Photo by Pexelson
What a world it would be if there was a fair patient-doctor ratio, diseases were diagnosed early on and treated in a timely manner, and no one died? Well, first of all, there would be a serious population imbalance, but that is a discussion for another day. For now, we know that technological advancements in healthcare can significantly improve the quality of life as well as help us realize the goal of maximum life expectancy and minimum fatality rates. Gone is the time when Artificial Intelligence (AI) was talked about in future tense. AI in healthcare has become the reality of today, attracting heavy funding from several investors.
The rise in consumerism in healthcare has necessitated the incorporation of high-end technology into the existing processes to deliver convenient patient care. Definitive Healthcare listed Telehealth, AI & Machine learning, and consumerism among the top 8 healthcare trends of 2019 in a recent webinar. They pointed out how 2019 will be the year of numerous M&As as huge healthcare systems and small hospitals join hands to deliver high-end services using improved technologies. The global investments in healthcare accounted for almost $9.5 Billion across 698 deals in 2018. With a primary focus on data analytics, wearables, mHealth applications, and practice management products, these investments are driving the growth of Health 2.0.
Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) have the potential to revolutionize the way care is given and received. The Healthcare IT industry is in a dynamic state and experiencing a paradigm shift. After years of resistance, healthcare is embracing the digital transformation by losing a little of its grip on the prevalent silos, if not completely breaking them up. Artificial Intelligence is, undoubtedly, promising for the healthcare industry, however, it is in a relatively early stage. Despite the many predictions, the change managers may face some blind spots in the process and have to be geared up to tackle their way around them.
The speculations, both positive and negative, around the implementation of AI in healthcare have made it necessary that we dive deeper to properly analyze its scope and promises. Detailing the industry-wide changes being made to welcome AI into the healthcare realms, let us try to have a dekko at the AI-driven future of healthcare.
Industry Debutants Bringing AI to Mainstream
As of the second quarter of 2018, more than $4 Billion had been invested in the AI startups. Investors are throwing money just hearing the name of AI in the sales pitch, trying to grab the opportunity by its hair before it turns its back on them.
These are the top 5 AI-based startups leading in healthcare technology:
BayLabs: Computer Vision for Clinical Decision Support
BayLabs is working to deploy deep learning algorithms with Ultrasound to detect red flags in cardiovascular imaging that a human eye might miss. By using AI, it is striving to reduce the frequency of errors and support clinical judgments while diagnosing symptoms, treating diseases, and prescribing medications.
CrossChx: Healthcare Identity Management
With a promise of automating about 78% of healthcare admin tasks, CrossChx develops AI bots that take over the cumbersome tasks of checking insurance eligibility, getting prior authorizations, scheduling appointments, and sending reminders. Automation of such tasks not only frees healthcare practitioners from these tiring activities but also decreases the average expenditure of healthcare systems by about 30%.
Prognos, Big Data Analysis (Big Data and Healthcare Marketing)
Having the largest database of diagnostics information sourced from more than 50 disease areas, Prognos utilizes 1000+ algorithms and big data analytics to predict a possible risk of disease onset.
AliveCor, Remote Patient Monitoring
The wearable industry is proliferating. By employing machine learning, AliveCor is using wearables as a remote electrocardiogram. This enables healthcare providers to effectively and actively monitor a patient’s condition, thus, reducing the need for readmissions due to lack of supervision and care.
Sensely, Virtual Care Provider
Taking advantage of telehealth, Sensely is making healthcare affordable with AI. With the help of AI, it analyses and interprets a patient’s symptoms, and provides a diagnosis accordingly. Doing so, it provides the comfort of healthcare-at-your-doorstep and reduces costs by preventing ER visits for trivial reasons.
Industry Veterans Joining Hands to Deliver Better Care
There is a continued rise in the M&A activities across industries for collectively imparting enhanced clinical efficiency, scalability, and value-based care. Leveraging the explosion of data, the healthcare industry is taking the first step to deliver superior patient experience.
The acquisition of Pillpack by Amazon, Aetna by CVS Health, or merger of Express Scripts and Cigna – all narrate the same story – the industry is prepared to ride the AI wave with absolute finesse. Cross-industry alliances and partnerships are becoming a trend as major players from other industries venture into healthcare. It is expected that post-acquisition of the online pharmacy Pillpack, Amazon is prepared to enter the healthcare insurance sector.
Walgreens has introduced “Find Care Now” digital health service to make healthcare easily accessible to patients across neighborhoods. With a vision to create a digital marketplace for patients, it acquired almost 2000 Rite Aid Stores, joined forces with Humana to offer senior-focus services in Kansas, allied with Home Chef for retail meal kits, purchased files of 185 Fred’s pharmacies, and partnered with Microsoft for creating an interconnected network of retail stores and health information systems.
