Indigenization of the median of markers for Down syndrome screening based on statistical analysis of medical big data

Abstract

Objective: To indigenize the median of Down syndrome (DS) screening markers for first and second trimester, and compare the impact of the indigenized and built-in median data on the efficiency of DS screening. Materials and Methods: Data derived from first and Second-trimester screening (FTS and STS) for DS, composed of selected pregnancies deemed to be normal, were examined in a retrospective study. Indigenization regression analysis was calculated by using five models to fit statistical the raw data. Multiple of median (MoM) values estimated by using indigenized medians were compared with those calculated by using built-in. Results: This study established a regression equation which is more suitable for the median of each screening marker in the local pregnant women. The changes of median MoM of screening markers were statistically significant after indigenization. For FTS, the detection rate was 100% when the false positive rate was 5%, and the cut-off value was 1/262. On the other hand, for STS, the detection rate of the model with indigenized parameters was 77.42%, which is 16.13% higher than that of built-in parameters. Conclusion: For the individual specific risk of pregnancy, when the indigenized parameters was used to calculate, is more accurately and screening effectiveness has been improved. This is a great reference significance for the current prenatal screening whether indigenized data should be used.

Keywords

• Down syndrome
• First-trimester screening
• Second-trimester screening
• Medical big data
• Detection rate