The Leicestershire Perinatal Mortality Study: a case study of multi-group discriminant analysis with complex sampling

Area of Science:

• Perinatal mortality research
• Epidemiological study design
• Statistical analysis in healthcare

Background:

• Traditional case-control studies often use two-group discriminant analysis.
• Perinatal deaths are heterogeneous, with risk factors potentially specific to subgroups.
• Non-random sampling of controls complicates analysis in perinatal mortality studies.

Purpose of the Study:

• To propose a novel seven-category classification for perinatal deaths.
• To apply multi-group discriminant analysis to heterogeneous perinatal mortality data.
• To address challenges posed by non-random control sampling in case-control studies.

Main Methods:

• Development of a seven-category classification system for perinatal deaths.
• Application of multi-group discriminant analysis for analyzing case-control data.
• Consideration of non-random sampling issues in control group selection.

Main Results:

• The proposed classification allows for a more nuanced understanding of perinatal death subgroups.
• Multi-group discriminant analysis provides a more effective analytical approach for heterogeneous data.
• The study highlights the impact of control sampling on the validity of findings.

Conclusions:

• A seven-category classification of perinatal deaths enhances the analysis of risk factors.
• Multi-group discriminant analysis is a suitable method for complex perinatal mortality data.
• Addressing control selection bias is crucial for accurate case-control study outcomes.