INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It has been reported that average Quality of Life (QoL) for patients with mild to moderate levels of asthma is close to population norms (1). It may be the case that the strong relationship found between impaired QoL and more severe asthma (2) does not apply to patients with mild asthma. On the other hand, asthma is a variable disease. Patients with mild asthma may experience periods of high morbidity, and QoL may be significantly impaired at this time. An understanding of variation in QoL among patients with mild disease may be important to our understanding of factors influencing use of health services. Differences in QoL among patients with chronic respiratory illness have been associated with different levels of physician contact (8) and risk of hospital readmission (9), and Osborne and colleagues (10) have suggested that the QoL differences between men and women with asthma might explain women's greater use of health services.

As part of a larger long-term study of family practice care of asthma we administered generic QoL scales (SF-12 and SF-36) and a specific respiratory QoL scale, the St. George's Respiratory Questionnaire (SGRQ) to 396 adult patients. We also assessed symptoms, lung function, and physician contact in this group. This report presents data on the generic and respiratory QoL scores of these relatively young people on early steps of management of asthma (11) and the relationship of their QoL scores to their current asthma symptoms. It also tests if differences in QoL scores are associated with differences in prospective physician contact, as has been found for older groups of patients with obstructive airway disease (8).

All patients in the study had been prescribed at least a bronchodilator for asthma symptoms by their family physician. Patients were thus at Step 1 or above of the British Thoracic Society Asthma Management Guidelines (11). Our study was designed to contribute to our understanding of quality of life and patterns of health service use among patients with moderate levels of asthma severity.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sample Selection

Ninety percent of practices in Grampian have computerized prescribing systems. Recording of repeat prescription drug prescribing on these computerized systems has been shown to be highly accurate (12). Fifty-five practices agreed to use their computerized patient record systems to identify patients 16 to 50 yr of age who had been prescribed a bronchodilator inhaler in the previous 3 mo. The computer search gave a list of 6,277 patients throughout Grampian (population, 300,000). From this list a 6% sample of 400 patients was chosen using a random number table.

We sent a letter inviting these patients to take part in the study. We replaced refusals and nonreplies (131) with the next name on the computer list, giving a final sample of 401 patients. Patients who agreed to take part were more likely to be women (230, 81% versus 171, 69%; p < 0.01) and slightly older (33.4 versus 29.7 yr; p < 0.01). Of the 401 who agreed, review of family practice records showed that five had nonasthma diagnoses. These were (1) sarcoidosis (one), chronic bronchitis without asthmatic features (three), and cystic fibrosis (1). These were excluded from final analysis. Of the remaining 396 patients, 344 had a specific asthma diagnosis. For 52 an explicit asthma diagnosis could not be found, but no alternative respiratory diagnosis had been entered and they were all prescribed asthma medication. At interview, these 52 patients defined themselves as having asthma. They did not differ from the majority in lung function, symptoms in the past month, sex, or age.

Health Status Questionnaires

The SF-36 (13) is a widely used generic QoL scale. It has been used successfully with patients with asthma (2). Van der Molen and colleagues (1) reported that the sensitivity of the SF-36 to differences between patients with asthma was higher than that of two specific respiratory scales, the LWAQ (14) and the AQLQ (15). The recently published SF-12 (16) is a shortened form of the SF-36. It groups 12 items from the SF-36 into two six-item subscales, one measuring physical functioning (PCS-12) and the other measuring emotional functioning (MCS-12). Jenkinson and colleagues (17) extracted the SF-12 scales from SF-36 data in a nonasthmatic population and concluded that the SF-12 gave results almost identical to the SF-36. Results for respiratory patients on the SF-12 have not been previously reported. The SGRQ (18) is an established respiratory QoL scale successfully used in asthma studies (19, 20) (9, 21). Although Harper and colleagues (21) have compared the SF-36 with the SGRQ in a COPD population, a similar comparison has not been made in an asthma population.

