BACKGROUND: The aim of this study was to examine documentation quality in physician staffed emergency medical services (EMS). This study compared simulated on-site care with the associated patient records written by EMS physicians. METHODS: For this study two standardized simulated case scenarios, ST segment elevation myocardial infarction (STEMI) and major trauma with traumatic brain injury were designed by an expert committee. Overall 29 EMS teams each consisting of 1 EMS physician and 2 paramedics ran through the scenarios on high fidelity patient simulators and each scenario was videotaped. The scenarios were stopped after 12 min for STEMI and after 14 min for major trauma independent of the actions carried out and each EMS physician then had 10 min to document this initial phase on standardized protocol sheets. The videotaped scenarios were analyzed by two independent investigators. Documentation of predefined contents and all drug dosages were checked against the simulated on-site care. The data were evaluated and classified as correct, incorrect or missing documentation although action performed. RESULTS: Written consent for data analysis was provided by 28 teams. Overall 20 parameters and actions in the STEMI scenario and 16 in the major trauma scenario as well as all drug dosages were evaluated. For the scenario STEMI 469 actions were analyzed of which 271 (58%) were correct, 94 (20%) incorrect and 104 (22%) had missing documentation. A total of 140 medications were administered of which 31 (22%) were documented incorrectly and 14 (10%) were not documented. For major trauma 401 actions were analyzed of which 244 (61%) were correct, 101 (25%) incorrect and 56 (14%) had missing documentation. In this scenario the teams administered 138 medications of which 31 (22%) were documented incorrectly and 16 (12%) were not documented. Infused amounts of crystalloids and colloids were mostly documented correctly in this case (35 correct /6 incorrect/8 not documented). Documentation of several clinical parameters was carried out predominantly correctly, e.g. initial blood pressure (STEMI: 25/2/1, major trauma: 21/4/2) and initial ECG rhythm (STEMI: 27/0/1, major trauma: 26/0/1). Documentation of other clinically relevant parameters was often performed incorrectly: 12-lead ECG in STEMI (5/9/12) and capnometry in major trauma (9/4/7). No team used a pain scale to assess the level of pain in the STEMI scenario but 12 EMS physicians documented an accordant value (numerical rating scale) on the patient records. Furthermore some parameters could be identified where documentation was mostly missing although they were measured, e.g. onset of symptoms in STEMI (5/4/15) and reduced level of consciousness and bradypnea in major trauma (9/2/17). CONCLUSION: Patient safety can be reduced if relevant preclinical data are not transmitted correctly to the admitting hospital. Therefore there is a need to improve documentation quality in EMS. Electronic documentation, training of EMS staff and quality management programs might offer solutions. Because of the small sample size further studies are needed to evaluate the validity of these results.