Powering Precision Medical Education Through AI-Generated Country-Specific Global Surgery Podcasts: A Low-or-Middle-Income-Country Case Study

Helen Please; Mike Nsubuga; Timothy Mwanje Kintu; Allan Bakesiga; Kelsey Stewart; Sergio M Navarro

doi:10.1093/bjs/znaf128.052

Abstract

Aim
Develop an AI-generated Global Surgery podcast series tailored to healthcare professionals’ needs in specific countries, focussing on Uganda as a pilot. This precision education model addresses disparities in global surgical education, historically driven by high-income countries (HICs). Podcasts were chosen for their scalability, accessibility, and self-directed learning.

Method
Building on prior data from our international trainee-based needs assessment (145 survey respondents from 18 countries; 94 from Low-and-Middle-Income-Countries, LMICs; 51 from High-Income-Countries, HICs). Generative AI (ChatGPT 4.0, OpenAI) and prompt engineering were used to develop a global surgery curriculum from survey data from Ugandan respondents (selected because Uganda had one of the highest response rates, n=31). The resultant curriculum was then adapted into a podcast format featuring country-specific themes using NotebookLM (Google), using a format of exercises, case studies, and expert interviews.

Results
The preliminary outputs of the survey and podcasts highlight diverse country-specific priorities; for HICs these include access to resources, research opportunities, and infrastructure development; for LMICs they emphasise the burden of surgical disease, use of local case studies, and foundational concepts in global surgery.

Conclusions
By tailoring content to regional needs and leveraging storytelling to foster engagement, podcasts can complement traditional curricula to address gaps in access and relevance. Limitations include current AI-generative software offering only US accents, which may not suit learners elsewhere, reflecting concerns about HICs bias in AI-generated content. Future research will validate the podcasts through focus groups with Ugandan surgical trainees, and explore speech-processing for regional accents, or blending human voice with AI-generated transcripts.

Original language	English
Number of pages	1
DOIs	https://doi.org/10.1093/bjs/znaf128.052
Publication status	Published - 19 Jun 2025
Event	ASiT Annual Conference 2025 - ICC, Belfast, United Kingdom Duration: 7 Mar 2025 → 9 Mar 2025 https://asit.org/conference/overview

Conference

Conference	ASiT Annual Conference 2025
Country/Territory	United Kingdom
City	Belfast
Period	7/03/25 → 9/03/25
Internet address	https://asit.org/conference/overview

Access to Document

10.1093/bjs/znaf128.052

Handle.net

Persistent link

Cite this

@conference{ba07bad381404145a92a38a72c430557,

title = "Powering Precision Medical Education Through AI-Generated Country-Specific Global Surgery Podcasts: A Low-or-Middle-Income-Country Case Study",

abstract = "Aim Develop an AI-generated Global Surgery podcast series tailored to healthcare professionals{\textquoteright} needs in specific countries, focussing on Uganda as a pilot. This precision education model addresses disparities in global surgical education, historically driven by high-income countries (HICs). Podcasts were chosen for their scalability, accessibility, and self-directed learning. Method Building on prior data from our international trainee-based needs assessment (145 survey respondents from 18 countries; 94 from Low-and-Middle-Income-Countries, LMICs; 51 from High-Income-Countries, HICs). Generative AI (ChatGPT 4.0, OpenAI) and prompt engineering were used to develop a global surgery curriculum from survey data from Ugandan respondents (selected because Uganda had one of the highest response rates, n=31). The resultant curriculum was then adapted into a podcast format featuring country-specific themes using NotebookLM (Google), using a format of exercises, case studies, and expert interviews. Results The preliminary outputs of the survey and podcasts highlight diverse country-specific priorities; for HICs these include access to resources, research opportunities, and infrastructure development; for LMICs they emphasise the burden of surgical disease, use of local case studies, and foundational concepts in global surgery. Conclusions By tailoring content to regional needs and leveraging storytelling to foster engagement, podcasts can complement traditional curricula to address gaps in access and relevance. Limitations include current AI-generative software offering only US accents, which may not suit learners elsewhere, reflecting concerns about HICs bias in AI-generated content. Future research will validate the podcasts through focus groups with Ugandan surgical trainees, and explore speech-processing for regional accents, or blending human voice with AI-generated transcripts.",

