Characterizing individual differences in heat-pain sensitivity.

Heat induced pain has been shown to follow a positively accelerating power function for groups of subjects, yet the extent to which this applies to individual subjects is unknown. Statistical methods were developed for assessing the goodness of fit and reliability of the power function for data from individual subjects with the aim of using such functions for characterizing individual differences in heat-pain sensitivity. 175 subjects rated ascending and random series of contact heat stimuli with visual analogue scales for pain intensity (VAS-I) and unpleasantness (VAS-A). Curve fitting showed excellent model fit. Substitution of model estimates in place of observed VAS scores produced minimal bias in group means, about 0.3 VAS units in the ascending series and 1.0 in the random series, on a 0-100 scale. Individual power function exponents were considerably higher for the ascending than for the random series and somewhat higher for VAS-A than for VAS-I (means: ascending VAS-I=9.04, VAS-A=9.80; random VAS-I=4.95, VAS-A=5.67). The reliability of VAS estimates was high (>==.93), and for the ascending series it remained so when extrapolating 4 degrees C beyond the empirical range. Exponent reliability was high for the ascending series (VAS-I=.92; VAS-A=.91), but considerably lower for the random series (VAS-I=.69; VAS-A=.71). Individual differences constituted 60% of the total variance in pain ratings, whereas stimulus temperature accounted for only 40%. This finding underscores the importance of taking individual differences into account when performing pain studies.