Levels of Evidence Rating System

Evidence Levels

Level 1: High Strength/Quality

  • Evidence consists of at least one randomized, double-blind trial without important limitations (i.e., large treatment effect); intent-to-treat analysis used, confidence intervals reported. If more than one trial is available, these trials have consistent results. or
  • Evidence consists of a meta-analysis of such trials with consistent results (i.e., low heterogeneity).

Evidence consisting of a non-blinded or single-blinded trial that meets study objective end points may be considered as Level 1 evidence in some cancer-related cases (e.g., NCI-sponsored cooperative group study or a multicenter trial).

Level 2: Moderate Strength/Quality

  • Evidence consists of at least one non-blinded or single-blinded, randomized clinical trial. or
  • Evidence consists of at least one non-blinded or single-blinded, non-randomized clinical trial. or
  • Evidence consists of a meta-analysis of randomized, controlled clinical trials with heterogenous results if reasons for heterogeneity in individual trials are adequately discussed. or
  • Evidence consists of at least one randomized, controlled clinical trial, but with important methodological limitations (e.g., large number of patients lost to follow-up and/or no intent-to-treat analysis and/or important data not recorded). or
  • Evidence is inconsistent (i.e., two or more randomized controlled trials with unexplained, widely varying estimates of treatment effects, even if results of individual trials would constitute strong Level 1 evidence when considered alone).

Evidence consisting of a non-blinded, non-randomized trial (i.e., a phase II study) may be considered as Level 2 evidence in some cancer-related cases (i.e., rare cancers or cancers with limited available treatment options).

Level 3: Low Strength/Quality

  • Evidence consists of observational studies, case reports, or case series; may also include randomized clinical trials with multiple serious deficiencies or study limitations.

Level 4: Opinion/Experience

  • Evidence consists of expert consensus panel reports or expert reviewers’ comments.

Strength of Study End Points

(listed in descending order)

  • Overall survival (total mortality)
  • Cause-specific mortality
  • Quality of life
  • Indirect surrogates
    • Overall response rate (ORR)
    • Progression-free survival (PFS)
    • Disease-free survival (DFS

AHFS Grades of Recommendation

  1. Recommended (Accepted)
    The drug or biologic should be used, is recommended/indicated, or is useful/effective/beneficial in most cases.
  2. Reasonable Choice (Accepted, with Possible Conditions) (e.g., treatment option)
    The drug or biologic is reasonable to use under certain conditions (e.g., in certain patient groups), can be useful/effective/beneficial, or is probably recommended or indicated.
  3. Not Fully Established (Unclear Risk/Benefit, Equivocal Evidence, Inadequate Data and/or Experience)
    Usefulness and/or effectiveness is unknown, unclear, or uncertain or is not well established relative to the standard of care.
  4. Not Recommended (Unaccepted)
    The drug or biologic is considered inappropriate, obsolete, or unproven; is not recommended, is not indicated, or is not useful/effective/beneficial; or may be harmful.