Maternal serum alpha-fetoprotein screening. II. Pitfalls in low-volume decentralized laboratory performance.

Assays of maternal serum alpha-fetoprotein are subject to the phenomenon of assay drift, which may be defined as incorrect increase or decrease of alpha-fetoprotein values from their true values. Low maternal serum alpha-fetoprotein weekly volume (for example, fewer than 500 specimens per week) will result in a greater than 47% probability that 10% assay drift will not be recognized. Further, laboratory reports to clinicians may lead to either misdirecting 43% more pregnant women (with positive drift) into further (possibly invasive) diagnostic procedures or the offer of further diagnostic services to 32% fewer gravidas at increased risk (with negative drift) than should be so managed. We address the problem outlined above and present the reasons for establishment of regional maternal serum alpha-fetoprotein screening programs operating at sufficient volume to permit the identification and control of assay drift.