Sensitivity and Geographic Representativeness of a Pioneer Occupational Health Surveillance System: Massachusetts ‘Teens At Work' Injury Surveillance System

BACKGROUND:  Annually, an estimated 230,000 teens suffer work-related nonfatal injuries nationwide. Since 1993, the Massachusetts Department of Public Health’s Occupational Health Surveillance Program has conducted surveillance of work-related injuries among youths under 18. Its Teens At Work (TAW) Injury Surveillance System identifies sentinel cases for follow-up and generates population-based summary data. Sentinel cases are ascertained through worker’s compensation (WC) and a sample of nine emergency departments (EDs); population-based data are based on WC records and the statewide emergency department visit dataset (EDD). This TAW surveillance system evaluation focused on the system’s sensitivity of detecting population-based cases and geographic representativeness of its population and sentinel data, two attributes identified as priorities by the program.

METHODS: The evaluation followed CDC’s 2001 Updated Guidelines for Evaluating Public Health Surveillance Systems. Frequencies and rates were calculated among 16-17-year-old workers, using the American Community Survey estimate as rate denominator. Cases that overlapped or were misclassified as work-related in the EDD were excluded. Sensitivity was evaluated using the Bureau of Labor Statistics’ Survey of Occupational Injuries and Illnesses (SOII) estimate as gold standard. Geographic representativeness of population-based data was determined by the proportion of Public Use Microdata Areas (PUMAs) with employed teens represented in EDD and WC sources. Geographic representativeness of sentinel cases was determined by the proportion of PUMAs with the highest rates represented in WC and sample ED sources.   

RESULTS: From 2005-2008, the TAW system detected approximately 3,200 work-related injuries among 16-17-year-old workers in Massachusetts after exclusions. Case overlaps comprised less than 5% of case counts. The rate was 15.4 cases/1000 workers. Compared to the SOII estimate of 1,682 injury cases, the TAW system captured 191% more cases. The TAW population-based data represented all PUMAs where teens were employed. Among the ten PUMAs with the highest rates, all ten were represented in WC but only four were represented in the sample EDs.  

CONCLUSIONS: Findings of the TAW surveillance system evaluation highlighted its strength at capturing work-related injuries among teens in Massachusetts and the importance of using multiple data sources. As the first system of its kind in the country, the TAW system showed high sensitivity compared to the official SOII estimates. The system also represented all Massachusetts areas where teens were employed. Recommendations for the system include 1) identify and account for case overlaps in annual population-based summary data and 2) improve geographic representativeness of sentinel cases by recruiting more sample EDs from PUMAs with the highest rates.