Page 64 - Read Online

Liu et al. Developing treatment guidance for myasthenia gravis

in the journal of Neurology. [1] without disagreement, instead, if the appropriate scores
have a median of 4-6, it could be either uncertain (if the
RAM was developed to combine the best available actual scores are within 4-6) or disagreement (if some
scientific evidence, even when the randomized scores are in 7-9 region and others are in 1-3 region),
controlled trials (the golden standard in evidence-based and the recommendation must be revised for future
medicine) are not available or cannot provide enough consensus. Similarly, if the appropriate scores have a
detailed guidance for everyday clinical practice. The median of 1-3, the recommendation is considered to
RAM is based on the collective judgment of experts with be inappropriate without disagreement.
the common objective to release statements regarding
the appropriateness of following a procedure, using It is also possible to reach an agreement about a
the multiple-rounds Delphi polling sessions to assess recommendation being appropriate, inappropriate or
the treatment rationality. [5] even uncertain. Facilitator manages voting but does
not vote. Cost and availability considerations are NOT
The procedure of RAM used at this stage and all options are assumed to be
There were 2 major concepts during the RAM: affordable and freely available.
appropriateness and agreement. The median
rating at each round is the appropriateness score RESULTS
and the summary of the appropriateness scores
for each recommendation is its agreement score. The application of RAM
Appropriateness ratings are collected from each panelist RAM is useful for rare cases with low morbidity, lack
to quantitatively assess the relative harm or benefit of a
particular intervention. Each recommendation is rated of RCT clinical research, pooling available research
on a 9 point scale: 1-3 are extremely inappropriate to evidence and expert experience to draw up a guideline,
inappropriate (i.e. risks > benefit); 4-6 are uncertain
(i.e. risks ≈ benefit); 7-9 are appropriate to extremely
appropriate (i.e. benefit > risks) [Table 1].

The detailed procedure of RAM is listed in the flow
chart of Figure 1.

Disagreements and uncertainty ratings assist in
determining “grey” areas for future research. Panel
consensus is NOT forced. Rather, the degree of
agreement is used to define the strength of the
recommendations. Agreement: ≤ 3/13 panelists or
≤ 4/14 panelists rate the recommendation outside
the 3 point region containing the median score of
appropriateness. Disagreement: ≥ 4/13 panelists or
≥ 5/14 panelists rate the recommendation in the 1-3
region and the same number in the 7-9 region of
appropriateness.

Therefore, if the appropriate scores on a particular Figure 1: Flow chart of the RAND/UCLA rationality approach to
recommendation have a median of 7-9, the develop a consensus (adopted from Dr. Sanders’ presentation with
recommendation is considered to be appropriate permission)

[5]
Table 1: RAND/UCLA appropriateness method rationality and its meaning of the 9-point score (adopted from )
Inappropriate 1 Extremely inappropriate (risk greatly exceed benefits)
2 Moderately inappropriate
3 Slightly inappropriate
Uncertain 4 May be inappropriate
5 Uncertain/equivocal (benefit and risk about equal)
6 May be appropriate (expected health benefits to an average patient exceed the expected health
risks by a sufficiently wide margin to make intervention worthwhile and the intervention is superior to
alternatives, including no intervention)
Appropriate 7 Slightly appropriate
8 Moderately appropriate
9 Extremely appropriate (benefit greatly exceed risks)
56 Neuroimmunology and Neuroinflammation ¦ Volume 4 ¦ March 30, 2017