Retrospective evaluation of correlation and agreement between two recovery scoring systems in horses

Recovery is one of the most important and risky phases in equine anaesthesia.1Several recovery scoring systems (RSS) have been developed to evaluate risk factors associated with morbidity and mortality during this perioperative phase. A descriptive scale2 and a composite rating scale3 are commonly used in equine practice and have been reported as reliable methods for qualifying recovery from general anaesthesia in horses4; however, a truly validated and universally used RSS is not available at present, complicating the comparison between different studies in which different scales are used. The aim of this study was to add information to the existing knowledge about recovery in horses, evaluating correlation and agreement between the two RSS used at the authors’ institution. Anaesthetic records of horses undergoing general anaesthesia between January 2013 and November 2015 were retrospectively reviewed. During this period, two types of recovery scores were used: Y, modified from Young and Taylor,2 and D, modified from Donaldson and others.3 System Y (Table 1) is a simple descriptive scale that employs a scale from 1 to 5 to score recovery quality in its entirety; it assigns several descriptors to each score, and high numerical values indicate ‘worst’ recovery. System D (Table 2) is a composite scoring system in which eight phases of the recovery are scored and then the values are summated to obtain a total score that is then matched to an overall descriptive score developed to rank the recovery on a scale from 1 to 5 according to the quality of recovery, with the highest number of the descriptive score indicating the worst recovery, similar to system Y.5 All horses included in the study had unassisted recoveries. For each horse, recoveries were scored with both systems by the same person, whose identity was not specified upon the anaesthetic record but that was either a diplomate of the European or American College of Veterinary Anaesthesia and Analgesia, a resident in Veterinary Anaesthesia and Analgesia, or a final-year student. Horses were included in the study if data about American Society of Anaesthesiologists (ASA) grade,6 type of surgery (emergency or elective) and both recovery scores were reported in the anaesthetic records. Demographic data of the horses were also recorded. Data were analysed using commercially available statistical software (SPSS, V.22.0, 2013). Correlation and agreement were evaluated with Spearman’s coefficient and k-weighted, respectively. These analyses were performed in all the cases grouped together and after subdividing the horses into groups based on ASA category and type of procedure. 10.1136/vr.104546