The SF-36 (13) is made up of 36 questions covering nine health domains: PF (physical functioning), RLP (role limitation-physical), RLM (role limitation-emotional), SF (social functioning), EV (energy and vitality), P (pain), HP (health perception), CIH (change in health during previous 12 mo). The SF-12 (16) extracts 12 items from the SF-36 questionnaire in two six-item subscales, PCS (physical functioning) and MCS (emotional functioning). The SF-36 scores range from 0 (maximum impairment) to 100 (no impairment), the SF-12 scores range from 10 (maximum impairment) to 70 (no impairment). The St. George's Respiratory Questionnaire (20) has three subscales: Symptoms, Impact, and Activity. Scores vary from 0 (no impairment in QoL) to 100 (maximum impairment).

Thus, a high score on the SF-36 and SF-12 scales indicates good QoL; a high score on the SGRQ indicates poor QoL.

Home Interviews

All patients had a home visit from a respiratory nurse. The SF-36 (including the SF-12) questionnaire was sent by post before the home visit. This questionnaire was collected when the research nurse made the home visit.

During the visit the research nurse completed the SGRQ with each patient. She carried out spirometry for each patient using a portable Vitalograph spirometer to measure FEV₁ and FVC. PEF was also measured using a Wright Mini Peak Flow Meter.

Asthma Symptoms

Patients were asked to rate the frequency of day cough or wheeze, and night disturbance caused by cough, wheeze, or other asthma symptoms, in the month before the interview. From this a three-category rating was derived: (0) no respiratory symptoms in the month before interview, (1) some days/nights with symptoms, but not every week in the previous month, (2) symptoms every week on one or more days/ nights in the previous month.

General Practice Data Collection

General practices in the study were visited and patient notes were reviewed by a different research nurse, who had no contact with the home visit patients. Clinical details for the patient were collected for the 12 mo before the home interview. These included respiratory diagnosis, whether inhaled steroid was prescribed in the previous year, family practice contact in the previous year, hospital outpatient appointments, and hospital admissions. Twelve months later practices were revisited and data were collected for this period.

Analysis

SPSS-PC was used to store and analyse all data. ANOVA, t test, and chi-square tests were used to test for significant relationship between QoL scores, reported symptom levels, family practice contact, and FEV₁.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Base Characteristics

Interviewees were 16 to 52 yr of age, with a mean age of 33 yr (SD, 9.9 yr); 43% were male (Table 1). Mean %predFEV₁ was 87.6 (SD, 18.0). Fifty-nine interviewees (15%) had FEV₁ below 70% predicted. Almost half those interviewed (187, 47%) reported atopic wheeze, defined as wheeze in response to atopic triggers, in the previous 12 mo. Eighty percent reported day or night asthma symptoms in the month before interview, 30% had doctor contact for asthma symptoms in the previous 12 mo, and 17% had had an oral steroid rescue course in the previous 12 mo. Twenty (5%) had been referred to specialist care in the 12 mo before interview. Three had had a hospital admission and seven had attended an Accident and Emergency department in the previous year.

There were no significant differences between men and women in frequency of symptoms reported (² = 1.7, df = 2, p = 0.4) or %predFEV₁ (t = 1.1, df = 395, p = 0.2). However, as Osborne and colleagues (10) found, women were significantly more likely than men to have had family practice contact for asthma in the previous 12 mo (37% versus 24%, ² = 7.5, df = 1, p < 0.001). Age was significantly related to lower FEV₁, but not to weekly symptoms or family practice contact (Table 2).

