Recommendations for Improvements of Clinical Trials in AML
| Problem . | Possible solutions . |
|---|---|
| High false-negative and false-positive rates | Increase in study size |
| Ill-defined historical control group | Explicit description (number of patients, type of study, diagnoses, treatment); adjustments for sampling variation and differences in case mix |
| Lack of control group | Use of explicitly described historical or concurrent control group; randomization, including multi-arm, multi-stage designs |
| Patient heterogeneity | Stratified trial; statistical adjustment (multivariate analysis) |
| Generalizability of treatment results, effect modification | Explicit description of inclusion/exclusion criteria, provision of information about total number of patients available for study vs those actually treated |
| Choice of surrogate endpoint that does not predict clinical benefit | Use of validated surrogates; validation of alternative endpoints before use |
| Delay in activation of phase 3 trial | Integrated phase 2/3 trial design; streamlining of internal and external groups and processes |
| Bias through early publication | Allowance of adequate follow-up time between completion of study accrual and publication; introduction of journal policies to discourage early publication |
| Problem . | Possible solutions . |
|---|---|
| High false-negative and false-positive rates | Increase in study size |
| Ill-defined historical control group | Explicit description (number of patients, type of study, diagnoses, treatment); adjustments for sampling variation and differences in case mix |
| Lack of control group | Use of explicitly described historical or concurrent control group; randomization, including multi-arm, multi-stage designs |
| Patient heterogeneity | Stratified trial; statistical adjustment (multivariate analysis) |
| Generalizability of treatment results, effect modification | Explicit description of inclusion/exclusion criteria, provision of information about total number of patients available for study vs those actually treated |
| Choice of surrogate endpoint that does not predict clinical benefit | Use of validated surrogates; validation of alternative endpoints before use |
| Delay in activation of phase 3 trial | Integrated phase 2/3 trial design; streamlining of internal and external groups and processes |
| Bias through early publication | Allowance of adequate follow-up time between completion of study accrual and publication; introduction of journal policies to discourage early publication |