Wearables Become the Engine for AI-pulled Wagon (AI and Telemedicine?)
The Global Digital Health market is expected to hit $504.5 Billion by 2025, with major revenues streaming from mHealth, wearables, and health-monitoring apps. As per Business Insider, key digital health solutions like EHR, digital therapeutics, telehealth, AI, wearables, and blockchain are the foundation of the industry’s digital awakening.
A recent report from Apple Heart Study conducted by Stanford University School of Medicine showed a promising impact of wearable technology in the healthcare system. Launched in November 2017, the study had a sample size of 400,000 participants and used data obtained from Apple watch. The wearable technology was used to detect heart pulse and identify any irregularities, which often turned out to be Atrial Fibrillation on subsequent testing. With a 71% positive predictive value, the algorithms can help diagnose the condition at an early stage so that the affected individual can seek required medical assistance on time.
The Apple Heart Study has, to some extent, proved that the hype around AI and predictive analytics isn’t for nothing. AI applications in healthcare, paired with machine learning or deep learning algorithms, can change the face of preventive healthcare by reducing deaths caused by delayed medical attention. Wearable technology can also allow doctors to keep a check on their patients through remote location and see to it that the patients are properly adhering to the prescribed medication. The active tracking feature of wearable devices alerts both patients and doctors when they need to pay each other a visit.
Cognoa, with its licensed Google-glass based AI technology, has devised a machine-learning based application system for children suffering from Autism Spectrum Disorder (ASD). Developed at Stanford University of Medicine, this ‘Superpower Glass’ has resulted in improved socialization skills by providing real-time feedback on social situations.
A smart fabric with in-built motion sensors has been developed by a team of researchers at Dartmouth College to prevent injury and help recovery among athletes and physical therapy patients. The high sensing capabilities of the fabric gather data from skin deformation, resistance changes, and pressure. A thin silver layer, embedded in the fabric for electrical conductivity, then communicates this data to a connected monitor. The monitor uses this information to determine the joint rotational angle, which can help instructors and therapists to provide treatment and coaching accordingly.
The wearable technology is still in its nascent stage, but it is speedily proliferating on the maturity curve. In addition to the trending smartwatches, we can soon expect to see smart monitoring devices in various forms for remote patient care taking the market by storm.
Breaking the ‘Iron Triangle’ in Healthcare with AI
The Iron Triangle concept was introduced by William Kissick in “Medicine’s Dilemma: Infinite Needs vs Finite Resources”. Access, Cost, and Quality –the three angles of an Iron Triangle are engaged in constant competition with one another. An attempt to improve one factor mostly harms another, leading to serious repercussions. This is why the Iron Triangle is often believed to be unbreakable.
Enhancing quality raises costs and limits access. Decreasing costs and improving access to may diminish quality. Inaccessibility becomes a concern in the light of high costs and quality. Attaining the equilibrium among these three factors is as difficult as penetrating an iron wall, hence the name.
As Artificial Intelligence in Healthcare is taking over the reins, one of the major benefits it brings is affordability. AI can effectively result in low costs, improved care, and easy accessibility by disrupting the traditional methods and getting access, cost, and quality to work in tandem. Enabling remote patient monitoring and telehealth, AI and Machine learning are efficient in rectifying the imbalance in doctor to patient ratio.
The Iron Triangle can be unlocked with AI because it solves the problem of ‘finite’ resources against ‘infinite’ needs.
Through digitization, encouraging engagement, and precision diagnostics, AI algorithms make operational processes affordable. By channeling communication between healthcare providers and receivers, AI drives high engagement and better patient management. Using machine learning and deep learning algorithms, AI can offer expert advice through accurate diagnosis.
AI Innovations Disrupting Healthcare
The University of Pennsylvania performed the first-ever robot-assisted spinal surgery in 2017 on a Chordoma-suffering patient. The doctors used Da Vinci’s robotic arms to successfully remove a rare tumor from the patient’s spine.
In Sweden, a baby boy was born after a robot-assisted uterus transplant. This was a first of its case as robotic arms were used to operate a keyhole surgery on the baby’s mother.
Robot-assisted surgery is one of the most promising applications of AI in healthcare, followed by virtual nursing assistants and administrative workflow assistance. As evident from the examples above, AI applications are already impregnating the healthcare silos. They can be highly efficacious in solving the staff shortage crisis in the industry.
The coming years will see a further increase in the AI-driven innovations in various sub-fields. Let’s look at a few:
- AI has shown and delivered the capabilities to transform mental healthcare. Offering the provision of continuous monitoring through smartphones and applications, Artificial Intelligence is especially useful in providing integrated group therapy and cognitive behavioral therapy to patients suffering from depression, substance abuse, or eating disorders. In a similar fashion, AI can also identify dysfunctional patterns that may cause brain damage leading to seizures.
