Table 2.
The American Academy of Neurology´s Classification of Evidence (I, II, III, IV) and level of confidence/classification of recommendations (A, B, C, U)
| A | Classification of evidence | |
| I | Triple-masked RCT in a representative population. Relevant baseline characteristics are presented, substantially equivalent or there is appropriate statistical adjustment. Additional criteria: A) concealed allocation, B) no more than two primary outcomes specified, C) exclusion/inclusion criteria clearly defined, D) at least 80% of participants completing the study | |
| II | RCT that lacks one or two Class I criteria A–D or prospective matched cohort study with masked/objective outcome assessment that meets B–D. Randomized crossover trials reporting either period and carryover effects or baseline characteristics of treatment order groups. Relevant baseline characteristics are presented, substantially equivalent or there is appropriate statistical adjustment | |
| III | Controlled studies (including studies with external controls) or crossover trial missing both period/carryover effects and presentation of baseline characteristics. Outcome is independently assessed, or independently derived by objective outcome measurement | |
| IV | Studies not meeting Class I, II, or III criteria | |
| B | Level of confidence/classification of recommendations | |
| A | High confidence: highly likely to be effective, ineffective or harmful (or [not] useful/predictive) for the given condition in the specified population. Requires at least two Class I studies | |
| B | Moderate confidence: likely to be effective, ineffective or harmful (or [not] useful/predictive) for the given condition in the specified population. Requires one Class I study or at least two Class II studies | |
| C | Low confidence: possibly likely to be effective, ineffective or harmful (or [not] useful/predictive) for the given condition in the specified population. Requires one Class II study or at least two Class III studies | |
| U | Very low confidence: data insufficient or conflicting, given current knowledge, treatment is unproven | |