Development of the Satisfaction with Inhaled Asthma Treatment Questionnaire

Focus group meetings

Two focus groups (afternoon, evening) were held with the aim of gaining an understanding of patients' perception of their treatment regimens and problems patients encounter using asthma medication and exploring patients' perspectives about the desirability of simplified treatment regimes. Inclusion criteria were: aged 18–75 yrs and diagnosis of mild, moderate or severe asthma, either well controlled or uncontrolled (as classified by their supervising general practitioner on the basis of recent peak flow records, admissions, use of treatment and personal knowledge of patient). Non-life-threatening comorbidity was acceptable. Suitable patients were identified from primary-care clinical databases and approached by letter and telephone. Both group meetings were conducted by the same moderator and followed a similar structure.

Preliminary instrument

The preliminary instrument was based on the findings from existing literature and the focus groups and from expert opinion. Predefined basic requirements were that the written format should be suitable for self-completion by the patient, and that it should be simple, brief, and easy to understand and administer. The preliminary instrument was tested in two samples of patients familiar with the use of asthma inhalers in order to assess the comprehensiveness, comprehensibility and relevance of items. Over a period of 2 weeks, those with asthma attending either a hospital- or primary care-based chest clinic were asked to complete the questionnaire whilst waiting to be seen. The preliminary instrument was modified to reflect patients' responses.

Psychometric testing

The validity, reliability and responsiveness of the revised instrument were tested in a larger group of patients with varying severity of asthma, all of whom were familiar with the use of asthma inhalers. Patients were identified and recruited from the patient databases in the same two centres as for the pilot test. Eligibility criteria were the same as for the focus group. Additional criteria included the use of a minimum of two asthma inhalers and capability of reading and understanding English at a level allowing them to complete the questionnaire. A questionnaire and prepaid response envelope were mailed to all of these patients and a reminder was sent to nonresponders after 2 weeks. Patients who agreed to complete the questionnaire again were sent the same instrument by mail ∼2 weeks after the first questionnaire.

The questionnaire incorporated additional questions relating to age, sex, employment status, time since diagnosis of asthma and the inhalers currently used, and a number of debriefing questions enquiring about the time required for completion, the degree of difficulty of questions, clarity of formulation and appropriateness and comprehensiveness of the items. The retest questionnaire also incorporated a question asking the patient's perception of change of asthma status since first completion.

Data analysis

Data from the preliminary and revised questionnaires were analysed using descriptive statistics to verify that there was variability of response categories to each question and to verify the number of missing data for each item. Tabulation of patient characteristics, as well as missing data and consistency checks were performed. Scores of the negatively phrased items were reversed, so that for all items a higher score indicated greater satisfaction. All items that were subject to a ceiling effect were eliminated.

Factor analysis (principal components with varimax rotation) assessed the data for evidence of a domain structure consistent with domains considered part of the satisfaction construct. Domain scores were constructed by summing the intrascale item ratings and dividing this by the number of items in the domain. An overall score was calculated by adding the scores of each item and dividing this by the number of items. It was not the authors' intention to explore the detailed performance of the instrument among asthma population subgroups, but to describe overall performance in samples of subjects drawn from clinical practice.

Internal consistency reliability was measured using Cronbach's alpha 19. Test/retest reliability (reproducibility 20) was assessed using intraclass correlation coefficients. Paired t‐tests were used to compare the scores of first and second assessments from patients who completed the questionnaire twice and who had indicated no change in their asthma condition since completion of the first questionnaire. Paired samples t‐tests were conducted for the individual items as well as the four factors and the global satisfaction item to test that the average of differences between both assessments differed from zero.

Construct validity was evaluated by examining Spearman's rank correlation coefficient between items and domains and between domain and overall scale scores and a global evaluation of patient satisfaction with medication. Results of the factor analysis were also used as an indicator of the construct validity of the measure. For the measures of construct validity (factor analysis and correlation coefficients) and internal consistency reliability, data from the first (n=131) and second (n=59) assessments were pooled (combined n=190).

Focus groups and preliminary questionnaire

Of 30 patients contacted by phone, 19 agreed to attend the focus group meetings. Of the 11 patients that declined, seven were unable to attend at the scheduled times. Seventeen patients (10 male) eventually attended the focus group meetings. The majority of patients had either mild or moderate controlled asthma, with only two severe cases participating in the study. Time since diagnosis of asthma varied from 3–52 yrs.

The general pattern that emerged from the focus groups was that patients had enough knowledge of their medication to allow its proper use and that the majority of patients reported having set times of the day at which they took their prescribed medication. Most patients said that they tended to adhere to the dose of maintenance medication prescribed. However, among the most experienced users there was a consensus that they had enough knowledge and experience to vary the dose of maintenance medication according to their individual needs, decreasing the dose when they were feeling stable and well, increasing the dose when they were feeling worse or susceptible to attacks.

With respect to rescue medication, patients expressed that they disliked using this medication and tried to avoid it, because they believed it to be harmful in large doses and/or “addictive”. Patients were asked about ease of use and convenience and the importance of having the appropriate inhaler available at the right time. The majority of patients reported that they rarely left home without taking rescue medication with them, even for short-term activities, such as going to local shops or to the cinema. A common fear was that inhalers may run out unexpectedly. Inhalers with an indicator showing the number of remaining doses were thought to be very convenient and were preferred by patients.

With regards to patients' preference for a specific type of inhaler, many patients had experience of different types of inhaler. For those who had, there was no consensus on the preferred type. Undesirable features of inhalers were reported to be a risk of side-effects, such as bad taste, dry throat and hoarseness.

