Background Relief of pain is a key activity of emergency medicine, however pain is often poorly monitored. As part of the evaluation of an electronic patient pain score recording device we needed to smooth a time series of pain scores in order to minimise ‘noise’ and false readings. In order to evaluate different methods we used the staff recorded pain score as a reference standard.
Patients used an electronic pain recording device containing buttons set out in the same way as a pain visual analog scale. The button corresponding to the currently level of pain was pressed in response to an audible prompt every 15 minutes. Pain scoring by ED staff continued in the normal way, recorded in the electronic health record.
Smoothing was undertaken using 20 and 30 minute bins, with either the median or the maximum patient recorded score being calculated for each bin. The staff recorded pain score nearest to 2 hours was paired with the patient recorded score in the same time window.
For each smoothing method a Bland-Altman plot was made of the paired results and Spearman’s correlation coefficient calculated.
There was little difference between the smoothing methods (table 1). However the most striking finding was that ED staff overall record a lower pain score (by 1 to 2 points) than is self-reported by patients. There were also very wide limits of agreement - implying variation between patient recorded and staff recorded pain scores (figure 1).
Better understanding of the dynamic of the interaction between patient and ED staff, including staff perceptions, is important in managing pain in emergency care. Our (perhaps naive) initial presumption that a staff recorded score forms a ‘reference standard’ may not be valid.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.