Data mining: a technique for discovering novel exposure and health outcome relationships.
Middendorf-P; Syamlal-G; Wang-M; Linch-K; Wood-J
American Industrial Hygiene Conference and Exposition, June 1-6, 2002, San Diego, California. Fairfax, VA: American Industrial Hygiene Association, 2002 Jun; :102
Data mining, or knowledge discovery, are techniques that are being applied to business data to increase competitiveness and profitability of companies. The data mining process involves large databases which have been developed, sometimes through combining several smaller databases, refining the data so that it uses the same variables, and then sifting through it to identify previously unsuspected patterns using statistical software models. Data mining techniques are likely applicable to databases beyond business databases, such as exposure and health outcome databases. The techniques of data mining may be useful in better defining relationships or elucidating new relationships between working conditions, exposure, and health outcomes. Two candidate databases for data mining are the U.S. DOL MSHA Coal's Management Information System (CMIS), and the The NIOSH Coal Workers Chest X-Ray Surveillance Program (CWXSP). CMIS is a national database of information on mine status, personnel, time and activity, respirable dust sampling entities and respirable dust and respirable quartz sample results measured by MSHA inspectors at surface and underground coal mines since 1970. The Coal Workers' X-ray Surveillance Program (CWXSP) is a NIOSH-administered occupational health program mandated by the Coal Mine Health and Safety Act of 1969. The primary objective of the CWXSP is to screen miners for coal workers' pneumoconiosis (CWP). Since 1970, all active underground coal miners have been required to have a chest radiograph at the time of hire and again three years later. Subsequently, they can volunteer for radiographs at approximately five-year intervals. In addition to the posterior-anterior chest x-ray, other information is collected, including miner identification, age, tenure, and specific job in the mine. The methods used to combine the databases into a single database useful for data mining, the techniques applied to the database to explore relationships, and the results of the exploration will be presented and discussed.
Data-processing; Models; Statistical-analysis; Respirable-dust; Sampling; Dusts; Coal-mining; Mine-workers; Pneumoconiosis
Work Environment and Workforce: Emerging Technologies
American Industrial Hygiene Conference and Exposition, June 1-6, 2002, San Diego, California