Developing and testing inter-rater reliability of a data collection tool for patient health records on end-of-life care of neurological patients in an acute hospital ward.
Gudrun Jonsdottir, Erna Haraldsdottir, Valgerdur Sigurdardottir and 4 others
PMID 37141442WHAT IT FOUND
A tool designed to extract end-of-life care data from neurological patients' records showed high reliability.
Abstractors agreed on 89% of items overall. Agreement was strongest for treatments and weakest for interpreting clinical signs from free text.
Key findings
01The overall percentage agreement between abstractors was 89%, with a Fleiss' kappa coefficient of 0.84, indicating substantial reliability.
02Reliability varied by category; 'Medically invasive and other treatments' had the highest agreement (95%), while 'Health professionals' contribution' had the lowest (83%).
03Items requiring interpretation of free text, such as 'Alertness/responsiveness 48 h before death', showed poor consistency with a kappa of 0.24 despite 83% percentage agreement.
STILL TO COME
How it was doneWhat they foundWhat it means for RNs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study only tested 32 of the 76 items in the NEOLCAT. Items related to communication with patients/families and spirituality were excluded from reliability testing because they are poorly documented in existing records. The records were from a single acute neurological unit in Iceland, which may limit generalizability to other healthcare systems. Data abstraction was manual and time-consuming (approx. 60 minutes per patient), which may not reflect automated extraction capabilities.
Declared interests
The authors declare no conflicts of interest. Funding was provided by the Icelandic Nurses' Association, the University of Iceland, and Landspitali National University Hospital. The funding sources were not involved in data analysis or interpretation.
The easy way to misread this
Do not assume the tool is ready for immediate clinical use to guide patient care. It was developed for research data extraction, not for bedside assessment or decision-making.