A combined measure of procedural volume and outcome to assess hospital quality of colorectal cancer surgery, a secondary analysis of clinical audit data

OBJECTIVE: To identify, on the basis of past performance, those hospitals that demonstrate good outcomes in sufficient numbers to make it likely that they will provide adequate quality of care in the future, using a combined measure of volume and outcome (CM-V&O). To compare this CM-V&O with measures using outcome-only (O-O) or volume-only (V-O), and verify 2010-quality of care assessment on 2011 data. DESIGN: Secondary analysis of clinical audit data. SETTING: The Dutch Surgical Colorectal Audit database of 2010 and 2011, the Netherlands. PARTICIPANTS: 8911 patients (test population, treated in 2010) and 9212 patients (verification population, treated in 2011) who underwent a resection of primary colorectal cancer in 89 Dutch hospitals. MAIN OUTCOME MEASURES: Outcome was measured by Observed/Expected (O/E) postoperative mortality and morbidity. CM-V&O states 2 criteria; 1) outcome is not significantly worse than average, and 2) outcome is significantly better than substandard, with 'substandard care' being defined as an unacceptably high O/E threshold for mortality and/or morbidity (which we set at 2 and 1.5 respectively). RESULTS: Average mortality and morbidity in 2010 were 4.1 and 24.3% respectively. 84 (94%) hospitals performed 'not worse than average' for mortality, but only 21 (24%) of those were able to prove they were also 'better than substandard' (O/E<2). For morbidity, 42 hospitals (47%) met the CM-V&O. Morbidity in 2011 was significantly lower in these hospitals (19.8 vs. 22.8% p<0.01). No relationship was found between hospitals' 2010 performance on O-O en V-O, and the quality of their care in 2011. CONCLUSION: CM-V&O for morbidity can be used to identify hospitals that provide adequate quality and is associated with better outcomes in the subsequent year.