Industry Focus: AI in Healthcare
No matter what your location, you probably know that check-ups are never hassle free. Doctors have 15 minutes (or less) to spend with each patient but — shockingly! — some patients have more than 15 minutes of information to share.
Wouldn’t it be nice if there were some way to collect that information more quickly, or make it easier to transfer? Wouldn’t it be better if there were a device that could take care of all the standard procedures — heart rate, temperature, and weight — before you even step into the doctor’s office?
Luckily, Artificial Intelligence (AI) developers have been asking the same questions and have begun to build solutions. With Healthcare AI, an emerging field in Digital Health, you may soon be able to save time and cost with your check-ups.
Like FemTech and other Digital Health sectors, Healthcare AI has become more than a branch of the tree. With over 100 startups in more than a dozen countries, the field has become its own industry. Online statistics source CBInsights has broken the market down into the following categories:
- Patient data and risk analytics
- Lifestyle management and monitoring
- Emergency room and surgery
- In-patient care and hospital management
- Drug discovery
- Virtual assistants
- Mental health
- Medical imaging and diagnostics
Every Healthcare AI app fits into one, if not several, of these segments. To encompass these blurred boundaries, we have grouped them according to the three locations critical to healthcare and wellness: the lab, the hospital, and the home.
AI in the Lab
Two things drive healthcare developments: the need for money and the need to make healthcare more accessible, less invasive, and less expensive. Both elements coexist in the lab, where research-based advancements positively affect both institutional funds and patient experience. With this in mind, we can conclude that the lab, although hardly seen and rarely mentioned, forms the root of healthcare.
Not surprisingly, accommodating research needs was the first goal of industrial-scale AI in healthcare. Giving life to data-mining programs that can find subjects, pair samples, and make predictions in less than a nanosecond, AI accelerated healthcare research and put hospitals at the centre of the digital age.
And that centre is microscopic—London start-up, Desktop Genetics, for example, has catalogued extensive CRISPR (clustered regularly-interspaced short palindromic repeats) data to facilitate the identification, cloning, and editing of genetic material. With this program, genetic scientists no longer rely on microscopes and markers to find genetic abnormalities and test genetic therapies. More than a database, Desktop Genetics leaves the matching to the machines and the science to the scientists.
Although not limited to genetics, matching is in Healthcare AI’s genes (pun intended!). We see this in Atomwise, a tool for drug discovery. Using a deep learning algorithm, the program compares the molecular structures of different drugs and diseases to determine what chemical combinations will provide the best treatment. This hours-long process would take years for humans, and has already formulated two successful treatments for Ebola patients. While not the most conventional relief worker, Atomwise has helped frame a positive future for healthcare in developing countries.
Medicine has always been tied to research, and research continues to drive it forward. With AI at its core, we predict that healthcare will progress even faster.
AI at the Hospital
To truly gauge the effectiveness of medical research, human trials are a must. This process is naturally long term, even more so when examining diseases as unique as cancer. With only 4% of cancer patients receiving any kind of clinical trial, many life-saving cancer studies never come to fruition.
Deep 6 AI, founded in 2013, plans to boost this number to 20% by 2021. Taking the gold in the Enterprise and Smart Data categories at South by Southwest (SXSW), the company has developed a program that combs through anonymous data to match patients to trials with the click of a button. These innovations mean more research success for doctors and more effective treatments for patients.
The human side of AI continues in the hospitals, which transpose research results into patient-centred healthcare. Qventus, formerly AnalyticsMD, meets humans and machines somewhere in the middle. Termed “air-traffic control,” it acts as a real-time decision maker for doctors and other healthcare practitioners. Crunching all the data recorded daily, the program alerts hospital staff to best practices, diagnoses, or potential hazards. In turn, it makes hospitals more efficient and visits more satisfying.
The hospital is where medical research comes to life. By analysing the effects of this research in real-time, AI humanizes healthcare and helps tailor it to the individual level.
