Doctors say anemia affects many patients, but it’s not so simple to diagnose. Lifespan Corp. researchers are looking for a better way.
That’s why doctors in emergency departments at Rhode Island Hospital and The Miriam Hospital participated in a study of an innovative smartphone app that might allow them to determine if a patient has anemia without time-consuming lab tests.
“It comes up a lot,” Dr. Selim Suner, an emergency medicine physician at Lifespan, said of the blood disorder. “We are confronted with this in an emergency department on a daily basis.”
Anemia, which means someone has a low hemoglobin count, is the most common blood disorder and affects more than 3 million Americans, according to the National Heart, Lung, and Blood Institute.
Anemia is often caused by a lack of iron, which can come from blood loss or poor diet. In developing countries, Suner says, anemia often comes from malnutrition or parasitic infections such as malaria or parasites in the gut, while in developed countries it comes from cancer or chronic illnesses such as renal failure.
There’s a wide variety of anemia symptoms, with general weakness being the most common, and it can make existing health concerns worse, says Dr. Gregory Jay, an emergency medicine physician at Rhode Island Hospital and The Miriam.
Testing a patient’s complete blood count is the standard for diagnosing anemia. But this can take hours and requires trained doctors, laboratory technicians and equipment that many hospitals don’t have. This means anemia often goes undetected in underdeveloped parts of the world that don’t have the resources to test for it, Suner says.
A new study published in the peer-reviewed journal PLOS One by Jay along with Suner and Dr. James Rayner, also a Lifespan emergency medicine physician, says the smartphone application’s ability to detect severe anemia quickly shows promise.
The study looked into how the app, developed and owned by Lifespan, can estimate a patient’s hemoglobin concentration using uploaded photos of the inner eyelid, also known as the palpebral conjunctiva.
The study involved taking photos with a dedicated smartphone of 435 emergency department patients at Rhode Island Hospital and The Miriam Hospital who had already gotten their blood drawn. Then the app would estimate hemoglobin concentration by assessing the hue – or pallor – of the patient’s palpebral conjunctiva.
Over many years, doctors have learned that the color of a patient’s conjunctiva can indicate how anemic they are. A healthy conjunctiva tends to be more red while the paler it is, they are more anemic. The conjunctiva is useful because it doesn’t have any melanin, so a person’s skin color can’t affect the color and it’s an area with a lot of blood vessels, Rayner says.
While Rayner says the app couldn’t accurately measure a person’s specific hemoglobin levels as well as a blood test, it proved to be an effective classifier of anemia, meaning it was able to tell if someone was extremely anemic.
This study is the most recent version of a series of methods Lifespan doctors are researching to improve an anemia diagnosis.
Suner says their research began in the late 1990s when doctors used digital cameras to test whether assessing a patient’s palpebral conjunctiva is an effective way of detecting anemia. Similar to the most recent study, researchers took photos of patients’ inner eyelids with a digital camera then used a mathematical formula to quantify the redness and estimate the hemoglobin levels, and then compared the results with the blood test.
The studies evolved with the technology, and researchers conducted studies with reflectance spectrophotometry, which calculates color by flashing light onto the surface of something and measuring the reflection. They did this by mounting an interferometer onto a slit lamp – a special microscope with a bright light the eye doctor shines into a patient’s eye for exams – to gather information on the lower eyelid. This proved to be more accurate than the digital camera, though the spectrophotometer’s large size proved challenging, Jay says.
Then cellphones with digital cameras became available, and that offered a new way for scientists to conduct their research.
In the latest study, the researchers say they used what’s called “raw” images, instead of the standard digital format called “JPEG” captured on digital cameras. These images have more data and can capture many more shades of color.
“It’s a big jump,” Suner said.
Though the most recent results are an improvement over past studies, there are still some accuracy issues.
The doctors say that work needs to be done to limit any reflections or shadows captured in photos of palpebral conjunctiva that may affect their results.
Knowing the patient’s hemoglobin from the blood test, Rayner says he hopes to use the image data to train a neural network to make predictions. A neural network is a kind of artificial intelligence that teaches computers to process information like a human brain.
Now, the app is on TestFlight, where app developers share their software with other leaders, and Jay says some improvements need to be made before they approach a developer. Also, developers have approached them to look at their technology and even do research on their devices.
“I think the app will eventually see the light of day,” Jay said. “Our efforts aren’t exclusive to our activities only.”
