- Excessive and unnecessary wait times in healthcare are hindering patient outcomes
- Artificial intelligence is moving us toward an “instantaneous healthcare system” today, and the pace of innovation is increasing rapidly
- AI can automate insurance claims, analyze complicated imaging or charts, book follow appointments, and a lot more
- Advances in AI are already speeding up healthcare and improving outcomes and rapidly improving with more data and usage
Much of the attention on healthcare has focused on lower patient outcomes even in the face of ever-escalating costs. And there seems to be no end in sight. But lost in that understandable outrage is another important story around healthcare: the agony and emotional stress of waiting.
A doctor visit can be cathartic, helping to put lingering health anxieties to rest or charting a path toward future wellness. But today’s physicians and their staff are under incredible strain, dealing with front- and back-office busywork. Put simply, medical practices have a lot less time to do the actual work – helping people live longer, more fruitful lives.
Today’s doctors perform miracles every day. But at nearly every step it feels like there’s just too much waiting. Waiting to get an appointment, waiting in the waiting room and in the exam room and, then, waiting for lab results or x-rays to come in, waiting for insurance approvals, waiting for referrals. It goes on and on – we can do better!
Imagine a world where waiting is a thing of the past. Where diagnoses are virtually instantaneous, where you show up for your appointment and see your doctor immediately, where pre-approvals are requested and granted before even walking out of your doctor’s office – this is imminent, and artificial intelligence is a major advancement that will help get us there.
My Own Story
Everyone has a healthcare story. For me, it was awaiting news about whether my three-week-old daughter would be approved to take an air ambulance from UCSF Benioff Children’s Hospital to the Children’s Hospital of Philadelphia to treat her congenital hyperinsulinism diagnosis. It was brutal: she underwent 44 days in neonatal intensive care units, over 500 blood sugar tests, one nuclear MRI, one pancreatectomy and a 24 hour fast before she was deemed cured. But the amount of advocacy and pushing required to get her the treatment she needed is far beyond what the average American can or should be expected to undertake.
The cherry on top of it all was the over $2 million in accrued costs (I shudder to think what we would have done if we were among the 27 million Americans without health insurance). That harrowing experience drives me every day to make healthcare more reliable, accessible and immediate.
Anxiety of Healthcare
Studies have shown that the anxiety of waiting for test results or diagnoses can actually lead to worse outcomes. At every step of the process, patients may feel like they are running into a brick wall when it comes to getting answers. Non-emergency MRI imaging takes anywhere from one to two weeks and lab results average two weeks. We need to do better and improve health outcomes faster.
Things that should be automatic, like getting doctor referrals, reaching insurance plan administrators and seeking pre-authorizations for necessary procedures are a tangled thicket of bureaucracy, leading to unnecessary delays and inferior patient outcomes. Everyone knows that patients benefit the sooner serious conditions can be addressed, so why are we still stuck waiting?
Waiting is associated with depression and anxiety, which can put further strain on the healthcare system. And delays affect minorities, women and those with lower incomes more directly, increasing costs and putting further burdens on the overall system.
Using A.I. to take out these inefficiencies, like the cost of administration, could help to drastically reduce the cost of healthcare, making it more accessible for more people. Using intelligent systems to drive out inefficiencies and get quicker results will lead to better outcomes for everyone. Lowering the cost of healthcare and making it more instantaneous should be a shared goal.
Healthcare Pain Points
I co-founded a software company to address pain points in healthcare because I believe artificial intelligence software along with improvements in care are leading us toward a brighter future of wellness. A.I. can already do incredible things, like scan x-rays and review medical charts faster and more accurately than we ever thought possible just a few short years ago, as well as cut down on paperwork and lessen the administrative burden on back-office staff. And with the pace of advancement, A.I. is getting better and less expensive to use every day.
We are in the midst of developing an ecosystem where lab results, forms, charts and appointments could all be completed in the time it takes you to hop off the exam table. That will make patient care better by making it more affordable and more effective. A.I. won’t – and shouldn’t – replace humans in healthcare, it will supercharge humans and dramatically improve health outcomes for our families and friends.
Consider all the inefficiencies in the healthcare system today. According to a recent study, in the fifteen largest U.S. cities, it now takes an average of 26 days just to schedule a physician appointment as a new patient. And we are now adding two days onto that wait every five years. At our most desperate moments – in the emergency room – wait times now exceed two hours. Worse, after you’ve been seen it can take agonizing days or weeks to get results back.
