Conversations in Health Literacy Using AI Technology for Osteoarthritis Patients (CHAT-OA)

CHAT-OA: Conversations in Health Literacy Using AI Technology for Osteoarthritis Patients

The goal of this clinical trial is to assess the use of a generative artificial intelligence large language model chatbot in improving decision making factors in patients with hip and knee osteoarthritis. The main questions it aims to answer are:

Does the use of an artificial intelligence chatbot have an effect on decisional conflict and anxiety related to decision making? Are changes in decisional conflict correlated with changes in patient reported outcomes? Are changes in decisional conflict correlated with health literacy? Participants will interact with an artificial intelligence chatbot prior to their clinic visit with an orthopaedic surgeon, using a structured prompt.

Study Overview

• Status: Active, not recruiting
• Conditions: Osteoarthritis; Osteoarthritis of the Hip or Knee
• Intervention / Treatment: Other: Artificial Intelligence Chatbot

• Study Type: Interventional
• Enrollment (Estimated): 100
• Phase: Not Applicable

Contacts and Locations

Study Locations

• United States
  • California
    • Sacramento, California, United States, 95817
      • University of California Davis Health

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• Adult
• Ability to read and write in English
• All races and ethnicities
• Clinical symptoms of hip and/or knee osteoarthritis

Exclusion Criteria:

• Inability to read and write in English

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Other
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: None (Open Label)

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
No Intervention: Standard of care; This arm will undergo standard clinical care without the intervention.
Experimental: Artificial Intelligence Chatbot; This arm will interact with an artificial intelligence chatbot prior to their clinic visit.	Other: Artificial Intelligence Chatbot; The intervention is a generative artificial intelligence large language model chatbot that has a structured prompt with fill-in-the-blank style questions that the participant will complete.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Decisional Conflict Scale	16 questions with Likert scale style answers ranging from 0 to 4. The raw score is converted to a score out of 100. The score ranges from 0 (no decisional conflict) to 100 (extremely high decisional conflict).	At three time points: Within one week after the clinic visit, 1 month after the clinic visit, 6 months after the clinic visit.
Beck Anxiety Inventory	21 questions with Likert scale style answers ranging from 0 to 4. The raw score is the sum of the answers, which is converted to a score out of 100. The score ranges from 0 (low anxiety) to 63 (high anxiety).	At three time points: Within one week after the clinic visit, 1 month after the clinic visit, 6 months after the clinic visit.

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Knee injury and Osteoarthritis Outcome Score for Joint Replacement	7 questions with Likert scale style answers ranging from 0 to 4. The raw score is the sum of the answers, which is converted to a score out of 100. The score ranges from 0 (total knee disability) to 100 (perfect knee health).	At enrollment
Hip dysfunction and Osteoarthritis Outcome Score for Joint Replacement	6 questions with Likert scale style answers ranging from 0 to 4. The raw score is the sum of the answers, which is converted to a score out of 100. The score ranges from 0 (total hip disability) to 100 (perfect hip health).	At enrollment
Patient-Reported Outcomes Measurement Information System Global Health Version 1.2	10 questions with Likert scale style answers ranging from 1 to 5 and 1 question about pain with answers from 0 to 10. The questions are grouped into 4 sub-groups corresponding to physical and mental health main groups, in addition to the pain score. The raw scores are a sum of the answers for both physical and mental health range from 4 (poor physical or mental health) to 20 (best possible physical or mental health). The raw score is then converted to a t-score using the associated tables with the range 16.2 (worst possible physical or mental health) to 67.7 (best possible physical or mental health). The t-score is interpreted in comparison to the general population that has a mean of 50 and standard deviation of 10.	At enrollment

Other Outcome Measures

Outcome Measure	Measure Description	Time Frame
Visual Analog Scale Pain Score	3 questions evaluating pain and different time points, each with answers ranging from 0 (no pain) to 100 (worst pain imaginable).	At enrollment
Health Literacy Single Item Screener	Both questions 1 and 2 have Likert scale style answers from 1 (limited ability to navigate healthcare scenarios independently) to 5 (adequate ability to navigate healthcare scenarios independently).	At enrollment
Literacy in Musculoskeletal Problems	9 questions each with 4 answers, in addition to an "I don't know" selection. The total score is the sum of the correct answers ranging from 0 (poor musculoskeletal health literacy) to 9 (excellent musculoskeletal health literacy).	At enrollment
Rapid Estimate of Adult Literacy in Medicine - Short Form	7-item word recognition test to provide clinicians with a valid quick assessment of patient health literacy. The test is administered in-person by a clinician. The participant is given a list of words by a clinician and asked to pronounce the words. The clinician listens and evaluates the number of words the participant pronounces correctly and the total score is based on the number of words that are correctly pronounced. A higher score is indicative of higher health literacy. The maximum score of 7 is correlated with high health literacy equivalent to high school education or greater, and a participants ability to read most patient education materials. The minimum score of 0 is associated with low health literacy equivalent to third grade education or below, and a participants inability to read most low-literacy patient education materials.	At the first clinic visit after enrollment
Flesch-Kincaid Readability Test	This test will be used to evaluate the readability of the chatbot outputs that are presented to the participants, and will be conducted by an investigator. This test analyzes free text and determines a readability score ranging from 0 (very difficult to read) to 100 (very easy to read), and corresponds to the education level required to read the text.	Within 1 month of AI chatbot use
Simple Measure of Gobbleygook	This test will be used to evaluate the readability of the chatbot outputs that are presented to the participants and it will be conducted by an investigator. This test analyzes free text and determines a readability score with a minimum of 4.1721 and a maximum that is theoretically infinite. The number of the score is equivalent to the approximate school grade level required to read the text.	Within 1 month of AI chatbot use
5-point Likert Safety-Harm Scale	This safety evaluation tool will be used by a physician to assess chatbot outputs that are presented to the participant with a Likert scale ranging from 0 (no likelihood of harm) to 5 (high likelihood of harm).	Within 1 week of AI chatbot use
Newest Vital Sign	This measure involves participants answering a series of questions based on their interpretation of a nutrition label. It includes 5 primary questions with an additional conditional question, making up to 6 questions in total, depending on the participant's responses. This assessment is designed to evaluate health literacy, specifically the ability to understand and apply nutritional information. The score is the sum of the correct answers, ranging from 0 (indicating limited health literacy) to 6 (indicating adequate health literacy).	At enrollment

Collaborators and Investigators

• Sponsor: University of California, Davis
• Investigators: Principal Investigator: Zachary C Lum, DO, University of California, Davis

Study record dates

Study Major Dates

• Study Start (Actual): January 6, 2025
• Primary Completion (Estimated): August 1, 2026
• Study Completion (Estimated): August 1, 2026

Study Registration Dates

• First Submitted: November 22, 2024
• First Submitted That Met QC Criteria: January 10, 2025
• First Posted (Actual): January 16, 2025

Study Record Updates

• Last Update Posted (Actual): April 14, 2026
• Last Update Submitted That Met QC Criteria: April 10, 2026
• Last Verified: April 1, 2026

More Information

Terms related to this study

• Keywords: artificial intelligence; health literacy; decisional conflict
• Additional Relevant MeSH Terms: Musculoskeletal Diseases; Arthritis; Joint Diseases; Rheumatic Diseases; Osteoarthritis; Osteoarthritis, Hip
• Other Study ID Numbers: 2169004

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No