Pain is a major cause for visiting a primary care physician. There are, however, few studies on the assessment of pain patients at the primary care level. The aim of this cross-sectional study was to investigate the concordance between general practitioners' (GPs') and patients' assessments of pain intensity and whether this assessment is influenced by the duration or intensity of pain. Seven hundred and thirty-eight patients aged 16 to 75 years, who were visiting a GP because of pain, participated. Both the patients and the GPs rated pain intensity using the horizontal 100 mm Visual Analogue Scale (VAS). Means and correlations were calculated using non-parametric tests. The VAS scales were arbitrarily divided into five grades (one unit = 20 mm) to investigate the concordance between GPs' and patients' assessments of pain intensity. Spearman 's correlation coefficient between GPs' and patients' assessments was 0.31 for non-chronic pain (of duration less than six months) and 0.20 for chronic pain. GPs evaluated graded pain intensity at least one unit lower than patients in 37% of the visits. In one-fifth of the visits (20.5%), the GP's rating was at least two units lower than the patient's rating. The more severe the pain as assessed by patients, the greater the non-concordance between patients' and GPs' assessments. There was considerable non-concordance between GPs' and patients' assessments of pain intensity. GPs tended to estimate their patients' pain intensity as clinically significantly lower than the patients themselves, particularly in chronic and severe pain.