Screening advances and diagnostic choice: The problem of residual risk

Objective: Over the past decade some authorities have suggested that advanced screening methodologies obviate the need for more invasive, diagnostic procedures. Data on Down syndrome (DS) births for Colorado from 1989 to 2005 were used to examine the implications of a decreasing use of amniocentesis. Methods: Publicly available, State of Colorado Department of Public Health data on DS birth rates for women were compared to amniocentesis use at Colorado's largest prenatal diagnostic center. Longitudinal changes on DS birth rates by maternal age (>35 and <35), and utilization of amniocentesis. Results: In Colorado, from 1989 to 2005, the rate for DS births for women 35+ rose considerably, while <35, rates remained stable (Cochran-Armitage test, p < 0.001). An autocorrelation-corrected test yielded a significant negative relationship between amniocentesis use (in 1,000s) and AMA DS rates (b = -11.30; p < 0.006; DW = 1.55). Confounding explanations involving sampling problems, socio-demographic factors, political conservatism and prevention orientation do not appear to account for these results. Conclusions: Replacement of definitive diagnosis with screening tests must be implemented with caution, particularly when using technologies with wide individual operator-dependent variability. Screening paradigms when performed with accuracy can markedly improve assessment of risks, but caution must be used in presenting negative screening results to women who still have a relatively high residual risk after a negative screen, and more generally in the displacement of technologies that provide definitive answers.