The 1996 and 1997 National Survey of Physician Asthma Management Practices : Background and study methodology

OBJECTIVES: To collect national baseline information on asthma management practices by physicians, and to compare these practices with the recommendations of the Canadian Asthma Consensus Conference (‘the guidelines’). DESIGN: Cross-sectional survey of representative samples of physicians in Canada in late 1996 and early 1997. POPULATION STUDIED: Five specialty types of physicians who manage asthma patients: respirology, pediatrics, internal medicine, allergy and clinical immunology, and general practice and family medicine. Stratified sampling by province and specialty was used to select physicians for the study. Weighting was used in the analysis to generalize the results to the national level for the five specialty groups of physicians. METHOD: Mailed questionnaire, self-administered by the respondent; three mailings of the questionnaires were used to increase the response rate. MAIN RESULTS: The frequency with which each of the five specialty types chose specific asthma management choices was determined, using weighted percentages representative of the specialty groups on a national basis. ANOVA determined the statistically significant differences among the five specialties in choosing particular asthma management actions. Then, logistic regression was used to calculate the odds ratios showing an association between the characteristics of the physician respondents and specific asthma management choices that they made in the survey. CONCLUSIONS: The data analysis demonstrated significant variations among physicians in asthma management practices, according to specialty type and other characteristics. The initial report was released in April 1998, and manuscripts for journal submissions are being prepared.

voir page suivante T he Laboratory Centre for Disease Control (LCDC) of Health Canada regards the implementation of the recommendations ('guidelines') from the 1998 Canadian Asthma Consensus Conference to be of vital importance for the effective control of asthma in Canada.As Canada's national public health agency for disease surveillance, prevention and control, LCDC identified the trends of increasing mortality and hospitalization rates due to asthma during the 1970s and 1980s.Asthma remains a prevalent condition.The National Population Health Survey  reported that 6% of the population two years of age and over fit the definition of having 'current asthma', nearly two million people in Canada (unpublished data, Statistics Canada).
In late 1994, following the LCDC's identification of the control of asthma as a public health priority, it formed the National Asthma Control Task Force, comprising 15 national health professional, patient and consumer organizations concerned with asthma in Canada.LCDC, specifically the Respiratory Disease Division, Bureau of Cardio-Respiratory Diseases and Diabetes, has served as the secretariat and funder of the Task Force, which is accountable to the Director General of the LCDC.
As its mandate, the Task Force has developed a National Asthma Control Strategy with the overall goal of reducing asthma mortality and morbidity in Canada.The strategy has 12 stated objectives with measurable 'targets'; specific objectives include achieving the "correct diagnosis of asthma" and the "appropriate management of asthma by physicians".The clinical practice guidelines, developed by the Canadian Asthma Consensus Conference in 1995 (1), and updated in 1998 and 1999, thus, comprise an essential means of realizing the key objectives of the National Asthma Control Strategy.
In 1996, the LCDC commissioned a national survey of how physicians manage asthma in their patients ('manage' comprising the clinical activities of diagnosis, treatment, education and follow-up of patients).The main objectives of the survey were to establish national baseline information on asthma management practices; to describe the variations in these practices according to the physicians' specialty type and other characteristics, and to compare the survey results with the practices as recommended by the 1995 Consensus Conference.

