713: Effect of changes in prenatal screening and diagnostic testing policies on resources utilization in a large, integrated health care system

Abstract

diagnostic testing policies on resources utilization in a large, integrated health care system Mary Norton, Carol Norem, Sanae Nakagawa, Steven Gregorich, Miriam Kuppermann Stanford University School of Medicine/Lucile Salter Packard Children’s Hospital, Obstetrics and Gynecology, Division of Maternal-Fetal Medicine, Stanford, CA, Kaiser Permanente, Northern California, Prenatal Genetic Screening Program, Oakland, CA, University of California, San Francisco, Department of Obstetrics, Gynecology, and Reproductive Sciences, San Francisco, CA, University of California, San Francisco, Medicine, San Francisco, CA, University of California, San Francisco, Department of Obstetrics, Gynecology & Reproductive Sciences, San Francisco, CA OBJECTIVE: Kaiser Permanente Northern California (KPNC) provides care to 35,000 pregnant women/year. In 2007, in line with American College of Obstetricians and Gynecologists (ACOG) guidelines, KPNC adopted a policy in which all women were offered prenatal screening, diagnostic testing, or no testing regardless of maternal age. Test options also changed, with the availability of first trimester screening in 2007 and integrated screening in 2009. We sought to determine the impact of these changes on resource utilization and testing outcomes. STUDY DESIGN: Data were obtained retrospectively from the KPNC cytogenetics lab and Prenatal Genetics Screening Program and used to determine rates of genetic screening for aneuploidy (first trimester screening, integrated screening and quad screening) and diagnostic testing (amnio and CVS) for the period 2006-2010. We also determined the absolute number of chromosome abnormalities detected, and the proportion of cytogenetic test results that were abnormal by maternal age. Dichotomous variables were compared using Chisquare or Fishers exact test. Rates of utilization, and testing outcomes, were compared for trend from the period before the policy implementation (2006) through 2010 using a linear repeated measures model. RESULTS: The number of deliveries decreased during this period, from 36278 to 34314, while the percentage of deliveries to women aged 35 years increased from 21.7% to 22.6% (p 0.004). The number of women who had screening increased from 73.9% to 75.4% (p35, the rate of screening increased from 51.6% in 2006 to 63.7% in 2010 (p35 who underwent diagnostic testing decreased from 40.9% to 26.8%. Among women aged 35, the rate of diagnostic testing decreased from 4.1% to 2.8% (p 0.0001). The percent abnormal diagnostic results increased from 5.9% to 8.2% (p 0.0002); the number of chromosome abnormalities identified overall was 7.2/1000 in 2006 and 6.7/1000 in 2010 (p 0.43). CONCLUSION: Adoption of ACOG guidelines and improvements in prenatal screening resulted in lower diagnostic test utilization with no decrease in detection of chromosome abnormalities. 714 A novel targeted sequencing approach improves noninvasive detection of chromosome ploidy in the first trimester Matthew Rabinowitz, George Gemelos, Milena Banjevic, Bernhard Zimmermann, Johan Baner, Brynn Levy, Matthew Hill Gene Security Network, Research and Development, Redwood City, CA, Gene Security Network, Statistics, Redwood City, CA, Columbia University Medical Center, Clinical Cytogenetics Laboratory, New York, NY OBJECTIVE: To improve noninvasive prenatal detection of fetal ploidy of chromosomes 18, 21, and X, particularly in samples with low fetal fraction, by using a targeted sequencing approach combined with parent genotypes and Hapmap data in a Bayesian Maximum Likelihood Estimation (MLE) algorithm. STUDY DESIGN: Maternal samples from four euploid and two trisomypositive pregnancies and respective paternal samples were obtained under an IRB-approved protocol from patients where fetal karyotype was known. Maternal cfDNA was extracted from plasma and roughly 10 million sequence reads were obtained following an enrichment procedure that targeted specific SNPs. Parent samples were similarly sequenced to obtain genotypes. RESULTS: The described algorithm correctly called chromosome 18 and 21 disomy for all euploid samples and normal chromosomes of aneuploid samples. Trisomy 18 and 21 calls were correct, as were chromosome X copy numbers in male and female fetuses. The confidence produced by the algorithm was in excess of 98% in all cases. CONCLUSION: The method described accurately reported the ploidy of all tested chromosomes from six samples, including samples comprised of less than 12% fetal DNA, which account for roughly 30% of first and early second trimester samples. The crucial difference between our MLE algorithm and published methods is that it leverages parent genotypes and Hapmap data to improve accuracy and generate a confidence metric. At low fetal fractions, all methods become less accurate and it is essential to correctly classify samples that do not contain sufficient fetal DNA to make a reliable call. Others have used chromosome Y probes to estimate fetal fraction of male fetuses, but concurrent parental genotyping enables estimation of fetal fraction for both sexes. Another inherent limitation of published methods using shotgun sequencing is that accuracy of ploidy calling varies among chromosomes due to differences in factors such as GC richness. Our targeted sequencing approach is largely independent of such chromosome-scale variations and yields more consistent performance between chromosomes. www.AJOG.org Academic Issues, Antepartum Fetal Assessment, Genetics, Hypertension, Medical-Surgical-Disease Poster Session V