QoL Scale Characteristics and Relation to Criterion Measures

Population norms on the SF-36 and norms from asthma studies for SF-36 and SGRQ are shown in Table 3. Column (a) shows British adults of working age (18 to 54 yr) (22). Column (b) shows scores for Americans with asthma in a Health Maintenance Organization (HMO) (10). Column (c) shows scores for a Netherlands asthma population. Column (d) shows mean scores for the UK sample in the current study. It can be seen that the UK group in the present study were close to population norms for all SF-36 subscales except for Physical Functioning and Health Perception. U.S. and Dutch scores were lower than UK general population or UK asthma, with U.S. scores generally worst. Column (e) shows SGRQ scores for patients with moderate asthma in a Dutch coordinated multicenter trial. Patients in the present UK study scored better (lower) on all SGRQ scales. It can be seen in Table 4 that in the present study sex (but not age) was significantly associated with scores on almost all QoL subscales. Women scored consistently worse than did men (approximately 10 to 12 % below male scores on the SF-36, 4 to 6% on the SF-12 and 14 to 44% [for Activity] on the St. George's scales). There were marked ceiling effects on the SF-36 subscales of Role Limitation and Social Functioning.

QoL and Symptoms

QoL scores in relation to symptoms are shown in Table 5. As frequency of symptoms in the previous month increased, QoL scores worsened on the SF-36, SF-12, and SGRQ. The final two columns show significance of contrast between symptom levels. On the SF-36 QoL was affected at all symptom levels for Physical Functioning, Mental Health, Energy and Vitality, Pain, and Health Perception. The SF-12 Physical subscale distinguished all levels of symptoms. All SGRQ scales distinguished all levels of symptom frequency in the previous month.

QoL and Prospective Health Service Use

Twelve months after the administration of the QoL scales, family practices were asked if the interviewee had attended for acute asthma in this time. Data were obtainable for 323 (82%) of the original interviewees. Ninety-three (29%) had attended their general practice for an asthma episode during this time. Patients who contacted a family practice in the 12 mo after interview had significantly worse QoL scores, at time of original interview, on all SF-36, SF-12, and SGRQ sub scales.

However, symptoms at interview also had a significant association with prospective physician contact. When symptom differences at time of interview were adjusted for, the association between SF-36 and SF-12 scores and prospective contact was no longer significant. The SGRQ scores continued to have a significant association with prospective physician contact. (Table 6).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, current respiratory symptoms were associated with impaired quality of life in patients with mild to moderate asthma whose overall average QoL scores were close to population norms. The impairment effect of current symptoms could be shown with both a generic QoL measure and a respiratory specific QoL measure. Any level of symptoms had some effect on QoL. The worse a patient's QoL the more likely it was that he or she would seek help from a family practice doctor for an asthma episode in the 12 mo after the study.

The patients identified in the study were relatively young. Although PEF and FEV₁ were below predicted levels only 15% had FEV₁ below 70% predicted, 59% had less than weekly symptoms, and 48% were at Step 1 of the BTS asthma management guidelines. Average QoL scores were equal or above population norms. However, both the respiratory and generic QoL scales showed that quality of life declined if patients had even occasional symptoms. Presence of any symptoms was related to significantly worse scores on all subscales of the SGRQ, on five scales of the SF-36 (Physical Functioning, Mental Health, Energy, Pain, and Health Perception), and on the PCS12 Physical scale of the SF-12.

Osborne and colleagues (10) examined 643 patients similar in age in a U.S. HMO. Their population showed almost identical levels of weekly symptoms (41%, as in the present study) and FEV₁ (19% below 80%predicted, compared with 15% below 70%predicted in the present study. But the U.S. QoL scores are lower than the UK scores, particularly for Social Functioning, Role Limitation: Physical and Emotional, and Pain. Van der Molen and colleagues (1) reported on 110 Dutch patients. Their symptoms and FEV₁ were worse than the American sample, but on the whole their QoL was greater. There may be cultural differences in the impact of respiratory symptoms on QoL. The study of Osborne and colleagues (10) found that QoL is worse in women than in men even though symptom severity is similar. We also found in the present study that women had worse QoL than did men, in spite of similar symptom severity.

