DELPHI PROCESS
Guidelines
should ideally be based on the current evidence available. When evidence is
lacking it is important to identify and critically appraise other levels and
sources of evidence and use this information to produce recommendations.
Although not as strong as evidence-based recommendations, so long as the
development process is transparent to the guideline users these other sources
of recommendations can be very helpful in bridging the gaps in evidence-based
guidelines.
Consensus methods
can be used where robust evidence is lacking as a means by which the evidence
available can be combined with clinical acumen and experience to provide a
practical and usable clinical tool. One way of producing a recommendation is
for a single person’s opinion to be written down based on their own personal
experiences. This recommendation would not be accepted by many clinicians as
their experiences may differ. Another way is to gather a group of experts
around a table and ask for them to write a recommendation. Group dynamics often
insure that the person with the loudest voice will have their viewpoint heard
and the recommendation may again be based on an individual’s opinion or a small
group.
A Delphi process (named after the Greek
oracle at Delphi) allows a large panel of interested
parties to have their opinions registered anonymously, analysed, and then fed
back to the same panel for further consideration. This way the most junior
opinion is given the same credit as the most senior one. When the whole panel
results are reviewed either one is at liberty to change their original opinion
or not (depending upon whether the group has helped them reconsider their
initial position or not). Consensus may be reached on specific questions after
one, two or three rounds of the Delphi panel or indeed it may never be reached. If no consensus is
reached then a specific recommendation will be difficult to make as opinion on
the matter is so divided. The Delphi panel process is probably the most robust consensus method
used in guideline development at this present time.
A Delphi panel will be used in this
guideline for recommendation formation where robust evidence is lacking. The
same Delphi panel will be used for both modules
of the guideline and will be established after the systematic review of the
literature has been performed.
The Delphi process methods will be published
before the questionnaires are sent out to the panel and the results will be
published after all rounds have been completed.