A prospective study was carried out from January to April 1993 in 42centers to assess the quality of care in adult heart surgery in France. Onehundred eight parameters were collected (i.e., risk factors, surgery,postoperative course) for 7181 patients. Eight simple, objective factorswere selected using a multivariate analysis. Point weighting was assignedto each factor. A simple cumulative score was thereby established toclassify patients within four incremental risk groups (respectivedistribution of patients and mortality were 38.6%, 26.5%, 21.3%, 13.5% and2%, 3.9%, 6.1%, 21.4%). This score was highly predictive for mortality andsevere morbidity. Objectivity and independence of the observer, coupledwith a reliable score, are needed for meaningful assessment of the qualityof care. Administrative supervision and/or comparison of raw center resultstherefore remain controversial. A national data base was developed as apart of this study, that includes anonymous information from two-thirds ofall cardiac surgery cases. Nation-wide results for France allow each centerfreely to assess its results. This self-assessment approach is, in ourview, the most accurate way of quality of care assessment. A pan- Europeansystem using a similar approach should be developed.