Juniper 902.907

Development and validation of a questionnaire to measure E.F. Juniper*, P.M. O'Byrne+, G.H. Guyatt*,+, P.J. Ferrie*, D.R. King* Development and validation of a questionnaire to measure asthma control. E.F. Juniper, P.M. O'Byrne, G.H. Guyatt, P.J. Ferrie, D.R. King. #ERS Journals Ltd 1999.
statistics +Medicine, McMaster University ABSTRACT: International guidelines on asthma management indicate that the pri- mary goal of treatment should be optimum asthma control. The aim of this study was to develop and validate the Asthma Control Questionnaire (ACQ).
Correspondence: E. Juniper, Dept of Clini- The authors generated a list of all symptoms used to assess control and sent it to 100 asthma clinicians who were members of guidelines committees (18 countries). They scored each symptom for its importance in evaluating asthma control. From the 91 responses, the five highest scoring symptoms were selected for the ACQ. In addition, there is one question on b2-agonist use and another on airway calibre (total ques-tions=7). The ACQ was tested in a 9-week observational study of 50 adults with symp- tomatic asthma. The ACQ and other measures of asthma health status were assessed In patients whose asthma was stable between clinic visits, reliability of the ACQ was high (intraclass correlation coefficient (ICC)=0.90). The questionnaire was very re- sponsive to change in asthma control (p<0.0001). Cross-sectional and longitudinalvalidity were supported by correlations between the ACQ and other measures of asth- ma health status being close to a priori predictions.
In conclusion, the Asthma Control Questionnaire has strong evaluative and dis- criminative properties and can be used with confidence to measure asthma control.
Eur Respir J 1999; 14: 902±907.
International guidelines indicate that the primary goal of asthma treatment is to achieve optimum control (minimi-zation of day and night time symptoms, bronchoconstric- tion and short-acting b-agonist use) and thus reduce therisk of life-threatening exacerbations and long-term mor- The authors defined asthma control as: "the full range of bidity [1±4]. COCKCROFT and SWYSTUN [5] have pointed clinical impairment that patients with asthma may exper- out that asthma "control" concerns the adequacy of treat- ience as a result of their disease. The range is from 'well ment, whilst "severity" concerns the underlying disease controlled', in which the patient is totally unimpaired and process: "the common perception that well-controlled unlimited (this may be achieved either spontaneously, as in asthma is synonymous with mild asthma and that poorly seasonal asthma, or by the use of medications) to 'extrem- controlled asthma is synonymous with severe asthma is ely poorly controlled', which is a 'life-threatening state'.
It was specified that the ACQ should: 1) Include symp- A number of questionnaires are described as measuring toms that asthma clinicians consider to be most important "asthma severity". In some, the term "severity" has been for assessing adequacy of asthma control. 2) Include a used appropriately [6, 7] and the instruments validated ag- measure of a) short-acting b2-agonist use and b) airway ainst other measures of airway pathology. No question- calibre. 3) Be applicable to all adults with asthma (17±70 naires have been specifically developed and validated to yrs). 4) Be reliable (give reproducible data when the clini- measure asthma control. Without such an instrument, a sim- cal state is stable and be able to discriminate between pa- ple, quantified method for identifying patients at risk and tients with different levels of asthma control). 5) Be re- for evaluating the effects of treatment has been lacking.
sponsive (be sensitive to small but clinically important The authors developed the Asthma Control Question- changes in asthma control). 6 ) Be valid (actually measure naire (ACQ) using recognized procedures for question- asthma control). 7) Be short and easy to complete.
naire specification, item selection and scaling [8] and theauthors have examined the measurement properties nec-essary for its use in clinical practice, clinical trials and cross-sectional surveys [9]. The aim was to develop asimple questionnaire that could be completed in the clinic The authors generated a list of all symptoms that might without daily recordings of symptoms, medication use be used by clinicians to assess the adequacy of asthma control by identifying treatment goals from international Table 1. ± Item reduction phase: highest scoring items Frequency (number of clinicians scoring each category) The first five symptoms were included in the Asthma Control Questionnaire (n=87; 4 clinians failed to respond to this question).
guidelines [1±3, 10], reviewing other asthma question- completed by the clinic staff. Patients recall their ex- naires and talking to asthma clinicians. Ten potentially periences during the previous 7 days and respond to important symptoms were identified (tables 1, 2).
each question using a 7-point scale (see Appendix). Theitems are equally weighted and the ACQ score is the mean of the 7 items and therefore between 0 (well con-trolled) and 6 (extremely poorly controlled).
