OBJECTIVE: The aim of this study was to characterise users of herbal medicines and assess the effect of socio-demographic characteristics, perceived health and chronic disease on the use of herbal medicine in a multi-ethnic Swedish health practice population. METHODS: A questionnaire was completed by 1433 (out of a total of 1776) patients aged 16 years and above who visited the Jordbro Health Centre (JHC) in Stockholm, Sweden, between 14 January and 30 June 2002. The results were linked to computerised medical records. RESULTS: Altogether 320 (22.3%) of this patient population used some form of herbal medicine. The bivariate analysis showed that the use of herbal medicine were more common among patients aged 45-64 years, females, high educated, patients born in Nordic countries or Europe compared to other age groups, males, low educated, patients born outside Europe and without chronic disease. In the logistic regression analysis when the effects of confounders were taken into account, females, high educated patients and patients with chronic disease had higher odds for use of herbal medicine than males, low educated and patients without chronic disease. The odds ratios (ORs) and 95% confidence interval (95% CI) were 1.95 (1.40-2.76) for female as compared with male patients; 2.10 (1.49-2.97) for subjects with a high level of education compared with subjects with a low educational level and 1.62 (1.15-2.29) for subjects with chronic disease compared with subjects without chronic disease. The common diagnoses were musculo-skeletal, respiratory and circulatory disorders, signs and symptoms and external causes according to the International Classification of Diseases (ICD10). There were no significant differences between users and non-users of herbal medicine regarding the number of consultations with any physician or the general practitioner (GP), contacts with the health care centre, use of prescribed medicines or number of days of sick leave during the past year. CONCLUSION: Females, well-educated patients and patients with chronic disease had higher odds for use of herbal medicine than others irrespective of other socio-demographic characteristics, and herbal medicine was seen to be used independently of conventional medicine.