Van der Molen and colleagues (1) concluded that quality of life was not significantly impaired in a study of a population with moderate asthma. We found that total QoL means for our UK asthma sample were close to population means of subjects with no chronic illness. However, the average QoL did not reveal the strong linear relationship between the presence of any symptoms and deterioration in QoL. When patients with mild asthma have symptoms their relative QoL is significantly affected. When they are symptom-free, QoL is as good as or better than the population average. This may contribute to our understanding of why many patients appear to respond to asthma only on an episodic basis, and find it difficult to follow regular prophylaxis.

The present study compared the SF-36 and SF-12 with the SGRQ in ability to reflect QoL impairment for patients with mild asthma. Although SF-36 scores worsened as symptoms increased, the SF-36 was less able than the SGRQ to show QoL differences between interviewees with no symptoms in the previous month and those with occasional but not weekly symptoms. Only five of the eight SF-36 scales showed differences between those with no symptoms in the previous month and those with less than weekly symptoms (Mental Health, Physical Functioning, Energy, Pain, Health Perception). Differences were found on the physical scale of the SF-12, but not on the emotional scale (MCS-12).

In contrast, although scores on the SGRQ for this study were markedly better (lower) than those reported in other studies using the SGRQ for patients with asthma, all SGRQ scales showed strong QoL effect of even occasional asthma symptoms. Jones and colleagues (18) have identified a 4-point difference on the SGRQ as indicating clinical significance. In the present study, patients with no recent symptoms and patients with some but not weekly symptoms had differences ranging from 4.6 points (Activity) to 10.7 (Impact). The SGRQ appeared to be well able to discriminate current symptomatic effects on QoL in this group of patients with mild asthma.

Was asthma correctly diagnosed in these patients? Diagnosis was taken from medical records. Although all patients were currently prescribed at least a bronchodilator and 52% were also currently prescribed an inhaled corticosteroid, a clinically validated diagnosis of asthma has not been demonstrated. However, they closely matched the characteristics of the 643 patients with asthma described by Osborne and colleagues drawn from an HMO. They were managed as asthmatic by their family practice physician. All but 2% of patients reported some respiratory symptoms in the previous year on the SGRQ.

Conversely, is this patient group correctly described as mild? Their mean PEF was 85%Predicted and mean FEV₁ 87%Predicted. Only 17% had had an asthma episode severe enough to require an oral steroid course in the previous 12 mo, and 20% reported no symptoms in the previous month. Only three had had a hospital admission in the previous 12 mo, and only 30% had family practice contact for an asthma episode. It seems not unreasonable to believe that this sample is representative of the mild to moderate asthmatic in the community.

The final question asked of the QoL measures was whether QoL impairment would relate to prospective health service contact. SF-36, SF-12, and SGRQ scores were all significantly different between patients who would go on to contact a doctor for asthma in the next 12 mo, and those who would not. But when symptoms at time of interview were controlled for, the predictive significance of the SF-36 and SF-12 was lost. This is at odds with the finding of Van der Boom and colleagues (8) among patients with COPD that QoL scores predicted prospective family practitioner contact but patient symptom score did not. The greater interpatient variability in symptoms in asthma than in COPD may enable the greater predictive power found for symptoms in the present study.

SGRQ scores continued to be significantly related to prospective physician contact when the effect of symptoms on prospective contact was removed. This suggests that family practice contact for asthma events is better predicted by respiratory-specific QoL differences than by QoL in general health domains. Although generic scales can discriminate QoL differences among patients with asthma, the aspects of QoL measured may not relate strongly to health service contact. It also suggests that a comprehensive questionnaire is more likely to pick up significant health effects of asthma than patient recall of symptom frequency in the previous month.

We conclude that the SF-36, particularly the Physical Functioning, Energy and Vitality, and Health Perception subscales, and the physical subscale (PCS-12) of the SF-12, can detect QoL differences between young patients with mild asthma who differ in current asthma symptoms, However, the QoL impairment measured by the SF-36 and SF-12 is not as strong a predictor of health service contact for these patients with mild asthma as QoL measured by the SGRQ. The SGRQ may be a more powerful QoL tool for assessing symptomatic effects and predicting health service contact among young patients with relatively mild asthma in family practice care.