AI at Home
Online subscription services look old-fashioned compared to Healthcare AI; however, developers have linked the two as they use biometrics to deliver health products and services tailored to you and only you.
VITL is one of those companies. Your “visit” begins with the AI chatbot LANA (Live & Adaptive Nutritional Advisor), who analyses your digestion, energy levels, mood, immune system, and antioxidant status to gauge your nutritional profile. The program then designs a personal vitamin pack, which you can have delivered to your door monthly. Highly successful in the start-up stage, VITL has revealed plans to integrate LANA with DNA, other diagnostic test results, and wearable device data.
Indeed, as wearables already dominate the Digital Health industry, they are a great place to start when looking for healthcare data. In Healthcare AI, they help bridge the gap between industrial-scale and personal-scale data sets.
Magnea, for example, has developed a wearable that classifies movement data into activities, Calories, and even dangerous situations. Designed for elderly patients, the program aims to reduce heart disease, diabetes, and ulcers — all illnesses exacerbated or caused by inactivity. No larger than a coin, the wearable syncs with an app that patients can share with doctors or family members.
In part because of its name, Healthcare AI seems to belong to a clinical setting, but thanks to advancements in wearable technology, it can merge seamlessly with your everyday life to improve your overall health.
Some Human-crunched Statistics
Instant healthcare decisions, new treatments in half the time, more resources for smaller institutions and independent researchers — these reasons, and more, are what make healthcare one of the most promising fields for AI. Making facilities more efficient while reducing patient cost, AI works in the best interests of both sides.
Investment patterns suggest that Healthcare AI will soon have (and, for the most part, already has) worldwide success. Silicon Valley still leads in the numbers, but the EU has picked up the pace. The UK, for instance, has taken 9.2% of new deals since 2012, and has amassed £998 million in Deep Tech funding since 2011. France and Germany trail close behind, having earned €499 million and €480 million during the same period respectively. At this rate, it is thought that 45% of the total deals in 2017 will go to non-US start-ups, the majority of which will be EU-based.
Imagining the future, however, is difficult without understanding the present. Already, the European DeepTech boom has launched Desktop Genetics into the international market. In 2016, Mattermark ranked them among the fastest-growing startups to present at SXSW. European investing has also benefited VITL, providing them the resources to not only develop LANA, but also to distribute their goods across the United States.
The good news doesn’t stop there. Worldwide, Healthcare AI investments are expected to reach €5.6 billion by 2021, improve patient outcomes by approximately 40% annually, and reduce healthcare costs by approximately 50%. It’s what Atomwise has already done for developing communities in Africa, magnified.
To consumers, Healthcare AI sounds irresistible. Thankfully, they aren’t alone: investors and developers feel the same way, and are willing to help this vision become a reality.
Issues Still in Development
Industry and International Compatibility Problems
AI has undoubtedly revolutionized our approach to healthcare. Its novelty, however, means that we only have limited insight into to how different Healthcare AI applications interact as a group.
Qventus, mentioned above, takes over the “air-traffic control” of healthcare institutions. Ada, which applies the same intelligent decision-making concept to individuals rather than institutions, could augment Qventus’ understanding of patients and accelerate professional treatment decisions. However, it will be years before either start-up merges their software with another company, should they agree to partnerships at all.
Compatibility between healthcare systems themselves also poses a foreseeable problem. Turning again to Ada and Qventus, we see that the former, based in London and Berlin, could encounter roadblocks when transferring advice to Qventus, currently designed for US institutions. Privacy policies would also create challenges, as HIPAA-compliant Qventus would also have to comply with Data Protection Act guidelines. Furthermore, any program would have to adjust to several different healthcare standards to attract interest on the international market.
AI in healthcare has become the norm across the globe. Unfortunately, its integration into the global market will have to wait for globalization itself.
A Farewell to Receptionists?
