Community-based epidemiological study of the relations of asthma and chronic bronchitis with occupation. Final report on HSE contract no. HPD/126/142/93

1. Background and aims: Incidence of occupational asthma is greater than other common occupational respiratory disorders today. In the UK, around 900 new cases of occupational asthma are reported each year under the Surveillance of Work-Related and Occupational Respiratory Disease (SWORD) scheme, with early results from the scheme suggesting an incidence of 22 new cases per million workers per year. Information gathered during the 1990/1991 Department of Employment Labour Force Survey found a prevalence often cases of occupational asthma per 20,000 population which is equivalent to 12,500 cases in the UK working population of 25 million. However, uncertainties exist about the reliability of self-reported data in the Labour Force Survey (LFS) and the completeness of reporting (and criteria for diagnosis) in SWORD.The present study was therefore designed to assess the reliability of self-reports of occupational asthma and chronic bronchitis attributed to occupation. The principal aims of the study were:(i) To assess the reliability of self-reports of occupational asthma and chronic bronchitis attributable to occupation;(ii) To assess the approximate frequency of clinically verifiable occupational asthma and chronic bronchitis attributed to occupation in the general population in an industrial area.A complementary study consisting of further analyses of the relationship between symptoms of asthma and chronic bronchitis and specific occupations and industries will be reported separately.2. Methods: Study sample-The study consisted of a cross-sectional survey, by postal questionnaire, of a sample of the working-age population in a defined geographical area of central Scotland. The target study population comprised all residents aged 16-64 inclusive in the contiguous districts of Livingston, Bathgate, Broxburn and Linlithgow in West Lothian and Grangemouth and Bo’ness in Central Region. Eligible subjects were identified using the Community Health Index of Lothian Health and Forth Valley Health Board and a random sample of 40,000 individuals was selected.Postal questionnaire survey: The postal questionnaire contained questions about symptoms of asthma and symptoms of chronic bronchitis, supplemented by further questions to identify subjects whose respiratory conditions were related to work. The first section of the postal questionnaire was designed to be as similar as possible to the trailer questionnaire on work-related illness used during the LFS, to allow for comparisons between LFS based and symptoms-based definitions. In addition, for those with respiratory symptoms, information was requested regarding the industry and occupation in which they worked at the time the symptoms first appeared. Full occupational histories were also collected.The principal mailing to the study sample of 40,000 individuals was followed by a second mailing of the complete questionnaire and a third mailing of an abbreviated version in an attempt to maximise the response. The study was also publicised in local papers, libraries, shopping centres and sports centres.The definition of asthma used in the study was based on previous IOM work, using combinations of positive replies to questions on respiratory symptoms, such as wheezing and chest tightness. The definition of chronic bronchitis was based on that used in the MRC respiratory symptoms questionnaire, with the separate MRC questions on cough and phlegm combined into one question on “”coughing up phlegm”” for simplicity. The prevalence of these symptoms was described. Comparisons were also made between the symptom-based classifications and classifications based on the modified LFS questions.The additional information on the pattern of symptoms away from work and on perception of work-relatedness of symptoms, plus information on places and activities in which symptoms worsen, were used to identify occupational asthma and chronic bronchitis attributed to occupation. The prevalence of these conditions within the study population was reported and tables and graphs used to highlight relationships between the conditions and age, sex and smoking habit. For those with work-related asthma, the industry and occupation associated with the onset of the condition were described.Clinical assessments: Following the examination of the questionnaire responses, a group of subjects admitting to respiratory symptoms were selected to undergo clinical assessment of their condition, Individuals were selected across a range of severity and included both those reporting non-work-related and those reporting work-related symptoms. Following examination at hospital by one of two consultant chest physicians, comparisons were made between the clinical diagnoses (classified into seven diagnosis groups) and the individuals’ self-reporting of symptoms, thus providing an assessment of the reliability of their self-reports.3. Results and discussion: Survey respons- The response rate for the postal survey was 50% of those delivered, with a total of 17487 questionnaires returned; particularly low response occurred among younger recipients and among males. This response was substantially lower than that which had been expected, based on the results of the pilot study carried