Search

This study evaluated the preparation of Shewhart's control charts using the concept of rational subgroups for monitoring the Salmonella antibody ELISA used for surveillance of Danish pig herds. Control charts were prepared for a buffer control sample, a negative serum sample and a positive serum sample. The quality control variables were the natural logarithm (In) of the uncalibrated optical density (OD) for the buffer control sample and the negative serum sample, and the calibrated OD (OD%) for the positive serum sample. Testing round (run) within laboratory robot was chosen as the subgroup, and separate control charts were prepared for five robots. The control limits were set at six times the standard deviation for In(OD) and three times the standard deviation for OD%. Evaluation based on a number of sensitising rules for control charts produced from historical data showed that use of the control charts could reveal systematic analytical errors.

Surveillance programmes based on laboratory screening tests are increasingly used to document freedom from disease in order to facilitate trade. The following aspects must be considered when designing such programmes: diseases to be selected; epidemiology of the diseases; unit of analysis (animal or herd); target age group (or target farm type); test characteristics and sample size. Issues related to these aspects are discussed and illustrated using the example of serological surveillance for exotic viral diseases in the pig population of Denmark. Sampling designs based on individual animal samples are compared with herd-based sampling (two-stage sampling). While the latter is likely to require a larger sample size, the increased level of information and the reliability of the results obtained are considered to be worth the expense. Issues related to the development of international standards for declaring freedom from disease are discussed. The authors conclude that international standards are desirable, providing that these standards represent scientifically valid principles.