Validating obstetric triage systems, what are we really measuring - A modified Delphi process introducing outcome measures for obstetric emergency triage systems.

BACKGROUND

Obstetric emergency triage is more complex than general emergency triage, since the pregnant woman, the fetus and labour status all must be assessed. It is a relatively new branch of triage and is not an integrated part of obstetric emergency care in Sweden. As in general emergency triage, there is no definition of true acuity for obstetric emergency patients. This makes validation of triage systems difficult and results in unclear capacity to identify patients requiring urgent attention. Predominately applied surrogate outcome measures do not reflect acuity at the time of triage and are often affected by organisational factors. The study aims to develop a set of weighted surrogate outcome measures representing acuity at the time of triage, enabling construct validation of obstetric triage systems.

METHODS

A four-round modified Delphi process was performed at a single tertiary obstetrics department. Seven obstetricians and three midwives participated in round 1, while only obstetricians participated in rounds 2-4 based on the profession's competence. The consensus level for rounds 2-4 was predefined at 100%.

RESULTS

A set of 31 immediate obstetrician-initiated interventions at the emergency department, for a patient presenting with an urgent condition, were defined. The interventions reflect acuity level at the time of triage and with minimum interference or influence by context. The outcomes were weighted at three levels, stratifying urgency in the most severe presentations of these conditions.

CONCLUSION

As true acuity in a patient seeking emergency care has not been defined, outcome measures reflecting true acuity at the time of triage should be applied when validating triage systems. Previous studies on validity in obstetric triage systems are scarce and inconclusive regarding internal and external validity. The outcome measures developed in this study may serve as a template for validating obstetric triage systems implemented in similar contexts.