A COMPARISON OF HEALTHCARE PRACTITIONER PREFERENCE FOR INFOGRAPHICS DEVELOPED USING TRADITIONAL METHODS VERSUS GENERATIVE ARTIFICIAL INTELLIGENCE
Author(s)
Gurinder S. Sidhu, MD, MHA, MBA1, Stephen Watt, MD2, Iqra Arham, PharmD, RPh, MBAc3, Shoaib Khan, MD3, Domenick Francis, PharmD3, Suki Fadel, PharmD4, Megan Markese, MD5, Amaryllis Soto, MBS3, Hanna Thomsen, PhD3, Michael J Harvey, MD6, Lori Klein, PharmD6, Alex Busch, PhD6.
1Pfizer Inc, Los Angeles, CA, USA, 2Pfizer Inc., New York City, NY, USA, 3Pfizer Inc., New York, NY, USA, 4Pfizer Inc., Fairfax, VA, USA, 5Pfizer Inc., Collegeville, PA, USA, 6Putnam Associates, Boston, MA, USA.
1Pfizer Inc, Los Angeles, CA, USA, 2Pfizer Inc., New York City, NY, USA, 3Pfizer Inc., New York, NY, USA, 4Pfizer Inc., Fairfax, VA, USA, 5Pfizer Inc., Collegeville, PA, USA, 6Putnam Associates, Boston, MA, USA.
OBJECTIVES: Generative artificial intelligence (GenAI) is widely recognized as a transformative tool that may significantly streamline medical communication content development while reducing time and cost. In this study, we surveyed healthcare practitioners (HCPs) across the United States for their opinions on infographics developed with and without GenAI assistance to determine which would be preferred.
METHODS: Infographics aimed at three COVID-19 risk groups comprising older adults (OA), patients with chronic kidney disease (CKD), and patients with diabetes were developed using either AI assistance or traditional (non‒AI-assisted) methods for review by HCPs; objective metrics (eg, word count, visual elements) were recorded for all infographics. HCPs evaluated infographics in a blinded manner across domains of clarity, accuracy, consistency, relevance, and quality. The survey included a single-infographic evaluation, a side-by-side comparison of AI-assisted and non‒AI-assisted versions, and open-ended responses to inform preference drivers. Paired t tests and binomial tests were used to compare scores.
RESULTS: Overall, AI-assisted infographics included more words and fewer visual elements (figures and icons) compared with non‒AI-assisted versions. When HCPs (OA=210, CKD=193, diabetes=197) viewed infographics individually, the non‒GenAI-assisted version of the OA infographic scored higher than the AI-assisted version across all domains. For CKD infographics, the GenAI-assisted version scored significantly higher regarding clarity and accuracy and as equivalent across the remaining domains. For diabetes infographics, GenAI-assisted and non‒GenAI-assisted versions scored equivalently. When infographics were compared side-by-side, HCPs preferred non‒GenAI-assisted versions across all three risk groups, but parity for some domains was observed for CKD and diabetes. Preference drivers included readability, visual simplicity, and a traditional layout.
CONCLUSIONS: In this study, parity was observed in some domains, but AI-assisted infographics were generally not preferred by HCPs compared with non‒AI-assisted versions. Results indicate that GenAI tools may facilitate medical content development if workflows can be optimized according to HCP preference drivers.
METHODS: Infographics aimed at three COVID-19 risk groups comprising older adults (OA), patients with chronic kidney disease (CKD), and patients with diabetes were developed using either AI assistance or traditional (non‒AI-assisted) methods for review by HCPs; objective metrics (eg, word count, visual elements) were recorded for all infographics. HCPs evaluated infographics in a blinded manner across domains of clarity, accuracy, consistency, relevance, and quality. The survey included a single-infographic evaluation, a side-by-side comparison of AI-assisted and non‒AI-assisted versions, and open-ended responses to inform preference drivers. Paired t tests and binomial tests were used to compare scores.
RESULTS: Overall, AI-assisted infographics included more words and fewer visual elements (figures and icons) compared with non‒AI-assisted versions. When HCPs (OA=210, CKD=193, diabetes=197) viewed infographics individually, the non‒GenAI-assisted version of the OA infographic scored higher than the AI-assisted version across all domains. For CKD infographics, the GenAI-assisted version scored significantly higher regarding clarity and accuracy and as equivalent across the remaining domains. For diabetes infographics, GenAI-assisted and non‒GenAI-assisted versions scored equivalently. When infographics were compared side-by-side, HCPs preferred non‒GenAI-assisted versions across all three risk groups, but parity for some domains was observed for CKD and diabetes. Preference drivers included readability, visual simplicity, and a traditional layout.
CONCLUSIONS: In this study, parity was observed in some domains, but AI-assisted infographics were generally not preferred by HCPs compared with non‒AI-assisted versions. Results indicate that GenAI tools may facilitate medical content development if workflows can be optimized according to HCP preference drivers.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
MSR72
Topic
Methodological & Statistical Research
Topic Subcategory
Artificial Intelligence, Machine Learning, Predictive Analytics
Disease
Infectious Disease (non-vaccine), No Additional Disease & Conditions/Specialized Treatment Areas