A comparison of four systems for scoring recovery quality after general anaesthesia in horses.

Abstract

REASON FOR PERFORMING STUDY: The recovery quality scoring systems (RQSSs) in current use have not been critically reviewed for reliability. OBJECTIVE: To examine reliability (reproducibility) of 4 RQSSs when applied to a ranked series. METHODS: A DVD incorporating the recordings of 9 horses recovering from general anaesthesia was evaluated by final year students over 5 days. On Day 1, each evaluator ranked recoveries from 1-9 (1 = best). Over the following 4 days, each evaluator scored the same recoveries using 4 different RQSSs (3 of them in common usage and previously published) applied in random order. The scores from each RQSS were ranked and plotted against the Day 1 ranking of each evaluator to establish the extent of agreement using generalisability theory. The same 9 recoveries were also ranked by 12 experienced equine anaesthetists and the Spearman Rank Correlation coefficient calculated to determine the agreement between experienced and inexperienced evaluators. RESULTS: The recoveries were evaluated by 117 students. All 4 RQSSs were equally reliable with low (<4%) interobserver variability. The main (80%) source of total variation arose from differences between horses. The overall ranking within each RQSS was strongly correlated with Day 1 ranking. There was strong correlation (r = 0.983) between the students' ranking and that established by experienced anaesthetists. Interobserver reliability was similar with all 4 RQSSs. CONCLUSION: All 4 RQSSs studied were similarly reliable. POTENTIAL RELEVANCE: The selection of a universally acceptable RQSS from amongst the 4 examined can be based on criteria other than reliability, e.g. ease of use. This will facilitate wider scale multi-centre studies in recovery quality after anaesthesia in horses.