Australian 'levels of evidence' (extended NHMRC evidence hierarchy)
This guideline updates the Australian National Health and Medical Research Council's original four-level evidence hierarchy to apply not only to treatment studies but also to diagnostic, prognostic, aetiological, and screening questions. It helps assess how likely a study is to be free from bias based on its design, providing a standard starting point for evaluating research across different clinical topics.
At a glance
Use when
Developing clinical practice guidelines; conducting systematic reviews or health technology assessments; evaluating non-interventional studies (diagnostic, prognostic, aetiological); standardizing evidence grading across institutions
Avoid when
Detailed risk of bias assessment is required (use critical appraisal tools instead); when synthesizing body of evidence (use GRADE or similar); when study applicability or indirectness is the primary concern
Inputs
Study design characteristics (e.g., RCT, cohort, case-control, case series), type of clinical question (intervention, diagnosis, prognosis, aetiology, screening)
Outputs
Assigned level of evidence (I–IV) indicating the likelihood of bias in a study
How it works
The extended NHMRC evidence hierarchy maintains the original Levels I–IV but expands their application beyond intervention studies. Level I remains systematic reviews of randomised controlled trials; Level II cohort studies; Level III case-control studies; and Level IV case series or expert opinion. The revision incorporates empirical evidence on study design and bias, integrates considerations of ethics and harms, and assigns levels to systematic reviews of lower-level evidence. It supports structured appraisal in guideline development, HTA, and systematic reviews across multiple clinical question types.
- HTA domains
- Clinical Effectiveness, Safety
- Assumptions
- Study design is a key determinant of risk of bias; higher-level designs (e.g., RCTs) are generally less prone to bias; systematic reviews of lower-level evidence can themselves constitute higher-level evidence; ethical constraints may limit certain study designs
- Strengths
- Broad applicability across clinical question types; grounded in empirical evidence on bias; maintains consistency with prior NHMRC framework; supports standardized evidence grading in guidelines and HTA; includes consideration of harms and ethics
- Limitations
- Does not replace detailed quality appraisal; does not account for all sources of bias within a design; may not fully capture nuances in complex or emerging methodologies; limited guidance on downgrading for indirectness or imprecision
- Also known as
- NHMRC Levels of Evidence, Australian Levels of Evidence, NHMRC Evidence Hierarchy
Questions this answers
- › What is the strength of evidence provided by different study designs for non-interventional clinical questions?
- › How can diagnostic, prognostic, or aetiological studies be ranked by risk of bias?
- › How should systematic reviews of lower-level evidence be classified?
- › What role does study design play in assessing evidence quality across different clinical domains?
- › How are ethics and assessment of harms incorporated into evidence hierarchies?
- › How can a consistent evidence grading system be applied across diverse health technology assessments?
References & sources
Similar by meaning
Beta record. Based on the original catalogue summary; primary-source enrichment pending.