Our Vision: Centralize, Standardize & Automate
Despite the availability of EMR systems, many healthcare providers are still on a paper-based system. This lack of centralized data leads to delays or the need for retests or additional screenings. Centralizing the functions that make the healthcare system run would provide a big boost to artificial intelligence systems allowing them to more quickly and automatically detect aberrations in a patient’s history.
Human-led standardization – systems that organize and record data in common ways – will be completely replaced by A.I. that can map out those commonalities. What if blood samples could be automatically analyzed 24 hours per day, with no human intervention, so that warning signs could be detected earlier? That’s how healthcare starts to become virtually instantaneous.
Machine learning and computer vision hold particular promise for healthcare, by analyzing millions or billions of points of data to recognize trends that humans might miss. Such systems might notice correlations that can influence preventative care, like certain work environments globally and a tendency to develop injuries or cancer. And chart and lab result analyses using computer vision will detect aberrations, even at a microscale, sooner and with higher accuracy – early diagnosis always leads to better outcomes. And with more data being fed into A.I. systems, they will rapidly improve.
Now imagine if A.I. systems, based on those analyses, could schedule a patient’s next scan or immediately request insurance preauthorization for the next procedure. This is not only possible, but imminent, as machines learn to “speak” with one another.
A.I. In The Back Office
There are many ways A.I. will improve the healthcare experience in the examining room, but a number of startups are also trying to tackle back office functions. For instance, Microsoft in April announced it would provide Epic Systems, the giant hospital software provider, with generative A.I. software for automating responses to patient queries and to identify trends by analyzing patient records.
It’s no wonder: Healthcare is a massive field – in the U.S. it reached $4.3 trillion in 2021, or 18.3% of gross domestic product, and is expected to grow by 50% by 2028.
At Infinitus Systems, the company I founded in 2019, we are tackling back-office functions that slow the entire patient experience down. Our A.I. addresses one aspect that drives doctors’ and hospitals’ staff up the wall: routine calls to insurers and pharmacies. By removing the human involvement from tedious, semi-structured and repetitive back-office tasks that are not clinical in nature, we are already freeing up physicians and their teams to spend more time with their patients, improving short- and long-term outcomes. (A.I. has already become so ingrained in my day-to-day work that I even used ChatGPT to brainstorm ideas as I wrote this piece.)
Today at Infinitus Systems, our A.I. makes tens of thousands of calls each month on behalf of care providers, allowing hospital and doctors’ staff to spend more time with patients and get them results faster. While our software mostly “talks” to humans today, we are fast approaching a time when intelligent systems are on either end of the transaction further cutting communication times.
I believe advancements in A.I. can lead us not just to better outcomes and faster more affordable care, but also a more compassionate healthcare system with a focus on patients and not on paperwork.
We’re not the only ones. Global healthcare A.I. investment is expected to reach $41.7 billion by 2027, compared with just $6.71 billion in 2021 and that was before the explosive growth and possibilities imagined by generative A.I. systems like ChatGPT. Among the hundreds of healthcare startups that have emerged in the past five years, many are testing new models relying on generative A.I. systems that will speed their success rate. And large hospital systems like the University of Pittsburgh Medical Center as well as Google and Microsoft are rolling out A.I. models to, among other things, more accurately address patient concerns.
At Johns Hopkins University they are testing a system that identifies sepsis, a body’s overreaction to an infection, on average six hours earlier than humans could, potentially saving thousands of lives per year. Because sepsis symptoms, like a fever, are similar to many other conditions it can be easy for humans to miss. Hopkins’ machine learning model can detect patterns from a patient’s history, symptoms and lab results and alert physicians. Around 1.7 million Americans develop sepsis every year and 250,000 of them will die – speeding up detection could save many of them.
Caption Health and PathAI are also among those working to automate medical imaging through A.I. Babylon Health seeks to automate patient intake and scheduling through chatbots and other systems, which help reduce errors and speed up the process. Others are working to synthesize data from bedside monitors, ventilators and other data sources, which could be instantaneously fed to nurses and doctors for quicker care.
At nearly every interaction the healthcare system A.I. can centralize, standardize and automate processes like data sharing, invoicing and supply chain management.
By combining similar text and voice technologies we could one day soon produce a system that converses and engages with patients and caregivers. Such an AI, trained on a patient’s medical history, could directly and instantaneously address concerns — “What’s this spot on my skin?” — and offer real solutions or tamp down the irrational fears we’ve all experienced after researching a symptom on the web.
All of these improvements won’t – and shouldn’t – replace doctors, but A.I.-led innovations are poised to help professionals do the job of providing care better. And far faster.
We have a long way to go, and, yes, some more waiting to do, but working together toward a goal of instantaneous healthcare is happening today.