Automated calculation of ‘early warning scores’

Smith and Oakley [1] have described important problems with the use of a patient ‘track and trigger’ system, and the consequent errors that arise because of current methods of recording and charting raw data, and of deriving an early warning score (EWS). They describe considerable variability in the recording of patient vital signs, with only 65% of observations sets containing a full set of data (systolic blood pressure, respiratory rate, temperature and heart rate), echoing the findings of the recent cluster-randomised controlled trial of Medical Emergency Teams [2]. Smith and Oakley also observed the ‘rounding up or down’ of physiological values, the recording of EWS without documentation of associated raw data, and the incorrect calculation of the EWS, some of which was thought to be due to the design of the EWS weighting system. We believe that many of these problems can be minimised or solved by using computer systems to capture patients' raw physiology at the bedside and to calculate EWS automatically. It is known that the use of an EWS can drive improvements in the collection of data [3]. A system that routinely prompts staff to collect all routinely measured variables at the same visit would enhance this further. We are currently implementing a personal digital assistant (PDA) based system that collects all commonly recorded vital signs data and automatically calculates EWS [4]. This removes the need for healthcare staff to know the weightings for individual physiological parameters, thereby reducing the possibility of the ‘data manipulation’ alluded to by Smith and Oakley. All data is automatically stored electronically on the main server of the hospital using a wireless local area network (W-LAN), with raw physiology data, EWS, vital signs charts and oxygen therapy records being instantaneously available to any member of the hospital health care team via W-LAN or the hospital intranet. All vital signs charts are legible, accurate, up-to-date, timed and dated, with auditable tracking of data to the level of patient location and the person in-putting the data using electronic signatures. During the development of this system, we have demonstrated improved accuracy of EWS calculation (as errors can only be due only to the input of raw physiology), and a reduction in the time required by nurses to chart and calculate an EWS, compared with the traditional pen and paper method [5]. At the same time, we found further data (unpublished) that complement Smith and Oakley's assertion that EWS accuracy decreased with increasing patient illness. Our study [5] used five patient scenarios that were always given in the same order. Scenarios 1 and 2 were of equal patient severity of illness (to attempt to accommodate any learning effect) and were the least severe; scenarios 3, 4 and 5 were of progressively increasing patient severity of illness. We found that the time taken to complete the scenarios increased significantly with severity of illness (the complexity of the scenario) when the scenarios were completed using pen and paper. There was no such increase when the PDA was used to complete the scenarios. In this latter case, the participants were simply entering physiological data values in the PDA as opposed to having to derive the EWS score from the raw data when using pen and paper. Our findings, and those of Smith and Oakley, demonstrate the importance of accuracy in collecting and charting of vital signs, and of calculating EWS using ‘track and trigger’ systems. Inaccurate scoring can have important clinical consequences and implications for the validation of early warning scoring systems. We believe that the use of appropriate, clinically relevant information technology can improve these processes considerably. The PDA system referred to, VitalPACTM, is a collaborative development of The Learning Clinic Ltd and Portsmouth Hospitals NHS Trust.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.