To report the utilization of services offered and pregnancy outcomes for a unique statewide prenatal triple marker screening program and to present a cost-benefit analysis. A state population of 32 million with considerable ethnic and age distribution and with a wide variety of delivery systems providing prenatal care was considered. The entire pregnant population who appeared for care before 20 weeks gestation, approximately one-half million per year during the years of 1995 to 1997, was included in the study.
Methods
Mandatory offering of serum testing, using alpha-fetoprotein from 1986 to 1995, and the addition of human chorionic gonadotropin and unconjugated estriol in 1995, with systematic follow-up of serum screen positives with ultrasound and amniocentesis. This study collected and analyzed the program data and reports of outcomes and collected similar information from the birth defects registry.
Results
Triple marker serum screening was accepted by 67.4% of the women eligible and yielded an initial positive rate of 7.3%. More than 90% of the initially screen positive pregnancies were seen at a prenatal diagnostic center. After correction of gestational age, 71.3% had amniocentesis. The overall amniocentesis rate among women screened was 2.6%. The Program’s detection rate was predicted to be 85% for neural tube defects, and, based on Monte Carlo modeling, was theoretically calculated to be 62% for Down syndrome. In practice, detection rates were 75% for neural tube defects and 41% for Down syndrome due to lower than expected amniocentesis acceptance rate. Nevertheless, at a 5% discount rate, the screening program was cost beneficial at a ratio of 2.69:1. The cost per case detected was $35,365 and per case prevented was $110,741.
Conclusion
It is possible to implement a cost-effective population-based screening in compliance with quality standards in a diverse ethnic population with a variety of health-care providers. Triple marker screening in the second trimester is a cost beneficial program even if utilization of all services is less than ideal. Genetics in Medicine, 1999:1(5):200–207.
