How Medical AI Is Changing Healthcare in 2022
Medical Artificial Intelligence is becoming one of the most disruptive factors in the healthcare domain. The market for AI in healthcare is projected to grow from $6.1 billion in 2021 to $64.11 billion in 2027 (growing by 45.12% annually).
How and why has AI gained such impressive traction for hospitals and medical providers? In this article, healthcare startups, clinicians, and medical practitioners will gain insights on:
- What Artificial Intelligence in healthcare is
- Benefits of AI in the medical field
- AI uses in healthcare
- Examples of AI in healthcare
What does using Artificial Intelligence in healthcare mean?
Before considering specific AI healthcare solutions, let’s outline what exactly medical AI is and hopes to do. AI in healthcare analyzes patient data and creates algorithms that build highly personalized, precise, and efficient medical treatment techniques.
Clinical AI identifies patterns and provides data-driven insights for medical professionals. These insights help improve patient outcomes, identify healthcare needs, and result in better and faster diagnoses.
Artificial Intelligence isn’t just one type of technology. It’s a group of many different technologies working together in tandem, such as machine learning (ML), natural language processing (NLP), deep learning (DL), and Big Data.
Healthcare AI is revolutionizing entire healthcare systems by:
- Transforming the diagnostic process
- Facilitating the development of new drugs
- Improving the quality of medical services
- Enhancing medical imaging analysis (ultrasound, CT, MRI)
- Optimizing costs and the quality of management
- Reducing disease frequency and improving treatment outcomes
- Increasing the speed and quality of supervisory and assessment activities
Top 8 roles of AI healthcare solutions that have already started to revolutionize medicine
By migrating from outdated solutions to up-to-date AI for healthcare, the medical industry can stay on top of fast-paced technology trends and utilize the full capabilities of healthcare software services.
Industry potential is not going unnoticed. Investment in artificial intelligence was up 108% year-over-year in 2021 and $66.8 billion in global funding for startups. Healthcare AI accounted for nearly one-fifth of total venture funding.
Now, let's dive deeper into a few of the areas that we think can best harness the enormous potential of AI for healthcare.
Improving medical data storing and managing
With the advent of cloud services and digital capacity, digital data keeps growing exponentially. The total amount of data created, captured, copied, and consumed globally was 79 zettabytes in 2021 and is expected to reach 181 zettabytes in 2025 – or 181 trillion gigabytes!
Medical institutions generate enormous amounts of data of all kinds. This data is usually scattered and unstructured and can be stored in dozens of clinics and medical records.
This makes it difficult to gather a patient’s history and make a diagnosis. Interpretation of tests, exams, and scans may also be inaccurate because of the volume of data.
Joshua Tauber from HealthReveal (now acquired by Accolade) notes other complications with medical data, “It’s often designed for business purposes, which doesn’t necessarily mean it’s clinically meaningful.”
He shares that addressing this issue was one of the goals when building HealthReveal, “We were taking data made for billing purposes and wrapping thoughtfulness around it in an attempt to find actual, clinical meaning.”
Because of traditional healthcare systems, medical experts often spend more time addressing administrative tasks, reviewing reports, and managing health IT and EHR systems than working with patients. It’s one of the reasons leading to physician burnout.
Additionally, the many patterns of disease outcomes can make patient data difficult to analyze. Even when physicians have access to all the data they need, it can often be difficult to attend to every detail and interpret it correctly.
A powerful AI-enabled cloud platform with access to medical databases can efficiently analyze medical records and patient data. Medical AI helps identify multiple diagnoses based on analyzed symptoms.
The search results are backed by millions of scientific information pages containing information on every rare disease. AI medical algorithms analyze symptoms, diagnose diseases, and suggest the safest and most effective medications based on a patient's profile.
With such large volumes of information, valuable medical data sometimes gets lost among the millions of databases. Custom healthcare software anchored by a powerful AI core could save medicine and pharma up to $100 billion a year.
AI can easily connect numerous data points that used to take years to process and analyze, thus improving efficiency in clinical trials, research, and decision-making in the doctor’s office.
Streamlining drug development and research
New drugs, vaccines, and medications require time and large investments during the R&D stages. Medical drugs also require a careful development approach, thorough tests, and clinical trials that don’t always guarantee 100% success.
