Estimating birth prevalence of Down's syndrome

Area of Science:

• Genetics and Reproductive Health
• Biostatistics
• Epidemiology

Background:

• Accurate maternal age-specific Down's syndrome prevalence data are essential for assessing environmental factors and guiding prenatal screening programs.
• Estimates should reflect populations without prenatal screening, typically using pre-screening era data.
• Under-ascertainment in existing datasets is a recognized issue not adequately addressed by current statistical models for live-birth prevalence.

Purpose of the Study:

• To develop an improved statistical model for estimating live-birth prevalence of Down's syndrome.
• To explicitly incorporate under-ascertainment into the prevalence estimation model.
• To analyze ascertainment rates across nine published studies.

Main Methods:

• Development of a novel statistical model accounting for under-ascertainment.
• Application of the model to data from nine previously published studies on Down's syndrome prevalence.
• Examination of ascertainment rates within the included studies.

Main Results:

• The developed model demonstrated a good fit for eight out of nine studies.
• The weighted estimate of Down's syndrome risk across maternal age distribution was 1.41 per 1000 live births (90% CI: 1.37-1.49).
• Excluding the one poorly fitting study did not significantly alter the estimated risks.

Conclusions:

• The proposed model predicts approximately 10% higher rates compared to models assuming complete ascertainment.
• These predicted rates align closely with estimates derived from studies known for high ascertainment levels.
• The findings suggest the model offers a more accurate representation of Down's syndrome prevalence by addressing under-ascertainment.