Krahn and associates relied on a variety of data sources, including surveys, government and private databases, market research and expert opinion. As a result, the cost-of-illness estimate is more comprehensive than those of studies conducted in the United States[1] and Australia.[2] Unfortunately, Krahn and associates do not discuss the validity and reliability of the specific data sources employed. Access to several sources permitted the researchers to present baseline and high estimates for each cost item. Although this appears to provide an estimate of precision, the authors provide no information on the accuracy of the results. For example, the calculations of the costs of inpatient care and of emergency visits used the mean cost for all admissions and for all emergency department visits, not those specific to asthma. To calculate the indirect costs associated with morbidity, the authors used activity loss information from the 197879 Canada Health Survey. Although the data were adjusted to control for population growth to 1990, no adjustment was made for the documented increase in morbidity due to asthma during this period.[3,4] No justification is provided for the weights of 1.0, 1.0 and 0.5 assigned to bed days, major-activity-loss days and cut-down days, respectively. These weights conflict with the values of 0.5, 0.4 and 0.3 used by Wigle and collaborators[5] in their estimation of the cost of long-term and short-term disability based on the same survey. The chosen weights should take into account the episodic nature of asthma and the probability that a sick worker can be replaced, mitigating the potential productivity loss. The use of the human capital method to value productivity losses remains controversial. This method assumes that the demographic profile of patients with asthma is identical to that on which national wage statistics are based. More recent methods are available for testing the assumptions in the human capital method concerning the opportunity cost of time.[6]
Assessments of the cost of asthma are complicated by the difficulty in ascertaining the diagnosis. The International Classification of Diseases, 9th revision (ICD-9) code used to identify asthma in administrative databases is also used for allergic bronchitis. The accuracy of self-reports of asthma in the Canada and Ontario health surveys also remains questionable.
However, the authors clearly recognize the limitations of their methods. Their study provides insight into the economic burden of a growing public health concern and points out the direction for future research and the importance of developing methods that improve accuracy and permit meaningful measurements of precision in economic assessments.
Wendy Ungar, MSc
Doctoral candidate
University of Toronto
Toronto, Ont.
References
The authors reported an incidence of 222 cases per year for Canada based on data from workers' compensation boards; it is unclear whether this incidence represents only true occupational asthma (OA) or also includes aggravation of asthma (AA), which is compensable in Ontario. We have recently examined the characteristics of more than 600 suspected cases of OA submitted to the Ontario Workers' Compensation Board from 1984 to 1988;[5] 235 workers received compensation for claims of OA and 234 for claims of AA. Between 1989 and 1993, the number of allowed claims per year in Ontario ranged from 94 to 126 for OA and from 30 to 42 for AA.[2]
It is unclear what the costs for workers' compensation throughout Canada represent; the authors indicate in the footnote to Table 2 that workers' compensation payments were treated separately because they were transfer payments and that they were not included in the column totals. One should take into account that the three main components of compensation are disability (time lost), medical expenses and rehabilitation. Did the cost estimates for worker's compensation include only time off work, or were estimates for health care (such as drugs, physician assessments and emergency visits) and vocational rehabilitation included? If not, the health care costs may have been included among the direct or indirect costs for asthma in general, suggesting that the costs for workers' compensation (and the proportion that this contributed to overall asthma cost) would be even higher.
The authors should be commended for including this work-related aspect of disease, which is often neglected.
Gary M. Liss, MD, MS, FRCPC
Ontario Ministry of Labour
Ka Sing Yeung, PhD
Lynn Elinson, MSc
Ontario Workers' Compensation Board
Susan Tarlo, MB, BS, FRCPC
Gage Research Institute
University of Toronto
Toronto, Ont.
References
We agree with Dr. Liss and associates about the importance of asthma's effect on workplace-related disability. We also agree that workers' compensation payments provide a gross underestimate of the economic importance of asthma in employed Canadians. Costs borne by the health care system (for inpatient care, emergency visits and some drugs) are not reflected in disability payments. Patients with brief exacerbations of asthma do not always apply for compensation. In addition, the level of reimbursement is affected by many factors other than costs to workers. We believe that our study accurately measures most of the cost of asthma affecting employed Canadians, but it does not provide separate reports for the employed and the unemployed. We provided workers' compensation data to facilitate comparisons with other studies, not to represent the full cost of asthma among those employed. With respect to the issue of the number of claims, the Ontario Workers' Compensation Board sent us a report in 1993 that indicated that the total number of claims for asthma (ICD-9 code 493) in 1989 was 61. Some cases of asthma may have been coded differently (e.g., ICD-9 code 506, respiratory condition from chemical fumes). If so, our estimates are conservative.
Murray Krahn, MD, MSc, FRCPC
Assistant professor
Departments of Medicine and Clinical Biochemistry
University of Toronto
Toronto, Ont.
Reference