Enhancing Multilingual Patient Education: ChatGPT's Accuracy and Readability for SSNHL Queries in English and Spanish
Bibliographic record
Abstract
Abstract Objective This study investigates ChatGPT's accuracy, readability, understandability, and actionability in responding to patient queries on sudden sensorineural hearing loss (SSNHL) in English and Spanish, when compared to Google responses. The objective is to address concerns regarding its proficiency in addressing medical inquiries when presented in a language divergent from its primary programming. Study Design Observational. Setting Virtual environment. Methods Using ChatGPT 3.5 and Google, questions from the AAO‐HNSF guidelines were presented in English and Spanish. Responses were graded by 2 otolaryngologists proficient in both languages using a 4‐point Likert scale and the PEMAT‐P tool. To ensure uniform application of the Likert scale, a third independent evaluator reviewed the consistency in grading. Readability was evaluated using 3 different tools specific to each language. IBM SPSS Version 29 was used for statistical analysis using one‐way analysis of variance. Results Across both languages, the responses displayed a native‐level language proficiency. Accuracy was comparable between sources and languages. Google's Spanish responses had better readability (effect size 0.35, P < .001), while Google's English responses were more understandable (effect size 0.67, P = .018). ChatGPT's English responses demonstrated the highest level of actionability (60%), though not significantly different when compared to other sources (effect size 0.47, P = .14). Conclusion ChatGPT offers patients comprehensive and guideline‐conforming answers to SSNHL patient medical queries in the 2 most spoken languages in the United States. However, improvements in its readability and understandability are warranted for more accessible patient education.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".