An evidence-based approach to forensic assessments of single stab injuries to the trunk

This thesis aims to strengthen the scientific basis for forensic assessments of single stab injuries to the trunk, which have traditionally been based upon professional experience rather than scientifically evaluated evidence. Using nationwide registries maintained by the Swedish National Board of Forensic Medicine, we conducted five studies including both fatalities and survivors of sharp force trauma, with the objectives of characterising injury patterns related to the injury causation (assaults or self-inflicted), grading injury severity, and develop systematic approaches to provide evidence-based forensic evaluations.Most homicide victims were young males (odds ratio [OR] 4.2, 95 % confidence interval [CI] 1.7–10.4), typically found outdoors (OR 18.6, 95 % CI 6.6–52.2), being under the influence of alcohol (OR 7.1, 95 % CI 2.9–17.7) or illicit narcotics (OR 4.3, 95 % CI 1.5–11.9), and mostly transported to hospital before death (56.4 %). In contrast, suicides mainly died in their homes (79.5 %), and some cases due to only minor injuries (NISS ≤ 8) (8.9 %). Based on the findings associated with homicides and suicides, respectively, we developed the Manner of Death in Stab Injuries to the Trunk (MODSIT) score, an evidence-based tool that in four models showed good to excellent performance in distinguishing manner of death. Receiver Operating Characteristic (ROC) analysis demonstrated excellent discriminatory ability (Area Under the Curve [AUC] 0.99, 95 % CI 0.97–1.00) for MODSIT score model 3 (injury characteristics, alcohol exposure and scene circumstances) distinguishing homicide from suicide.To reduce the risk of classification bias inherent in research of fatalities, survived assaults verified by perpetrator confession or witness testimony were also analysed. Compared to homicides, survivors of assaults more often sustained stab wounds to the abdomen (OR 1.9, 95 % CI 1.1–3.1), left axillary region (OR 3.7, 95 % CI 1.5–8.8), or back (OR 3.7, 95 % CI 1.9–7.2), but less frequently to the ribcage (OR 0.2, 95 % CI 0.1–0.3).Finally, we demonstrated in a ROC analysis that the trauma scoring system New Injury Severity Score (NISS) provided excellent predictive performance (AUC 0.94, 95 % CI 0.92–0.96) discriminating fatal from non-fatal injuries in the absence of life-saving intervention, supporting its potential application as a systematic tool in forensic life-threat assessments.Together, these contributions provide characteristic findings and methodological tools that can complement expert opinion in forensic medicine. This thesis advances forensic medicine towards more consistent and evidence-based practice with the potential to strengthen fairness and reliability in the legal system.