- Voice-based search is getting mainstreamed in every industry and healthcare is no exception. Using voice-first technology, AI enables smart speakers through Natural Language Processing.
- Healthcare providers can receive alerts if one of their patients is facing domestic violence. This is quite unconventional, but then, so is AI. By identifying anomalies in the reported history of a patient with the current injuries, AI algorithm can raise a red flag so that the care providers can intervene wherever required.
- By facilitating the delivery of remote care, AI can prevent fatalities caused due to late assistance. AI-driven tools can detect an impending stroke by detecting irregularities in the pulse data obtained from the wearable device of the patient and notify the user to schedule an appointment with their care provider.
- Soon, hospitals will be using natural language processing (NLP) and machine learning (ML) for billing and medical coding purposes. This will take a heavy chunk of the administrative burden off the providers.
- AI algorithms can help diagnose diseases like glaucoma and diabetic retinopathy earlier through a detailed analysis of patterns in pixels.
- Clinical diagnostics will be the most affected by AI impregnation. It can be especially helpful in developing countries where a disease outbreak is a common occurrence. With the help of deep learning tools, the process of quantifying malaria parasites in blood samples can be automated, resulting in faster detection and thus, faster treatment.
- AI tools can skim through a person’s SM posts and EHR notes to identify words that might indicate the possibility of suicide and self-harm. By doing so, it can send out alerts so that the person can receive the required attention in time.
AI in Healthcare – Opportunity or Trap?
Allowing the shift from ‘caring for the sick’ to ‘preventing the sickness’, Artificial Intelligence is critical in improving results and reducing expenses. Thanks to AI, there is a decline in the cost of clinical research, drug development, hospital care, and health insurance.
The benefits of AI sprawl from automated operations, preventive care, to precision surgery and better patient outcomes. Technology in healthcare, by reducing the overall spending, can effectuate annual savings of about $150 Billion in the next seven years. With the potential of addressing about 20% of the unfulfilled clinical needs, AI in healthcare will attract major investments from both the public and private sector. While scalability remains an issue for AI applications, there are three major areas that will get the most funding – Digitization, Engagement, and Diagnostics. Through enabling digitization, AI in healthcare streamlines operational processes. While encouraging interaction between patients and healthcare providers, AI-based applications support proactive dealing of deadly diseases. Its analytic system can offer health advice to the patients and notify them when they need an expert’s opinion.
The AI spectrum is not completely black and white. It has a few intermittent shades of grey that need to be taken into consideration before taking a full-blown approach to its adoption. In a survey, one-fourth of the sample size said that they would not use AI-powered health services. This apprehension basically stems from the absence of a human-level cognition.
It is rather easy to manipulate the AI algorithms, as even a slight tweak in the data can corrupt the produced outcome. As AI learns and grows by feeding on data, the susceptibility of errors made by AI systems is extremely high, given the frequent occurrences of data breaches and cybersecurity threats. The inability of AI to take a judgment call in such situations turns out to be one of the biggest risks that its implementation will expose the healthcare industry to. Some bad players in the industry may even take advantage of this shortcoming to leverage financial gains and rip off patients based on an orchestrated diagnosis.
Being a machine, at the end of the day, AI cannot compare to human doctors. One minor defect can cost the lives of the patients. Although AI is being implemented to decrease the possibility of medical errors, the same tool can produce reverse effects due to unintended misdiagnosis. This dark side of Artificial Intelligence creates the need for two things – one, regulatory compliance for AI in healthcare, and two, impeccable healthcare IT software testing.
The openness to accept AI requires that authorities take care of the legal aspects and set down proper guidelines and regulations. An AI tool keeps feeding on new data and information and applies ML algorithms to constantly evolve based on that new information. Therefore, even if the authorities lay out compliance rules for a particular AI application or system, they will be irrelevant for the evolved algorithm. Finding a way around this will be a challenge, that too a gigantic one.
Implementing AI-based systems demand seamless flow of data, uninterrupted interconnectivity between a series of devices, and defect-free showcasing of outcomes. Cybersecurity and data privacy being the main concerns take the highest priority. To ensure that all this comes together and drives the desired results, AI developers need to incorporate optimal software quality assurance services. Healthcare domain testing for AI applications may prevent fatal blunders as the stakes involve life and death. Intricate and detailed healthcare application testing should be carried out to identify potential vulnerabilities and prevent glitches in terms of security and performance of the AI systems.
There is no doubt that AI in healthcare will open many windows of opportunities, but it should be seen that those windows do not provide entrance to the dark forces.