Counting Concepts: Estimating the population-at-risk for regulatory action
National Occupational Injury Research Symposium
Morgantown, West Virginia
October 15-17, 1997
Alex Botkin, Director, Industry Analysis
CONSAD Research Corporation
121 N. Highland Avenue
Pittsburgh, Pennsylvania
Background of regulatory analysis needs
In its deliberations and rulemaking OSHA is bound by the authority and direction of the executive, legislative, and judicial branches of government. The judicial branch rules on suits brought by interested parties to OSHA's regulatory rulings. Several of these such as the Cotton Dust case which defined: the roles of economic benefit-cost, and technological feasibility analyses; and the Benzene cases, which required OSHA to describe the health hazards posed by chemicals and to make a determination of the existence of significant risk to workers, have made major impacts on the manner in which OSHA promulgates regulations.
The executive branch participates through Presidential Executive Order (EO) 12866 which implements the administration's regulatory review policy through the Office of Management and Budget (OMB) with respect to OSHA and other Federal agencies. Under this EO, OSHA must conduct an economic analysis and show the cost-effectiveness of proposed regulations.
Legislative action can be regulatory in nature as well mandating the construction industry lead standard, prohibiting work on ergonomics, imposing procedural requirements under the Regulatory Flexibility Act, the Small Business Regulatory Fairness Act, and the possible rewriting of the basic OSH Act itself. Underlying all these considerations is the Administrative Procedures Act (APA) which requires that interested parties be allowed to submit comments and provide evidence relating to proposed standards.
The Office of Regulatory Analysis at OSHA is charged with the development of regulatory economic analysis information concerning the costs, benefits, and technological feasibility of proposed and final OSHA regulations, as well as the relative cost-effectiveness of various regulatory and nonregulatory alternatives. Its task of assessment brings its work in contact with the entire range of regulatory history and requirements.
The analyst conducting an assessment constructs an economic cost model which integrates the prospective regulatory language, employment data, industrial or technological structure, exposure data, exposure control and abatement strategies, work practices, and finance and market condition information. The model must be accepted by the participants in the regulatory process as a valid representation of the true state of affairs. The model must reflect the critical elements of the impacts and allow for sensitivity analysis in order to estimate alternative regulatory strategies and outcomes. It must also be able to accommodate additional information provided by members of the regulated community.
This paper discusses the elements in the estimation of the population-at-risk for an economic analysis, and presents examples of regulatory analysis model development which illustrate ways in which the analysis is shaped by the availability of data.
The modeling is affected by three principal parameters: the activities that comprise the exposure or hazard situations, the determination of the number of workers in typical activities, and the employers of the workers. These micro elements which comprise the model are aggregated and then adjusted to be congruent with a national level, macro dataset which serves as a control total.
In order to meet the needs of modeling in a regulatory environment, the population-at-risk must be linked with its employers. While it is possible to use other databases such as occupational projections from the Bureau of Labor Statistics to create national estimates, those databases do not have a direct link to employers by size and industrial classification. Occupational titles also have weaknesses because the tasks implied by the title may not reflect the full scope of actual activities undertaken. It is work activity that is the source of the occupational risk.
Activities
The activity, as used in this context, is a comprehensive event involving an exposure or safety hazard. It is similar to a homogenous exposure group (HEG) or workplace exposure zone, described in industrial hygiene exposure assessment literature. An activity can represent the exposure to a chemical or an unsafe workplace condition. It can be a complex industrial process or the work assignment of a single individual. Understanding the activity permits the analyst to establish a baseline describing the conditions in the workplace that would be altered by imposition of a new regulation.
In general, the analyst strives to characterize the activity at its most granular level. The definition typically includes:
The granularity is, however, limited by the availability of information describing the activity. Furthermore, the detailed analysis may also be limited to those practices affected by a new regulation rather than a description of all aspects of the activity.
Workers
The number of workers, the second element, is estimated based on the activities already established. A common unit of measure is the crew, composed of some a number of workers appropriate to the task. The activity is usually narrowly defined to permit the assumption that the work roles of the crew members are interchangeable or the exposures are closely correlated such that further definition is unwarranted. It is also possible to have a crew consisting of just one worker. The crew is also needed as a unit of measure for cost estimation of safety equipment such as direct reading instruments or retrieval equipment. The frequency and duration of the activity in total, as well as its hazardous components, are also determined. Depending upon the regulation being analyzed, it may be necessary to understand the roles of the workers in the crew to assess the need for additional personnel to fulfill such roles as supervisor, competent person or attendant. Worker turnover is also estimated to account for the full number of workers who may participate in the crew.
