Introduction Most Irish emergency departments (EDs) do not record diagnosis in a standardised way at the conclusion of each patient care episode. This lack of data hampers service planning and the transformation of ED funding from a block grant system to one based on activity.

Accordingly, this pilot study aimed to test whether clinician-assigned ED discharge diagnosis coding could be implemented in an Irish ED, and whether coded data would be of sufficient quality to support operational needs and the proposed model of activity-based funding (ABF).

Methods Following a training programme, discharge diagnosis coding of all ED attendances at the Midland Regional Hospital Tullamore was monitored over thirty-one weeks, using the ICD-10-AM ED short list as a data dictionary and the Integrated Patient Management System (iPMS) ED coding module as the data recording system. The number, spread and accuracy of coded patient care episodes and opinions of clinicians involved regarding system usability were evaluated.

Results 12,479 patient care episodes (61% of total ED visits) were assigned a discharge diagnosis. 67% of available codes were used. Use of less specific symptom codes, at 13.3%, was within international norms. An audit of coding accuracy found 97.1% accuracy of clinician-assigned codes.

Clinicians rated the usability of iPMS at 63.9 on the System Usability Scale (low-marginal acceptability). They reported insufficient codes in the data dictionary to fully capture clinical activity, but that the process was not too time consuming for daily use, nor one more appropriately performed by administrative staff.

Conclusion Discharge diagnoses can be assigned by emergency medicine clinicians at the point of patient discharge or referral using a standardised list of diagnoses and current information technology infrastructure. The results of the study support the proposed model of ABF for EDs, and will be used to inform further roll-out of discharge diagnosis coding.