Current literature, whilst limited, highlights inconsistencies in the efficacy of vehicle torso side airbags (tSAB). The objective of this study is to further investigate this using data from the German In‐depth Accident Study (GIDAS). Collisions involving vehicle–vehicle/light‐truck interaction, whereby an injured occupant is seated nearside and in the front row, was used. From this, a matched cohort study is undertaken, in which collisions with a torso side airbag (exposure) deployed are paired with similar collisions without a tSAB (non‐exposure) on the basis of;​ occupant, vehicle and collision characteristics. This pairing process is automated by a Genetic Matching algorithm that calculates a matrix of optimal weighting values for the selected matching criteria and is capable of returning multiple non‐exposure collisions per exposure collision. A conditional logistic regression then outputs the Odds Ratios for a given injury severity. Results show that for collisions with a deployed tSAB, the risk of overall thoracic injury was not reduced (MAIS1+, OR=1.08, 95%CI [0.65‐1.82]), however a non‐significant reduced risk for rib fracture occurred (MAIS1+, RR=0.67, 95%CI [0.29‐1.54]). When stratified by change of velocity, a strong protective effect for the thoracic region occurred for dV range similar to Euro N‐CAP testing (MAIS1+, RR=0.48, 95%CI [0.18‐1.27]), however an increased risk of injury occurred for greater speeds (MAIS1+, RR=1.63, 95%CI [0.59‐4.4]). Additionally the occupant may be more susceptible to light thoracic injury at mid‐low impact speeds (MAIS1+, RR=2.48, 95%CI [0.93‐6.16]).