From Expert Opinion to Evidence: Understanding Delphi Consensus
Randomised controlled trials, systematic reviews and observational studies remain fundamental to evidence-based medicine, yet they do not answer every clinically relevant question. Evidence may be incomplete, heterogeneous, rapidly evolving or insufficiently granular to inform specific decisions, for example about treatment pathways, patient selection, referral criteria or standards of care.
In these circumstances, expert judgement often becomes necessary. However, there is a substantive methodological distinction between expert opinion and structured expert consensus. The former may reflect the experience of individual clinicians or small groups. The latter applies a predefined and transparent methodology to assess the extent to which informed experts converge around a particular clinical position.
Several structured consensus methodologies exist, including the Delphi method, Nominal Group Technique (NGT) and Structured Expert Elicitation (SEE). While these approaches differ in design and application, they share a common objective: transforming individual expert opinion into a more systematic and reproducible form of evidence. Among these methods, Delphi has become the most widely used approach in healthcare research and publication, particularly when consensus is sought across geographically dispersed experts.
Different approaches may be appropriate for different objectives. NGT is often used to prioritise topics, identify key challenges or generate recommendations within a facilitated group workshop. SEE is commonly applied when quantitative estimates are required in the absence of data, such as forecasting disease burden or estimating clinical outcomes. Delphi is particularly well suited to healthcare evidence generation because it can systematically evaluate agreement across a broad range of clinical, treatment and policy questions while minimising the influence of group dynamics.
Expert opinion is informative but methodologically limited
Clinical expertise provides an important interpretive layer within healthcare decision-making.
Experienced clinicians may recognise patterns not yet fully characterised in the literature, understand practical limitations associated with existing guidance, and identify points of uncertainty within evolving treatment landscapes.
However, individual expert opinion is inherently vulnerable to variation. Clinical perspectives may differ according to geography, specialty, healthcare system, patient population and individual experience. Informal group discussion may also be influenced by hierarchy, dominant personalities and professional relationships.
As a result, the existence of expert agreement does not establish that a conclusion has been generated through a sufficiently robust evidentiary process. The question is therefore not simply whether experts agree. It is whether that agreement has been elicited, measured and reported in a manner that is systematic and reproducible.
Structured consensus transforms expert opinion into evidence
Structured consensus methodologies are designed to formalise expert judgement and generate more robust expert opinion evidence.
Rather than relying on informal discussion, they introduce methodological controls around panel selection, statement development, voting procedures, iteration, feedback and the definition of consensus.
The Delphi method is particularly well suited to healthcare research because it combines independent expert input, structured iteration and predefined consensus criteria. Delphi panels involve successive rounds of independent rating by a selected panel of experts, with controlled feedback provided between rounds. Participants are given the opportunity to reconsider their responses in light of the aggregated group responses while remaining insulated from direct peer influence. This process transforms unstructured expert opinion into a systematic and transparent form of evidence.
When is structured consensus appropriate?
Structured consensus is most useful when there is a clearly defined evidentiary or clinical uncertainty. Examples may include situations in which:
- published data support the efficacy of an intervention but do not clarify its optimal place within a treatment pathway
- clinical guidelines differ in their recommendations
- practice varies substantially across specialties, centres or geographies
- emerging therapies or technologies have entered clinical use ahead of detailed guidance;
- there is no consistently applied definition for a clinical concept
- or available studies do not adequately address the complexity of real-world decision-making.
In these settings, structured consensus can help establish how a defined expert community interprets the existing evidence and how consistently that interpretation is shared.
For healthcare organisations, structured consensus studies can be particularly valuable when evidence generation needs extend beyond the scope of traditional clinical research. Medical Affairs teams frequently use expert opinion evidence to understand treatment pathways, patient selection, monitoring approaches, diagnostic challenges and standards of care. Market Access teams may use structured consensus to characterise unmet need, document variation in clinical practice, define treatment positioning or support evidence packages in areas where published data remain limited.
What determines the credibility of a Delphi consensus?
The credibility of expert opinion evidence generated through Delphi depends not simply on whether consensus is achieved, but on whether the process used to achieve it is rigorous, transparent and reproducible. A credible Delphi provides enough methodological detail for readers to understand how the consensus was generated and to assess the legitimacy of its conclusions.
Appropriate panel composition
The credibility of the consensus is closely linked to the expertise represented within the panel. Participant selection should therefore be aligned with the research question and based on clear, defensible criteria. Depending on the topic, this may include specialty, professional role, geographical setting, care setting or relevant experience.
Panel size alone is not a marker of methodological quality. A large but poorly constituted panel may provide less meaningful evidence than a smaller panel whose composition is carefully justified.
Predefined consensus criteria
Consensus criteria should be established before the results are known. The protocol should specify the threshold for consensus, how agreement or disagreement will be assessed, and how equivocal or unresolved responses will be handled.
Pre-specification strengthens methodological credibility by reducing the possibility that analytical decisions are influenced by the findings. It also makes the process more reproducible for other researchers.
Independent participation and iteration
Anonymised or independent voting is a defining feature of Delphi methodology. It helps minimise the influence of hierarchy, professional status and interpersonal dynamics on individual responses.
The iterative structure is equally important. Participants can review aggregated feedback and reconsider their positions across successive rounds. This means that consensus is not simply measured at one point in time, but explored through a structured process of reflection and reassessment.
Transparent reporting of disagreement
Not every statement should be expected to reach consensus. Persistent disagreement may reveal uncertainty, variation in practice or areas where further evidence is needed. Reporting non-consensus transparently is therefore an important part of a credible Delphi rather than an indication of methodological failure.
Reproducibility and publication
The credibility of a Delphi study depends on a transparent and reproducible methodology. Clear reporting of participant selection, study rounds, consensus criteria and analysis allows readers to assess how the conclusions were reached and enables other researchers to build on the approach.
Publication is an important part of this process. It brings the methodology and findings into the wider scientific record, allowing independent scrutiny, citation and comparison with future research. Peer-reviewed publication also gives the work a permanent, recognised place in the evidence base.
Conclusion: from expert judgement to expert opinion evidence
Many important healthcare questions sit between established evidence and clinical implementation. In these situations, expert opinion is often needed, but not all expert opinion carries the same evidentiary weight. Structured consensus can help address these gaps with greater methodological rigour than informal expert engagement alone.
The value of expert opinion is strengthened when it is generated through a transparent, predefined and reproducible process. Structured consensus methods, particularly Delphi, provide a practical and scientifically credible approach for addressing questions that cannot yet be answered by conventional evidence alone.
Delphi does not replace clinical trials, observational research or systematic reviews. Instead, it complements them by generating robust expert opinion evidence where uncertainty remains and clinical decision-making cannot wait for further data.
At Triducive, we design and deliver consensus programmes to generate robust expert opinion evidence. Triducive has experience in publishing over 70 manuscripts in peer-reviewed journals. Our team of healthcare experts and scientific writers is ready to provide guidance and support at any stage of the evidence generation process.
Contact us for more information.