When talking about Healthcare AI, it’s necessary to discuss not just what it contributes, but also what it takes away. As Healthcare AI gains leverage in triage and point-of-care (POC) diagnostics, it reduces the need for lower-level administrative and entry-level healthcare positions.
Nurses, for instance, may lose some of their stake in the healthcare community. Responsible for triage, vital statistics, and preliminary questions in emergency rooms and office visits, deep-learning devices – which transmit twice as much information in half the time – may cover this role. Even in in-patient care, nursing positions are at risk: with the introduction of programs that administer medication based on biometric evaluations of patient pain and smart technology that adjusts room temperature and aids patient mobility, human nurses may lose their place.
That said, even in the age of AI, healthcare technology is still at the risk of malfunctions, loss of connection, or disrepair. In addition, it also fails to provide the empathy and emotional care unique to human beings. While institutions that adopt Healthcare AI typically leave room for extensive IT departments in their budget, they must still rely on human help to respond to emotional situations and to fill in until technical difficulties are resolved. Although the nature of the work may change, nurses and administrators will have a place in healthcare until we devise error-free—as well as emotionally-connected—modes of operation.
Career Opportunities in a Fast-growing Field
Experts predict that the Healthcare AI industry will grow at a rate of 40% per year. With this growth comes new careers and new job opportunities. As AI involves humans as much as machines, the field offers a chance for both technology- and service-minded people alike to get involved in something totally original.
While rooted in Deep Tech, Healthcare AI is a multidisciplinary field that requires several different skill sets. The ideal candidate has health and tech experience, but any of the following abilities will catch the eyes of Healthcare AI developers:
- Technology design
- Graphical modelling
- Data learning theory
- Cognitive science
New Careers Mix Traditional Skills and Tech
Healthcare AI pairs machine learning with medical data to improve patient experience, reduce care costs, and increase doctor efficiency. As such, the fact that Deep Tech experts and medical professionals have taken the lead in Healthcare AI start-ups shouldn’t surprise you.
Academia has yet to merge the techno-medical partnership, but there’s no reason why you can’t work around the system. If you envision a career in medicine, you can expect to transition from a multi-person team to a partnership with one, virtual personal assistant. Although a basic understanding will suffice, technical knowledge works on a “more is more” paradigm: your relationship with technology — and, indeed, your performance in the field — will not suffer from knowing Healthcare AI inside and out.
In marketing, the same is true. Healthcare AI developers will seek out for tech-savvy communicators to sell their products to medical institutions and research groups. For them, a marketing rep will have to know not only the needs of hospitals, but also how AI can meet those needs.
Patient Access Associate
In research, teams license Healthcare AI programs to assist with studies. Patient access associates work with these teams to assure that they comply with company standards, and that the public understands their reported results. Working as the link between the lab and the world, patient access associates are:
- Good communicators
- Aware of healthcare research policy
- Expert data interpreters
- Detail-oriented quality analysts
Reasoning Technology Specialist
Reasoning technology specialists provide technical support for AI programs and assist users with their function. They also perform direct installations of AI programs to an institution’s servers, and assure its compatibility with institution-owned (and, in some cases, individual) devices. They work cooperatively with users of all ability levels to achieve the maximum benefits of Healthcare AI programs. They are, additionally:
- Skilled translators of technical details
- Experienced OS users
- Well-trained coders
- Patient problem-solvers
Machine Learning Specialist
Machine learning specialists combine a knowledge of machine learning and deep learning frameworks with large datasets to train AI software to recognize patterns, make decisions, or communicate to clients. They have an extensive education in computer science, as well as a strong background in math and statistics that affords them an immense capacity for quantitative reasoning. Machine learning specialists are also:
- Resilient in the face of difficulty
- Open interpreters of technical jargon
- Deep data analysts
- Cognizant of healthcare models and policy
One of the fastest-growing sectors of the Digital Health field, Healthcare AI looks especially promising for developers, marketers, researchers, and even patients. Follow our newsletter, Twitter or Facebook page for constant updates about careers, as well as developments in the field.