ANTIPRURITIC EVALUATION 61 Table I WEST-itch Esthesiometer Number of Force (g) Stimulus monofilaments each monofilament 5 4 60 4 6 20 3 8 4 2 10 1 1 12 0.25 level is specific to each individual. For example, some individuals report more itch at lower-force stimulus levels. The minimal itch-level criterion was met when the subject rated one (or more) stimuli (the five force levels of the esthesiometer) at a level of itch of 3 or more for each site. In order to continue to completion in the study, subjects were required to meet this minimal itch criterion. This minimal itch-level criterion was required because some minimum pretreatment itch level is necessary to measure antipruritic effects. The 30% scale value (i.e., 3) was selected as the minimum level because it provides a clear distinction from a "no-itch" state but still allows the inclusion of subjects who are experiencing itch below the mid-scale value (5). Further, it was assumed that few consumers would use anti-itch medications for levels of itch rated under 3 on the itch scale and that, therefore, inclusion of these subjects would not be representative of product-use conditions. At one minute postapplication, the subject was restimulated with the stimuli (the specific levels of the itch esthesiometer, ! to 5, that had produced at least a level-3-rated itch on both sites during the preapplication evaluation of itch intensity). Each site was stimulated in order, and the subject's attention was focused on the level of itch by having him/her verbally rate the level of itch produced. The same post-test procedure that was done at one minute postapplication was repeated at thirty minutes. A topical antibiotic cream (Bacitracin ©) was then applied to the test sites to minimize the pos- sibility of infection from the skin stripping. PROCEDURE The procedures for both studies were similar. After the subject passed the minimal-itch criterion, each product was weighed (0.2 g for Study I and 0.4 g for Study II) in a separate room, and brought into the testing room. The products were blinded to recognition, identified only by the randomized-site code for each subject, and were given to the tester on plastic mittens. The products were simultaneously applied, one to each site, according to the randomized application code, for two minutes of gentle stroking by two testers (one for each site). By two minutes, for most of the subjects, the products appeared completely rubbed into the skin. For a few subjects, the products did not rub in entirely, and the remaining product was gently relocated to the edge of the test site.

62 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS STUDY II: DATA REDUCTION AND ANALYSIS Since the objective of Study II was to track general and time-specific effects of two products, the maximal itch rating for itch subject at pretreatment, posttreatment 1 minute and posttreatment of 0.5 hours was subjected to a repeated Analysis of Variance (ANOVA) measurement. Data from the two pretreatment sites were subjected to a product moment correlation in order to determine the within-subject reliability of the procedure. RESULTS All statistical analyses were performed using the SPSS TM for Windows. TM STUDY I The paired t-test computed to determine whether an oil-in-water vehicle could reduce itch resulted in a t-value of 10.33 (df = 33), which is significant at less than the .001 level. The mean maximum itch rating changed from 5.18 (SEM = .353) in the pretreatment condition to 1.62 (SEM = .213) in combined posttreatment conditions. Another paired t-test was computed to determine whether the addition of 6% ben- zocaine to the vehicle would further reduce itch ratings. The results of this analysis indicated that the benzocaine significantly reduced itch relative to the vehicle alone in intact skin. After subtracting the pretreatment levels, the combined mean maximum rating for the benzocaine showed an itch-reduced difference on 0.42 (SEM = .20) relative to the vehicle (t = 2.07 df = 30 p .05). STUDY II Intrasubject reliability was determined by correlating the pretreatment test sites and additionally by computing a Cronbach's alpha. One subject was eliminated from the analysis because her values represented a significant multidimensional outlier. The results of these analyses indicated that there was a high level of reliability, with the tO • SITE 2 to Figure 3. Intrasubject reliability. Pretreatment itch ratings at two test sites (r = .92).