Assessment of systemic lupus erythematosus diagnoses within Quebec's health administrative databases
Bibliographic record
Abstract
Background: Systemic lupus erythematosus (SLE) is a chronic, relatively uncommon autoimmune disease that has a relapsing-remitting course, with clinical manifestations in various organ systems (cutaneous, renal, and other). To control disease, immunosuppressive drugs are often required. Health administrative databases are useful for studying SLE because of their wide population coverage, and could potentially be used to study SLE incidence, prevalence, clinical manifestations, and medication use. However, because the diagnoses in these administrative databases are not necessarily clinically confirmed, SLE case ascertainment is a methodological challenge. First, some of the methodological issues were examined in this thesis. Second, clinical manifestations and the association between early antimalarial drug use and future renal manifestations were examined in a cohort of SLE patients. Methods: The initial SLE case definition was a previously-used algorithm that identified subjects as having SLE if they met one of the following criteria: one SLE hospital discharge code, one rheumatologist SLE claim and/or two SLE non-rheumatologist claims at least eight weeks apart but within two years. Alternative algorithms were formed by modifying one or more of the initial algorithm's parameters. Incidence and prevalence estimates were determined using each alternative algorithm and compared to the initial estimates. The effect of using different data period lengths for detecting patients was also examined. Kaplan-Meier (K-M) analyses were performed to assess documentation of clinical SLE manifestations and use of selected immunosuppressant medications, within an incident SLE cohort identified by the initial algorithm (described above). The observation interval began four years prior to SLE diagnosis and continued up to eight years after SLE diagnosis. Cox proportional hazards regression analyses were used to examine the association between early antimalarial drug use and renal manifestations. Results: With the initial algorithm, the 1998 yearly incidence was 6.0 cases per 100,000 (95% confidence interval (CI), 5.5–6.6). When parameters from the initial algorithm were changed, the 1998 incidence varied to between 4.4 and 7.4/100,000. The prevalence also changed from 65.5/100,000 (95% CI: 63.7–67.4) with the initial algorithm, to between 47.8–79.1/100,000 with the alternate algorithms. When the length of the data period changed from fifteen years to five years, the 2001 yearly incidence was overestimated by 38.3% (5.7/100,000 initially and 7.9/100,000 with only five years of data) and the prevalence was underestimated by 29.9% (the new estimate being 46.0/100,000, 95% CI: 44.4–47.5).Over-all, 66.2% (95%CI: 63.4–68.9%) of incident patients (within the SLE cohort assembled using the initial algorithm) had evidence of at least one SLE manifestation within the period under examination. The most common manifestation was cutaneous involvement, present in 30.0%. Within the sub-cohort of incident SLE patients covered by RAMQ drug insurance, 87.2% (95% CI: 84.2–90.3%) had received at least one of the medications under study, by the end of the study interval. No association was found between early antimalarial drug use and subsequent renal manifestations.Conclusion: Varying the case definition and data period can change incidence and prevalence estimates considerably, so all features, including the time period in which the data spans, should be selected carefully and explicitly stated. The majority of incident SLE patients had evidence of SLE manifestations or used medications which would provide possible confirmation of SLE case status. This additional information can be used in future health services administrative database research to understand SLE, and help compensate for the databases' lack of clinical confirming data.
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 | 0.003 | 0.018 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.004 | 0.009 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
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".