To identify, using a consensus development process, a list of common conditions likely to be ambulatory care sensitive (ACS); i.e. conditions for which practicable improvements in access to timely and effective ambulatory care in the English National Health Service would either reduce the incidence of the condition or avoid substantial proportions of current hospital admissions. Three panels of clinicians each reviewed about a third of an initial list of 174 conditions commonly recorded as hospital discharge diagnoses for residents of the North West Thames Region. For each condition, panellists estimated the proportion of cases currently admitted to hospital for which, given timely and effective ambulatory care: onset of disease could have been prevented; admission for existing disease could have been prevented; admission, once indicated, should take place within 48 hours. After an introductory meeting to discuss and clarify the task, panel members completed three rounds of a questionnaire, with postal feedback between each round, and a second meeting to discuss interim results before the final round. Seventeen general practitioners (GPs) and 17 hospital specialists working in the region comprised the panels. The panels considered that for 30 of the 174 conditions at least 70% of admissions to hospital could be avoided, either by prevention of disease onset or timely and effective ambulatory care, though predominantly through the latter. For each of a further 66 conditions, 50-69% of admissions could be prevented. Within-panel agreement between hospital specialists and GPs was generally good, although the GPs tended to give slightly higher scores for avoidability of admissions than the specialists. There was marked convergence of scores in succeeding rounds. Although a consensus-based list of ACS conditions cannot be definitive, the clear view of the panels was that the scope for avoiding admission through better ambulatory care is very substantial.