Objectives. This study sought expert consensus about which categories of patients from 248 Medical Priority (MPDS) ambulance dispatch codes might be appropriate for a nonemergency response or for whom dispatch of an ambulance might be appropriately denied if the patient were referred to a more suitable health care provider. Methods. A Delphi technique was used. Ten physicians, from the specialities of emergency medicine, general practice, and pre-hospital care formed the expert panel but were blinded to each other's identity. Participants received a written description of the operation of the MPDS and the Delphi technique and voted independently by mail. Results. Using majority voting, 54 dispatch codes (22%) were recommended for a nonemergency response/referral. This equates to 12.44% of annual emergency calls in a typical UK ambulance service (n = 9,021; 95% confidence interval, 12.21 to 12.69%). The kappa statistic (chance-corrected proportional agreement) between members of the expert panel was 0.62 (substantial). Conclusions. The recommended dispatch codes for non-emergency response or referral represent a significant proportion of emergency ambulance calls. Theoretically, the implementation of nonemergency responses could have the benefit of reducing accidents involving emergency ambulances and could lead to improved response times for critically ill patients by freeing up resources. It could also support the targeting of patients to appropriate health care providers on first contact with the health service. However, given the poor reliability of expert opinion, further research using clinical outcome data is required to validate the recommendations made in this article before changing existing ambulance response systems.