Response time performance is related to increased survival for a relatively small group of patients with critical emergencies. Effectively utilizing current resources is a challenge for all emergency medical services (EMS) systems for reasons of cost-effectiveness and safety.

The objective of this study was to identify opportunities for improving ambulance response-time performance in an urban EMS system using fixed deployment.

This was a qualitative and quantitative case study which consisted of structured interviews with policy makers, managers, and workers in a fire department EMS division, as well as analysis of dispatch data and observation of dispatch operations.

The current computer-aided dispatch (CAD) system does not identify the closest ambulance to the emergency, and therefore, dispatchers must guess which unit is closer when units are not within their stations or “first due” areas. There is no means to track how often dispatchers guess correctly or how often the closest ambulance actually is dispatched to the emergency.

Temporal and geographic patterns were identified. Opportunities also were identified to improve response time performance through the use of dynamic deployment and peak-load staffing.

The results suggest that there were opportunities for improving ambulance response times by implementing strategies such as peak-load staffing and dynamic deployment. However, the most important improvement would be the implementation of a policy to send the closest ambulance to the emergency. More research is needed to identify how prevalent the failure to send the closest ambulance is within EMS systems that use fixeddeployment response strategies and computer-aided dispatch systems that are incapable of tracking unit locations outside of their stations.

Paramedics accurately estimate the closest trauma hospital for ground transport.

Ground ambulance scene transports of trauma system patients to six participating trauma hospitals in Multnomah County, Oregon from 1 January 1986 to 1 January 1987 were studied. Transports involving multiple patients or pediatric patients were excluded.

A retrospective analysis was performed on consecutive patient transports to be taken to the closest trauma hospital as required by protocol. The availability of each hospital to receive trauma patients was monitored continuously by a central communications facility. Paramedics were provided hospital availability data at the time of patient system entry. When several hospitals were available, the paramedics were required by protocol to select the “closest” hospital. Subsequently, the vector distance from the trauma site to each of the available hospitals was measured using a grid map. This method was validated by odometer measurement (r2 = 0.924). Chisquare analysis was used to analyze hospital bypasses to specific hospitals.

Of the 1193 eligible patients entered into the trauma system, 160 (13%; 95% CI = 11–15%) transports bypassed the closest available hospital for a receiving hospital ≥1 mile more distant. There were 11 (1%; 0–2%) patients transported to a hospital more than five miles more distant. Of the 132 patients with a trauma score (TS) <12, 15 (11%; 6–18%) were taken to a hospital one mile or further beyond the closest hospital. None (0%; 0–2%) were transported more than five miles past the closest hospital. Of the six hospitals, three were bypassed more than one mile significantly more often then they received bypass patients. One hospital received such patients four times more than it was bypassed (p <.001).

While paramedics generally can identify the closest hospital for trauma patient transport, some systematic hospital bypass errors occur. If a community wants assurance of an equitable patient distribution among participating trauma hospitals and assignment of the closest geographic hospital for injured patients, then map vector distance determination to identify the closest available hospital should supplement paramedic dispatching.