IN THE WRONG PLACE?: GEOGRAPHIC VARIATION IN U.S. OCCUPATIONAL INJURY / ILLNESS RATES
MetadataShow full item record
BACKGROUND: Around the world and across U.S. counties, workers and businesses operate in a diverse landscape of demographics, economy, culture, policy and industry. This dissertation presents four papers exploring geographic variation in U.S. occupational injury/illness rates. METHODS: The literature on geographic variation in occupational injury/illness is reviewed and categorized. Three papers examine geographic variation in the OSHA Data Initiative (ODI), 1997-2001, a database of high injury/illness industries. The first presents surveillance tools including mapping, spatial statistics, and ranking. The second uses multilevel regression to examine social determinants of county-level variation in lost workday injury/illness rates (LWDII). Finally, a case study of the meat processing industry uses mapping and regression to explore risk factors associated with both establishment location and high-LWDII establishments. RESULTS: 1) There is a small, uncoordinated literature using geographic methods to examine occupational injury/illness. 2) There is geographic variation in occupational injury/illness rates. The sample mean LWDII was 7.22 per 100 workers (county range: 0, 25.2). The five highest rate states were Vermont (9.77), West Virginia (9.76), Michigan (9.67), Maine (9.54) and Kentucky (8.99). Rates were low throughout the South. 3) Geographic variation was associated with social risk factors. In regressions, high rates were positively associated with county poverty, percent Caucasian, unionization, strong safety net, and industry hazard. Meat establishment locations were associated with county percent African American, non-college educated, longterm job gain, and urbanicity, plus iii state-level anti-union policy, medium union membership, and slightly reduced OSHA inspections. By contrast, high-LWDII meat establishments were associated with county percent Caucasian, low income, high school education, and longterm job loss. 4) There is suggestive evidence of substantial, biased underreporting in the ODI. CONCLUSIONS: Explanations for the findings are discussed. Recommendations focus on addressing underreporting, generating more county-level occupational injury/illness data, promoting county-level surveillance, increasing geographic research in occupational injury/illness, piloting programs for geographic targeting, and changing business and worker incentives and capacity for prevention.