By 2025, 90% of U.S. hospitals will use artificial intelligence to save lives and improve their quality of care.
This should come as no surprise to anyone in the industry.
AI reduces the human effort spent on administrative and analytical tasks, empowering physicians and hospitals to focus on areas which truly require human intervention. In fact, half of U.S. hospital executives are actively investing in new AI applications.
Efficient AI solutions already save hospitals millions of dollars every year, but we’ve only just begun to see their potential impact. Robots might not replace doctors, but hospitals equipped with AI tools willoutperform non-AI driven hospitals—and soon.
Investing in AI-driven healthcare presents a significant opportunity for early adopters. Right now, we’re at the sweet spot: there is enough proven technology on the market to confidently invest on the ground floor, but we’ve still only scratched the surface of future AI applications.
So, which areas are ripe for AI disruption?
In this article, I will explore the top eight most exciting AI innovations in healthcare. And if you want to stay ahead of the curve, make 2018 the year when you experiment with one or more of these ideas.
1. Automatically Monitor Patients in Intensive Care
New AI analytic tools can monitor patients to predict complications at the earliest possible moment.
Inpatient monitoring is a significant resource drain: in some hospitals, nurses spend more than 60% of their time monitoring patients. Even with this oversight, postoperative complications remain a significant cause of mortality in inpatient settings.
AI augments specialist care by analyzing patient vitals and other data to identify anomalies before they’re obvious to nurses and doctors.
For example, aspiration pneumonia is a major risk for intubated patients. To address this issue, an Israeli startup called Art Medical produces sensor-enabled smart feeding tubes which monitor patients for complications like pneumonia, without the need for constant human oversight. At-risk patients trigger an alarm which is sent to the central nursing station.
The Cleveland Clinic has also used Microsoft’s AI assistant, Cortana, to monitor patients in the ICU. When an individual patient’s data indicates a potential issue, the AI-assistant notifies care providers. This allows clinicians to detect the earliest signs of cardiac arrhythmia.
Robotic monitoring solutions like these reduce “alarm fatigue” from false alarms, helping care providers manage their time effectively. More importantly, they keep vulnerable patients from slipping through the cracks. Within the next decade, smart monitoring devices with the ability to flag at-risk patients are set to become as ubiquitous as cardiac monitors.
Want to get ahead of the game? Consider investing in intensive care monitoring. This is big business; vast quantities of data can be collected from in-hospital patients to improve the overall quality and effectiveness of care. Plus, startups currently operating in this space today have barely scratched the surface of this technology’s potential.
2. Automate Patient Outreach with Chatbots
AI-powered chatbots will take over administrative tasks ranging from referrals and patient satisfaction surveys to appointment booking and billing.
Thanks to AI-equipped chatbots, anyone with a smartphone or computer can be referred to a specialist, verify their insurance, describe health symptoms, and set up medical appointments (and follow-ups) in advance.
Many of these applications have dramatic effects on both patient outcomes and the provider’s bottom line.
Unlike human employees, well-trained AI is highly accurate and can work around the clock. But it doesn’t stop there: specifically-designed chatbots can save hospitals millions of dollars on top of reduced labor costs.
For example, a single missed appointment costs the provider an average of $200. Smart appointment booking, which can predict no-shows before they happen, could collectively save U.S. clinics and hospitals up to $150 billion annually.
Healthcare execs are rushing to implement these solutions to reach more patients and take advantage of the efficiency of robot medical assistants. And the numbers back it up: the AI chatbot market is currently projected to hit $1.5 billion by 2024.
Whether you buy an off-the-shelf chatbot for your healthcare business or decide to invest in one, here’s one thing we are sure about: no matter what you think now, in 5 years chatbots will be a cornerstone of hospital management. It’s not a question of whether this will happen, but a question of who has the foresight to implement it today.
3. Improve Hospital Efficiency with AI Command Centers
Hospitals generate massive amounts of data that can be leveraged with AI to increase operational efficiency.
As the U.S. moves towards value-based care models, hospitals increasingly need to minimize lengths of stay to reduce costs. Modern medical centers are also logistically complicated and require sophisticated technology to manage resource allocation and patient flow.
AI can help with both of these issues by streamlining care management systems and finding potential inefficiencies in real time.
For example, Johns Hopkins Hospital uses an AI command center equipped with predictive tools to make the hospital’s operations more efficient. The system pulls together data from multiple IT systems at once, prioritizing activity in specific sections of the hospital based on need. This allows the hospital to assign beds faster, discharge patients more quickly, and accept more complex and time-consuming cases.
