A novel method to assess test‐quality of phase‐3 trials

Objectives To investigate new drug development programs, we piloted a method to assess “test‐quality” of phase‐3 trials. Methods . 40 phase‐3 trial results, announced in 2001, were evaluated using standardized design, performance, and trial endpoint criteria, and available FDA reviews. Trials were reliable if scientifically designed, conducted according to GCP ( presumptively reliable if GCP unconfirmed), and employed FDA‐accepted endpoints; unconvincing if endpoints were unacceptable; and inconclusive if the design was scientifically weak or GCP was violated. Results . 34 trials (85%) were evaluable: 100% were scientifically sound, employing randomization, adequate blinding and power, and employed acceptable endpoints. GCP compliance was confirmed in 50%, unconfirmed in 47%; 3% were non‐compliant. Using our taxonomy, 97% of trials were presumptively or definitely reliable , 3% were inconclusive , and none were unconvincing . Trial sponsors reported 82% of trials as “successful,” of which 57% were FDA approved/approvable by 2003. Implications . High test‐quality of publicly announced phase‐3 trials in 2001 was confirmed. About one‐half of “successful” trials led to FDA approval within 2 years, usually in public companies or partnerships with prior drug approvals (81%). A “quality and completeness” assessment of pre‐phase‐3 programs will be applied to further investigate ability of phase‐3 trials (publicly announced or private) to reliably demonstrate drug effectiveness. Clinical Pharmacology & Therapeutics (2004) 75 , P8–P8; doi: 10.1016/j.clpt.2003.11.030