摘要

The goal of this study is to test the validity of RESPOND, a web-based decision support system to assess and manage anemia in cancer patients as per the European Organisation for Research and Treatment of Cancer (EORTC) guidelines. The intraclass correlation metrics for the algorithmic definitions were reported previously. Reported here are the concurrent validity, the extent to which clinicians' anemia management is guidelines-congruent when using the system; and discriminant validity, the extent to which clinicians practice in congruence with guidelines when vs. when not using the system. Hybrid matched design with precohort (retrospective; clinicians not using RESPOND) and postcohort (prospective; clinicians using RESPOND) of anemic patients matched on cancer type and chemotherapy regimen and followed up over 4 months after treatment initiation with erythropoietic proteins (34 patients per cohort; total N = 68). Congruence scores quantified the extent to which anemia management was congruent with the EORTC guidelines (range 0-10). Hemoglobin (Hb) increased significantly for both cohorts, but the postcohort group showed more rapid rate of Hb increase over time (p < 0.006), higher Hb by visit 4 (p = 0.007), and greater Hb increase by visit 4 (p = 0.006). Concurrent validity was high with mean postcohort congruence scores of 8.18 +/- 1.38. Discriminant validity was inferred from statistically significant differences in mean congruence scores between cohorts (p < 0.001) and from the postcohort having odds ratios of 3.64 for patients to reach Hb a parts per thousand yenaEuro parts per thousand 11 g/dL and 2.91 to achieve Hb a parts per thousand yenaEuro parts per thousand 12 g/dL. RESPOND, a validated computerized clinical guidance system with an incremental effect beyond the pharmacotherapeutic effect of erythropoietic proteins, offers clinicians accurate and safe guidance in managing anemia in cancer patients.

  • 出版日期2010-7