STUDY METHOD
The questionnaire for this study -called the Physician Asthma Management Survey -was designed by the LCDC (Respiratory Disease Division), in collaboration with a steering committee representing the following specialties: respirology, family medicine, emergency medicine, pediatrics, allergy and clinical immunology, nursing and epidemiology, and professional medical education.(See the Acknowledgements at the end of this article for the names of the steering committee members and the LCDC co-investigators.)Five specialties of physicians who manage asthma patients were surveyed: respirology, pediatrics, internal medicine, allergy and clinical immunology, and general practice and family medicine.Emergency room physicians had previously conducted and published their own national survey (2).Data collection: The questionnaire described three clinical scenarios: a 17-year-old female patient with a mild to moderate exacerbation of asthma, an 18-year-old male with a more severe exacerbation, and a four-year-old female with symptoms suggestive of asthma but not yet diagnosed.For these scenarios, physicians were asked about their diagnostic and treatment decisions.They were also asked direct questions about the use of various methods for diagnosing, treating, educating and following-up patients with asthma.
The questionnaires were mailed to 4489 physicians in the five specialty types in all 10 provinces and two territories during late 1996 and early 1997.The sample size of each category of physicians was chosen such that the study could detect, for a particular outcome, a difference of 10% in prevalence among the categories with a 95% CI, assuming an overall response rate of 60%.
In anticipation of a potentially low level of response by physicians to a mailed, self-administered survey, a number of steps were taken to maximize the response rate.Included with the questionnaire was a bilingual cover letter from Health Canada that included the signatures of the steering committee members and their respective national medical organizations.Two weeks after the first mailing, a reminder card was sent; three weeks after that, the questionnaires were mailed a second time.
After the second mailing, the response rate was 35%.To improve the response rate further, a shortened version of the original questionnaire was developed in consultation with the steering committee.Pretesting of the abbreviated questionnaire showed that it could be completed by physicians in 5 mins.The third mailing of the survey consisted of both the short form questionnaires and the original (long form) questionnaires being sent to separate groups of physicians who had not replied to the first two mailings.Personalized cover letters (addressed to the recipients) were included with the questionnaires.The final overall response rate was 52%.Data analysis: Price Waterhouse Consultants (Toronto, Ontario) were contracted by LCDC to mail the questionnaires, to collect the responses, and to conduct initial data analysis and interpretation.LCDC subsequently conducted further analysis of the survey data.
Two types of data analysis were completed.The first was a descriptive analysis that described the weighted percentages of each of the five specialty types who chose various responses on the survey, ie, the various 'asthma management choices'.The percentages were weighted according to the sampling scheme, so that they were representative of the specialty groups on a national basis.
ANOVA was then used to determine whether there were any statistically significant differences in the percentages of the five specialties in making a particular asthma management choice.The Bonferroni test for multiple comparisons was used at a 5% level of significance (P<0.05), because many hypotheses were being tested at the same time.
The second type of data analysis was used to determine  any statistically significant associations, as estimated by odds ratios, between the characteristics of the physician respondents and the asthma management choices that they made in the survey.Using the SAS (Statistical Analysis System (software system) (SAS Institute, Cary, North Carolina), a logistic regression model was used to calculate the odds ratios.Table 1 describes the physician characteristics that comprised the 'independent variables' in the logistic regression model.The left-hand column describes the variables, and the right-hand column describes the possible characteristics.Table 2 shows the asthma management choices that were the 'dependent' or 'outcome' variables in the regression model.The left-hand column describes the baseline variables used by both Price Waterhouse Consultants and the LCDC in their respective analyses; the right-hand column shows the additional variables used in the LCDC analysis.A large series of odds ratios were generated by logistic regression models, and their statistical significance was indicated by P.
NEXT STEPS Price Waterhouse Consultants submitted their report to LCDC in August 1997, and the main results were discussed at the meeting of the National Asthma Control Task Force in November 1997.Selected findings were presented at the Third National Conference on Asthma and Education ('ASED 3'), held in Vancouver in November 1997.The consultant's report was subsequently released by LCDC in April 1998 for dissemination among the member organizations of the Task Force (unpublished data, Price Waterhouse Consultants: Physician Asthma Management Practices in Canada: Report of a National Survey).
Papers describing the key findings of this national survey are being written by the LCDC and members of the steering committee for submission to major medical journals.Specifically, there will be a 'main paper' highlighting the overall results of the survey, and several 'side papers' describing findings specific to the specialty groups (eg, respirologists, pediatricians and allergists).

CONCLUSIONS
The findings of the national Physician Asthma Management Survey are expected to help identify the specific topic areas of asthma management that require more physician education and the specific groups of physicians to which the educational interventions should be targeted.The survey findings also serve as the baseline for comparison with any future national survey of physician asthma management practices.Such surveys can help to evaluate the effectiveness of the dissemination and implementation of the updated Canadian Asthma Consensus Guidelines, and the achievement of the objectives of the National Asthma Control Strategy.

ACKNOWLEDGEMENTS:
The authors thank the following individuals who contributed significantly to the design and methodology of the study, including the content of the questionnaires and the sampling scheme; the data collection, analysis and interpretation; and the writing or review of manuscripts describing the results of the survey. Laboratory

TABLE 2 Dependent (outcome) variables used in logistic regression modelling of associations between physician characteristics and asthma management choices made in the National Survey of Physician Asthma Management Practices Outcome variables used by both Price Waterhouse and Laboratory Centre for Disease Control
Centre for Disease Control, Health Canada, Bureau of Cardio-Respiratory Diseases and Diabetes: Dr Felix Li, former Bureau Director; Ms Louise McRae, Program Officer; Dr Charles Mustard, former Chief, Respiratory Disease Division; Dr Frank Mo, Statistical Analyst; Ms Lynne Belle-Isle, former Senior Analyst.Steering Committee, Physician Asthma Management Study (external members): Dr Robert Beveridge, Canadian Association of Emergency Physicians; Dr Louis-Philippe Boulet, Canadian Thoracic Society; Dr Benjamin Chan, Price Waterhouse Consultants; Dr Lisa Cicutto, Canadian Nurses Respiratory Society; Dr Eric Leith, Canadian Society of Allergy and Clinical Immunology; Dr Ian Mitchell, Canadian Paediatric Society; Dr Grahame Owen, College of Family Physicians of Canada; Dr John Parboosingh, Royal College of Physicians and Surgeons of Canada.