The principal objective of the study was to test hypotheses related to prevention of occupational disease and injury in construction workers. The purpose of the specific aim addressed herein was to estimate the morbidity impacts of an Owner-Controlled Insurance Program (OCIP) and state workers' compensation reform. Historically, construction has been the most hazardous industry and has only recently been surpassed by manufacturing. Study of construction worker morbidity has been hampered by the multiplicity of small employers, each with its own separate-workers' compensation insurance plan, who are present on construction sites for variable and often short periods of time. Construction of Denver International Airport (DIA) provided a unique opportunity to describe the magnitude of injury on a major construction project for which complete data on injury and hours at risk were available for over 32,000 employees working 31 million hours. For construction project owners and states wishing to reduce the incidence and burden of work-related injury and illness, knowledge about the effects of particular project organizational features, such as OCIPs, and of legal reform of workers' compensation systems upon injury and payment rates is important. These are potentially effective tools within the domains of project ownership and government. Comprehensive payroll data for all workers, who were paid standard Davis Bacon wages, allowed calculation of person hours at risk by job classification. Complete reporting, facilitated by a single workers' compensation plan covering all contracts and by an on-site medical clinic and designated provider system, allowed us to determine both total and lost-work-time (L WT) injury rates per 200,000 hours at risk. Workers' compensation payment rates were calculated and compared with expected loss rates, derived by the National Council on Compensation Insurance. All rates were calculated by relevant time periods. An apparent lack of validity of nationally determined injury rates did not allow us to make meaningful comparisons with DIA rates, leaving unanswered the question of the effects of an OCIP on injury rates. Moreover, no similar data for a similar project without an OCIP with which to compare the DIA experience existed; we could therefore draw no conclusion about the effects of an OCIP on worker morbidity. Without information about the workers' compensation experience of similar projects in states without legal reform during the same period, it is not possible to attribute the reduction in DIA payment rates we found for the pre-reform and post-reform periods to workers' compensation reform.