Subjects and methods. One hundred asthma clinicians,representing 18 countries, were asked to participate.
They were clinicians who had served on international asthma guidelines committees [1±3, 10], asthma clini-cians with measurement expertise and other opinion lea- ders in asthma management. Each clinician was sent a Fifty adults (17±70 yrs) were enrolled from previous list of the 10 symptoms and asked both to score (5= studies, local media notices and asthma clinics. They were extremely important, 0=useless) and rank each symptom required to have symptomatic asthma with an ACQ score for its importance in evaluating asthma control. Parti- >0.5 at enrolment. Patients were excluded if they had cipants were told that questions concerning airway cali- evidence of chronic obstructive pulmonary disease, an ill- bre and short-acting b2-agonists use would be included ness with symptoms similar to those of asthma (e.g.
in the final questionnaire. They were asked to select the cardiac), recurrent chest infections, or were unable to com- municate in English. The study was approved by theMcMaster University Faculty of Health Sciences Ethics Results. Ninety-one clinicians returned the questionnaire.
Committee. All patients signed an informed consent form.
Importance and ranking of the 10 symptoms are shownin tables 1 and 2. Seventy-four per cent of respondentspreferred prebronchodilator forced expiratory volume in one second (FEV1) % predicted as the measure of airwaycalibre.
In this 9-week, observational study, patients were seen in the clinic at enrolment and after 1, 5 and 9 weeks. At Asthma control questionnaire. The ACQ includes the each visit, spirometry was measured before and 20 min five highest scoring symptoms, one question about b2- after bronchodilator, and patients completed the ACQ. For agonist use and another about FEV1, the latter being validation purposes, patients also, completed the Asthma Table 2. ± Item reduction phase: highest ranking items The first five symptoms were included in the Asthma Control Questionnaire (not all clinicans included all symptoms in their ranking).
*: rank order of importance is, 1=most important, 10=least important.
Categorizing patients. Conceptually, testing the ACQ's measurement properties required defining a group of pa-tients who remained clinically stable between consecu- tive clinic visits (weeks 1±5 and 5±9) and another group who experienced change in their asthma control. For each time period, each patient was categorized using the clinician's global rating of change: stable group = scores of -1, 0 or +1; unstable group = scores -7± -2 and + 2± Evaluative properties. Responsiveness of the ACQ was examined in three ways. Firstly, for patients in the un- stable group, it was determined whether the ACQ could Data are presented as Pearson correlation coefficients. A priori detect within-patient change using a paired t-test. Sec- predictions: 1) Change in Asthma Quality of Life Question- ondly, it was assessed whether the ACQ could detect naire: r=0.4±0.8; the highest correlations should be with the differences between stable and unstable patients using an symptom domain (r=0.6±0.8) and the lowest with the environ- unpaired t-test. Thirdly, the responsiveness index (mean mental domain (r=0.4±0.6). 2) Change in "the physical health" change (D)/standard deviation of change (SDD) was cal- domain of the Medical Outcomes Survey Short Form-36 (SF- culated [14]. To ensure that the contribution of two 36): r=0.2±0.4. 3) Change in the 5 additional asthma symptoms observations by some patients did not result in an over- (cough, chest tightness, sputum, coloured sputum and clearing estimate of the precision of responsiveness, the authors the throat): r=0.4±0.6. 4) Clinician's global rating of change: inflated the variance by the quantity 1 + (n-1) r where r is the intraclass correlation coefficient (ICC) of thechange scores and n=2 (No. of observations per sub- Quality of Life Questionnaire (AQLQ) [11] and the Medi- ject) [15]. For longitudinal validity, a priori predica- cal Outcomes Survey Short Form-36 (SF-36) [12] and tions were made about the amount of correlations that scored the five least important asthma symptoms (table should be expected to be observed if the ACQ truly 1), responding to each question on a 7-point scale. For measures change in asthma control. The predictions were one week before each follow-up visit, patients recorded based on results from previous studies and clinical ex- prebronchodilator peak expiratory flow (PEF) each morn- ing and daily use of short-acting b2-agonist. At each fol-low-up visit, a clinician rated change in the patient's Discriminative properties. Reliability of the ACQ was asthma control since the previous clinic visit (+7 = a very determined from patients in the stable group. If a patient great deal better, 0 = no change, -7 = a very great deal was stable between both weeks 1±5 and weeks 5±9, a worse) [13]. The clinician was blinded to the ACQ data single observation was selected using a random number and used only spirometry, diary, AQLQ and SF-36 data generator. Reliability was estimated as the within-subject plus a consultation with the patient.
