Validation of The 30 Second Asthma Test as a measure of asthma control.

BACKGROUND
The primary objective of asthma management is to help patients establish and maintain optimal disease control. Simple and efficient tools are needed to assess patient-reported symptoms so that they can be used with or without airway function to evaluate asthma control.


OBJECTIVE
The objective of the present study was to evaluate the validity of The 30 Second Asthma Test (GlaxoSmithKline Inc, Canada), based on the Canadian Asthma Guidelines, by estimating its relationship with criterion measures of control.


RESULTS
The discriminative and diagnostic validity of The 30 Second Asthma Test was examined in a sample of 81 patients with a confirmed diagnosis of asthma. Based on a cut-off score of two or greater on The 30 Second Asthma Test, the overall agreement with specialist ratings was 65%, and 58% with per cent predicted forced expiratory volume in 1 s. The 30 Second Asthma Test scores distinguished between groups of patients who were classified based on the change in intensity of therapy.


CONCLUSION
The results support the use of The 30 Second Asthma Test as a brief screening tool for asthma control.

A sthma is one of the most prevalent conditions affecting Canadians. According to the 1996 National Health Survey (1), asthma affects 6% of adults and 12% of children, and its prevalence is increasing. Asthma poses a heavy burden on national health care expenditures, and reduces the quality of life of affected individuals and their families (2).
The primary goal of asthma treatment is for patients to maintain optimal control to prevent exacerbations and unnecessary hospitalizations. Routine, formal assessments of control by physicians are infrequent, a shortcoming that may explain why poor asthma control is much more common than it should be (3,4). Consequently, physicians and patients tend to overestimate the adequacy of control (3,4). Many patients perceive their asthma as controlled, even if they experience significant limitations in leisure and work activities (3)(4)(5). As a result, patients often do not convey their limitations to physicians, which may increase the gap between actual and optimal asthma care (3).
Standardized measures of asthma control can help to guide appropriate changes in therapy. Spirometry is viewed as the most accurate means of evaluating adequate asthma control (6)(7)(8)(9). Unlike the situation in specialty clinics, family physicians are limited in their capacity to administer spirometry tests to all patients. Therefore, there is a need for alternate tools that can be readily administered at home and in primary care practice. The Canadian Asthma Consensus Report (10) defined acceptable asthma control using both clinical symptoms and physiological parameters (Table 1). These variables included daytime symptoms, night-time symptoms, use of rescue bronchodilators, limitations in physical activity and missed school or work. The 30 Second Asthma Test (GlaxoSmithKline Inc, Canada) was developed based on the clinical symptoms outlined in the asthma guidelines (Table 1) (11). The properties of this measure have never been formally examined. The objective of the present study was to evaluate the validity of The 30 Second Asthma Test™ for assessing asthma control compared with spirometry-based assessment, specialist judgments and changes in intensity of therapy.

METHODS
The present study was approved by the Institutional Review Board of the Montreal General Hospital (Montreal, Quebec).

Study population
Patients with a confirmed diagnosis of asthma were recruited from the Montreal General Hospital asthma clinic ( Table 2). Patients were excluded if they were unable to communicate in either English or French.

Measures and variables
The 30 Second Asthma Test™ consists of five questions on the presence or absence of nocturnal symptoms, daytime symptoms (coughing and wheezing), the use of rescue medications, symptom interference with daily activities and absenteeism from work or school. The maximum total score of five represents the unweighted sum of positive responses.

Criterion measures of control
Specialist ratings of patient level of control were evaluated using a five-point Likert scale ranging from very good to poor. Medications used by the patient were recorded at the time of the visit, including changes in the intensity of treatment, which was classified as unchanged, increased or decreased.
Forced expiratory volume in 1 s (FEV 1 ) was evaluated during the patient visit. The best FEV 1 measurement over the previous 12 months was also recorded. Asthma control was evaluated using FEV 1 estimates by calculating the ratio of patients' FEV 1 results at the time of the visit to the best recorded FEV 1 observed over the previous 12 months (FEV 1[control] = FEV 1[today] /FEV 1[best] ). At the time of the visit, patients were classified as in control if the FEV 1 ratio was greater or equal to 90%, and out of control if the FEV 1 ratio was less than 90%, as has been suggested in the Canadian Asthma Guidelines (10).
To characterize the study population, age, sex and smoking history (patients who smoked over 10 packs/year) were obtained from patient medical records.

