Preliminary evaluation of subjective scoring systems for assessment of postoperative pain in horses

This study examined the performance of two subjective pain scoring systems for evaluating equine postoperative pain, and investigated differences in pain scoring tendencies of veterinarians and grooms. Fifteen horses were included in the study. Group 1 (n = 8) had chronic lameness and was admitted for elective arthroscopy under general anaesthesia, on one or two femoropatellar, femorotibial or tibiotarsal joint or digital flexor tendons. The anaesthetic protocol for each horse was similar but not standardized. Multi‐modal peri‐operative analgesia comprised: romifidine (100 µg kg^?1 IV); flunixin (1.1 mg kg^?1 IV); ketamine (2.2 mg kg^?1 IV); morphine (0.12 mg kg^?1 IV); phenylbutazone (4 mg kg^?1 IV/PO). Group 2 (n = 7) included pain free controls. At 6 hours post‐recovery from anaesthesia (PR) (group 1) or at 20.00 hours (group 2 with one limb bandaged), horses were filmed undisturbed in their stables for 90 seconds (dynamic behaviour, DB); thereafter, the surgery site and pharynx of each horse were palpated (and filmed) in a standardized manner (interactive behaviour, IB). Two observer groups, seven veterinarians and eight grooms, watched video footage of each horse and assigned pain scores using a visual analog scale (VAS) and a numerical rating scale (NRS). Observers assigned a pain score (VAS and NRS) for DB and IB separately and overall. Statistical analysis (Minitab 13.0, Wilcoxon signed rank and Mann–Whitney U‐tests) investigated differences in pain scores attributed to groups 1 and 2 horses, compared pain scores assigned by veterinarians and grooms, and examined differences in the performance of VAS and NRS techniques. There were significant differences in the pain scores assigned by veterinarians and grooms to groups 1 and 2 horses. When using DB or IB separately (but not combined) to score perceived pain, grooms assigned higher scores to group 1 than group 2 (U = 81.5, p < 0.05; U = 82.0, p < 0.05) using the VAS. There was no difference in NRS scores attributed by grooms to groups 1 and 2. Using DB and IB separately or combined, there was no difference in pain scores attributed to groups 1 and 2 by veterinarians using either VAS or NRS scoring systems. Using separate VAS scores for DB (W = 32.5, p < 0.05) and IB (W = 26.5, p < 0.05) and using combined (DB + IB) VAS scores, grooms awarded higher pain scores (W = 27.0, p < 0.05) than veterinarians to group 1. Using the NRS, vets and grooms did not score pain differently for group 1. For group 2, grooms scored pain significantly higher than vets when using the VAS to score IB separately (W = 21.0, p < 0.05); no other differences between grooms and veterinarians in pain scoring of group 2 (NRS or VAS, DB and IB separately or combined) were identified. The performance of subjective pain scoring systems for assessment of equine postoperative pain varies according to the scale used, the behaviour evaluated (dynamic or interactive) and the observer group. While data suggest that grooms distinguished post‐surgery horses from controls more successfully than vets and assigned higher pain scores to these horses, the specific behavioral criteria on which scores were assigned requires future investigation and identification.