P257 Real-world evidence of TNF inhibition in axial spondyloarthritis: can we generalise the results from clinical trials?
Bibliographic record
Abstract
Abstract Background The development and utility of management guidelines assumes that clinical trial findings are generalisable. Seldom is data available to test this. We aimed to determine, in the British Society for Rheumatology Biologics Register for Ankylosing Spondylitis (BSRBR-AS), the proportion of patients commencing TNF inhibition (TNFi) that would/would not have been eligible for clinical trials that led to TNFi treatment guidelines, and whether treatment response differed between the trials and this real-world population. Methods Biologic-naïve spondyloarthritis patients were recruited from across Great Britain. Data was obtained from clinical records, and participants completed postal questionnaires. Participant characteristics were extracted from the placebo-controlled randomised trials in the NICE Health Technology Assessment: TNF-alpha inhibitors for ankylosing spondylitis and non-radiographic axial spondyloarthritis (TA383). Descriptive statistics were used to examine differences, including treatment response (ASAS-20), between BSRBR-AS participants who would/would not have been eligible for the clinical trials, and the trial participants. Results 816/2420 (34%) BSRBR-AS participants were commencing TNFi. They were younger (mean age 44 versus 50yrs) with shorter disease duration (15 versus 22yrs), more active disease (BASDAI 6.4 versus 4.0), and poorer function (BASFI 6.2 versus 3.8). Fourteen clinical trials were identified. Compared to trial populations, fewer BSRBR-AS participants were male (67% versus 71%; difference: -4.1% (95%CI: -7.8%, -0.4%)) and fewer were HLA-B27 positive (76% versus 82%; difference: -6.6% (-10.6%, -2.6%)). BSRBR-AS participants were 6yrs older than trial participants, with longer symptom duration. They reported similar disease activity (BASDAI: 6.4 versus 6.2; difference 0.2 (-0.3, 0.7)), although significantly poorer function (BASFI: 6.2 versus 5.1; difference 1.1 (0.5, 1.8)) and spinal mobility (BASMI: 4.2 versus 3.3; difference 1.0 (0.8, 1.1)). Only 333 (41%) of BSRBR-AS participants commencing TNFi would have been eligible for any of the relevant trials. Ten trials reported ASAS20 response criteria, and 864/1401 participants reported a positive treatment response (61.7%). Follow-up data was available for 318 (39%) BSRBR-AS participants, of whom 163 (51.3%) achieved an ASAS20 treatment response (difference: 10.4% (4.4%, 16.5%)). There was no difference in ASAS20 response between those who would/would not have been eligible for clinical trials (50% versus 52%; difference 2.0% (-9.4%, 13.4%)). Conclusion In this real-world population, although the likelihood of meeting response criteria was unrelated to factors determining trial eligibility, the proportion of patients responding to TNFi was lower than in the clinical trial literature. Could this be explained by selection bias? Although fewer BSRBR-AS participants provided follow-up data than in the clinical trials, to account for the observed difference participants lost to follow-up would have to be one-third more likely to achieve ASAS20 response than those who provided follow-up data. We believe this is unlikely. These findings have important implications for the generalisability of trial results, and also for the cost-effectiveness of TNFi agents. Disclosures G.T. Jones: Grants/research support; GTJ is/was a grant holder for research funded by Pfizer, AbbVie, UCB and Celgene., GTJ is/was involved in research that received financial support from Novartis. L.E. Dean: Grants/research support; LED is/was involved in research that received financial support from Pfizer, AbbVie, UCB and Novartis. E. Pathan: None. G.J. Macfarlane: Grants/research support; GJM is/was a grant holder for research funded by Pfizer, AbbVie, UCB and Celgene., GJM is/was involved in research that received financial support from Novartis.
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 | 0.294 | 0.665 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.008 | 0.009 |
| Bibliometrics | 0.007 | 0.008 |
| Science and technology studies | 0.001 | 0.007 |
| Scholarly communication | 0.012 | 0.009 |
| Open science | 0.007 | 0.006 |
| Research integrity | 0.010 | 0.009 |
| Insufficient payload (model declined to judge) | 0.018 | 0.004 |
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; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
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".