standard deviation and related to the total standard dev-iation as an ICC. For cross-sectional validity, data from Medications. Patients whose asthma was adequately the second clinic visit (week 1) were used and once ag- controlled continued on their established asthma medica- ain a priori predications were made about the level of tions throughout the study. Patients whose asthma was correlation that should be expected if the ACQ truly not adequately controlled at week 1 and/or week 5 were advised to increase their medication as recommended bytheir asthma physician.
All fifty patients (18 males, 32 females) completed the study. Their mean age was 37‹13 (SD) yrs and pre-bron- General approach. A health status instrument that is re- chodilator FEV1 % pred. was 77.2‹18.8. Twelve patients quired to measure change over time (e.g. clinical trials used short-acting b2-agonists alone; 34 needed regular and clinical practice) must have good evaluative proper- inhaled steroids plus short-acting b2-agonists; three took ties which are responsiveness (the ability to detect im- inhaled steroids plus both long and short-acting b2-agon- portant within-patient changes, even if they are small) ists; and one required all three medications plus an oral and longitudinal construct validity (appropriate correlati- ons between changes in the new instrument and changesin established health status measures) [9]. An instrument that is required to distinguish between people at a sin-gle point in time (e.g. surveys and impairment assess- Thirty-six patients contributed 50 observations to the ment) must have good discriminative properties which unstable group. Symmetry of improvements (n=26) and are reliability (high ratio of variance between-patients deteriorations (n=24) allowed the combination of the data to variance within-patients) and cross-sectional con- by changing the sign of those who deteriorated. The ACQ struct validity (appropriate correlations between estab- was able to detect change in these patients (mean change lished measures and the new instrument) [9]. The ACQ 0.73‹0.54, p<0.0001). There was minimal change in the was tested for both evaluative and discriminative proper- stable group (mean change 0.01‹0.24, p>0.05) and the ACQ was able to detect the difference between the stable and unstable groups (p<0.0001). The responsiveness index To meet the authors' specifications of brevity and com- was 1.35. Correlations between changes in the ACQ and pletion in the clinic, the ACQ includes only those symp- changes in other measures, in general, agreed well with a toms that most commonly reveal lack of control in the majority of patients and does not include daily PEFmeasurement. Nevertheless, it is believed that there is animportant role for the ACQ in the management of in- dividual patients in both general practice and tertiary care.
