December 2013

Using OSHA inspection data to analyze respirator protection program compliance

Several million American workers wear respirators on a regular basis, and the Occupational Safety and Health Administration (OSHA) requires that nonagricultural firms have a respiratory protection program. This article uses the OSHA inspection data base to examine all inspections in manufacturing in 47 states from 1999 through 2006; the examination starts with 1999 because an expanded OSHA respiratory program standard became effective in late 1998. The article identifies all inspections and all establishments at which respiratory protection (RP) violations were cited, and it compares the prevalence of violations by industry with the prevalence reported in a recent Bureau of Labor Statistics survey of respirator use. Multivariate analyses are used to identify the roles of industry, establishment size, union status, and employee participation in the inspection on noncompliance at the inspection level and for repeated inspections at the same establishment. The authors find that the pattern of noncompliance across industries mostly mirrors the survey findings about the prevalence of requirements for respirator use, although the chemical industry has fewer violations than expected. The probability of citing an RP violation is similar across establishment size categories, except for a large drop for establishments with over 200 workers. The presence of a worker accompanying the inspector increases the probability that a respiratory program violation will be cited; the presence of a union slightly decreases it.

Despite recent advances, knowledge is limited about the extent to which workers who are exposed to noise, toxic materials, and other risks are protected by personal protective equipment (PPE). This article explores some of the insights about respiratory protection that can be gained from an analysis of the inspection data from the Occupational Safety and Health Administration (OSHA).

This analysis complements a study that the Bureau of Labor Statistics carried out in 2001 titled Respirator Use in the Private Sector (RUPS). The survey was carried out in cooperation with the National Personal Protective Technology Laboratory (NPPTL), a division of the National Institute for Occupational Safety and Health (NIOSH).1 That survey collected extensive information from respondents at companies about the types of respirators used and the conditions of their use and asked specific questions about the companies’ compliance with some aspects of the OSHA respiratory protection standard (1910.95). By looking at OSHA inspection findings about respirator use, this paper provides another way to estimate the use of respirators as well as insights into noncompliance. The questions addressed in this study are the following:

  • How many establishments had violations of the OSHA Respiratory Protection (RP) standard?2
  • How did those numbers vary among establishments in different industries and size categories, and how did the numbers vary with the presence of labor unions?
  • How closely do the findings from inspection data match those in the RUPS report?
  • Which particular provisions of the RP standards were violated and what do these violations tell us about the patterns of noncompliance?
  • What are the characteristics of the establishments that have repeated violations of the RP standards?

There are several reasons to believe that the value of OSHA inspection data could be substantial:

  • OSHA inspects approximately 100,000 workplaces each year; the extent of coverage exceeds what most surveys or special research efforts can provide.
  • OSHA health standards typically stipulate exposure limits. When those limits are exceeded, OSHA requires that engineering controls and administrative controls be considered. If implementing them is not feasible, then respiratory protection programs are required. The standards that are cited are often very specific, thereby aiding in the diagnoses of problems at the worksites.
  • Data on violations cited have a great deal of validity in that it is unlikely that the violations did not occur. OSHA does not have to rely only on compliance self-reports; such self-reports would be likely to underreport problems. On the other hand, OSHA may be unable to detect some instances of noncompliance.
  • The data are public—including the name and address of each inspected workplace—which allows potential followup by either NIOSH or OSHA for research or prevention purposes. For example, knowledge of where violations have occurred could help NIOSH design research of several types, such as choosing facilities to investigate to see whether defects in respiratory protection programs are related to physiological changes in workers. When these RP programs are required, employers are often required to maintain medical records. Knowledge of where violations were cited in one inspection but not in a subsequent one could be used to identify workplaces where NIOSH might study how improvements in protection came about.
  • The OSHA data often, although not always, provide exposure measurements at workplaces where toxic substance exposures have been present. Knowing exposure levels helps us interpret the information on PPE use.
  • The OSHA data are continually updated, hence providing the potential for longitudinal studies.
  • The OSHA data include a set, although limited, of establishment characteristics (industry, size, location, union status, past inspection history) that may be useful for predicting outcomes.

Only one prior study has used OSHA inspection data to look at the quality of respirator programs.3 It reviewed inspection data from 1976 to 1982 in the states where federal OSHA operated the enforcement program. Thus it does not provide information about the much more detailed respiratory protection standard adopted in 1998. It did calculate that 12 percent of health inspections cited 1 or more violations of the RP standard in effect during the 1976–1982 period. Also, by estimating that the compliance with the RP was examined in only 37 percent of health inspections, the study concluded that 34 percent of the inspections in which respirator programs were reviewed resulted in a citation.4

The limitations of OSHA data for surveillance are reviewed in the discussion section of this article. One obvious issue is that the OSHA data are not collected through a random process and so representativeness cannot be assured. Nevertheless, a number of researchers have used the data for studies of exposure prevalence and trends.5 As a study on noise reported, “The exposure measurements in IMIS [Integrated Management Information System, OSHA inspection data] can be a useful surveillance tool, but interpretation of the data is dependent on the sampling strategies used by OSHA compliance officers and consultants applicable at the time the measurements are collected.”6 This conclusion applies to exposure data for other hazards as well. The data’s usefulness for characterizing exposures also depends upon the volume of data. The study concluded that the number of samples was adequate to characterize noise exposures in manufacturing but not in other sectors. For this study, our 1999–2006 data set includes over 30,000 respiratory program violations at manufacturing firms.


1 NPPTL’s role in surveillance has been endorsed by several reports, including the 2008 report of the Institute of Medicine Review Committee on Personal Protection Control Technology.

2 This standard (1910.134) has the same number as the earlier OSHA respiratory protection standard but is far more detailed in its requirements. See Federal Register, January 8, 1998, effective October 5, 1998.

3 F.S. Rosenthal and J.M. Paull, “The quality of respirator programs: an analysis from OSHA compliance data,” American Industrial Hygiene Association Journal 46(12) (1985), pp. 709–715.

4 The article reports a figure of 27 percent (not 34 percent), but this figure excludes cases that we include. Calculated using our definition of inspections with RP violations, the figure would be 34 percent.

5 Recent published studies using IMIS data have looked at exposures to formaldehyde (Lavoue 2008), silica (Yassin et al. 2005, Linch et al. 1998), wood dust (Teschke et al. 1999), beryllium (Henneberger et al. 2004), noise (Middendorf 2004), chromium (Lurie and Wolfe 2002) and lead, silica, and perchloroethylene (Gomez 1997).

6 Paul J. Middendorf, “Surveillance of occupational noise exposures using OSHA's Integrated Management Information System,” American Journal of Industrial Medicine (November 2004), pp. 492–504.

1next page

View full article
About the Author

John Mendeloff

John Mendeloff is a professor at the University of Pittsburgh and the director of the RAND Corporation Center for Health and Safety in the Workplace.

Maryann D’Alessandro

Maryann D'Alessandro is the director of the National Personal Protective Technology Laboratory, Centers for Disease Control and Prevention.

Hangsheng Liu

Hangsheng Liu is a policy researcher at the RAND Corporation.

Elizabeth Steiner

Elizabeth Steiner is a research assistant at the RAND Corporation.

Jessica Kopsic

Jessica Kopsic is a technical research analyst at MDRC.

Rachel Burns

Rachel M. Burns is a project associate at the RAND Corporation.