A simplified (modified) Duke Activity Status Index (M-DASI) to characterise functional capacity: a secondary analysis of the Measurement of Exercise Tolerance before Surgery (METS) study

Background: Accurate assessment of functional capacity, a predictor of postoperative morbidity and mortality, is essential to improving surgical planning and outcomes. We assessed if all 12 items of the Duke Activity Status Index (DASI) were equally important in reflecting exercise capacity. Methods: In this secondary cross-sectional analysis of the international, multicentre Measurement of Exercise Tolerance before Surgery (METS) study, we assessed cardiopulmonary exercise testing and DASI data from 1455 participants. Multivariable regression analyses were used to revise the DASI model in predicting an anaerobic threshold (AT) >11 ml kg min and peak oxygen consumption (VO peak) >16 ml kg min, cut-points that represent a reduced risk of postoperative complications. Results: Five questions were identified to have dominance in predicting AT>11 ml kg min and VO peak>16 ml.kgmin. These items were included in the M-DASI-5Q and retained utility in predicting AT>11 ml.kg.min (area under the receiver-operating-characteristic [AUROC]-AT: M-DASI-5Q=0.67 vs original 12-question DASI=0.66) and VO peak (AUROC-VO2 peak: M-DASI-5Q 0.73 vs original 12-question DASI 0.71). Conversely, in a sensitivity analysis we removed one potentially sensitive question related to the ability to have sexual relations, and the ability of the remaining four questions (M-DASI-4Q) to predict an adequate functional threshold remained no worse than the original 12-question DASI model. Adding a dynamic component to the M-DASI-4Q by assessing the chronotropic response to exercise improved its ability to discriminate between those with VO peak>16 ml.kg.min and VO pea