- An algorithm has been developed by researchers that predicts anemia up to 70%.
- It is possible to calculate blood hemoglobin concentration by taking smartphone images of the inner eyelid.
- An app for smartphones could be used to screen for anemia among those who live in remote areas.
More than 5%TrustedSource of Americans and 25% of the world’s population are affected by anemia. This medical condition is characterized by low blood hemoglobin and symptoms such as fatigue, dizziness or headaches, shortness in breath, difficulty concentrating, and dizziness.
Malnutrition, parasitic infections, and underlying diseases are common causes of severe anemia. It can lead to death and morbidity in vulnerable populationsTrusted Source. This includes children, older adults, and people with chronic diseases.
Recent research published in PLOS ONETrustedSource has shown that predicting anemia using a smartphone photo of the inside of the lower eyelid is accurate to 72%.
Anemia is often diagnosed by healthcare professionals using sensitive laboratory equipment. This means that there is an disproportionate amount of anemia in rural areas with limited access to healthcare.
The study authors believe that there is a need to create a point-of-care tool that can identify anemia. It should be affordable, noninvasive, and accessible. Preexisting technology would be the ideal tool.
Two-phase research was conducted by researchers to determine if a smartphone camera could be used to detect anemia. The first phase consisted of taking images of the lower eyelids in 142 patients at an emergency department with a smartphone.
Researchers chose the inner lower eyelid (also called the palpebral conectiva) because of its unique features.
- It is easy to photograph.
- There are no opposing colors between the blood vessels and the conjunctival surfaces.
- It is extremely close between the surface of the blood vessels and the surface.
- The area’s blood flow is not affected by temperature or other environmental factors.
Researchers were able to create an algorithm to maximize color resolution by zooming in on one small area of each photo. They also developed a predictive model that compares the skin and whites to hemoglobin levels.
The second phase consisted of testing the algorithm using smartphone images from 202 patients at the emergency department. The model was able to predict anemia accurately 72.6% of the time, according to the results. The model was able to predict severe anemia, which would require a blood transfusion, with a higher accuracy of between 86%-94.4%.
Dr. Dr. Selim Suner of Brown University and Rhode Island Hospital explained that anemia is a condition in which people need iron supplements. These are easy and cheap to take. Suner stated, “Making the diagnosis can be difficult.”
Dr. Girish Nadkarni is the clinical director at Mount Sinai Health System’s Hasso Plattner Institute for Digital Health. He said, “Using a smartphone for screening anemia is advantageous due to the decentralized nature — it avoids the need to draw blood — as well as the time and effort savings that this entails.”
According to the results of the study, flash photography is not required in order to produce acceptable images for anemia detection. The authors also note that raw images are data directly from the camera sensor, without any processing or compression like JPEG.
Variable image quality was one of the potential limitations noted by researchers. This could be due to the person retracting their eyelids during recording. It is not known if the lighting was uniform, or if it played any role in image quality.
In 2019, 36% of all the world’s population used smartphones. According to trends, while those who are wealthy are more likely than others to have smartphones, this trend is steadily increasing in countries with lower socioeconomic status.
These results have set the stage to develop an app within a smartphone that can not only take the image, but also analyze the elements in the image to predict hemoglobin content in real-time,” the authors state.
“This is a particularly attractive opportunity for developing nations, which may not have a well-connected telecommunication network but may have a sparse, rudimentary and poorly distributed medical system.
The future development will focus on designing a user interface that makes it easy for lay people to take suitable photographs. One in which lighting, focus and area of interest can all be optimized. Further algorithm development and validation of model predictions are also necessary for imaging tools.
Dr. Suner says that this study shows that anemia prediction with a smartphone can be done. This project and others could have a positive impact on large populations and contribute to global health.