People frequently use mobile apps in their personal and work lives, letting those applications give them information ranging from budgeting assistance to baking tips. Could a mobile health app diagnose illnesses too? Here, Emily Newton looks at how such applications might help patients and clinicians alike by making diagnostic tools as accessible as someone’s smartphone.
When someone has a medical problem, they understandably want physicians to provide the right diagnosis. Failing to do that could mean the person begins a treatment that could cause more harm than good. Additionally, if the patient has a disease that spreads quickly, not correctly recognising the ailment fast enough could lead to avoidable health complications.
Misdiagnosis also occurs more often than many people think. A research paper that looked at the global issues stemming from incorrect diagnoses in primary care mentioned approximately 5% of United States patients get misdiagnosed annually. It also cited how there was a 3.6% prevalence of diagnostic errors at 12 Malaysian primary care facilities.
Complications can also result in countries that do not have the necessary tools to make prompt, accurate diagnoses. Moreover, physicians are human, and they will make mistakes. Fortunately, the ubiquitous smartphone could help reduce those errors.
Screening for Viral Infections Without Technical Knowledge
The lack of resources can make misdiagnosis more prevalent. In addition to sourcing the laboratory equipment used to test patient samples, it’s also necessary to find people with the training to handle someone’s bodily fluids and read the test results.
Researchers from Harvard Medical School and Brigham and Women’s Hospital sought to improve access with a smartphone. Their system tests for the Zika, Hepatitis B and Hepatitis C viruses with a phone, an external catalytic microchip device and specialised software.
The process involves placing bodily fluid into a channel on the microchip device, then pouring a small amount of hydrogen peroxide onto the sample, causing bubbles to form in it. The assessor then points their smartphone camera at the fluid, and a deep learning algorithm within the phone’s software looks for viruses.
This approach requires a virus concentration threshold of 250 particles per millilitre in viral loads. A concentration at or above that amount causes a positive test, while a sample with a lower viral load gets a negative result. This test gives conclusions in less than an hour and is nearly 99% accurate in detecting viruses. The researchers stipulated that a person can use any smartphone model for the test, and they can interpret the results without medical knowledge.
The team behind this effort envisions potentially sending the software and external catalytic microchip device to countries designated as disease hot spots. That way, some diagnoses could happen without specialists on the scene.
Diagnosing Skin Cancer
Skin cancer is a common disease in humans, but it’s often highly treatable if spotted early. That’s why people at an elevated risk of developing it have regular checks in dermatologists’ offices. Those doctors may take biopsies and send them off for closer examination.
One method of doing that is by using a spectrophotometer, which measures the amount of light passing through a solution. Spectrophotometers also have companion software for data processing. People interested in buying used equipment often find that the lack of accompanying software is the main barrier preventing them from sourcing a used spectrophotometer. Software unlocks the device's potential, and it provides opportunities to use spectrophotometers for diagnosing illnesses.
More than a decade ago, scientists investigated using spectrophotometry to diagnose cutaneous melanoma. More specifically, the researchers used a video spectrophotometer imaging system to examine lesions from patients suspected to have cutaneous melanoma. The results showed that spectrophotometry detected melanoma with a sensitivity of 66.6% and a specificity of 76.2%.The team concluded spectrophotometry was a useful tool for diagnosing melanoma. However, dermatoscopes provided more accuracy, they said.
More recently, researchers are interested in smartphone applications that use algorithms to spot skin cancer. A recent systematic review warned that people should be careful about relying on them for such diagnoses. The findings emphasised that many relevant studies were small or of poor quality. Additionally, no app can detect all cases of melanoma or other skin cancers.
A mobile health app could help patients become more proactive, though. Tools exist that help people take pictures of their moles and other concerning features, then track if they change over time. Even if these apps don’t make direct diagnoses, they could help physicians reach faster, more accurate conclusions with information collected by their patients.
Finding Asymptomatic COVID-19 Cases
Scientists are still learning about the role of asymptomatic patients in spreading COVID-19. However, recent evidence suggests that as many as one in five people may carry the virus without experiencing symptoms. Moreover, such individuals can spread the virus to others, but not as readily as people with symptoms.
Opinions vary on how much asymptomatic individuals contribute to overall virus cases, but most scientists agree they play a role to some degree. Scientists at MIT wondered if a mobile health app could identify COVID-19 in people who don’t feel sick.
The researchers trained an artificial intelligence algorithm with cough-sound data from 4,256 subjects. The trained model achieved a sensitivity of 98.5% and a specificity of 94.2% when validated with diagnosed individuals who also took COVID-19 lab tests. Moreover, the model provided a 100% sensitivity rate with a specificity of 83.2% in asymptomatic cases.
Thus, this approach could supplement current COVID-19-curbing measures in a free, non-invasive and scalable way. It could be particularly advantageous for people who must take extraordinary care to avoid spreading the novel coronavirus. For example, a nursing home worker might use the app each day before starting a shift, helping to prevent an outbreak in the facility.
A Mobile Health App Is a Helper, but Not a Cure-All
These examples show why medical professionals and patients may soon consider mobile health app options instrumental in accelerating diagnostics and screening for ailments. These platforms won’t replace human evaluations, but they could provide valuable help to medical practitioners and the patients who rely on them.
Author: Emily Newton is the Editor-In-Chief of Revolutionized, a magazine exploring innovations in science and industry that shares ideas to promote a better tomorrow.