Co-designed training boosts doctor confidence in coding hospital-acquired adverse events

Ireland’s national hospital database may be missing a significant share of hospital-acquired adverse events, and a team of researchers argues that a communication gap between frontline doctors and clinical coders is a key reason why. In a study published in BMC Health Services Research, Rachel McCauley, Anna Connolly, Anne Matthews, Marcia Kirwan and colleagues describe the development and pilot testing of an educational intervention co-designed by the two groups whose interaction ultimately determines the quality of the country’s hospital data. The work focused on three adverse events that the evidence suggests are underreported in administrative records: delirium, pneumonia and urinary tract infections acquired during hospital stays.

The Hospital In-Patient Enquiry system, known as HIPE, is Ireland’s national administrative database recording activity in publicly funded acute hospitals. It is not a passive archive: the information it holds feeds into healthcare planning, funding allocation and patient safety monitoring. Every coded diagnosis in the system traces back to clinical documentation written by doctors at the bedside, which is then translated into standardised codes by administrative health data coders. The researchers point out that hospital-acquired adverse events can be underrepresented in such datasets for several reasons, including challenges in clinical documentation, uncertainty about how to identify these events, and limited communication between clinicians and the coders who process their notes.

To address this, the team employed a four-stage co-creation framework. The first stage was an evidence review, in which the researchers examined literature and policy documents to identify existing challenges and current education related to administrative data recording. The second stage involved data collection and stakeholder consultation. The third was co-production, in which the findings were used to develop educational resources collaboratively, and the fourth was prototyping, in which draft materials were piloted with new non-consultant hospital doctors and assessed for acceptability and perceived value.

The stakeholder consultation revealed how deep the mutual misunderstanding runs. Twelve non-consultant hospital doctors took part in three focus groups, with additional discussions held with consultants, while seventy-five administrative health data coders and managers completed a mixed-methods national survey, supplemented by consultation with policymakers. Both groups reported barriers to data production, and a central finding was a limited understanding of each other’s roles in the data production process. Doctors, in other words, did not fully grasp what coders needed from their documentation, and coders lacked visibility into the clinical realities shaping what doctors wrote. The researchers used these insights to develop targeted educational resources designed to tie in with existing clinical processes rather than requiring doctors to adopt entirely new workflows.

The pilot evaluation involved ten newly qualified non-consultant hospital doctors. Before the intervention, all participants reported low confidence in completing documentation to support coding of hospital-acquired adverse events. After working through the co-designed resources, all reported increased confidence, and the researchers recorded high acceptability of the materials among the group. The authors emphasise that this was a small pilot assessing perceived confidence and acceptability, not a trial measuring whether documentation quality or coding accuracy actually improved in routine practice.

Even so, the study demonstrates the feasibility and acceptability of a co-designed educational intervention aimed at improving patient documentation and the accurate recording of hospital-acquired adverse events. According to the researchers, the participation of both doctors and coders supported the co-creation of resources that addressed real-world documentation and coding challenges, rather than problems imagined by an education team working in isolation. This participatory approach, they suggest, is what allowed the materials to align with the practical constraints of clinical work.

The implications extend beyond the pilot group. Administrative datasets such as HIPE underpin decisions about where resources flow within the health system and how patient safety is monitored nationally. If conditions like hospital-acquired delirium, pneumonia and urinary tract infections are systematically underrepresented, then planning and safety signals built on those data may be correspondingly distorted. The authors note that while further evaluation across multiple sites is required, the findings suggest that targeted, co-produced educational tools may contribute to improvements in administrative data quality, enhanced patient safety monitoring and healthcare planning.

The study stops short of claiming clinical benefits, and the researchers are careful about the limits of what they can conclude. The pilot measured self-reported confidence and acceptability among a small group of newly qualified doctors at a single point in the intervention’s development. Whether the increased confidence translates into measurably better documentation, more complete coding of the three target conditions, or improved detection of hospital-acquired adverse events in the national dataset remains to be tested. The authors state that further evaluation across multiple sites is required before firmer claims can be made about the intervention’s effect on data quality at scale.

Subject of Research: A co-designed educational intervention to improve documentation and coding of hospital-acquired adverse events in Ireland's national hospital database

Article Title: Addressing barriers to accurate discharge data quality through a co-designed educational intervention

Article References: McCauley, R., Connolly, A., Matthews, A., & Kirwan, M. (2026). Addressing barriers to accurate discharge data quality through a co-designed educational intervention. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15475-1

Image Credits: AI Generated

DOI: 10.1186/s12913-026-15475-1

Keywords: hospital-acquired adverse events, HIPE, clinical coding, clinical documentation, health data quality, co-designed education, non-consultant hospital doctors, administrative health data, patient safety monitoring, Ireland, Addressing, barriers