The accuracy of predicting cardiac arrest by emergency medical services dispatchers: the calling party effect.

OBJECTIVES To analyze the accuracy of paramedic emergency medical services (EMS) dispatchers in predicting cardiac arrest and to assess the effect of the caller party on dispatcher accuracy in an advanced life support, public utility model EMS system, with greater than 90,000 calls and greater than 60,000 transports per year. METHODS This was a retrospective analysis from January 1, 2000, through June 30, 2000, of 911 calls with dispatcher-assigned presumptive patient condition (PPC) or field diagnosis of cardiac arrest. Sensitivity and positive predictive value (PPV) of the PPC code for cardiac arrest by calling parties were calculated. Homogeneity of sensitivity and PPV of the PPC code for cardiac arrest by calling parties was studied with chi-square analysis. Relevant proportions, relative risk ratios, and associated 95% confidence intervals (95% CIs) were calculated. Student's t-test was used to compare quality assurance scores between calling parties. RESULTS There were 506 patients included in the study. Overall sensitivity for dispatcher-assigned PPC of cardiac arrest was 68.3% (95% CI = 63.3% to 73.0%) with a PPV of 65.0% (95% CI = 60.0% to 69.7%). There was a significant difference in the PPV for the EMS dispatcher diagnosis of cardiac arrest depending on the type of caller (chi(2) = 17.34, p < 0.001). CONCLUSIONS A higher level of medical training may improve dispatch accuracy for predicting cardiac arrest. The type of calling party influenced the PPV of dispatcher-assigned condition.