Enhancing Diagnostic Accuracy of Ophthalmological Conditions with Complex Prompts in GPT-4: A Comparative Analysis of Global and LMIC-Specific Pathologies (Preprint)
Bibliographic record
Abstract
Background: The global incidence of blindness has continued to increase, despite the enactment of a Global Eye Health Action Plan by the World Health Assembly. This can be attributed, in part, to an aging population, but also to the limited diagnostic resources within low- and middle-income countries (LMICs). The advent of generative artificial intelligence (AI) within health care could pose a novel solution to combating the prevalence of blindness globally. Objective: The objectives of this study are to quantify the effect the addition of a complex prompt has on the diagnostic accuracy of a commercially available LLM, and to assess whether such LLMs are better or worse at diagnosing conditions that are more prevalent in LMICs. Methods: Ten clinical vignettes representing globally and LMIC-prevalent ophthalmological conditions were presented to GPT-4-0125-preview using simple and complex prompts. Diagnostic performance metrics, including sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), were calculated. Statistical comparison between prompts was conducted using a chi-square test of independence. Results: The complex prompt achieved a higher diagnostic accuracy (90.1%) compared to the simple prompt (60.4%), with a statistically significant difference (χ2=428.86; P<.001). Sensitivity, specificity, PPV, and NPV were consistently improved for most conditions with the complex prompt. The simple prompt struggled with LMIC-prevalent conditions, diagnosing only 1 of 5 accurately, while the complex prompt successfully diagnosed 4 of 5. Conclusions: The study established that overall, the inclusion of a complex prompt positively affected the diagnostic accuracy of GPT-4-0125-preview, particularly for LMIC-prevalent conditions. This highlights the potential for LLMs, when appropriately tailored, to support clinicians in diverse health care settings. Future research should explore the generalizability of these findings across other models and specialties.
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 | 0.008 | 0.053 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 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".