FIGURE 1 Panels generated using Viscovery (Viscovery Software GmbH; Vienna, Austria). The Viscovery program placed all patients on a specific position on all maps based on their multidimensional response profile. The more subjects resemble in terms of their response to pulmonary rehabilitation the closer they are on the map. Contrarily, the more they differ the further they are away from each other. When looking at an outcome measure of pulmonary rehabilitation, patients “raise a red flag” if they had a very good response, “a green flag” when the response was good to moderate, and “a blue flag” when the response was poor. In this way the maps can be interpreted. Using the topology of the self-organising map, the Viscovery program could identify four different clusters of patients with chronic obstructive pulmonary disease (COPD) with a significantly different multidimensional response profile. C1: cluster 1 “very good responder” (n=378); C2: cluster 2 “good responder” (n=742); C3: cluster 3 “moderate responder” (n=731); C4: cluster 4 “poor responder” (n=217). a) Left panel shows the multidimensional response outcome. All other attribute pictures in a) are the absolute change in Medical Research Council (MRC)dyspnoea grade, 6-min walk distance (6MWD), cycle endurance time (constant work-rate test; CWRT), Canadian Occupational Performance Measure, Performance (COPM-P), Canadian Occupational Performance Measure, Satisfaction (COPM-S), Hospital Anxiety and Depression Scale, anxiety (HADS-A), Hospital Anxiety and Depression Scale, depression (HADS-D), and St. George's Respiratory Questionnaire total score (SGRQ-T). b) Left panel shows the proportion of clinically relevant outcomes (exceeding at least 1× minimally clinical important difference (MCID)). All other panels are the proportion of patients per outcome showing a clinically relevant improvement (exceeding at least 1× MCID). c) First panel shows the proportion of clinically relevant outcomes (exceeding at least 2× MCID). All other panels are the proportion of patients per outcome showing a clinically relevant improvement (exceeding at least 2× MCID).