Real-World Application of an AI-Assisted Digital Workflow for Clinical Observational Data Collection in Dermatology...

Area of Science:

• Medical Informatics
• Dermatology Research
• Artificial Intelligence in Healthcare

Background:

• Traditional paper-based data collection in dermatology residency research is inefficient and error-prone.
• Electronic data capture (EDC) platforms are often expensive and difficult to implement in resource-limited settings.
• There is a need for cost-effective and accessible solutions to streamline research data management.

Purpose of the Study:

• To evaluate a low-cost, AI-assisted workflow for dermatology residency research data collection.
• To assess the impact of this digital workflow on data entry time and accuracy compared to traditional methods.
• To determine the feasibility of using widely available tools like Google Workspace and AI for research data management.

Main Methods:

• A prospective, single-center simulation study involving 100 hypothetical patient records.
• Comparison of paper-based data collection with manual digitization versus a digital workflow using Google Forms, Google Sheets, and AI validation (ChatGPT).
• Analysis of mean entry time per record, error rates per field, and absolute risk reduction using statistical tests.

Main Results:

• The AI-assisted workflow reduced mean entry time per record from 24.4 minutes to 7.5 minutes, a significant decrease (p<0.001).
• Error rates dropped from 8.54% to 2.39%, representing a 72% relative reduction in errors (p=0.013).
• No new errors were introduced by the digital workflow.

Conclusions:

• An AI-assisted workflow using Google Workspace substantially decreases time and errors in dermatology research data collection.
• This approach is affordable, quick to implement, and scalable for institutions with limited resources.
• Adopting such digital workflows can enhance research productivity, data quality, and resident well-being, fostering a stronger research culture.