Effectiveness was measured by 4 variables: objective response (%), duration of response (weeks), delay until progression (weeks) and median survival (months). The objective response was 57.1 for docetaxel, 28.9 for paclitaxel and 16 for vinorelbine. The duration of response was 28 weeks for docetaxel and paclitaxel, and 21 weeks for vinorelbine. The delay until progression was higher for docetaxel (21 weeks) and paclitaxel (18 weeks), and lower for vinorelbine (12.9 weeks). The median survival was the same for both strategies (12 months).
The toxicity was measured by 7 variables: infection/febrile neutropenia with hospitalisation (%), infection/febrile neutropenia without hospitalisation (%), cutaneous reactions (%),arthralgies-myalgies (%), hydrolic retention leading to a break in the treatment (%), hydrolic retention without stopping the treatment (%), severe neurotoxicity without stop in the treatment (%).
The infection/febrile neutropenia with hospitalisation was greater with docetaxel (17.9) than with paclitaxel (2) or vinorelbine (3). The infection/febrile neutropenia without hospitalisation was lower with docetaxel (1.9) than with paclitaxel (2) and was absent with the vinorelbinestrategy. Cutaneous reactions were present only with docetaxel (7). Arthralgies-myalgies were observed only with paclitaxel (16). Hydrolic retentions were higher with docetaxel (1.9 and 2.9). Finally, severe neurotoxicity was greater for the paclitaxel strategy. The above measures formed the input variables to theMarkov model, which identified 53 possible states.