How COVID-19 Accelerated Use Of AI In Healthcare

It is a frigid morning in June of the year 2030, and you are reluctant to get out of your warm bed. When you wake up, you notice that you have a sore throat and a stuffy nose. After the alarm buzzes for the second time, you slowly get up and walk to the bathroom.

While you brush your teeth, an infrared sensor built into in the mirror of the bathroom automatically takes your temperature.

On your way to the kitchen for your first cup of coffee, you receive an alert from your personal Artificial Intelligence (AI) Health Assistant on your smartphone notifying you of some abnormalities that it picked up in your saliva sample via your smart toothbrush point-of-care (POC) device. And also that you are running a low fever.

While you are drinking your coffee, the health Assistant informs you that luckily you tested negative for the newest variation of the Severe Acute Respiratory Syndrome (SARS) Coronavirus that are doing the rounds. Still, you have tested positive for a specific strain of the seasonal flu.

The Health Assistant suggests that you take some paracetamol, use a decongestant for the stuffy nose and a throat spray with an anti-inflammatory and anaesthetic function for the sore throat. It also warns you not to use ibuprofen due to your renal impairment.

But if you prefer to consult with your general practitioner there is is a video call time slot available this afternoon, in which case the details of the test and symptoms will be forwarded to the doctor.

“Afterwards he prescribes some medicine for symptomatic relief. The prescription is automatically and digitally transmitted to the pharmacy. Within an hour, the medication is delivered to your doorstep by a drone after you received a notifying message on your smartphone.”

Due to individual comorbidities that you have, you decide to consult the doctor via a video call instead. After studying the data and discussing your symptoms, the doctor requests you to connect to your Home Health System to check your blood pressure and listen to your heart.

Afterwards he prescribes some medicine for symptomatic relief. The prescription is automatically and digitally transmitted to the pharmacy. Within an hour, the medication is delivered to your doorstep by a drone after you received a notifying message on your smartphone.

Does it sound very futuristic or farfetched? No, the future is not as far off as it seems. In the next decade or so many households will have personal AI Health Assistants in their homes to take care of the family’s daily health. And since most biomarkers present in blood and urine can also be detected in a sample of saliva, these Health Assistants could be equipped with AI, the newest nanotechnology, molecular diagnostic and biomarker detection capabilities to make an accurate diagnosis of a variety of oral and systemic diseases.

People would prefer such an easy-to-use oral diagnostic test. To substantiate the idea that an automated oral diagnostic test would be preferred as compared to a more invasive alternative, just consider the success of the oral thermometer to detect fever, which has replaced its predecessor, the rectal thermometer.

And now during the COVID-19 pandemic, it seems that the digital non-contact infrared thermometer has replaced the oral thermometer. Covid-19 certainly accelerated the adoption of digital technology.

Traditionally, most diseases are diagnosed by the physician obtaining the medical history of the patient and the symptoms reported by the patient, a physical examination, and a chemical analysis of blood and/or urine samples.

“OVER THE LAST FEW YEARS, SALIVA IS INCREASINGLY BEING USED FOR THE DIAGNOSIS OF ORAL AND SYSTEMIC DISEASES SINCE IT REQUIRES A LESS PERVASIVE PROCEDURE THAN THE DRAWING OF BLOOD.”

The patient’s samples are typically sent to a clinical diagnostic laboratory for determination of the levels of a series of markers including ions, antibodies, hormone levels, and a variety of disease-specific biomarkers. After some time, the laboratory report is returned to the physician, the results and treatment communicated to the patient.

Over the last few years, saliva is increasingly being used for the diagnosis of oral and systemic diseases since it requires a less pervasive procedure than the drawing of blood. It is often used in situations where healthcare staff are working with pediatric and geriatric patients, or when access to health care is limited in remote geographic areas where phlebotomists (a technician drawing blood) are unavailable.

