School
School of Public Health
Document Type
Dissertation
Department (new version)
Public Health
Date of Award
1993
Degree Name by School
Doctor of Public Health (DrPH) School of Public Health
Abstract
This study examined the effectiveness of OSHA's safety program assessment method in predicting workplace safety performance. The OSHA system for evaluating safety programs involves 22 indicators which are rated by state consultants following a comprehensive safety survey of each evaluated company. Program ratings form the basis for the consultant's recommendations regarding management of safety within each company. In addition, these program ratings are among the criteria which determine a company's eligibility for exemption from routine OSHA inspection. The effectiveness of the rating system, however, has never been evaluated. A sample of 120 companies was selected from 328 companies receiving initial onsite consultative visits by the Alabama safety and health consultation agency (Safe State) in 1990. The sample included thirty companies for each of four consultants. Company sizes ranged from 10 employees to 500 employees. The sample included a diversity of industry types with a range of low, moderate, and high injury experience. Intra-rater agreement and inter-rater agreement for these evaluations were estimated using correlational analyses. Program evaluations performed by the consultants in 1990 were repeated in 1991 using only file documentation. Intra-rater agreement was estimated by comparing these two ratings. Programs in 40 randomly selected companies were rated by the author as the basis for estimating inter-rater agreement. Poisson multiple regression techniques were used to evaluate the influence of the 22 indicators in predicting lost-workday cases. The full regression model included 29 independent variables, with lost-workday cases as the outcome variable and company size as the rate multiplier. The reduced model was determined using forward, step-wise procedures. The study found intra-rater agreement high for each of the four consultants. Inter-rater agreement among the consultants varied, but was generally poor. Safety program assessment scores assigned by state consultants to Alabama companies were found to correlate with the lost-workday incidence reported by these companies, but the predictive power of these correlations was limited. Only 9 of the 22 indicators were found to account for a significant portion of the variation in lost-workday case incidence.
ProQuest Publication Number
ProQuest ID
9407444
ISBN
979-8-208-64792-9
Recommended Citation
Weems, William Howard, "OSHA program assessments as predictors of safety performance" (1993). All ETDs from UAB. 7187.
https://digitalcommons.library.uab.edu/etd-collection/7187
Comments
DrPH