Exploring the Effectiveness of AI Generative Models for Diabetic Patients

Diabetes Mellitus, Type 2 Clinical Trial

Official title:

Exploring the Effectiveness of AI Generative Models for Diabetic Patients

Clinical Trial Details — Status: Recruiting

Administrative data

• NCT number: NCT05883072
• Other study ID #: AI Chatbot Evaluation
• Status: Recruiting
• Start date: January 5, 2023
• Est. completion date: December 31, 2026

Study information

• Verified date: June 2023
• Source: Pakistan Council of Scientific and Industrial Research
• Contact: Ghulam Mustafa
• Phone: +923322340395
• Email: mustafabme[@]gmail.com
• Is FDA regulated: No
• Study type: Observational

Clinical Trial Summary

We plan to explore the usability of Generative AI-Chatbot for Diabetic Patient

Description:

Diabetes is rapidly spreading, affecting a significant number of adults, with a staggering total of 537 million diabetic individuals. This condition gives rise to various complications that can lead to diabetic retinopathy, foot ulcers, cardiac problems, and kidney damage. However, many of these complications can be mitigated by providing patients with accurate information concerning their diet, stress management, and weight control.

The recent advancements in Generative Artificial Intelligence-based chatbots have demonstrated their efficacy as intelligent assistants across various aspects of human life. In this study, we aim to assess the effectiveness of these Language Models in assisting patients. Our research plan entails the interaction between patients and chatbots like ChatGPT, both with and without human support, followed by evaluations of these interactions by specialists. Additionally, we will gather feedback from patients regarding their experiences and perceptions of the chatbot interactions.

Recruitment information / eligibility

• Status: Recruiting
• Enrollment: 300
• Est. completion date: December 31, 2026
• Est. primary completion date: December 30, 2024
• Accepts healthy volunteers: No
• Gender: All
• Age group: 18 Years to 100 Years

Eligibility

Inclusion Criteria:

• Present physically in Pakistan

• Adults (18 years or older)

• Diabetic Patient

Exclusion Criteria:

• Adults unable to consent

• Individuals who are not yet adults (infants, children, teenagers)

• Prisoners

Related Conditions & MeSH terms

• Diabetes
• Diabetes Mellitus
• Diabetes Mellitus, Type 2
• Diabetic Retinopathy

Intervention

Behavioral:
• Exploring AI-Chatbot
All participants will be provided access to AI-Chatbot and will be asked to enquire their daily life problems related to diabetes. They will also be asked to review the replies of the Chatbot after their interaction.

Locations

Country	Name	City	State
Pakistan	CDLE, PCSIR Lbas Complex	Karachi	Sindh

Sponsors (1)

Lead Sponsor	Collaborator
Pakistan Council of Scientific and Industrial Research

Country where clinical trial is conducted

Pakistan, 

Outcome

Type	Measure	Description	Time frame	Safety issue
Primary	Usability of the Chatbot for diabetic patient	To assess the usability of the Chatbot, we will employ the mHealth App Usability Questionnaire (MAUQ) to gather feedback from patients following their interaction. Our study will utilize Table 4 of this questionnaire, which consists of three sections: ease of use, interface, and satisfaction and usefulness. Specifically, we will focus our evaluation on 10 out of the 18 questions presented in this table. The selected questions are S1, S2, S6, S7, S9, S11, S12, S13, S14, and S18. Patients will provide their responses on a scale of 1 to 5, where "1" indicates very poor and "5" denotes very good.	One time
Primary	Internet Speed	Minimum downloading speed of internet will be measured during the chat. This will be recorded in Mega bits per second.	One time
Secondary	Analyzing Chat response generated by AI Chatbot	Likert scale will be used to evaluate the chat response of Chatbot. The following parameters will be evaluated by the specialists for each response namely Clear, Complete and Correct. Clear and Completeness will be evaluated on a range of 1-5, where 1 means poor quality and 5 means very good quality response.; The correctness of each response will be further analyzed as Safe and latest. It will be evaluated in binary terms i.e. Yes or No.	One time