Operations-focused AI can also manage OR, Cath lab, and GI lab scheduling. By predicting how much time each procedure will take, these new apps reduce the amount of time high-volume procedure rooms go empty. In addition to reducing wait times and maximizing the use of operating rooms and lab facilities, AI assistants reduce the likelihood that other surgeries will have to be rescheduled unexpectedly. Because ORs and labs are often the most lucrative aspects of a hospital system, making them more efficient directly impacts your bottom line.
Interested in leveraging hospital data to improve communication and workflow efficiency? Consider looking into scheduling and operations-focused AI. At DAP, we’re already working with many companies building apps to streamline hospital management; these solutions will only become more popular in the coming years.
4. Remotely Monitor Chronic Conditions
Smartphone apps and wearable devices use AI to help monitor patients with chronic conditions.
Many patients with chronic conditions can’t self-monitor outside of the hospital. In the case of diabetes, this can mean repeated, expensive hospital visits or even deadly blood sugar crashes. Doctors, meanwhile, spend valuable time manually monitoring these patients’ symptoms.
Tempo, a Dutch medical device company, replaces this process with non-invasive stickers and an AI platform which can forecast and manage blood glucose levels automatically—and remotely. Strikingly, initial studies saw a 9% reduction in blood sugar lows in patients using these stickers compared to those actively monitored by physicians.
Another startup, Sensely, also uses AI to monitor patients with chronic conditions. Their application uses Bluetooth smart devices to take patient vitals, flagging those potentially at risk and helping them to book medical appointments.
Such applications mark only the very beginning of massive innovation in health tech. Physicians can treat virtually any disease cheaper and more effectively with the help of AI-assisted tools like these wearable devices and smartphone apps. If you’re looking to get ahead of the game, consider looking into one of these applications, or building your own.
5. Enhance Robotic Surgery & Improve Surgical Outcomes
AI applications improve the accuracy, safety, and efficiency of both robotic and manual surgery.
In 1992, a 64-year-old man had a successful hip replacement from RoboDoc, a robotic surgical tool. In the nearly three decades since then, robot-assisted surgery has grown by leaps and bounds.
In fact, it is projected that by 2025, as much as 80% of surgical procedures will be performed at least partially by robots.
Robotic surgery offers many well-known benefits, including the ability to make microscopic stable movements and allow surgeons to work remotely and perform minimally invasive surgeries with shorter recovery times.
But what many don’t know is that the most common robotic surgery applications are already partially automated using AI—and this is one of the most promising areas for future innovation.
For example, the da Vinci system, the most well-known robotic surgeon, uses computer-aided vision and reduces human hand tremors from the surgeon controlling it using machine learning; though it is the longest-running big player (FDA approved in 2000), other start-ups and research groups are also rapidly innovating in this space.
In late 2017, the semi-autonomous surgical robot “STAR” (Smart Tissue Autonomous Robot) actually outperformed surgeons on a tumor-removal task. Such advances will only become more common as more investment and innovation takes place within robotic-surgery.
Even when surgeons work without robotic assist, AI can help automate and regulate surgical procedures to quantitatively assess surgeons’ skills. It can also reduce human error by prompting surgeons with safety checklists and suggesting actions to streamline surgical processes.
Robotic surgery is a relatively mature field in which AI improvements are poised to make a significant and immediate impact for those with the ability to research and design the tools. Minimally invasive and outpatient surgeries fueled by these technologies are already seeing massive growth worldwide.
In short, if you aren’t already among those actively investing in AI-driven surgery, now is the time to begin looking into ways to leverage the power of AI to improve surgical outcomes.
6. Strengthen Hospital IT Security
AI helps security systems identify critical issues using human behavior and analytics.
90% of U.S. hospitals experienced a cyber attack within the last two years. You read that correctly. Ninety percent.
The costs of these incidents are astronomical, and the loss of institutional trust this causes is hard to overestimate. In the past five years alone, security breaches have cost the healthcare industry over 30 billion dollars. In the future, cybercrime will only become more common as hospital systems and their data move online and become more lucrative targets.
In simpler times, password systems and anti-virus software were enough. Those days are long gone.
Healthcare environments make security upgrades particularly challenging because data cannot simply be locked down; clinicians, billing, and administrative staff all still need to access patient records., Yet, breaches most commonly come from internal sources.
Luckily, AI applications are quickly expanding to meet IT security needs. Smart software can augment traditional metrics like passcodes with biometrics. AI can “learn” the patterns typical of an IT system to model normal human behavior. When an anomalous, potentially malicious pattern occurs in the system, the AI app can identify this much faster than even a highly trained employee, stopping attacks before they become a problem.
