Aim of study was to evaluate if the subjective assessment of average pain over an 8-hour period is a valid estimate of pain intensity in cancer patients and whether it can accurately describe pain intensity during fentanyl TTS (f-TTS) analgesia. 44 consecutive patients with cancer pain treated with f-TTS (24 males and 20 females), recorded on 0-10 numerical scales the intensity of their pain at hourly intervals, the average of these 8 scores was then calculated (S1); at the 8th hour, they were requested to rate average pain intensity over the last 8 hours (S2). Cases who presented with hourly pain scores which were at least 2 points above S1, were classified as patients having pain peaks (PPP). Patients were also classified as TTS stable (TTS-S) if the same dose of f-TTS was used for more than three days or TTS unstable (TTS-U) if f-TTS dose had been changed within three days. Intraclass correlation coefficients between S1 and S2 was 0.85 (95% CI=0.77 - 0.93); 12 patients had pain peaks, but the comparison with others did not show any difference neither on S1 nor on S2 nor on the difference between them. Comparing TTS-S (18) with TTS-U (26) patients, average pain score (TTS-U mean - SD-3.5-1.5), TTS-S-3.2-1.8) and patientrated 8-hour average pain (TTS-U,=3.9-1.4, TTS-S,=3.2-1.5) did not show any difference, while more TTS unstable patients used at least one dose of p.r.n analgesics (61% vs 28% p-0.036). 38% of TTS unstable patients had pain peaks versus only 11% of TTS-S patients (p=0.08). The study shows that average pain over previous 8 hours is a valid estimate of pain intensity and gives an fair description of fentanyl TTS (f-TTS) analgesia even if hourly measurement gives a more sensitive measure of not completely stable pain.