Background/Purpose:

EULAR and ACR have
jointly funded a project to improve existing classification criteria for SLE;
this abstract reports on the early phase of this project.  Classification criteria
are being developed using multicriteria decision
analysis methods.  The goal of the initial phase of this project was to
generate a broad set of items potentially useful for the classification of SLE
and select a list of candidate items for use in forced-choice decision
analysis.  The aims were to include items that
encompass a broad spectrum of SLE, including patients whose disease is in the
early stage as well as those in the late stage, and that are in accordance with
criteria used for diagnosis of SLE in clinical practice.

Methods:

Items of potential utility for SLE classification criteria
were identified by a large international group of lupus clinicians in
independent exercises via response to online surveys.  In the first
survey, respondents were asked to evaluate existing ACR and SLICC
classification items and invited to propose additional, nontraditional clinical
items.  Submitted novel candidate items were summarized,
de-duplicated, and the number of total items reduced via a Delphi process.

Results:

135 individuals were invited to participate in each survey.
In the first survey, 120 (89% response rate) proposed a total of 196 unique
novel items for consideration.  Combining items with significant overlap
resulted in 159 items, including six that each contained one or more closely
related modifiers.   Nine items that are difficult to routinely
assess in current clinical practice (eg, “elevated
MCP1”) were removed from consideration.  The resulting 150 items were
evaluated in a second online survey.  Items achieving a threshold
significance score were retained for inclusion in SLE case scenarios to be
evaluated by lupus clinicians in the next phase of the project. 

Conclusion:

This approach to the development of SLE classification
criteria represents application of state-of-the-art methods for criteria
development.   The resulting core item list will be assessed for
discriminative characteristics in specific clinical scenarios to create a
weighted item list. The performance (sensitivity, specificity) of the weighted
item list will then be compared to existing classification criteria using data
collected from SLE patients and controls.