Employers
The final unit of analysis is typically used to link the activities and workers to a national level data set in order to calculate aggregate costs and health benefits for a particular group of employers. In most analyses the standard industrial classification (SIC) of the employing firm or establishment is used to define the correct industrial group. National level statistics, such as County Business Patterns or Dun & Bradstreet data, serve as control totals on the number of affected workers and establishments. These data also permit the estimation of the differential economic impact of the proposed regulatory action on specific sectors of industry, and size classes of firms.
Other classification schemes are also possible. For construction industry impacts, the construction project, which conceptually is an aggregation of individual activities, can serve as the definer of the national total. In other instances, the annual consumption of a chemical and the number of office buildings have served as control totals as is illustrated in the following examples.
General Industry Confined Spaces Standard
As with many occupational safety and health hazards, the confined space is a concept. It is instantly recognizable to informed safety professionals, but is of little importance in any other business or economic context and thus not enumerated or examined in statistical data. A panel of over 50 safety professionals and engineers, identified confined space entry situations and provided definitions and estimates of many modeling parameters for specific industrial sectors.
Initial interviews and inquiry with the experts revealed that confined spaces were of three generic types:
The entry activity itself was also classifiable into categories such as inspection, testing, maintenance, cleaning, repair, process needs, and refitting. The experts also provided estimates on:
The establishment data for confined space activities related to processes were based on County Business Patterns. Facility confined spaces were based on Department of Energy data on commercial buildings. The number of product confined spaces was based on average production rates estimated for various classes of establishments and products.
A large-scale survey approach was not used because it would have been difficult, if not impossible, to adequately and economically capture some of the more infrequently-performed work activities reflected in the analysis.
Methylene Chloride Standard
After considerable initial research, application groups (activities)for the methylene chloride standard were defined for the purposes of an industry-wide survey. The survey results from approximately 1300 respondents provided the initial estimates by application group of the number of workers directly involved in the activity, engineering controls, and the average amount of solvent purchased each year.
The challenge came in estimating the number of affected establishments. Since many of the activities using methylene chloride, such as metal cleaning or adhesive use, were ancillary to the purpose of the establishment, typical control totals such as establishment counts could not be used. The solution was to develop estimates of total solvent flow by application group. This limited the total number of firms, based on the average firm=s purchases of solvent, to the maximum amount of solvent available for consumption. For the final rule, additional data received at public hearings were incorporated into the model.
Construction Industry Confined Spaces Standard
Currently, work is continuing on the standard for confined spaces in construction. The overall modeling strategy for this analysis, is similar to its general industry counterpart in the use of industry experts. In construction, however, the project is a more critical metric than the firm in establishing the population-at-risk. The type of project influences the number of potential confined spaces. Also, a construction firm may bid on a variety of projects and perform work that is not limited to its industry category. Construction workers are similarly flexible in their employment. Because some spaces are only confined during construction, a different database was needed to reflect the hazards.
There are two principal data bases available for the modeling of the construction industry. The analysis was modeled using the construction industry database of F.W. Dodge, which identifies over 209 types of projects in 34 major groups. The Census of Construction, based on monthly permit data, estimates the volume of projects by type of project based on dollar value not the number of projects. Since the number and types of confined space entries are not related to dollar value, the Census data would not have been appropriate.
The F.W. Dodge categories were regrouped based on their similarity in the presence and type of confined spaces. The population-at-risk was directly related to the volume of construction activity. Subsequently, for cost analysis purposes, confined space entries and projects were assigned to different construction SIC groups based on the activities that took place during the entry.
Conclusion
The methods used for calculating the population-at-risk are similar in most ways to any other analytical objective. The most important requirement for regulatory economic analysis, however, is the linking of the stream of costs and benefits derived from a particular population to an employer. The employees directly benefit from a healthy and safe work environment. So, too, does the employer, who for what is generally a modest investment, can then benefit from a more productive workforce and work environment.
In each of the three examples discussed, the analyst was faced with insufficient publicly- available quantitative industry data to directly identify:
A combination of strategies was used to overcome the data deficit and to develop a comprehensive analysis.
The financial and economic dimensions of regulatory action are dependent upon the conditions found in the workplace. For purposes of regulatory and epidemiological research as well, the processes, activities and even job titles must have a consistent taxonomy to enable comparisons across multiple sources of data. The resources needed to create these databases are enormous. A consistent and coordinated program of industry analysis and documentation could yield substantial benefits in the future identification, assessment and abatement of occupational safety and health hazards.
Notes: