Psychometric Assessment of the Rat Grimace Scale and Development of an Analgesic Intervention Score

et al. (2014) Psychometric Assessment of the Rat Grimace Scale and Development of an Analgesic Intervention Score. PLoS ONE 9(5): e97882. doi:10.1371/journal.pone.0097882 Psychometric Assessment of the Rat Grimace Scale and Development of an Analgesic Intervention Score Vanessa Oliver 0 Debbie De Rantere 0 Rheanne Ritchie 0 Jessica Chisholm 0 Kent G. Hecker 0 Daniel S. J. Pang 0 Cheryl McCormick, Brock University, Canada 0 1 Department of Veterinary Clinical and Diagnostic Sciences, Faculty of Veterinary Medicine, University of Calgary , Calgary, AB , Canada , 2 Department of Zoology, Faculty of Biological Sciences, University of Calgary , Calgary, AB , Canada , 3 Hotchkiss Brain Institute, University of Calgary , Calgary, AB , Canada Our limited ability to assess spontaneous pain in rodent models of painful human conditions may be associated with a translational failure of promising analgesic compounds in to clinical use. If measurement of spontaneous pain behaviours can be used to generate an analgesic intervention score their use could expand to guide the use of analgesics, as mandated by regulatory bodies and ethical and welfare obligations. One such measure of spontaneous pain, the Rat Grimace Scale (RGS), has recently been described and shown to exhibit reliability. However, reliability of measurement scores is context and content specific, and further testing required to assess translation to a heterogenous setting (different model, raters, environment). The objectives of this study were to perform reliability testing with the Rat Grimace Scale in a heterogenous setting and generate an analgesic intervention score for its use. In a randomised, blinded study, sixteen adult female rats received one of three analgesia treatments (0.05 mg/kg buprenorphine subcutaneously, 1 mg/kg meloxicam subcutaneously, 0.2 mg/kg oral buprenorphine in jelly) peri-operatively (telemetry unit implantation surgery). Rats were videorecorded (before, 1-6 and 12 hours post-operatively) and images collected for independent scoring by three blinded raters using the RGS, and five experts based on ''pain/no pain'' assessment. Scores were used to calculate inter- and intra-rater reliability with an intraclass correlation coefficient and generate an analgesic intervention score with receiver operating characteristic curve analysis. The RGS scores showed very good inter- and intra-rater reliability (0.85 [0.78-0.90 95% CI] and 0.83 [0.76-0.89], respectively). An analgesic intervention threshold of greater than 0.67 was determined. These data demonstrate that the RGS is a useful tool which can be successfully employed in a heterogenous setting, and has the potential to guide analgesic intervention. - Funding: Daniel Pang was supported by an NSERC Discovery Grant. Vanessa Oliver was supported by the University of Calgary Faculty of Veterinary Medicine as part of the Investigative Medicine student rotation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. The recognition and assessment of spontaneous behaviours associated with pain in laboratory species has been identified as an area requiring further investigation in biomedical research and veterinary medicine [15]. This need has emerged from the poor predictive ability of animal models of pain [1,3], the ethical and regulatory obligations of providing appropriate analgesia [69] and the fundamental requirement to understand and assess the efficacy of analgesic agents in animals [2,10]. Until recently, pain scales used in animals, many of which have been adopted from medicine have not undergone rigorous assessment for their reliability and validity [1114]. As appreciation for the importance of scale assessment for validity and reliability using psychometric methods grows, pain scales are being developed with these principles in multiple species including dogs, cats, mice and rats [4,5,1419]. Most recently, based on apparent evolutionary conservation of facial expressions and successful employment of facial action coding in non-verbal humans, facial expression scoring has been applied in rats, mice and rabbits [15,16,20,21]. The practical utility and widespread adoption of these scales require evidence of the validity (does a scale measure what it claims to measure) and reliability (measurement error associated with a scale) of these scores. Aspects of validity (construct and content) have been addressed for the Rat (RGS), Mouse (MGS) and Rabbit Grimace Scales [10,15,16,21,22]. Reliability of pain scale scores can be assessed using internal consistency, inter- and intra-rater reliability. Internal consistency reflects the degree to which scale items are inter-related, while inter- and intra-reliability quantify the ability of a scale to return similar measures between different raters and the same rater at different times, respectively. A single reliability study is insufficient to allow generalisation to a heterogenous setting (different study populations, environments and raters), and Sotocinal et al. (2011), in their initial validation of the RGS, encouraged others to assess their scale [16]. Through repeated use and publication, this will eventually allow reliability generalisation, a concept similar to meta-analysis, as applied to measurement scales [23]. Identification of an analgesic intervention score for a pain scale dramatically increases its utility, expanding its use from an observational research tool to one facilitating decision making and intervention [24]. This moves towards fulfilling the aforementioned goals of providing appropriate analgesia and assessment of the efficacy of analgesic agents. At a fundamental level, identification of an analgesic intervention score allows personnel involved in animal work to fulfil a duty of care towards nonverbal subjects [14]. The aims of this study were to further assess reliability of the RGS and identify an analgesic intervention score. Materials and Methods Adult female Sprague-Dawley rats (284420 g), obtained from surplus stock at the University of Calgary, were scheduled for surgical implantation of a telemetric radio-transmitter device (4ET-S2 Radio Transmitter, Data Sciences International, Saint Paul, MN, USA) as part of an unrelated study. Animals were randomised to receive one of three analgesia treatments perioperatively or enter a sham treatment group. Animals were maintained in a 12 hr12 hr light-dark cycle (lights on at 0700) and housed in pairs or groups of three in micro-filter cages (48627620 cm [Ancare Corp., Worcester, MA, USA]) and were provided both fresh water and food (Prolab 2500 Rodent 5p14, Lab diet, PMI Nutrition International, St Louis, MO, USA) ad libitium. Plastic tubing (PVC Pipe, provided by the Health Science Animal Resource Centre, Calgary, Alberta CA) wood shavings (Aspen chip, NEPCO, Warr (...truncated)