Evidence of Educational Bias in Cognitive Screening of Adults with Sickle Cell Disease: Comparison of Available Tools and Possible Strategies for Mitigation
Bibliographic record
Abstract
Background: Cognitive impairment is a dreaded complication of sickle cell disease (SCD) that impacts quality of life, school performance and employment. In 2020, the American Society of Hematology issued a strong recommendation that clinicians supervising the care of adults with SCD conduct surveillance for cognitive impairment using simplified signaling questions (DeBaun, 2020). However, guidance on the optimal screening strategy is lacking and several available tools are biased by language and education. The Rowland Universal Dementia Assessment Scale (RUDAS) was specifically designed for cognitive screening in multicultural populations (Storey, 2004). In the general elderly population, RUDAS is less biased by education than the Montreal Cognitive Assessment (MoCA) (Naqvi, 2015). Hypothesis: In adults with SCD, performance on the RUDAS is less influenced by educational attainment when compared to the MoCA. Our primary aim was to estimate the prevalence of suspected cognitive impairment using RUDAS and MoCA in adult SCD patients. The secondary aims were to examine for the presence of educational bias and to develop mitigation strategies in case of such a bias. Methods: Study design: cross-sectional study at UMGRR clinic at Henri Mondor Hospital, Créteil (France). Inclusion criteria: out-patients ≥18 years-old; all SCD phenotypes. Exclusion criteria: inability to obtain informed consent and/or follow study instructions. Intervention: Cognitive screening was performed using the RUDAS (translated to French by Philippe Desmarais), MoCA (third alternative version) and an additional visuospatial task of copying overlapping triangles (from the French BEC96 assessment). RUDAS and MoCA scores <28 and <26, respectively, were considered suggestive of cognitive impairment per previous studies (Basic, 2009 and Nasredinne, 2005) and patients were referred for definite neuropsychological evaluation. Survey on demographics and screening for depression and anxiety using Hospital Anxiety Depression Scale (HADS) were completed by the participants. Educational attainment was scored based on the number of years of schooling for the highest completed diploma. Statistical plan: linear regression was performed to identify possible associations between RUDAS, MoCA and social determinants of health. Results: Among the first 45 consecutive adult SCD patients undergoing routine cognitive screening, the median age was 39 (range 19-67). RUDAS and MoCA scores suggestive of mild cognitive impairment were found in 33/45 (73.3%) and 29/45 (64.4%) participants, respectively. There was a strong correlation between both tests (r=0.48, p=0.001). Both RUDAS and MoCA scores increased significantly with increasing level of education (r=0.36, p=0.015 and r=0.39, p=0.007, respectively), but were not significantly influenced by the HADS score. RUDAS and MoCA test items most biased by education were visuoconstructional tasks. Tasks assessing executive functioning and language were also biased in MoCA. Substituting the 3D visuospatial task of the RUDAS by a 2D task reduced the educational bias (r=0.20, p=0.045). Adding 1 point for highest level of education £ 12 years after kindergarten did significantly mitigate the effect of education on the RUDAS but only partially for the MoCA (r=0.23, p=0.131 and r=0.30, p=0.047). Conclusions: Overall, these results suggest there is an educational bias in the neurocognitive screening of adult SCD patients using available tools such as the RUDAS and MoCA. Although RUDAS was less biased overall, visuospatial assessment remained biased. The task often considered more "culture-fair" is still subject to the impact of educational potential (Statucka, 2019). We provide different strategies to mitigate education bias when assessing with RUDAS. Thus, the RUDAS adjusted by the educational level allows to systematically identify SCD patients in need of comprehensive neurocognitive testing. Prospective validation is ongoing. Disclosures Forté: Canadian Hematology Society: Research Funding; Pfizer - Global Medical Grants: Research Funding. Soulieres:Novartis: Research Funding; BMS: Membership on an entity's Board of Directors or advisory committees. Kuo:Pfizer: Consultancy, Research Funding; Celgene: Consultancy; Alexion: Consultancy, Honoraria; Novartis: Consultancy, Honoraria; Bioverativ: Membership on an entity's Board of Directors or advisory committees; Agios: Consultancy, Membership on an entity's Board of Directors or advisory committees; Bluebird Bio: Consultancy; Apellis: Consultancy. Bartolucci:Roche: Consultancy; Innovhem: Other; AGIOS: Consultancy; Bluebird: Consultancy; Emmaus: Consultancy; Addmedica: Research Funding; Fabre Foundation: Research Funding; Novartis: Research Funding; Bluebird: Research Funding; GBT: Consultancy; ADDMEDICA: Consultancy; HEMANEXT: Consultancy; Novartis: Consultancy.
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 | 0.074 | 0.136 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.005 | 0.004 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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".