The study examined how age and sex influence injury severity in vehicle occupants involved in frontal and side collisions, using the Maximum Abbreviated Injury Scale (MAIS) as the outcome variable. Data were drawn from a large German crash database and filtered to include only front‑seat occupants who were belted, not involved in rollovers or underruns, and for whom the Estimated Energy Severity (EES) and MAIS were known. The frontal impact sample was reduced to 376 occupants, while the side impact sample contained 1,034 occupants after filtering.

MAIS scores were grouped into three categories—0 (uninjured), 1 (minor injury), and 2+ (moderate to severe injury)—to ensure sufficient numbers in each group. Exploratory analysis revealed that median age increased with injury severity (36 years for MAIS 0, 41 years for MAIS 1, and 50.5 years for MAIS 2+ in frontal impacts). In side impacts, the median age was 47 years for MAIS 0, 43 years for MAIS 1, and 45 years for MAIS 2+. Sex distribution showed a slightly higher proportion of uninjured males and a higher proportion of severe injuries among females in frontal collisions. In side collisions, the pattern was less clear, but drivers comprised about 80 % of occupants across all severity groups.

Key predictor variables considered for both frontal and side models included age, sex, principal direction of force (for frontal impacts), EES, seating position, collision partner type (car/LGV, narrow object, wide object, other/unknown), and whether the impact occurred on the occupant’s side of the vehicle (for side impacts). Height and weight were excluded due to missing data in 56–60 % of cases. An interaction term between age and sex was tested but did not improve model fit.

Model development followed an iterative likelihood‑ratio approach. Variables that did not significantly improve fit were removed, and the classification rate (percentage of occupants correctly assigned to the correct MAIS group) and pseudo‑R² were monitored throughout. The final side‑impact model retained age, sex, and collision partner type as significant predictors. Age had a small but significant effect on the odds of a moderate or severe injury (odds ratio 1.021 per year, p = 0.012). Male occupants were less likely to sustain a minor injury (odds ratio 0.426, p = 0.048) but showed no significant difference for severe injuries. Collision with a narrow object dramatically increased the odds of a severe injury (odds ratio 14.3, p < 0.001), whereas collisions with wide objects or other/unknown partners had weaker effects. The odds ratio for a minor injury associated with a narrow object was 0.099, indicating a protective effect against minor injuries but a strong risk for severe outcomes.

For frontal impacts, the final model also highlighted EES as a key driver of severity: median EES rose from 21 km/h for uninjured occupants to 43 km/h for those with MAIS 2+. The direction of force was most frequently 12 o’clock for severe injuries (78 % of MAIS 2+ cases). Collision partner type again influenced severity, with narrow and wide objects associated with higher rates of severe injuries.

Body‑region models were constructed using the same predictors as the overall MAIS models, allowing direct comparison of which variables were significant for head, thorax, abdomen, lower extremities, and other regions. These models were not built through variable selection but rather applied the final predictor set to each region’s MAIS score.

The report does not provide explicit information on project partners, roles, timeframe, or funding sources. The project is identified by the code PPR2048, indicating a German research initiative, but further collaboration details are not disclosed in the excerpt.