BACKGROUND: Studies investigating the factors associated with need for total hip replacement should ideally be based on prospective investigation of new attenders in primary care. AIM: To determine the incidence of listing for total hip replacement, and its predictors, among attenders in primary care with a new episode of hip pain. DESIGN OF STUDY: Prospective multicentre cohort study. SETTING: One hundred and ninety-five patients (mean age = 63 years, 68% female) with new episode of hip pain, attending primary care between November 1994 and October 1997. At the first visit, patients were evaluated for indices of pain and disability, range of hip movement, and radiographic changes of osteoarthritis. METHOD: General practitioner participants were recruited from the membership of the Primary Care Rheumatology Society to recruit all consecutive attenders with a new episode of hip pain. Annual follow-up was carried out to determine which patients were being 'put on a waiting list' for total hip replacement. RESULTS: Seven per cent of patients were put on a waiting list for total hip replacement within 12 months and 23% of patients within four years. At presentation, pain duration, pain severity, (including the need to use a stick) and restriction of internal rotation were the major clinical predictors of being put on a waiting list. Radiographic predictors of osteoarthritis performed similarly to the clinical measures. A simple scoring system based on both radiographic severity and two of the clinical measures was derived that identified groups at high likelihood of being put on a waiting list (sensitivity = 76%) with a low false-positive rate (specificity = 95%). CONCLUSION: New primary care attenders with pain are frequently accepted for total hip replacement soon after their first attendance--a decision that can be predicted by simple clinical measures.