author = "Helen Please and Mike Nsubuga and \{Mwanje Kintu\}, Timothy and Allan Bakesiga and Kelsey Stewart and Navarro, \{Sergio M\}",

year = "2025",

month = jun,

day = "19",

doi = "10.1093/bjs/znaf128.052",

language = "English",

note = "ASiT Annual Conference 2025 ; Conference date: 07-03-2025 Through 09-03-2025",

url = "https://asit.org/conference/overview",

}

TY - CONF

T1 - Powering Precision Medical Education Through AI-Generated Country-Specific Global Surgery Podcasts

T2 - ASiT Annual Conference 2025

AU - Please, Helen

AU - Nsubuga, Mike

AU - Mwanje Kintu, Timothy

AU - Bakesiga, Allan

AU - Stewart, Kelsey

AU - Navarro, Sergio M

PY - 2025/6/19

Y1 - 2025/6/19

N2 - Aim Develop an AI-generated Global Surgery podcast series tailored to healthcare professionals’ needs in specific countries, focussing on Uganda as a pilot. This precision education model addresses disparities in global surgical education, historically driven by high-income countries (HICs). Podcasts were chosen for their scalability, accessibility, and self-directed learning. Method Building on prior data from our international trainee-based needs assessment (145 survey respondents from 18 countries; 94 from Low-and-Middle-Income-Countries, LMICs; 51 from High-Income-Countries, HICs). Generative AI (ChatGPT 4.0, OpenAI) and prompt engineering were used to develop a global surgery curriculum from survey data from Ugandan respondents (selected because Uganda had one of the highest response rates, n=31). The resultant curriculum was then adapted into a podcast format featuring country-specific themes using NotebookLM (Google), using a format of exercises, case studies, and expert interviews. Results The preliminary outputs of the survey and podcasts highlight diverse country-specific priorities; for HICs these include access to resources, research opportunities, and infrastructure development; for LMICs they emphasise the burden of surgical disease, use of local case studies, and foundational concepts in global surgery. Conclusions By tailoring content to regional needs and leveraging storytelling to foster engagement, podcasts can complement traditional curricula to address gaps in access and relevance. Limitations include current AI-generative software offering only US accents, which may not suit learners elsewhere, reflecting concerns about HICs bias in AI-generated content. Future research will validate the podcasts through focus groups with Ugandan surgical trainees, and explore speech-processing for regional accents, or blending human voice with AI-generated transcripts.

AB - Aim Develop an AI-generated Global Surgery podcast series tailored to healthcare professionals’ needs in specific countries, focussing on Uganda as a pilot. This precision education model addresses disparities in global surgical education, historically driven by high-income countries (HICs). Podcasts were chosen for their scalability, accessibility, and self-directed learning. Method Building on prior data from our international trainee-based needs assessment (145 survey respondents from 18 countries; 94 from Low-and-Middle-Income-Countries, LMICs; 51 from High-Income-Countries, HICs). Generative AI (ChatGPT 4.0, OpenAI) and prompt engineering were used to develop a global surgery curriculum from survey data from Ugandan respondents (selected because Uganda had one of the highest response rates, n=31). The resultant curriculum was then adapted into a podcast format featuring country-specific themes using NotebookLM (Google), using a format of exercises, case studies, and expert interviews. Results The preliminary outputs of the survey and podcasts highlight diverse country-specific priorities; for HICs these include access to resources, research opportunities, and infrastructure development; for LMICs they emphasise the burden of surgical disease, use of local case studies, and foundational concepts in global surgery. Conclusions By tailoring content to regional needs and leveraging storytelling to foster engagement, podcasts can complement traditional curricula to address gaps in access and relevance. Limitations include current AI-generative software offering only US accents, which may not suit learners elsewhere, reflecting concerns about HICs bias in AI-generated content. Future research will validate the podcasts through focus groups with Ugandan surgical trainees, and explore speech-processing for regional accents, or blending human voice with AI-generated transcripts.

U2 - 10.1093/bjs/znaf128.052

DO - 10.1093/bjs/znaf128.052

M3 - Conference Abstract

Y2 - 7 March 2025 through 9 March 2025