Abstract # 2551 Poster # 133:

Scheduled for Sunday, September 20, 2009 06:30 PM-09:00 PM: Session 14 (Mission Bay Ballroom CDE) Poster Presentation


DEVELOPING A RELIABLE LABORATORY PRIMATE ALOPECIA SCORING SYSTEM FOR INTERFACILITY COLLABORATION AND ON-LINE TRAINING

C. M. Crockett1, K. C. Baker2, C. K. Lutz3, K. Coleman4, M. A. Fahey5, M. A. Bloomsmith6, B. McCowan7, J. Sullivan8 and J. L. Weed9
1Washington National Primate Research Center, University of Washington, Box 357330, Seattle, WA 98195-7330, USA, 2Tulane National Primate Research Center, 3Southwest National Primate Research Center, 4Oregon National Primate Research Center, 5New England National Primate Research Center, 6Yerkes National Primate Research Center, 7California National Primate Research Center, 8Wisconsin National Primate Research Center, 9Division of Veterinary Resources, Office of Research Services/NIH/DHHS
line
     

Of the potential flags of impaired well-being in laboratory primates, alopecia has received particular regulatory focus, making it a top priority for behavioral managers. Consensus on a reliable alopecia scoring system used by multiple facilities can better determine prevalence and correlated factors, and guide consistent management plans. Existing methods for scoring human burn victims and skin disorders generally produce low interrater reliability. Behavioral managers at National Primate Research Centers are perfecting a reliable NHP alopecia scoring system for online training. A preliminary scoring scheme estimating percentage of hair missing was discarded in favor of estimating percentage of body surface affected, more consistent with veterinary practice. Seven raters independently scored 31 macaque photographs using a modified rule of 9s method for estimating body surface percentages, classified into 6 ordinal “Aloscore” categories. The photographs were also scored for pattern of alopecia in the affected areas. Raters had high agreement in Aloscore [Kendall’s coefficient of concordance: W=0.997, Χ2(30) =209.27, p<0.001]. Each rater’s Aloscores were compared with the modal scores; Cohen's Kappas averaged 0.95 [range=0.94-0.97]. Concordance in pattern classification, although highly significant, was considerably lower [Kappa=0.45; z=14.83; p<0.001]. Each rater’s pattern scores were compared with the modal scores; Cohen's Kappas averaged 0.77 [range=0.63-0.92]. The numerical Aloscore system is readily trained and produced high agreement. Pattern definitions are undergoing modification to improve reliability prior to adoption.