OBJECTIVES: Clinical trials are a prerequisite for evidence-based medicine. Trial representativity by internal and external validities is an assumption for a transfer into the real world. METHODS: A questionnaire addressing several forms of bias was established. Randomized controlled trials for adjuvant treatment of pancreatic adenocarcinoma with an evidence level of lower than 2 were selected and evaluated for internal and external validities focusing on selection bias. RESULTS: Four selected trials (European Study Group for Pancreatic Cancer 1 [ESPAC-1], Charité Onkologie Clinical Studies in GI Cancer 001 [CONKO 001], Radiation Therapy Oncology Group 97-04 [RTOG 97-04], and Adjuvant Chemoradioimmunotherapy of Pancreatic Carcinoma [CapRI]) proved to be evaluable. External validity was to a large extent given in the trials, whereby Radiation Therapy Oncology Group 97-04 included significantly more patients with better prognostic markers (negative for lymph nodes and T1/T2 tumors). Their inclusion and exclusion criteria are limited to the (unavoidable) restrictions for safety reasons and clinical restrictions. Comparison of participants with nonparticipants and with the target population via literature research proved to be preventive against selection and recruitment biases. A possible detection bias in the Charité Onkologie Clinical Studies in GI Cancer 001 trial cannot, however, be excluded. CONCLUSIONS: The analyzed trials have acceptable internal and external validities despite some minor criticisms. The protocols could be transferred into the clinical routine. It is recommended to open the inclusion criteria of trials wider and to implement the comparison with the target population.
OBJECTIVES: Clinical trials are a prerequisite for evidence-based medicine. Trial representativity by internal and external validities is an assumption for a transfer into the real world. METHODS: A questionnaire addressing several forms of bias was established. Randomized controlled trials for adjuvant treatment of pancreatic adenocarcinoma with an evidence level of lower than 2 were selected and evaluated for internal and external validities focusing on selection bias. RESULTS: Four selected trials (European Study Group for Pancreatic Cancer 1 [ESPAC-1], Charité Onkologie Clinical Studies in GI Cancer 001 [CONKO 001], Radiation Therapy Oncology Group 97-04 [RTOG 97-04], and Adjuvant Chemoradioimmunotherapy of Pancreatic Carcinoma [CapRI]) proved to be evaluable. External validity was to a large extent given in the trials, whereby Radiation Therapy Oncology Group 97-04 included significantly more patients with better prognostic markers (negative for lymph nodes and T1/T2 tumors). Their inclusion and exclusion criteria are limited to the (unavoidable) restrictions for safety reasons and clinical restrictions. Comparison of participants with nonparticipants and with the target population via literature research proved to be preventive against selection and recruitment biases. A possible detection bias in the Charité Onkologie Clinical Studies in GI Cancer 001 trial cannot, however, be excluded. CONCLUSIONS: The analyzed trials have acceptable internal and external validities despite some minor criticisms. The protocols could be transferred into the clinical routine. It is recommended to open the inclusion criteria of trials wider and to implement the comparison with the target population.
Authors: Chakshu Sharma; Karim M Eltawil; Paul D Renfrew; Mark J Walsh; Michele Molinari Journal: World J Gastroenterol Date: 2011-02-21 Impact factor: 5.742