Pilot testing

The preliminary questionnaire comprised 12 items focusing on convenience and ease of use for each inhaler used, effectiveness and potential side-effects, and patients' confidence in and satisfaction with inhalers. Two additional questions related to the degree of difficulty and comprehensiveness of the questionnaire.

Forty-seven patients (40% male, 60% female, average age 56 yrs (range 25–78 yrs) who had asthma for an average of 17 yrs (range 1–60 yrs)) completed this questionnaire. This pretesting phase revealed a number of problems with the construction of the questionnaire. Although the majority of patients indicated that the questions were easy to answer (65%) or at an average level of difficulty (33%), the amount of missing data was substantial. The instruction to answer the questions separately for each inhaler used was not understood by one-half of the sample. This problem, as well as other comments from patients, led to significant revision of the instrument. All questions were changed into a statement mode, with response categories for all items revised into a 7‐point scale (1 “strongly disagree” to 7 “strongly agree”). A number of items were added, based on the feedback received from the patients, to make the questionnaire more comprehensive. To minimise the chance of acquiesce bias, some statements were phrased positively and others negatively in a random order. For the formatting of the questionnaire, design principles were applied as outlined by Mullin et al. 21. The revised SATQ consisted of 35 items.

Psychometric testing

A total of 131 patients with asthma participated in the psychometric evaluation of the revised instrument representing a response rate of 60% (219 were approached to participate), a response rate in line with similar postal surveys in primary care 22, 23. The demographic and other patient characteristics of responders are shown in Appendix 1, table 4⇓. Responders were of a similar sex distribution but older than nonresponders (females 68% response, males 57%, p>0.05; responders (mean±sem) 45.4±1.3 yrs versus nonresponders 36.5±1.6 yrs, p<0.001; 54% were in employment). Of those who were unemployed, 10% indicated that their unemployment was related to their asthma. The average age was 45 yrs (range 18–70 yrs) and the majority of respondents (91%) had been using asthma inhalers for >1 yr. After 2 weeks, 95 (73%) patients indicated a willingness to fill out the questionnaire again, and 59 (62%) retest questionnaires were completed and returned. The 35‐item questionnaire showed few missing data: 59 of a potential total of 6,650 responses (0.9%).

Construct validity

Factor analysis was performed on the first 34 items in the questionnaire. The last item measures the overall satisfaction and was excluded from the factor analysis. Principal component analysis used for initial factor extraction followed by varimax rotation showed nine factors with an eigenvalue >1. The screen plot indicated that more than four factors added little to the variance explained and was therefore considered unwarranted. The four-factor solution accounted for 48.3% of the variance (in order 16.9, 14.4, 8.9 and 8.1%). Item factor loadings for the varimax-rotated solution are shown in Appendix 1, table 5⇓. Eight items showed factor loadings of <0.35 on any factor or had high factor loadings on more than one factor and were excluded. The final version of the questionnaire (Appendix 2) thus consisted of 26 items.

The items loading on the first factor (k=8) reflected contents relating to the effectiveness of asthma medication. Items loading on the second factor (k=7) related to ease of use of asthma inhalers and treatment compliance. The third factor (k=6 ) related to the burden of asthma medication in general and inhalers in particular. The items loading on the fourth factor (k=5) related to potential side-effects and worries about taking the right amount of medication.

Internal consistency and test/retest reliability

The results of the internal consistency and test/retest reliability analyses are presented in table 1⇓. Cronbach's alpha values ranged 0.71–0.88 (exceeding the 0.70 criterion proposed in the literature 24, 25) for the four scales, indicating evidence of reliability of the scales and no redundancy of the items. Intraclass correlation coefficients were used as a measure of test/retest reliability (reproducibility). The four scales ranged 0.66–0.74. Paired samples t‐tests for the individual items, as well as the four factors and the global satisfaction item, showed no significant differences for any of the comparisons (data not shown).

Table 2⇓ shows the distribution of the four scales, the overall score (average of the four domain scores) and the global satisfaction item. Although the percentage of patients that reported the highest possible score (per cent ceiling) was relatively small, the majority of patients were at the upper end of the distribution, indicating good satisfaction with the management of their asthma. There was no difference between patient groups defined by age or sex with respect to average scores (males 5.41 versus females 5.30, nonsignificant (ns); <45 yrs 5.31 versus >45 yrs 5.36, ns). Patients in this study were generally satisfied with their medication; mean domain score values indicate they were most satisfied with the effectiveness and ease of use of their medication and least satisfied with medication burden and side-effects and worries. Patients also reported high satisfaction scores on the global item, with nearly one-half of the patients reporting the highest possible score.

Spearman's rank correlation coefficients between domains are shown in table 3⇓. Significant correlations were found between domain and overall scale scores and patients' global satisfaction with their medication. The strongest relationship was between satisfaction with effectiveness and global satisfaction. The lowest correlation was between the domain measuring ease of use and with side-effects/worries about medication.

Patient evaluation of measure

Patient responses to questions concerning the degree of difficulty and comprehensiveness of the instrument are shown in Appendix 1, table 6⇓. The majority of patients did not experience any problems with the difficulty or phrasing of the items. Twenty-seven patients provided feedback regarding the formulation of questions (Appendix 1) resulting in the removal of one question from the questionnaire. No aspects of patients' satisfaction already addressed in the SATQ were identified. Most (60%) of the patients estimated their completion time for the questionnaire to be <7 min.