Background: Patient explanation is an important factor for improving patient safety and satisfaction. However, evaluating patient explanations presents practical challenges.
Objectives: The purpose of this study was to develop methods to evaluate patient explanations using clinical documentation data from electronic medical records (EMRs) provided by healthcare professionals.
Methods: Nursing records were selected for analysis because they are the most detailed EMRs amongst those provided by all healthcare professionals. Therefore, electronic text data from nursing records of an acute hospital dated May 2015 to November 2015 were collected from an EMR server. All 2365 records were text-mined and 2620 records were found to include patient explanations. Records were broken into 16 348 sentences and 246 940 words. Words were parsed by word class and analysed using correspondence analysis methods.
Results: Of the 2620 records, 529 (20.2%), 513 (19.5%), 1121 (42.8%), and 767 (29.7%) records contained explanation documentation regarding the patient’s condition, admission/discharge, medical procedure(s) (e.g. injection and operation), and patient reaction to the explanation (e.g. understanding and anxiousness), respectively. In addition, 36.8% of explanation documentation on the patient condition accompanied patient reaction to the explanation, 40.2% on medical admission/discharge, and 35.0% regarding medical procedure(s). Results of correspondence analysis indicated that documented explanations of the patient condition tend to include the words 'family' and 'talk', admission/discharge explanations tend to include the words 'outpatient' and 'consultation', and medical procedure explanations tend to include the words 'question' and 'anxiousness'.
Conclusions: This study suggests that documentation data from EMR provides an opportunity for quantitative analysis of the quality of patient explanations. Although this study was limited to a case study, the analysis methods in this study can be applied to all hospitals that use EMR.