Due to extensive research, the science around using saliva for diagnostic purposes has developed tremendously, and significant success has been achieved with the diagnosis of numerous diseases such as:

• Cardio Cardiovascular disease – Salivary biomarkers have been incorporated into point-of-care (POC) devices for the rapid assessment and identification of cardiovascular disease (CVD). Elevated salivary lysozyme (an enzyme) levels have shown a significant association with hypertension, an early stage of CVD.

• Renal disease – A series of salivary markers were associated with end-stage renal disease. Salivary tests could also be used by patients to decide when dialysis is required.

• Systemic malignancies – Successful studies detected certain biomarkers in patients with breast cancer, ovarian tumours, and pancreatic cancer.

• Diabetes – A unique proteomics (study of proteins) signature identified in saliva obtained from people with type-2 diabetes could assist with the diagnosis of diabetes.

• Autoimmune diseases – Oral biomarkers derived from the salivary proteome (complement of proteins) assists in the diagnosis of primary Sjögrens’s syndrome (pSS) and other rheumatoid factor diseases such as Lupus and Scleroderma.

• Oral infections – Detection of oral diseases caused by cancer (oral squamous cell carcinoma), fungi (candida), viruses (human papillomavirus, Epstein-Barr virus), and bacteria.

• Human Immunodeficiency Virus (HIV) – the detection of antibodies to HIV in saliva is as sensitive and specific as a blood test. The ability to accurately detect antibodies to HIV led to the discovery of the potential to detect antibodies to many other viral and bacterial pathogens.

The scenario of the personal Health Assistant driven by artificial intelligence (AI) is thus much closer than we may imagine.  As medical science and computer science are more and more integrated in the Fourth Industrial Revolution (4IR), we will enter an era where AI will play an indispensable role in the everyday health care of individuals and families.

The more we connect wearables, biosensors, smart home detectors and other Internet-of-Things (IoT) devices to the Health Assistant, the more complete the data and health picture will be.

There is little doubt that the Covid-19 pandemic and subsequent lockdown accelerated the digitalisation of health care.

“In California, computer scientists developed systems that can remotely monitor the health of the elderly in their own homes and make alarm if they fall ill due to Covid-19 or any other illness.”

Everywhere around the globe, innovative applications of artificial intelligence (AI) have been implemented. In South Korea, it was used with great success to warn people if they get too close to a confirmed Covid-19 case.

In China, Alibaba developed an AI algorithm that can diagnose suspected Covid-19 cases within 20 seconds with a 96 per cent accuracy. Numerous countries developed AI tracking and tracing systems to prevent the spreading of the Coronavirus.

In California, computer scientists developed systems that can remotely monitor the health of the elderly in their own homes and make alarm if they fall ill due to Covid-19 or any other illness.

In South Africa, several medical practitioners started to consult their patients via video conferencing facilities to limit the spread of the virus. One of the large medical schemes began to offer online consultations for people who are concerned that the Coronavirus may infect them. And in Muizenberg, geriatric patients do their physiotherapy session via Zoom.

The COVID-19 virus caught most governments and health systems by surprise. This resulted in a lack of reliable data and models, poor decisions, slow responses, inadequately distributed personal protection equipment (PPE) and medical supplies such as ventilators, a severe shortage of hospital and ICU beds, and not enough medical staff in hotspots.

AI certainly made a difference – not everywhere, but in individual pockets of excellence, where countries were generally better prepared for a pandemic. But now data is flowing, making the use of AI and machine learning possible, thus enhancing the modelling and decision-making capabilities of health care officials and citizens alike.

Whether it is a personal AI Health Assistant at your home, or deep neural networks (a subtype of AI) used to accurately interpret medical scans, pathology slides, eye exams, and colonoscopies at high speed, we will see much more use of artificial intelligence in health care after the Covid-19 pandemic. The pandemic has substantially influenced the way we approach health care.

Professor Louis C H Fourie is a futurist and technology strategist.

This article was first published in iol.co.za.