Objective: We investigated the association between teammate familiarity and workplace injury in the emergency medical services (EMS) setting. Methods: From January 2011 to November 2013, we abstracted a mean of 29 months of shift records and Occupational Safety Health Administration injury logs from 14 EMS organisations with 37 total bases located in four US Census regions. Total teammate familiarity was calculated for each dyad as the total number of times a clinician dyad worked together over the study period. We used negative binomial regression to examine differences in injury incidence rate ratios (IRRs) by familiarity. Results: We analysed 715 826 shift records, representing 4197 EMS clinicians and 60 701 unique dyads. We determined the mean shifts per dyad was (5.9, SD 19.7), and quartiles of familiarity were 1 shift worked together over the study period, 2–3 shifts, 4–9 shifts and ≥10 shifts worked together. More than half of all dyads worked one shift together (53.9%, n=32 739), 24.8% of dyads 2–3 shifts, 11.8% worked 4–9 shifts and 9.6% worked ≥10 shifts. The overall incidence rate of injury across all organisations was 17.5 per 100 full-time equivalent (FTE), range 4.7–85.6 per 100 FTE. The raw injury rate was 33.5 per 100 FTEs for dyads with one shift of total familiarity, 14.2 for 2–3 shifts, 8.3 for 4–9 shifts and 0.3 for ≥10 shifts. Negative binomial regression confirmed that dyads with ≥10 shifts had the lowest incidence of injury (IRR 0.03; 95% CI 0.02 to 0.04). Conclusions: Familiarity between teammates varies in the EMS setting, and less familiarity is associated with greater incidence of workplace injury.