AI Tools for Dyslexia

(PRECISE Trial)

Recruiting at 1 trial location

Overseen By Soaptarshi Paul, MD

Age: 18+

Sex: Any

Trial Phase: Academic

Sponsor: Montefiore Medical Center

Disqualifiers: Not internal medicine physician

No Placebo Group All trial participants will receive the active study treatment (no placebo)

Compensation: Varies

Number of Visits: 1

Duration: 90 minutes

56 Participants Needed

What You Need to Know Before You Apply

What is the purpose of this trial?

This trial tests two AI tools, GPT-4 and OpenEvidence, to evaluate their effectiveness in aiding doctors with disease diagnosis and management. It serves as a direct comparison to determine which tool better assists doctors in their work. The trial aims to enhance the use of these tools in real-life medical settings. Internal medicine residents and attending physicians seeking to improve their decision-making skills may find this study suitable. As an unphased trial, participants have the opportunity to contribute to pioneering research that could transform medical decision-making.

Will I have to stop taking my current medications?

The trial information does not specify whether you need to stop taking your current medications.

What prior data suggests that these AI tools are safe for clinical decision support?

Research has shown that GPT-4, also known as ChatGPT, and OpenEvidence serve as tools for gathering information and aiding decision-making in medical settings. GPT-4 provides valuable support for individuals with conditions like dyslexia. However, experts caution that using ChatGPT for safety advice can be risky.

OpenEvidence, in contrast, compiles and disseminates information from reliable medical studies. It relies on trusted sources to offer evidence-based advice, ensuring the safety of its recommendations. Although specific studies have not highlighted problems with either tool, they primarily assist doctors rather than operate independently, suggesting general safety.

In summary, both GPT-4 and OpenEvidence are well-regarded in medical settings, with no major safety issues reported. ^{1 2 3 4 5}

Why are researchers excited about this trial?

Researchers are excited about using AI tools like GPT-4 and OpenEvidence for dyslexia because they offer a fresh approach to a condition traditionally managed through educational interventions and specialized tutoring. Unlike conventional methods, GPT-4 uses advanced natural language processing to provide personalized reading support and feedback, potentially adapting to individual learning styles more effectively. OpenEvidence, on the other hand, serves as an innovative research tool, allowing users to access and process information in a way that can enhance understanding and retention. These AI-driven methods might offer quicker, tailored assistance for dyslexia, promising to make learning more accessible and efficient.

What evidence suggests that this trial's treatments could be effective for dyslexia?

This trial will compare the effectiveness of GPT-4 and OpenEvidence in supporting individuals with dyslexia. Research has shown that GPT-4 helps students with dyslexia better understand their reading, leading to positive changes in behavior and emotions. Specifically, studies have demonstrated that students using GPT-4 improve their reading skills. Participants in this trial may use GPT-4 as their research tool.

OpenEvidence, used by participants in another arm of this trial, is designed to simplify medical information, aiding doctors in decision-making. Although specific results for dyslexia are not available for OpenEvidence, computer-based treatments in general have greatly improved reading and spelling for people with dyslexia. Both tools offer new ways to support individuals with dyslexia and their caregivers in various situations. ^{5 6 7 8 9}

Who Is on the Research Team?

Shitij Arora, MD

Principal Investigator

Montefiore Medical Center

Are You a Good Fit for This Trial?

This trial is for internal medicine physicians who are interested in how AI tools can help with clinical decisions. It's not specified who can't join, but typically those without a medical background or familiarity with such tools might be excluded.

Inclusion Criteria

Internal medicine residents

Internal medicine attending physicians

Exclusion Criteria

Not meeting Inclusion Criteria

Timeline for a Trial Participant

Screening

Participants are screened for eligibility to participate in the trial

1-2 weeks

Treatment

Participants complete a 90-minute Zoom session solving clinical cases using either OpenEvidence or ChatGPT

90 minutes

1 visit (virtual)

Follow-up

Participants' clinical reasoning performance and decision confidence are evaluated post-session

15 minutes

What Are the Treatments Tested in This Trial?

Interventions

• GPT-4
• OpenEvidence

Trial Overview The study compares two AI systems: OpenEvidence and GPT-4, to see which one better supports doctors making diagnoses and treatment choices. Physicians will use these tools during the trial to evaluate their effectiveness.

How Is the Trial Designed?

Group I: OpenEvidence Active Control 1 Intervention

Participants in this arm will use OpenEvidence as their research tool

Group II: ChatGPT Active Control 1 Intervention

Patients in this arm will use Chat-GPT as their research tool

Who Is Running the Clinical Trial?

Montefiore Medical Center

Lead Sponsor

Trials

468

Recruited

599,000+

Citations

1. journals.sagepub.com journals.sagepub.com/doi/10.1177/20965311251358269

A Study on Using ChatGPT to Help Students With Dyslexia ... The results indicate that, with the assistance of CHATTING, students with dyslexia demonstrate positive improvement in behavioral, emotional ...

2. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC12385150/

The Effectiveness of Artificial Intelligence-Based ... A ChatGPT-based intervention for Arabic-speaking children with dyslexia (n = 60) resulted in a significant improvement in reading comprehension ...

3. dl.acm.org dl.acm.org/doi/full/10.1145/3706468.3706483

Got It! Prompting Readability Using ChatGPT to Enhance ... The results reveal that ChatGPT significantly improves the readability of the government letters with a 20% increase in comprehensibility scores ...

4. on.dystinct.org on.dystinct.org/chatgpt-learning-disability-assistive-technology-expert-opinion/

ChatGPT for students with Dyslexia? Expert Opinion Based on the concerns addressed above, ChatGPT may not be the most beneficial for individuals with dyslexia.

5. biyoklinikder.org biyoklinikder.org/TIPTEKNO24-Bildiriler/03.pdf

Assessing ChatGPT's Accuracy in Dyslexia Inquiry Building on prior work showcasing chatbots' usefulness in healthcare, this study innovatively tests the OpenAI Assistant's efficacy as a dyslexia inquiry tool.

6. sciencedirect.com sciencedirect.com/science/article/pii/S0925753523001868

The risks of using ChatGPT to obtain common safety- ... The experts concluded that there is potential for significant risks when using ChatGPT as a source of information and advice for safety-related issues.

7. rockymountainada.org rockymountainada.org/resources/research/chatgpt-and-disability-benefits-concerns-and-future-potential/

ChatGPT and Disability: Benefits, Concerns, and Future ... ChatGPT allows individuals with disabilities to receive assistance in the employment search in a safe and non-judgmental environment, which may be a barrier for ...

8. towardsdatascience.com towardsdatascience.com/gpt-4v-has-directional-dyslexia-2e94a675bc1b/

GPT-4V Has Directional Dyslexia Since autumn 2023, GPT-4 Turbo has gained "vision" capabilities, meaning it accepts images as input and it can now directly participate in VQA ...

9. dyslexic.ai dyslexic.ai/p/newsletter-254-chatgpt-gamechanging-accommodation-neurodivergent-thinkers

Newsletter 254: ChatGPT as a Game-Changing ... AI tools like ChatGPT aren't just helpful; they're essential accommodations for neurodivergent and dyslexic individuals.