Use of multiple data sources to enumerate work-related amputations and carpal tunnel syndrome in Massachusetts, California, and Washington

Background: Given evidence of undercounting in the Survey of Occupational Injuries and Illness (SOII), the Bureau of Labor Statistics funded Massachusetts, California, and Washington to assess feasibility of using multiple data sources to enumerate work-related amputations and carpal tunnel syndrome.

Methods: Researchers across states collaborated to develop common surveillance case definitions and criteria for identifying cases. Data sources used varied by state depending on availability,legal access and size of population studied. Sources included SOII, workers' compensation claim and statewide hospitalization data (all states; emergency department data(MA); medical records (WA, MA; CA sample). Cases were matched across datasets and demographic/employment characteristics of cases by data source described. Total enumerated cases and the subset of SOII eligible cases were compared to SOII estimated cases during the study periods (MA,CA:2007-2008; WA 2006 - 2008.)

Results: Multisource surveillance consistently indentified more cases than any single source and more than the SOII estimates. This was partially explained by capture of case ineligible for SOII. However, all states found that SOII underestimated the number of SOII eligible cases. Estimates of the undercount varied by health condition and state and ranged from 30 - 64% for amputations and from 13-72% for CTS. Changes in case status over time and reporting of cases as other types of injuries in the SOII explained some but not all undercount.

Conclusion: Multisource surveillance substantially improves our capacity to characterize the burden of work-related injuries and illness but poses practical challenges that vary by state and currently limit aggregating data across states.