Procedure
The 30 Second Asthma Test™ was completed by the patient at the time of the visit to the asthma clinic, before seeing the specialist physician. Assessment of FEV 1 was performed in the pulmonary function laboratory. The specialist physician assessed the degree of asthma control at the time of the visit and made changes, if any, to the intensity of drug therapy. The specialist physician knew the value of the FEV 1 , the clinical history and results of the physical examination, but was not shown information from The 30 Second Asthma Test™.

Data analyses
The internal consistency of the items on The 30 Second Asthma Test™ was estimated using the Cronbach's alpha coefficient.
The screening accuracy of The 30 Second Asthma Test™ was calculated using the specialist ratings and per cent predicted FEV 1 . Patients were classified as in control, according to the specialist ratings, if asthma control was classified as very good, good or acceptable, and out of control if classified as fair or poor. For The 30 Second Asthma Test™, in control was defined as a score of less than two, and out of control was defined as a score of greater than or equal to two. The sensitivity, specificity, and positive and negative predictive values were calculated for a cut-off score of two on The 30 Second Asthma Test™. To evaluate optimal cutpoints on The 30 Second Asthma Test™, the receiver operator characteristic (ROC) curve was constructed using specialist ratings, and the area under the curve was examined. Optimal sensitivity and specificity were estimated from the ROC curves, defined as the values corresponding to the point on the ROC curve lying in the left uppermost part of the curve.
Control status based on scores of The 30 Second Asthma Test™ was also evaluated across groups based on the change in intensity of therapy (unchanged, increased and decreased).

RESULTS
Seventy-five per cent of patients had at least one of the five symptoms (Table 3). Daytime symptoms were the most frequently reported (53%), followed by the use of rescue medications and limitations in physical activities (40%).
The internal consistency of the items on The 30 Second Asthma Test™ was moderate (Cronbach's alpha = 0.70) and the individual coefficient alpha for each item was 0.60 or greater.
The correlation between specialist and patient ratings of asthma control was moderate (r=0.5) and significant  (P<0.001). When the cut-off criteria of two or more symptoms were used, 51% of patients were classified as out of control based on The 30 Second Asthma Test™, and 26% were classified as out of control based on the specialist ratings ( Table 4).
The sensitivity and specificity of The 30 Second Asthma Test™ (cut-off score of two or greater) using the specialist ratings as the criterion measure were 81% and 60%, respectively. The area under the ROC curve was 0.79 ( Figure 1). The optimal sensitivity and specificity, defined as the values corresponding to the point on the ROC curve lying closest to the left uppermost part of the curve, were 68% and 79%, respectively, which corresponded to a cut-off score of three on The 30 Second Asthma Test™.
The sensitivity and specificity of The 30 Second Asthma Test™ using FEV 1(control) as the criterion measure were 65% and 55%, respectively ( Table 5). As expected, if the intensity of therapy was unchanged or decreased, a larger proportion of patients were classified as in control compared with out of control based on The 30 Second Asthma Test™ (Table 6).

DISCUSSION
The Canadian Asthma Guidelines (11) were developed to standardize and highlight the importance of evaluating asthma control routinely. These guidelines were the basis for the content and development of The 30 Second Asthma Test™. There was support for the validity of The 30 Second Asthma Test™, which performed as expected in relation to specialist ratings, FEV 1(control) and change in intensity of therapy. A cutoff score of two or higher was an indication of poor control as evidenced by the sensitivity analysis. The patient ratings corresponded better with specialist ratings than FEV 1(control)

