Using principal component analysis to describe wound status.

Principal component (PC) analysis was used to assess the status of pressure ulcers over time in 37 subjects. Content validity was established by literature review and expert opinion. The primary variables in the wound healing model are ulcer surface area, exudate amount, and surface appearance. A linear function based on PC analysis of these values was established at each time point with an appropriate weighting of 2:3:3. This function explained 55% to 65% of the variation in the data. Other variables were considered in the model as well. However, because of the correlation of these variables with the original three, the modeling was not improved. A decreasing function from week 0 to week 8 established a good linear fit to the data and reasonable discrimination between time points, given the limitations of the sample size. There is good discrimination among the time points, at least for the earlier times compared with the later times. The procedure, although promising, requires validation with a larger data set.