Can primary care patients be identified within an emergency department workload?

To quantify the proportion of emergency department (ED) discharges that could have been managed in primary care, and to determine the consistency with which healthcare professionals assess cases as 'primary care appropriate'.

1200 case notes from 30,493 Wellington Hospital ED discharges over a 12-month period were randomly selected. 180 were assessed by an expert panel of 12 healthcare professionals comprising GPs, ED specialists, and nurses. Level of agreement was measured between professional groups. The panel adapted an international clinical protocol on ED appropriateness. The remaining 1020 cases were assessed according to these adapted criteria.

The panel considered 37% of all ED presentations (49.7% of those not admitted) 'primary care appropriate,' and 50% of all ED presentations (68.1% of those not admitted) 'primary care appropriate' if laboratory and radiological facilities were immediately available. These figures were similar when the full 1200 cases were reviewed by researchers. There was poor to moderate agreement between individuals and professional groups about which were appropriate (Kappas 0.35-0.45). In 15% of cases, professionals gave a different response to the same case on different occasions.

Our study has shown retrospectively that a significant number of ED presentations may have been managed in primary care. However this was determined with knowledge of investigation results and final diagnosis, by clinicians who had poor agreement about individual cases. In reality, triage nurses in ED have only the presenting complaint. This study shows how variable clinicians can be in determining primary care appropriateness even with full clinical details.