REAL-WORLD APPLICATION OF THE SLE RISK PROBABILITY INDEX TO TRIAGE ANA POSITIVE PATIENTS: A PRAGMATIC RETROSPECTIVE SINGLE-CENTER STUDY
Bibliographic record
Abstract
PV141 / #368 Poster Topic: AS17 - Miscellaneous Background/Purpose Antinuclear antibodies (ANA) testing is crucial for identifying patients with systemic lupus erythematosus (SLE). However, its specificity is suboptimal and may result in unnecessary referrals to Rheumatology. Triage tools that enable early differentiation between SLE and non-SLE among ANA-positive patients could expedite initial assessment and improve patient outcomes,[1] making them particularly valuable in lupus referral centers. The SLE Risk Probability Index (SLERPI) is a simple-to-use, machine-based model utilizing 8 clinical and 6 laboratory variables, designed to assist SLE diagnosis, with potential clinical utility for this purpose.[2] Methods We retrospectively reviewed consecutive referrals that were received between January 1st and October 31st, 2024 to rule out SLE. Referrals for patients younger than 16 years old or without documented positive ANA (1:80 or greater by immunofluorescence) or with a prior diagnosis of connective tissue disease were excluded. A single reviewer analyzed each consultation request from the referring clinicians and scored SLERPI items individually, using only the information included in the referral. Based on this scoring, the total SLERPI score was classified as positive or negative, using a cut-off of >7 points, as previously described. Patients were classified as either non-SLE or diagnosed with SLE based on clinical documentation from their charts. Additional collected data included initial triage priority (Urgent, Semi-Urgent, Routine) assigned by the referral center, and a retrospective triage reassessment by the principal reviewer. Descriptive statistics were used. Results Fifty referrals meeting the selection criteria were identified. Of these, 44 were from general practitioners and 6 from specialists, with 90% showing positive ANA by immunofluorescence at titers ≥1/160. Initial triage classified 20% of cases as Semi-Urgent and 80% as Routine. On average, 8.7 out of 14 SLERPI items were undocumented by the referring provider. Only 4 items—Arthritis, Platelet levels, Leukocyte levels, and Proteinuria—were documented in ≥50% of referrals, while 8 out of 14 items were documented in ≤20% of referrals. Among the 50 charts reviewed, 5 had a positive SLERPI score (>7), while 45 were negative. Four patients were diagnosed with SLE, and 46 were deemed non-SLE. The preliminary analysis demonstrated a positive predictive value of 60% for a SLERPI score >7 to identify SLE patients, with a high negative predictive value of 98%, for an 8% SLE prevalence (Figure 1). SLERPI score >7 showed a positive likelihood ratio of 17.3 and a negative likelihood ratio of 0.26 for SLE identification. A SLERPI score >7 would have reclassified 6 non-SLE patients as Routine and prioritized 1 SLE case as Semi Urgent (Figure 2). The SLERPI score’s negative likelihood ratio was 0.13 for identifying patients ultimately classified as Routine, including those initially marked as Routine who remained so after review and those reclassified from Semi-Urgent to Routine. Figure 1. Figure 2. Conclusions SLERPI shows potential as an effective triage tool for distinguishing SLE from non-SLE among ANA-positive patients. With a high negative predictive value of 98% for SLE diagnosis and a negative likelihood ratio of 0.13, the SLERPI score could help prioritize referrals. This may, in turn, accelerate early rheumatology assessment in newly diagnosed SLE patients, thus improving outcomes. This approach might also allow for more efficient resource allocation in lupus referral centers. References: [1.] Adamichou C. Ann Rheum Dis 2021;80(6):758-66. [2.] Floris A. Arthritis Care Res 2020;72:1794-9.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.011 | 0.027 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".