To date, efforts to improve occupant protection in side impact crashes have concentrated on reducing the injuries to occupants seated on the struck side of the vehicle arising from contact with the intruding side structure and/or external objects. Crash investigations indicate that occupants on the struck side of a vehicle may also be injured by contact with an adjacent occupant in the same seating row. Anecdotal information suggests that the injury consequences of occupant-to-occupant impacts can be severe, and sometimes life threatening. Occupant-to-occupant impacts leave little evidence in the vehicle, and hence these impacts can be difficult for crash investigators to detect and may be underreported. The objective of this study was to evaluate the risk of impact injury from occupant-to-occupant impacts in side impact vehicle crashes. The study examined 9608 crashes extracted from NASS/CDS 1993--2006 to investigate the risk of occupant-to-occupant impacts. The study computed relative risk ratio of serious injury (MAIS 3 or greater) for drivers with and without an adjacent front seat passenger present. This approach avoids uncertainties in the coding of occupant-to-occupant contact. The NASS data showed an 8% increased injury risk for struck side drivers in cases where a belted front seat passenger was present. If the front seat passenger was unbelted, struck side driver injury risk was found to be 30% higher than for struck side drivers without a front seat passenger. A series of 6 full scale vehicle side impact crash tests, both mobile deformable barrier to vehicle and vehicle-to-pole, were conducted to assess injury risk and determine the occupant kinematics which lead to occupant-to-occupant impact. Limitations of the biofidelity of current ATDs to simulate occupant interaction were noted. Occupant interaction indicating risk of serious head injury to both the driver and front seat passenger was observed in vehicle-to-pole side impact. The results show that despite the introduction of countermeasures to protect struck side occupants from contact with intruding structure or external objects, these occupants may be severely injured by impacting adjacent occupants. The feasibility of a potential countermeasure, developed to offer protection for two adjacent occupants as well as a single occupant seated on the non-struck side, was investigated through analysis of the dummy injury responses produced in pole side impact tests, with and without the countermeasure installed. The countermeasure was observed to reduce the risk of head injury from occupant interaction.