In control Out of control Total
The 30 Second Asthma Test™, n (%) In control score < 2 33 (41) 7 (9) 40 (50) Out of control score ≥ 2 27 (34) 13 (16) 40 (50) Total, n 60 20 The 30 Second Asthma Test™ had a sensitivity of 65%, specificity of 55%, positive predictive value of 33%, negative predictive value of 83% and accuracy of 58%. FEV 1 Forced expiratory volume in 1 s  estimates, which is in keeping with other studies (3,12,13). Thirty-four per cent of patients had more than two symptoms on The 30 Second Asthma Test™ but were classified as in control according to FEV 1(control) . This supports the fact that symptom assessments are capturing a different aspect of control that is not reflected in airway function tests alone.
Previously developed measures of asthma control assess a combination of the items found in The 30 Second Asthma Test™ with variations between measures in the aspects of the symptoms evaluated (eg, some ask about getting less done as opposed to missing work or school), the time period evaluated and the frequency of symptoms ( Table 7). The only other measure that has been compared with criterion measures of control as we did in the present study is the Asthma Control Test (ACT) (13). As opposed to asking about missed school or work, or limitations in physical activity as in The 30 Second Asthma Test™, the ACT has an item that asks "During the past four weeks, how often did your asthma keep you from getting as much done at work, school or at home?" The two items on physical limitations and their consequences on The 30 Second Asthma Test™ helps to discriminate between patients who have some limitations in physical activity and those who are significantly more impaired and are unable to attend school or work as a result of their asthma. The ACT, however, has an item that asks patients to rate their overall asthma control, which may help physicians to identify patients who have a disparity between their perceived level of control and actual symptoms. Future work is needed to compare The 30 Second  Asthma Test™ to other self-report measures of control, such as the ACT, to identify items and scoring scales that would be most discriminative and sensitive to changes in control. Based on the ROC curve, the optimal cut-off score for The 30 Second Asthma Test™ was three. This corresponded to a higher specificity at the cost of reducing the sensitivity of The 30 Second Asthma Test™, which would decrease the number of false-positive cases (patients identified as not well controlled). Given the consequences of missing a patient who is not well controlled, the recommendation of a cut-off score of two, with a sensitivity of 81%, provides an appropriate balance between the risk of falsely labelling a patient as well controlled as opposed to poorly controlled.
The specialist ratings of control were made by two physicians who were considered to be experts in asthma care. The specialists weighed all the information at their disposal, such as past medical history recorded in the chart, prior assessments of the same patient (specialist ratings were obtained at follow-up but not initial visits), questioning the patient, physical examination, as well as current and past spirometric results. In the absence of a 'gold standard' of asthma control, we thought that the best approach for assessing the screening accuracy of The 30 Second Asthma Test™ was to use the summary ratings of experienced specialists.
One possible limitation of the present study is that 34% of patients had smoked over 10 packs/year. This raised the question of whether patients were truly asthmatic. For patients to qualify for the study, the physician specialists had to provide the criteria for why they thought a diagnosis of asthma was justified. All patients in this study met the criteria for a diagnosis of asthma based on one or more of the following: response to a bronchodilator, a positive methacholine challenge test, variability in FEV 1 and/or response to therapy. We also examined the sensitivity of The 30 Second Asthma Test™ among nonsmokers and found that the sensitivity was 76%, somewhat lower than the full sample. The specificity increased to 77% among nonsmokers.
The validity and simplicity of The 30 Second Asthma Test™ support its use in the clinic to provide an objective assessment of patients' symptoms that can be monitored over time. Self-reported measures ensure that an appropriate patient history is taken, and adds to the comprehensiveness of airway function and physician assessments of control. The use of The 30 Second Asthma Test™, in conjunction with patient education, may also prompt patients to seek physician care when their asthma is not well controlled. The items also serve as a reminder of the asthma guidelines, which are not always adhered to by physicians (3). Future work is needed to examine the responsiveness of The 30 Second Asthma Test™ by comparing it with criterion measures of change and comparing its performance with other patient-reported measures of control.
FUNDING: Dr Sara Ahmed was supported by a fellowship award from the Canadian Institute of Health Research. The present work was supported by a grant from GlaxoSmithKline Canada.
Validation of The 30 Second Asthma Test™