Bringing a new drug to market costs about $2.8 billion dollars over 12+ years. Moreover, the regulatory challenges of bringing new treatments to market are often extremely hard to overcome. These include:
- Research stage: analyzing thousands of potential formulas to select drug candidates and studying immune response.
- Preclinical stage: laboratory analysis to identify relevant antigens to develop the concept and structure of the drugs.
- Clinical trials: testing the medicines on test groups with different characteristics.
- Regulatory validation and certification: checking the safety of the medicines and compliance with legal regulations.
- Production and quality control: drugs mass production and quality tests.
Adopting AI and ML for drug discovery and research can streamline these processes by selecting only the right candidates. AI predictive analytics in healthcare can also simulate trillions of potential drug interactions with their biological targets.
For example, Atomwise currently uses an AI engine to produce drug formulas faster and with better quality. Berg Health developed a similar project to enhance drug treatment around the globe.
AI can reduce the time to develop new drugs by analyzing the molecular structures of existing drugs and proposing new ones according to given requirements. New Medical AI technologies give hope that we’ll be able to develop medicines faster and for diseases that cannot be cured today such as multiple sclerosis, Alzheimer's, and others.
Administration process automation
Even before the pandemic, the healthcare system faced a global shortage of senior and mid-level medical personnel. According to the World Health Organization, some countries will need more than 18 million healthcare staff by 2030 for people around the world to have access to healthcare services.
With population growth, longer life expectancy, and changing clinical disease patterns, this shortage will most likely only get worse. These factors only increase the demand for highly-skilled healthcare professionals and make access to medical services care more difficult.
Physician burnout costs in the US healthcare system
The average doctor spends 8.7 hours per week on administration, which prevents them from focusing on more specialized tasks and lowers their career satisfaction. With medical AI, doctors no longer need to dedicate as much time to filling out medical records or analyzing patient history.
David Yakobovitch, General Partner at DataPower Ventures, agrees, “The healthcare industry’s digital transformation of medical records coupled with the FHIR standard for datasets allows startups to offer innovative products and improve outcomes with better, faster, cheaper solutions.”
These solutions can maximize healthcare workers’ already limited bandwidth and help address the current shortage and the widening gap in the future. With digital processes and data storage, medical workers can instead focus their time and efforts on solving critical diagnostic issues and developing treatment plans.
Modern AI technologies can ultimately help the healthcare system to increase patient and medical staff satisfaction, reduce the cost of medical services, and improve the quality of medical care.
Leveraging AI-based surgical robots
AI robot-assisted surgery is gaining traction thanks to increased efficiency and its “minimally invasive” nature. The global AI-based surgical robots market size was valued at $5 billion in 2020 and is expected to reach $17.2 billion by 2028 at a CAGR of 17.2%.
Clinics utilize AI-powered robots to assist doctors in various procedures, from minimally invasive to open-heart, and there are several medical solutions that prove to be promising and effective.
For example, cognitive surgical robotics makes it possible to reduce hospital stay by using specific tools in each individual surgery depending on a patient's data. The miniature HeartLander robot allows heart surgery to be performed through small incisions.
There are also tools that decrease the length of surgical treatments and eliminate the human fatigue factor.
The Da Vinci robotic system allows surgeons to control a robotic instrument from a computer console and as a result, perform a number of complex procedures more efficiently.
The AI-powered Da Vinci medical robot in practice
AI can help surgeons in many other ways as well, including:
- Supervising the doctor's work by serving as insurance in case of inattention or exhaustion
- Improving assistance for surgeons by reminding them about the procedure steps
- Creating precise and minimally invasive tissue incisions
- Reducing pain for patients by selecting the optimal incision and suture geometry
- Eliminating the human fatigue factor and increasing the efficiency of surgical treatment
Using predictive analytics in healthcare to improve the diagnostic process and forecast diseases
Approximately 251,000 people die each year because of preventable medical errors — about 9.5% of all deaths annually in the US alone. In light of that, precision medicine is one of AI’s most valuable healthcare assets.
AI medical predictive analytics can:
- Gather a complete medical history and test data
- Predict and diagnose diseases at a faster rate than most medical professionals
- Analyze the current health condition
AI in healthcare can quickly process medical data and strengthen clinical decision support. It helps save doctors’ time, increase the accuracy of diagnoses, and enable a timely treatment plan.
