Multicenter, Prospective Study Evaluating the Accuracy of an AI-based Smartphone Application for Non-invasive Anemia Detection

Overview

Title: Multicenter, prospective study evaluating the accuracy of an AI-based smartphone application for non-invasive anemia - AnemiAI trial Short Introduction: Anemia is a common clinical condition, often underdiagnosed, with significant impact on surgical, vulnerable, and underserved populations with limited access to laboratory testing. The NiADA technology uses artificial intelligence to estimate hemoglobin levels from images of the lower palpebral conjunctiva, offering a non-invasive, rapid, and accessible alternative to conventional diagnostics. Purpose: To evaluate the diagnostic accuracy of the NiADA application for non-invasive anemia detection, compared to the standard laboratory method (venous blood draw). Study Design: Following informed consent, participants will undergo clinical data collection and non-invasive hemoglobin estimation using the NiADA application, operated by a trained healthcare professional. The lower palpebral conjunctiva will be imaged using a compatible smartphone camera, and the app result will be recorded immediately. Within four hours of the app assessment, a venous blood sample will be collected for laboratory hemoglobin analysis, following institutional protocols. Both results - from the app and laboratory - will be recorded on the REDCap platform. Additionally, patients will complete a brief acceptability questionnaire. Execution time, technical errors, and task completion rates will also be monitored. All centers will follow a standardized data collection protocol to ensure methodological consistency and data integrity. Study Population Sample Size (n) Adult patients (≥18 years), hospitalized or in outpatient care, with a clinical request for hemoglobin testing. Inclusion Criteria Age ≥18 years Any gender or ethnicity Clinical indication for venous hemoglobin testing