out five years earlier, which achieved a response rate of 80%. However, comparisons of responders to the first and second mailings with responders to the third mailing (as the best representatives of the non-responders for whom we had health data) showed few differences. In particular a small increase in the prevalence of respiratory symptoms (statistically non-significant) among responders to the third mailing did not support the hypothesis of preferential, or early, response among those with ill-health.Conversely, comparison of response rate and prevalence of asthma and chronic bronchitis in the 18 postcode sectors included in the study sample showed a clear pattern of higher prevalence in areas with lower response. The postcode sectors also varied in economic status and it is likely that both the low response rate and the high prevalence of symptoms are due to some extent to the socio-economic status of the areas. This is likely to have led to some over-estimate of the prevalence of illness among the study responders. However, we conclude that, although there is some evidence of non-response bias, the information available to us suggests that its effect is unlikely to be large in the present study. In addition,the questionnaires which were returned to us were, on the whole, fully completed and the routing of the questions was followed accurately, suggesting that those who did participate in the study provided considered and logical answers.Prevalence of respiratory symptomsThe prevalence of asthma in the study sample, defined using combinations of respiratory symptoms, was 8%. The prevalence was similar among males and females and higher in the oldest, and to a lesser extent, youngest age groups. Prevalence of all asthma was higher among current smokers than among ex- and non-smokers but this association was no longer apparent after exclusion of those individuals with asthma who also had symptoms of chronic bronchitis. There was good agreement between the prevalence of symptoms-defined asthma and replies to the direct question on asthma attacks in the last 12 months. The prevalence of symptoms of chronic bronchitis in the study sample, defined as persistently coughing up phlegm, was 15%. This prevalence was higher among males than females and higher among current smokers.Prevalence of both asthma and chronic bronchitis were higher in the present study than in previous IOM studies of occupational cohorts, using the same or similar symptom definitions. This higher prevalence is likely to reflect (at least in part) the “”healthy worker effect””; whereby currently employed individuals are by definition healthy enough to work, while the general population also includes the severely ill and those people who are disabled from working. Some of the increased prevalence of chronic bronchitis may also be due to the combination of the separate MRC questions on cough and phlegm into joint questions on “”coughing up phlegm”” in the current study. Using the standard MRC definition requires the respondent to answer positively to both persistent cough and persistent phlegm rather than to only one combined syndrome as here, although conversely the syndrome of “”coughing up phlegm”” may be seen by some individuals as a more severe condition than that of “”coughing”” and “”bringing up phlegm”” .Work-related illness – comparison with the Labour Force Survey (LFS)Ten percent of respondents to the current study reported an illness caused by, or made worse by, their work in the last 12 months compared to 6% in the LFS. Only 41 individuals (0.24%) reported work-related chronic bronchitis (and five reported emphysema) compared to 0.2% reporting chronic bronchitis or emphysema in the LFS.Self-reports of work-related asthma specifically were more frequent in the present study than in the LFS; with 104 individuals (0.6% of the study sample) reporting asthma caused by, or made worse by, work compared to 0.2% in the LFS. This excess is apparent even after adjustment for the excess of work-related illness overall, with proportionally twice as many work-related asthma sufferers in the present study (6% of those with any work-related illness) as in the LFS (3% of those with any work-related illness).While the possible role of non-response bias among the study respondents in this excess cannot be excluded, it is unclear why any such bias should affect asthma prevalence preferentially. The questionnaire title “”A study of lung health”” and the other questionnaire sections could have influenced the reporting of respiratory illness but it seems unlikely that this should affect asthma reporting more than chronic bronchitis. It is, however, possible that individuals with various non-specific respiratory symptoms tended to ascribe them to “”asthma”” rather than “”chronic bronchitis””. In addition, the present study was also of a very different study population to the LFS. While the LFS results refer to the average findings across households in England and Wales, our results refer to individuals in a small area of East Central Scotland. Other possible reasons for the excess are the method of questionnaire administration (our study was self-administered while the LFS was interviewer administered) and the selection of the study sample (our study targeted individuals aged 16-64 while the LFS included all members of selected households aged 16 or over). “”