“Before AI, teams of hundreds of researchers would attempt to build a profile by researching everything they could about an individual. But that doesn’t scale when you consider that there are about 59 million healthcare practitioners in the world – and that at any given moment, some of the information that was collected could be rendered obsolete through a job change,” says Julie Stern, SVP of Engineering and Chief Information Security office at H1.
Stern shares that “with AI, H1 is able to keep that information current for more than 10 million of those practitioners and the numbers are growing every week.”
Another industry leader, IBM Watson Health, uses AI to tackle diseases. This could be identifying potential problems with the vascular system, recognizing cancer, or determining whether a patient is likely to have blood clots. IBM Watson can quickly process new information and make data-driven predictions.
Watson for Oncology can analyze the meaning and context of structured and unstructured data in clinical notes and reports that may be critical when selecting a treatment pathway. By combining attributes from a patient’s file with clinical expertise, external research, and data, the program identifies potential treatment plans for a patient.
For example, IBM's Artificial Intelligence analyzed 20 million scientific articles on cancer in 10 minutes and used them to make the right diagnosis for a patient.
Several hospitals in the U.K. already use a similar AI-powered healthcare platform called DeepMind Health. It processes information about a person's health and shares its findings with a physician, who then makes a final diagnosis.
AI systems like Ada can communicate directly with people and provide recommendations. The medical app communicates with patients and asks them about symptoms and complaints. It then gives recommendations, including what kind of doctor to visit, and suggests contacting a specialist for a remote consultation.
Sense.ly is a Medical AI program that monitors the condition of people who suffer from chronic diseases or have recently had long-term treatment. The app can structure data about a person's condition, send it to a specialist, and make recommendations. The system can also remind users to take medicines or visit a doctor.
However, programs that rely on large amounts of patient or provider data face a unique challenge.
“Healthcare presents security concerns because users typically don't want their data freely available. We've seen more research around Federated Learning where the AI engine is pushed to an edge device and doesn't require a centralized data storage" shares Maryam Farooq founder and director of New York Artificial Intelligence and co-founder and COO of aggregate intellect.
Genetic analysis systems work in a similar way, helping to understand the primary cause of a disease. One of the human genome testing platforms called Sophia Genetics identifies a patient's susceptibility to various diseases and brings notable insights to the doctor's attention.
Enhancing medical imaging analysis
About 50% to 63% of U.S. women will receive at least one “false-positive” mammogram over 10 years. In some cases, a few radiologists can disagree on their interpretation of the results while looking at the same mammography.
Visual pattern recognition software based on an AI core is forecast to be 5% to 10% more accurate than the average doctor. In contrast, AI algorithms and deep learning diagnosed breast cancer faster than 11 pathologists.
As AI in healthcare and deep learning techniques continue to flourish, they will inevitably streamline such diagnostic fields as:
- Radiology (CT, MRI, and mammography interpretation)
- Pathology (microscopic and cytological diagnoses)
- Dermatology (rash identification and pigmented lesion evaluation for potential melanoma)
- Ophthalmology (retinal vessel examination to predict the risk for diabetic retinopathy and cardiovascular disease)
Artificial intelligence systems help automate routine processes in hospitals, speed them up, and make them more efficient. This applies to the visualization of a variety of medical images like ultrasound, CT scans, and MRIs.
It can take a while for a doctor to examine and compare this data. AI technology in healthcare can do the task several times faster than a human.
A medical AI imaging diagnostic platform detects the virus in the early stage with 96% accuracy
From her time as a chair on the Healthcare Track of the AI Summit New York, Maryam Farooq reflects, “We are starting to see the emergence of Graph Neural Networks in things like bioinformatics and neuroimaging.”
This is how the NanoxAI system works. It is based on a neural network that has been trained on several hundred thousand images of patients with lung issues. The medical app helps to conduct diagnostics in the pulmonology field.
IBM created a medical imaging AI platform based on a trainable neural network called Arterys. It can analyze data, extract insights, and create visualizations of the human heart.
