Little information is currently available concerning the relationship between results obtained in humans from surgical test wounds and results from wound models. Therefore, to evaluate human wound healing parameters, tubings of expanded polytetrafluoroethylene were implanted in a subcutaneous test wound in the arm of 47 volunteers and 20 patients undergoing hernia repair. The surgical patients also had implants left in the surgical wound cavity. After 10 days the deposition of collagen in the tubings as expressed by hydroxyproline content was 30% higher in the surgical wound than in the test wound, p < 0.05. The amount of collagen deposited in the tubing within the surgical wound did not correlate with measurements in the test wound, whereas a significant correlation of proline levels was found between the two sites, p < 0.05. Deposition of proline and total protein in the model was equivalent. In both wound types age negatively correlated with levels of protein, but not collagen. The variability of the results was 40% lower in the subcutaneous test wound than in the surgical wound. There was no significant difference in hydroxyproline deposition between the volunteers and the patients undergoing hernia repair. In patients undergoing minor surgery without signs of compromised healing the expanded polytetrafluoroethylene test wound in the arm reflects the deposition of non-collagenous protein, but not collagen, within the surgical wound.