Thirty-six patients contributed 50 observations to the Not only does the ACQ allow clinicians to become familiar with the goals of asthma management, but it can also 0.90. Correlations between the ACQ and the other out- identify patients with poor control and, more accurately come measures matched the a priori predictions quite well than recall, evaluate the effects of interventions. In addi- tion, completion of the ACQ in the waiting room may saveconsultation time and group monitoring may be used toenhance disease management and thus reduce resource For categorizing patients into the stable and unstable Although there is a plethora of asthma symptom ques- groups, it would have been ideal for several clinicians, tionnaires, this is the first to be specifically developed and blinded to the ACQ data, to have independently assessed validated to measure asthma control. The authors first each patient at each clinic visit and a consensus taken as to examined the goals of asthma treatment [1±3, 10] and then whether each patient had changed. This was not feasible sought consensus from international opinion leaders con- and left the choice of reviewing recorded data at a later date cerning the symptoms and measures of airway calibre that for a group decision or having one clinician make the de- are most important for assessing asthma control. Agree- cision at the time of the clinic visit. The latter was selected ment was sufficiently high for the authors to have con- because a group decision would have had to rely heavily fidence in selecting the top scoring items for the ACQ.
on recorded symptoms, b2-agonist use and airway calibre, The technique used to develop the ACQ is unique in that i.e. all the data recorded in the ACQ. The approach taken a large number of leading asthma clinicians from around makes it less likely that spuriously high correlations re- the world participated in identifying the questions that sulted from ACQ data exercising undue influence on the should be included. The authors interpret the high response rate as an indication of the importance they place upon this A limitation of this study is the relatively small and pos- instrument to help disseminate the concept and goals of sibly homogeneous sample. Although the authors endea- optimum asthma management. The high level of con- voured to enrol patients with a wide range of asthma sensus among clinicians removes any ambiguities that may severity and socioeconomic background, most were Cau- have existed concerning the meaning of optimum asthma casian. Testing of the Asthma Control Questionnaire in other settings will increase confidence in the general ap- This study has shown that the ACQ has strong mea- surement properties both as an evaluative and as a dis-criminative instrument and can be used with confidence inboth clinical trials and cross-sectional surveys. The ACQ isneeded for research studies to measure the primary goal of Acknowledgements. The authors thank the fol- asthma treatment, to identify populations at risk and to lowing clinicians who completed the item reduc- facilitate comparison of results across studies.
tion questionnaire, many of whom also gaveexcellent suggestions: E. Adelroth, R. Antic,W.C. Bailey, N.C. Barnes, P.J. Barnes, R. Beas- ley, A.B. Becker, E. Bel, L-P. Boulet, J. Bous-quet, R.A.L. Brewis, J.R. Britton, A.S. Buist, W.W. Busse, A. Cartier, K.R. Chapman, T.J.H.
Clark, G.M. Cochrane, D.W. Cockcroft, G.K.
Crompton, R. Dahl, R.J. Davies, J. Dolovich, S.R. Durham, D. Evans, J.M. Fitzgerald, P.G.
Gibson, S. Godfrey, M. Gotz, F.E. Hargreave, B.D.W. Harrison, D.J. Hendrick, S.R. Hilton, T.W. Higenbottam, S.T. Holgate, W.F. Holmes, P.H. Howarth, G.H. Koeter, L.I. Landau, D.J.
Lane, T.H. Lee, H. Levison, A.J. Lockhart, H.
Magnussen, J-L. Malo, G.B. Marks, R.J. Martin, A.D. Milner, C. Mitchell, H.J. Neijens, H.S.
Nelson, E. Neville, M.T. Newhouse, A.J. New- man-Taylor, R.L. Page, M.R. Partridge, R.A.
Pauwels, M.G. Pearson, S. Pedersen, T.A.E.
Platts-Mills, D.S. Postma, J.F. Price, H.H. Rea, Data are presented as Pearson correlation coefficients. A priori C.E. Reed, R. Rodriguez-Roisin, A.R. Rubinfeld, predictions: 1) AQLQ: r=0.4±0.8. The highest correction should M. Rudolf, R.E. Ruffin, G.F. Ryan, J.M. Samet, be with the symptom domain (r=0.6±0.8) and the lowest with J.P. Seale, M.R. Sears, G.G. Shapiro, A.L. Shef- the environment domain (r=0.4±0.6). 2) "physical health" do- fer, P. Sherwood Burge, M. Silverman, P.J. Sterk, main of the Medical Outcomes Survey Short Form-36 (SF-36): A. Szczeklik, A.E. Tattersfield, N. Thomson, J.H.
r=0.4±0.6. 3). The five additional asthma control symptoms: Toogood, E.H. Walters, J.O. Warner, S.T. Weiss, A.J. Woolcock, M. Yeung, P. Zimmerman.