Empowering telemedicine with AI healthcare algorithms
Remote consultations increase access to high-quality medical care, especially in low-density areas. In addition, online telemedicine helps reduce healthcare costs and source a second opinion on test results to clarify a diagnosis and treatment plan.
AI makes telemedicine much more accessible. Providers use it to conduct remote diagnoses, collect medical scores, and handle patient information.
For example, Yakobovitch highlights a DataPower Ventures and Remedy co-investment, “Take Ash Wellness, the industry's leading remote diagnostic testing service. They’ve built a digital health platform with an end-to-end system that facilitates everything from physical kitting to communication with lab partners.”
In identifying some of the platform’s unique offerings, he continues, “Ash features applications and API integrations for digital health service providers, creating a seamless, and delightful consumer experience -- all paid by insurers, public, and private health practices,” continues Yakobovitch.
The benefits of AI telehealth during the pandemic and beyond
Google developed an automated analysis system that detects diabetic retinopathy from retinal photos. This AI algorithm helps doctors avoid the administrative tasks and challenges of diagnosis to focus on treatment.
Many large companies employ medical AI to recognize symptoms during telemedicine meetings. Before an online consultation, it can predict diagnoses and suggest a specific doctor to a patient. AI-powered telemedicine reduces the workload of medical professionals while allowing patients to monitor their condition more carefully.
Streamlining virtual nursing assistants
Physicians, nurses, and other clinical staff often have to take care of many patients at once while conducting monotonous and repetitive tasks.
Smart virtual nurse-bots in healthcare could save $20 billion annually and 20% of nurses’ time. Healthcare virtual nursing assistants are already working alongside human nurses in US hospitals. They ask patients about their health and provide advice, tips, and other information.
For instance, digital assistant Sally is a smiling woman in a white coat standing alongside nurse Walt. Sally and Walt are animated avatars, virtual personal health coaches from iCare Navigator's AI-based patient interaction and education platform.
iCare Navigator uses electronic patient records and leverages machine learning to build personalized patient relationships. The app determines when a patient will be most ready to receive information about their health and manage their treatment.
It turned out that 74% of patients preferred to get discharge instructions from a virtual nurse rather than from a human.
Molly is another AI virtual nurse avatar used at the University of California, San Francisco, and the U.K. National Health Service. Molly asks patients about their health, assesses symptoms, and makes recommendations for the most effective treatment based on symptoms.
An AI-powered virtual nursing assistant Molly asks you to check your blood pressure
Instead of searching for symptoms found on the Internet, a person can get help from a virtual nurse. Smart virtual nurses not only provide a patient with medical advice about common diseases but also help make an appointment to see a doctor.
They are available 24/7 and ready to answer questions in real-time. This is one of the major applications of clinical AI, which is increasingly being used to raise awareness and improve self-management skills for patients with chronic diseases.
The adoption of AI technologies has reshaped every industry and sector. However, implementing AI in the medical field is perhaps the most disruptive and crucial task that can affect hundreds of millions of people. In the long run, the possibilities of AI are virtually limitless.
However, for now, computers can’t simulate complex processes of a human’s higher nervous system: creativity, emotions, mindset, etc.
Christina Farr, a health-tech investor from OMERS Ventures, shares a similar hesitancy, “I believe we’re still in the early innings when it comes to using AI in healthcare. There’s a huge amount of variation in how advanced or sophisticated a company’s technology actually is.”
She continues, “The medical community remains skeptical, particularly around companies pushing for clinical (versus operational) use-cases. I’d say I’m optimistic, but cautiously so!”
We may see companies address this variability in a few years or even decades as more powerful AI emerges. However, computers have already learned how to solve problems based on the rules and algorithms set by a human.
Medical AI improves the accuracy of diagnoses, the availability of doctors, and the systematization of medical data. Larger companies can afford substantial budgets and highly-skilled employees to ensure custom healthcare software development.
On the other hand, implementing cutting-edge technology is expensive, and people do not yet fully trust artificial intelligence. Farooq suggests, “the biggest hurdle for most ML applications comes down to industry adoption – how do you create a trust for a tool for both patients AND providers?”
Simpler and cheaper AI systems will make medicine more accessible, and quality marketing and positive feedback will show customers that AI has many benefits and potential use cases. It’s a great chance to find the right approach for an audience and cover a profitable niche.