Article Text

61 Accuracy of euroscore calculation on dendrite
  1. M Koh,
  2. R Jeganathan,
  3. A Graham
  1. Royal Victoria Hospital, Belfast, UK


Introduction EuroSCORE is the most widely used cardiac surgical risk model in assessing individual patient risk and institutional outcome. In compliance with the Society for Cardiothoracic Surgery, we have used on-line Dendrite Clinical Systems for data collection and EuroSCORE calculation (Figure 1). However accuracy of data entry is not ascertained and subjective nature of some variables in EuroSCORE poses potential risk of observer variation.

Objective We conducted an audit study to evaluate accuracy of EuroSCORE calculation on Dendrite and to identify factors causing discrepancy in EuroSCORE.

Method A retrospective assessment was carried out on data entry and calculated EuroSCORE on Dendrite over two discrete audit periods (consecutive 25 cases in October 2014 and 39 cases in December 2014). The team was encouraged to be aware of importance of accurate data entry in-between the two periods.

Logistic EuroSCORE and EuroSCORE II were calculated independently by a single observer and compared with the scores recorded on Dendrite.

Results Incidence of discrepancy in Logistic EuroSCORE and EuroSCORE II was significantly decreased from 28.0% in the first period to 17.9% in the second period and 64.0% to 25.6% respectively. Also significant reduction in the number of variants with discrepancy per case was observed from 1.04 to 0.36. Four factors were identified as cause of discrepancy: simple mistakes, misapplication of EuroSCORE definition, observer variation and system error of Dendrite. Incidence of simple mistakes and misapplication of definition was significantly decreased over the audit periods from 36% to 7.7% and 36% to 2.9% respectively, which largely contributed to the improved accuracy of EuroSCORE. Observer variation was observed in NYHA classification and poor mobility. Dendrite was noted to have system errors in EuroSCORE calculation related to three variants; pulmonary hypertension, weight of intervention and surgery on thoracic aorta. This problem caused discrepancy in seven cases of the entire study cohort (10.9%).

Abstract 61 Figure 1

Comparison of Dendrite Scores with Logistic EuroSCORE & EuroSCORE II

Conclusions Discrepancy in EuroSCORE was common in the first audit period, mainly due to simple mistakes and misapplication of definition; however, quality of data entry was significantly improved over the audit periods by encouraging the awareness of the team.

Observer variation is inevitable in NYHA and poor mobility due to subjective nature of these variables, although its influence on the entire score was rather small. Score calculator on Dendrite has several critical system errors; hence not reliable particularly in complicated cases with multiple risk factors including pulmonary hypertension, aortic surgery and multiple procedures. These discrepancies may have major implications for national reporting of surgeon specific mortality.

Abstract 61 Figure 2

Factors causing discrepancy in each variable

Statistics from

Request permissions

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.