Circle the number of the response that best describes how you have been during the past week On average, during the past week, how often were you woken by your asthma On average, during the past week, how bad were your asthma symptoms In general, during the past week, how limited were you in your activities In general, during the past week, how much shortness of breath did you In general, during the past week, how much of the time did you wheeze? On average, during the past week, how many puffs of short-acting bronchodilator (eg. Ventolin) have you used each day? To be completed by a member of the clinic staff (Record actual values on the dotted lines and score the FEV1 % predicted in the next #The Asthma Control Questionnaire is copyrighted. It may not be changed, translated or sold (paper or software) without the validate a new quality of life instrument. In: Spilker B. ed.
Quality of Life and Pharmacoeconomics in Clinical Tri- British Thoracic Society, Research Unit of the Royal als. Second Edition. Raven Press Ltd, New York, 1995; College of Physicians of London, King's Fund Centre, National Asthma Campaign. Guidelines for management Guyatt GH, Kirshner B, Jaeschke R. Measuring health of asthma in adults: I chronic persistent asthma. Br Med J status: what are the necessary measurement properties? J Clin Epidemiol 1992; 45: 1341±1345.
US Department of Health Services. International Consen- Hargreave FE, Dolovich J, Newhouse MT. The assess- sus Report on Diagnosis and Treatment of Asthma. Beth- ment and treatment of asthma: a conference report. J Al- esda: National Heart, Lung, and Blood Institute, 1992.
lergy Clin Immunol 1990; 85: 1098±1111.
Thoracic Society of Australia and New Zealand. Wool- Juniper EF, Guyatt GH, Ferrie PJ, Griffith LE. Measuring cock A, Rubinfeld AR, Seale JP, Landau LL, Antic R, quality of life in asthma. Am Rev Respir Dis 1993; 147: Mitchell C, Rea HH, Zimmerman P. Asthma management plan 1989. Med J Aust 1989; 151: 650±653.
Bousquet J, Knani J, Dhivert H, et al. Quality of life in Ernst P, Fitzgerald JM, Spier S. Canadian asthma con- asthma. 1. Internal consistency and validity of the SF-36 sensus conference: summary of recommendations. Can questionnaire. Am J Respir Crit Care Med 1994; 149: Cockcroft DW, Swystun VA. Asthma control versus asth- Juniper EF, Guyatt GH, Willan A, Griffith LE. Deter- ma severity. J Allergy Clin Immunol 1996; 98: 1016±1018.
mining a minimal importafit change in a disease-specific Burney PGJ, Chinn S, Britton JR, Tattersfield AE, Papa- quality of life instrument. J Clin Epidemiol 1994; 47: 81± costa AO. What symptoms predict the bronchial respon- sive to histamine? Evaluation in a community survey of Guyatt GH, Walter S, Norman G. Measuring change over the bronchial symptoms questionnaire (1984) of the in- time: assessing the usefulness of evaluative instruments. J ternational union against tuberculosis and lung disease.
Juniper EF, Guyatt GH, Feeny DH, Griffith LE, Ferrie PJ.
Venables KM, Farrer N, Sharp L, Graneek BJ, Newman-Taylor AJ. Respiratory symptoms questionnaire for asth- Minimum skills required by children to complete health- ma epidemiology; validity and reproducibility. Thorax related quality of life instruments: comparison of instru- ments for measuring asthma-specific quality of life. Eur Juniper EF, Guyatt GH, Jaeschke R. How to develop and

Source: http://www.efficas.my/doctor/pdf/Quality%20of%20Life/Juniper%20asthma%20control%201999%20.pdf