Patient Perception of Plain-Language Medical Notes Generated Using Artificial Intelligence Software: Pilot Mixed-Methods Study
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
BACKGROUND: Clinicians' time with patients has become increasingly limited due to regulatory burden, documentation and billing, administrative responsibilities, and market forces. These factors limit clinicians' time to deliver thorough explanations to patients. OpenNotes began as a research initiative exploring the ability of sharing medical notes with patients to help patients understand their health care. Providing patients access to their medical notes has been shown to have many benefits, including improved patient satisfaction and clinical outcomes. OpenNotes has since evolved into a national movement that helps clinicians share notes with patients. However, a significant barrier to the widespread adoption of OpenNotes has been clinicians' concerns that OpenNotes may cost additional time to correct patient confusion over medical language. Recent advances in artificial intelligence (AI) technology may help resolve this concern by converting medical notes to plain language with minimal time required of clinicians. OBJECTIVE: This pilot study assesses patient comprehension and perceived benefits, concerns, and insights regarding an AI-simplified note through comprehension questions and guided interview. METHODS: Synthea, a synthetic patient generator, was used to generate a standardized medical-language patient note which was then simplified using AI software. A multiple-choice comprehension assessment questionnaire was drafted with physician input. Study participants were recruited from inpatients at the University of Colorado Hospital. Participants were randomly assigned to be tested for their comprehension of the standardized medical-language version or AI-generated plain-language version of the patient note. Following this, participants reviewed the opposite version of the note and participated in a guided interview. A Student t test was performed to assess for differences in comprehension assessment scores between plain-language and medical-language note groups. Multivariate modeling was performed to assess the impact of demographic variables on comprehension. Interview responses were thematically analyzed. RESULTS: Twenty patients agreed to participate. The mean number of comprehension assessment questions answered correctly was found to be higher in the plain-language group compared with the medical-language group; however, the Student t test was found to be underpowered to determine if this was significant. Age, ethnicity, and health literacy were found to have a significant impact on comprehension scores by multivariate modeling. Thematic analysis of guided interviews highlighted patients' perceived benefits, concerns, and suggestions regarding such notes. Major themes of benefits were that simplified plain-language notes may (1) be more useable than unsimplified medical-language notes, (2) improve the patient-clinician relationship, and (3) empower patients through an enhanced understanding of their health care. CONCLUSIONS: AI software may translate medical notes into plain-language notes that are perceived as beneficial by patients. Limitations included sample size, inpatient-only setting, and possible confounding factors. Larger studies are needed to assess comprehension. Insight from patient responses to guided interviews can guide the future study and development of this technology.
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.
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 | 0.011 | 0.006 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.003 |
| Insufficient payload (model declined to judge) | 0.001 | 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 it