The Role of AI in Supporting At-Home Diagnostics

Some industry leaders and proponents of artificial intelligence (AI) support its integration into health care, but there is some skepticism among patients about its use. However, according to a recent MIT study, 75% of health care facilities that utilized AI reported improved capacity to manage illnesses. So, although some individuals are hesitant towards integrating AI into health care, AI is proving its usefulness within the industry. 

Additionally, medical institutions and public health could go even further by leveraging AI to support telehealth services in the areas where patients already interact with digital portals (EHRs) and utilize asynchronous care such as at-home diagnostics. Through image recognition, natural language and machine learning, AI can be used to improve remote health care outcomes and reduce workload for providers without alienating patients. 

At-home diagnostics and the care continuum 

At-home diagnostic testing enables providers to check patients for diseases, screen populations in public health initiatives and offer remote medication maintenance. Patients collect a sample at home, mail it in a secure container to a lab, lab technicians result the sample and a physician reviews the results, which also appear in a digital portal for the patient to view. 

Popular at-home diagnostics include HIV/STI testing, colorectal cancer screening and HbA1c diabetes testing. These services are particularly useful for marginalized or vulnerable patient populations who cannot take the time off from work to attend an in-person appointment. Additionally, it enables patients with medical trauma, who may not want to enter an office, to maintain more control over their health care experience.

The role of AI in remote health care

AI technologies have rapidly improved health care research and patient outcomes. Providers around the world already use AI in telehealth to provide real-time analysis of interactions with patients in a remote setting, and make recommendations that can lower employee fatigue without disrupting patient-provider relations. 

AI also supports patients while reducing burden for physicians through advancements in  remote patient monitoring (RPM), a common and popular approach to managing chronic illnesses such as cardiac diseases, diabetes and asthma. Data from remote patient monitoring devices is sent to physicians to reference and alter care plans. Clinical-grade wearables and sensors have an AI software component that not only transmits raw data, but consolidates it for easy interpretation. 

Image recognition for physical symptoms 

Image classification is a certified learning problem that defines a set of targets and trains a model to recognize them using labeled example photos. AI can power image classification for physical symptoms in health care such as rashes, skin cancer and lesions. Google is already training AI to perform medical imaging in radiology, pathology and dermatology. 

AI company HeHealth uses image recognition to scan images submitted by AMAB patients to screen for physical manifestations of STIs, such as genital warts and sores from syphilis. These AI powered suggestions can then be confirmed or disproved by asynchronous telemedicine techniques, including physician-ordered at-home diagnostics. The process would be a lower lift for physicians, but would not impair patient outcomes.

Natural language for expedited telehealth

AI can assist physicians in sorting through clinical data in a patient file to extract higher quality insights, especially in a telehealth appointment where time is limited and physical barriers are present. The AI can screen for flags like patient family history to determine if at-home diagnostics for certain types of cancers is recommended or flag appropriate STI panels for patients with preferred sexual behaviors. AI can also scrub data to uncover anything that might be missed or note improperly coded patient conditions for the physician. This type of assistance supports the provider as he or she creates a more accurate care plan.

Machine learning algorithms for discrete testing

Not all patients who choose remote health care want to speak directly to a physician. Some patients, especially those from communities with a history of medical trauma or discrimination, may seek health care at a distance. With the help of chatbots and virtual assistants, healthcare organizations can improve patient experience by streamlining their access to basic healthcare services. It offers certain patient populations who wish to access discreet at-home testing services the platform to speak with a chatbot rather than a physician. 

A physician could review the patient's medical file and any of the questions asked by the chatbot asynchronously. After reviewing, they could then recommend any at-home diagnostic testing services based on their clinical judgment. The patient never enters an office or speaks directly to a physician, but the physician can still review all the relevant information to assign any recommended testing.  

Of course, there’s still a lot to be done from a regulatory lens before this becomes widely adopted by the medical community. There are currently no universal regulations in place regarding medical liability and much of medical AI is still speculative. But the possibilities are there, and its potential to revolutionize the way care is delivered once those standards are in place opens the door to patients, providers and health systems alike! 

AI as a remote diagnostics tool

While some in the industry may be wary of advancements in AI healthcare technology, there are many ways to utilize the pros of artificial intelligence that do not disrupt the patient-provider relationships. Patients who want to work directly with a physician can still do so via synchronous methods like video conferencing or audio-only telehealth. AI will be particularly helpful for at-home diagnostics because of its ability to simplify diagnostic processes through machine learning and natural language processing, which will decrease the burden on clinicians and increase access to at-home diagnostic care for patients.