Patterns of female genital-anal injury following consensual and non-consensual intercourse: a systematic review with exploratory likelihood ratio analysis

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Abstract This systematic review aimed to compare female genital-anal injury patterns following sexual assault (SA) and consensual sexual intercourse (SI), and to explore the feasibility of a likelihood ratio (LR) framework for location-specific injury interpretation. Web of Science, Scopus, Science Direct, and PubMed databases were searched following PRISMA guidelines. Studies reporting genital-anal injury prevalence in females following SI and/or SA were included. Location-specific injury probabilities were calculated for 15 anatomical sites, and exploratory LR values were derived with approximate 95% confidence intervals (CIs) and a sensitivity analysis for zero-cell handling. Eight studies comprising 4,350 females were included (SA: n = 3,332; SI: n = 1,018). Overall injury prevalence was significantly higher following SA (25.4%; 95% CI: 23.9–26.9%) compared to SI (2.5%; 95% CI: 1.7–3.7%). The posterior fourchette demonstrated the highest injury probability in both groups (SA: 8.0%; SI: 1.0%). Exploratory LR values ranged from 1.0 to 13.8, with the labia (LR = 13.8; 95% CI: 5.7–33.1), vagina (LR = 12.3; 95% CI: 4.6–32.7), and hymen (LR = 12.0; 95% CI: 5.0–28.8) showing the strongest discriminatory potential. Sensitivity analysis demonstrated that LR values for sites with zero SI injuries were highly sensitive to the continuity correction applied, ranging over an order of magnitude. While location-specific injury data demonstrate theoretical feasibility for LR-based forensic assessment, substantial methodological heterogeneity across studies and wide confidence intervals limit current applicability. Purpose-built national injury databases with standardised examination protocols are essential before LR approaches can be validated for forensic practice.