One fish, two fish, we QC fish: controlling data quality among more than 50 organizations over a four-year period.

Abstract

EPA is conducting a National Study of Chemical Residues in Lake Fish Tissue. The study involves five analytical laboratories, multiple sampling teams from each of the 47 participating states, several tribes, all 10 EPA Regions and several EPA program offices, with input from other federal agencies. To fulfill study objectives, state and tribal sampling teams are voluntarily collecting predator and bottom-dwelling fish from approximately 500 randomly selected lakes over a 4-year period. The fish will be analyzed for more than 300 pollutants. The long-term nature of the study, combined with the large number of participants, created several QA challenges: (1) controlling variability among sampling activities performed by different sampling teams from more than 50 organizations over a 4-year period; (2) controlling variability in lab processes over a 4-year period; (3) generating results that will meet the primary study objectives for use by OW statisticians; (4) generating results that will meet the undefined needs of more than 50 participating organizations; and (5) devising a system for evaluating and defining data quality and for reporting data quality assessments concurrently with the data to ensure that assessment efforts are streamlined and that assessments are consistent among organizations. This paper describes the QA program employed for the study and presents an interim assessment of the program's effectiveness.