The app analyzes your fingernail beds to reveal hemoglobin levels. (Alexander Sobol/Shutterstock)
In a nutshell
- A new smartphone app can accurately estimate hemoglobin levels and screen for anemia, just by analyzing a photo of your fingernails, offering a noninvasive alternative to traditional blood tests.
- In real-world use across 1.4 million tests, the app showed 89% sensitivity and 93% specificity for detecting anemia, with even greater accuracy when personalized for people with chronic conditions like kidney disease.
- The app’s widespread use has enabled researchers to map anemia trends across the U.S. and uncover socioeconomic patterns, highlighting its potential for both individual health monitoring and large-scale public health insights.
ORANGE, Calif. — Smartphones have become essential tools for everything from banking to fitness tracking. Now, a new app has emerged that could transform how we monitor our health. Researchers have developed a smartphone application that can screen for anemia, a condition affecting over 2 billion people worldwide, using nothing more than a photo of your fingernails.
The research, published in the Proceedings of the National Academy of Sciences, was conducted by researchers from multiple American universities. Their app has already been used more than 1.4 million times across the United States, demonstrating accuracy levels comparable to traditional blood tests when detecting low hemoglobin levels, the hallmark of anemia.
This allows patients to skip the needle, lab visit, and waiting period typically required for anemia testing. Instead, you simply open an app, take a picture of your fingernails, and receive an estimate of your hemoglobin levels within seconds. For the millions living with chronic conditions that require regular hemoglobin monitoring—like chronic kidney disease, cancer, or nutritional deficiencies—this technology could be life-changing.
How the App Works
The color of your fingernail beds directly correlates with the amount of hemoglobin in your blood. Hemoglobin is the protein in red blood cells that carries oxygen throughout your body and gives blood its red color. When hemoglobin levels drop, as they do in anemia, the fingernail beds appear paler. The app uses artificial intelligence to analyze this paleness and calculate an estimated hemoglobin level.
In real-world testing with over 9,000 users who also reported their actual blood test results, the app demonstrated impressive accuracy. When using a hemoglobin threshold of 12.5 g/dL to identify anemia, the app achieved 89% sensitivity and 93% specificity, numbers that rival traditional screening methods.
The app also has a “personalization” feature. For people with chronic anemia, the app can be calibrated to their individual baseline by entering lab-confirmed hemoglobin results alongside the app’s estimates. This personalization dramatically improved accuracy, especially for patients with chronic kidney disease.
Beyond individual benefits, the widespread adoption of this app has created unique opportunities for public health research. With location data collected from over 900,000 tests, researchers have created an “anemia map” of the United States, revealing geographical and demographic patterns that weren’t previously possible to observe.
Android users consistently showed lower hemoglobin levels than iPhone users, both in app estimates and self-reported lab results. Since Android phones tend to be less expensive than iPhones (by about $200 on average among the top 10 phone models in the study), this suggests a link between socioeconomic status and anemia.
Crossing Barriers to Healthcare Access
Additional analysis also revealed that counties with higher median incomes, larger populations, higher percentages of Black residents, and more primary care physicians per resident had more app users. This information could help target anemia interventions to populations most in need.
The team’s research included both real-world data from general app users and a controlled clinical study with chronic kidney disease patients. In the clinical study, 31 participants with chronic anemia due to kidney disease completed a protocol where they had both app estimates and traditional blood tests over multiple visits.
While the app isn’t meant to replace medical care or emergency clinical decisions for severely anemic patients, its accuracy makes it valuable for screening and monitoring. After personalization, the app’s performance approached levels seen in FDA-cleared medical devices for non-invasive hemoglobin measurement.
Traditional blood testing for anemia involves trained personnel, expensive equipment, lab access, and often insurance coverage. These are all significant barriers for many people worldwide. This smartphone-based approach requires only a device that billions already carry in their pockets.
For someone with chronic anemia who needs monthly blood tests, this app could mean fewer needle sticks, fewer clinic visits, and more timely awareness of changing hemoglobin levels. For populations in remote or underserved areas, it could mean access to anemia screening that was previously unavailable.
Healthcare continues to move beyond clinic walls and into homes and communities. This smartphone-based anemia detection app is another example of making healthcare tools accessible, personalized, and non-invasive.
Paper Summary
Methodology
Researchers developed a smartphone app that uses images of fingernail beds to estimate hemoglobin levels based on their pallor (paleness). The original algorithm was trained on over 2,000 participant-acquired images paired with clinical blood test results. The app was then made publicly available on both Apple and Google app stores, allowing users to take “fingernail selfies” for anemia screening and optionally input their lab-confirmed hemoglobin levels. For the clinical study portion, researchers followed 35 chronic kidney disease patients through eight clinic visits, where both app-based and standard blood hemoglobin measurements were taken. The first four visits were used to create personalized algorithms for each patient, which were then tested in the subsequent visits.
Results
The smartphone app has been used more than 1.4 million times across the United States. When comparing app estimates to self-reported lab results from 9,061 users, the app demonstrated a mean absolute error of ±0.72 g/dL and achieved 89% sensitivity and 93% specificity for detecting anemia using a threshold of 12.5 g/dL. Personalization significantly improved accuracy—for chronic kidney disease patients, the mean absolute error decreased from 1.36 to 0.74 g/dL after calibration. Geographic analysis revealed that Android users had lower hemoglobin levels than iPhone users (12.2 vs. 12.9 g/dL), suggesting a link between socioeconomic status and anemia. The researchers also created county-level “anemia maps” showing hemoglobin distribution patterns across the U.S.
Limitations
Key limitations of the real-world data include inability to verify correct app usage or accuracy of self-reported lab tests, and potential sampling bias since users are more likely to continue using an app that works well for them. The clinical study had limitations including consistent imaging parameters, small sample size for comorbidity analysis, and the fact that images were taken by study personnel on only two smartphone models in controlled clinical settings, which eliminates much of the variability seen in at-home testing environments. The study also notes that the app is not intended for emergency clinical decisions in severe anemia cases, as its lower limit was set at 7 g/dL.
Funding/Disclosures
Funding for the study was provided by the National Heart, Lung, and Blood Institute of the NIH via award 2R44HL139250-02A1 and by AstraZeneca, which sponsored the chronic kidney disease clinical study. Several authors, including lead author Robert Mannino, are employed by Sanguina, Inc. and have an interest in the technology. Mannino and three other authors are equity holders in Sanguina, Inc. The technology presented has been patented (US Patent number 12268498).
Publication Information
The paper “Real-world implementation of a noninvasive, AI-augmented, anemia-screening smartphone app and personalization for hemoglobin level self-monitoring” was published in the Proceedings of the National Academy of Sciences (PNAS) on May 13, 2025. The authors include Robert G. Mannino, Julie Sullivan, Jennifer K. Frediani, and others, with Wilbur A. Lam as the senior author.
