AI Is a New Must for Medical Software
Article Aug 08, 2018 | by Yaroslav Kuflinski
It appears that artificial intelligence is an apple of discord in disruptive technologies. On the one hand, even those who create new products based on AI or implement it across their internal processes, are worried. “Artificial intelligence and robots will kill many jobs,” predicts Jack Ma, CEO of Alibaba. IBM’s Ginni Rometty supports this grim statement, claiming that “100 percent of jobs will be somehow affected by technology.”
On the other hand, there is healthcare AI. Despite being one of the most cautious domains when it comes to tech disruptions, the industry supports AI incorporation into all things medical software development, from EHRs and remote patient monitoring to population health solutions. The reason why clinical stakeholders welcome artificial intelligence is that it is used for dealing with bottlenecks and pitfalls across the care continuum rather than for replacing human health specialists.
Pre-care: AI addresses patient concerns
Providers aren’t ever-present in patients’ lives, and they can’t help when they don’t know their assistance is needed. Patients, in their turn, can be uncertain whether their health complaint requires an office visit.
AI can bridge this gap, playing a role of a patient’s personal health advisor, able to understand the set of symptoms and output relevant suggestions. Depending on what is bothering a patient, the options vary from taking painkillers or antipyretics to scheduling an appointment with a Primary Care Physician or even calling emergency.
This way, patients can clarify their health status in the comfort of their homes and either take a good nap or seek further medical assistance. Some AI-powered systems also compose a list of reported symptoms, making it easier for a patient to recall them in the doctor’s office.
However, it is important to note that these technologies function more as a source of reference than as full-fledged medical diagnostic tools. Their performance in peer-reviewed studies against physicians as a diagnostic tool indicates it is likely to remain so for the foreseeable future.
Examples: Your.MD, Babylon, Ada, Ask NHS, KidsMD.
Care delivery: AI cuts on inefficiencies and medical errors
Decluttered waiting rooms
Patients are always being tested for their patience. It all starts at a receptionist’s front desk, where he or she either gives patients paper charts to fill in or makes them dictate all needed information. Then the patient has to sit down in a waiting room for some time, from a few minutes to several hours. The patient’s plans can be ruined because no one has a full day to spend on waiting for a 20-minute appointment (at best).
AI offers numerous ways to solve the problem of cluttered waiting rooms and unbearable periods of plain sitting and doing nothing. There are custom apps that allow patients to get through check-ins by first answering 20 questions to help the AI to rule out the unnecessary ones and offer more focused enquires. These interactive systems are more patient-centric and they require less time to gather all the needed information.
Other solutions create one-stop systems for patient convenience; they are mostly used in small-sized practices. Such apps are installed on patients’ devices. They allow synchronization with smart wearables to collect patient data and compose it into reports for fast check-ins and streamlined office visits as well as offer telemedicine functionality via chat and video calls to reduce wasted time significantly.
Examples: Carbon Health, DrChrono, Forward.
Improved diagnostics and treatment
Medical errors are considered the third leading cause of death in the U.S. after heart disease and cancer. Therefore, it is an absolute must to create tools that can decrease the rate of preventable medical errors by supporting evidence-based decision-making.
For AI, diagnostics and treatment mark the area where it can truly shine because of its powerful data processing capabilities.
Artificial intelligence can grasp patient-reported symptoms in line with a particular patient’s EHR history, offer a number of possible conditions, and suggest on further steps, such as test and procedures. It can assist in automating medical image analysis (X-ray, CT, MRI, etc.), streamlining genome sequencing, and improving pathology accuracy.
After a diagnosis is made, the algorithms can provide a few treatment plan variations with calculated dosages, considering the patient’s allergies, available clinical trials, drug interactions, and innovative approaches. AI also powers robot-assisted surgery to ensure unpreceded accuracy in micro-invasive procedures, such as neurosurgery.
Examples: Infervision, Optellum, IDx-DR, da Vinci, GNS Healthcare, Oncora Medicals, Amara Health Analytics.
Follow-up patient support: AI extends care outside hospital walls
Patients are people, and people are flawed. Forgetting to take a pill, skipping a workout, breaking healthy nutrition routines; everyone has been guilty of these crimes at some point. But if a patient gets discharged from a hospital after an acuity, starts recovering from a surgery, develops a chronic condition, or plans to become a mother soon – there’s not much room for discipline gaps.
It is highly important for patients to act according to their doctor’s recommendations, but adopting new habits and reshaping their regimen can be overwhelming. Of course, patients can’t get 24/7 human support so artificial intelligence can be really useful to meet them where they stand and carefully guide them through all new experiences.
The possibilities are enormous, they are usually packed into mHealth apps and can include:
• Processing patient-reported data (mood, sleep, pain, weight, blood glucose, temperature, etc.) and creating reports on health status changes
• Offering tailored therapeutic education
• Notifying patients about taking medications
• Creating nutrition guidelines and suggesting relevant recipes
• Composing workout sets for a particular condition or goal
With AI and its capabilities, patients won’t feel left out having to deal with their new reality alone. The algorithms can get a grasp on a patient’s overall wellbeing and will notify the provider in case something is a bit off – a missed pill, negative trends in temperature, or bad mood for several days in a row.
Examples: Wellframe, Ginger.io, DIABNEXT.
Not instead of but together
We really like to think of AI in healthcare as a friendly force that augments human powers way beyond the natural limits, allowing doctors to save even more lives.
Knowing everything that artificial intelligence can bring into care delivery – early diagnosis, more tailored treatment, predictable outcomes and post-discharge patient support – we can be sure that AI is becoming a must in medical software and isn’t going anywhere anytime soon.
Yaroslav Kuflinski is an AI/ML Observer at Iflexion. He has profound experience in IT and keeps up to date on the latest AI/ML research. Yaroslav focuses on AI and ML as tools to solve complex business problems and maximize operations.
We spoke to Andrew Howley from Adventure Scientists,a pioneering not-for-profit organization that seeks to unite skilled adventurers with scientists keen to receive valuable data from remote areas, to learn more about the initiative and the impact their projects are having in the scientific community and beyond.READ MORE
If you work in science, chances are you spend upwards of 50% of your time analyzing data in one form or another.However, it's easy to get lost when it comes to the question of what techniques to apply to what data. This is where data mining comes in - put broadly, data mining is the utilization of statistical techniques to discover patterns or associations in the datasets you have. Here we provide an overview of the critical steps you'll need to get the most out of your data analysis pipeline.