Some AI-driven IT security apps are already on the market (take, for example, Cognetyx’s Ambient Cognitive Cyber Surveillance and the more famous IBM’s Watson for Cybersecurity), and AI internet security promises to be an area of rapid future growth. According to Gartner, 75% of security software applications will include AI features by 2020.
Unfortunately, cyber attacks are a matter of if, not when. Especially if you’re already innovating in the medical IT arena, incorporating AI into your existing infrastructure is key to proactively ensure that your systems are equipped to handle future security threats, and keep sensitive patient data safe from cybercriminals.
7. Discover New Drugs Faster
AI also has promising applications in drug discovery, which must constantly outpace disease evolution.
Drug discovery is a complicated, time consuming, and expensive process which requires constant biological discoveries of new compounds. Only a small percentage of these compounds make it through the testing process, which costs over 2.5 billion USD per drug and takes up to a decade to complete.
AI promises to drastically reshape this entire process, and big pharma companies are already signing multi-million dollar contracts with AI startups to speed up the drug discovery process.
Way back in 2015, Atomwise designed an AI drug-discovery app which could predict how molecules are likely to stick together. Using this method, they were able to discover two drugs that could treat Ebola in a few weeks instead of years.
More recently, Pfizer partnered with IBM’s Watson in 2016 to create an application which can read the enormous and rapidly expanding body of cancer literature to identify possible combinations of compounds which would not be obvious to human readers.
If you’re involved with pharmacological development and testing processes in any way, consider investing in AI-supported drug discovery methods in 2018. Drug discovery is already big business, and likely to see continued demand as the need for new antibiotic, epidemic-prevention, and cancer-treating drugs accelerates. Even a small improvement in drug-approval rates could mean several billion dollars in profit, and rapid response to new disease strains could save millions of lives.
8. Use Machine Learning to Aid in Diagnostics and Clinical Decision-Making
AI supercomputers can process tens of billions of medical images or patient records in seconds. The end result? Enabling medical professionals to make faster and more accurate decisions.
Diagnostics is one of the most expensive and complicated processes in healthcare and is often prone to human error—with potentially fatal consequences.
Due to its complexity, it has also proved difficult to automate. However, vendors and major academic centers worldwide are seeing success this year with a wide array of artificial intelligence neural networks used for clinical diagnosis and decision-making.
The tech startup Aidoc uses AI to comb through hundreds of brain CT scan slices per second, searching for irregularities hundreds of times faster than a radiologist. Any anomalies it finds are flagged as high priority for the radiologist. Even preliminary descriptions of imaging and measurements of lesions can be completed by the computer.
IBM’s Watson, one of the most famous supercomputing AI applications, offers both an image-analysis solution and a machine learning algorithm that pulls relevant information from medical records. Both of these products assist providers with diagnosis and decision making.
Finally, Google’s DeepMind health project is developing algorithms for use in health record aggregation; since individual patient records become very large and include information from a variety of sources over time, this technique improves physicians’ ability to diagnose and monitor diseases based on patient history.
What does this mean for industry leaders in the U.S.? In short, diagnostics is and will remain an area ripe for new investment and innovation.
Beyond the economic benefits of AI implementation, these algorithms can ease decision fatigue and the intense case burdens of medical staff, allowing them to focus on the complex and unusual cases which truly require creative human problem-solving.
Adding AI to existing clinical procedures also promises to substantially reduce human error. The end result: dramatically improved patient outcomes.
In this article, we’ve covered some of the most promising areas of AI investment and development in 2018, including:
- Automating patient monitoring with smart inpatient AI applications
- Using chatbots to automate patient communication and administrative tasks
- Managing hospital operations with centralized AI command centers
- Monitoring chronic conditions with wearable devices and apps
- Enhancing robotic surgery
- Improving IT security to protect vulnerable and sensitive data
- Speeding up the drug-discovery process
- Augmenting clinical decision-making with AI image and record-analysis algorithms
According to a recent journal article published in Brazilian Radiology, “if AI is the new electricity, data is the new coal.”
Hospital systems generate millions of gigabytes of data every single day. Using modern, artificially intelligent-driven solutions, healthcare providers can leverage this virtually unlimited information to provide cutting-edge, individualized, and precise medicine.
Already, healthcare leaders are seeing significant savings from implementing AI solutions to improve patient-doctor communications, hospital administration, diagnostics, and clinical decision-making. This year, we predict continued growth with a tangible impact on patient health.