituted early in the course of the disease in Sweden. Thus, the therapeutic tradition and guidelines in Sweden favor classification into a more severe GINA category. However, the severity of disease among our patients corresponds well with their low scores on the SF-36.
The finding that the clinician’s ratings correlated poorly with changes in AQLQ(S) scores or pulmonary function tests emphasizes the need for other means of evaluating the effects of asthma therapy than his/her general impression in the routine health care of asthma patients. The patient’s global rating correlated reasonably well with the symptoms domain of the AQLQ(S) and with the ACQ, suggesting that symptom control is most important for the patient’s global rating. In agreement with this interpretation, it has been shown that patients adjust their medication according to symptoms rather than PEF measurements.22
In summary, the present results suggest that the Mini-AQLQ and ACQ instruments are valid and useful instruments for research and for monitoring of the quality of asthma care at the group level in primary health care. Our results indicate that the main priority of the patient may be symptom control, and that they feel better than one would expect, based on their disease classifications according to GINA and their SF-36 scores. Furthermore, the clinician’s ratings of asthma severity reflected patient-reported outcomes poorly. Thus, brief health-related QoL questionnaires may be a valuable aid in the management of individual patients in routine clinical practice. The Mini-AQLQ performed well enough to be a suitable instrument for further validation regarding its possible usefulness in the care of individual patients with asthma in primary care.
Acknowledgements
The authors thank research nurse Lena Wahlberg for monitoring of the study sites and for entering results into the database. The enthusiasm and willingness to cooperate among the physicians and nurses of the participating health care centers are much appreciated.
Footnotes
Abbreviations: ACQ = Asthma Control Questionnaire; AQLQ = Asthma Quality of Life Questionnaire; AQLQ(S) = Asthma Quality of Life Questionnaire with standardized activities; GINA = Global Initiative for Asthma; GP = general practitioner; ICC = intraclass correlation; Mini-AQLQ = Mini-Asthma Quality of Life Questionnaire; PEF = peak expiratory flow; QoL = quality of life; SF-36 = Short Form-36 Health Survey
This study was supported by the Vårdal Foundation, the Karolinska Institute, the Stockholm County, and the Drug and Therapeutics Committees of the Stockholm and Sörmland Counties.
Received for publication February 25, 2005. Accepted for publication October 2, 2005.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
