Comparing traditional and AI-enhanced strategies for developing patient decision aids: a multiple case study
Bibliographic record
Abstract
OBJECTIVES: Our objective was to develop and test prompts designed to generate balanced, evidence-based information from artificial intelligence (AI) for the development of patient decision aid (DA) content. We compared the outputs of this AI-enhanced strategy with those produced by an experienced human team using a traditional development approach. METHODS: We conducted a comparative, mixed-methods, multiple-case study, with each case being a DA. Eight DAs were randomly selected from the Ottawa Inventory, stratified by author type (commercial, academic, public institution, professional association). We then followed a systematic process involving two researchers working independently. One researcher described the topics of the selected DAs and extracted their content by listing the available options with their benefits and harms. The other researcher-blind to the DA-used the topic description to generate AI-enhanced DA content by iteratively refining the prompt structures based on the International Patient Decision Aids Standards until the generated content stabilised. Quantitative analyses compared the number of options, benefits and harms generated by the traditional and AI-enhanced strategies, while qualitative analyses examined the differences in content. RESULTS: The selected DAs targeted different populations (older adults, women, the general population, children) and were produced in Canada, the UK, the USA or Australia. One type of DA (n=6) focused on a specific option (eg, whether to get vaccinated against COVID-19), the other (n=2) focused on improving an outcome (eg, treating attention-deficit/hyperactivity disorder symptoms). For option-focused DAs, 66% of the benefits/harms were generated by the AI-enhanced strategy only and 6.2% by the traditional strategy only. For outcome-focused DAs, 47% of the options were generated by the AI-enhanced strategy only, and 4% by the traditional strategy only. An evidence search confirmed that the options generated only by the AI-enhanced strategy were indeed beneficial, ruling out hallucinations. However, the AI-enhanced strategy did not suggest optimal combinations. Qualitative analysis showed that AI-enhanced content was generally richer. CONCLUSIONS: This study provides practical guidance on leveraging AI to improve the efficiency of DA development and improve their quality.
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
Direct model labels (unvalidated)
Per-model category and study-design labels from the labeling rounds. They are machine output, unvalidated, and the disagreement between models ships as data. No study design here is MEDLINE-validated yet.
| Model arm | Categories | Study design | Confidence |
|---|---|---|---|
| gemma | no category Domain: not available · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Qualitative | low |
| gpt | no category Domain: not available · Genre: Empirical About the Canadian research system: no · About a Canadian topic: no | Other design | low |
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 | 0.002 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 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, unvalidatedLabeled directly by 2 models reading the full record.
The models disagree on parts of this classification; every voice is preserved in the section at the end of the page.
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".