Have you ever felt like your doctor was too busy entering notes into their computer to give you their full attention? Wondered why a specialist you’ve waited months to see asks you the exact same questions your family doctor asked, despite the existing documentation? Spent painful hours or even weeks waiting for test results and wondered what takes so long?
You’re not alone. The shortcomings of today’s healthcare system can leave you feeling frustrated at best, or at worst, like a number instead of a person.
Fortunately, Artificial Intelligence (AI) is reshaping healthcare to deliver more efficient, patient-focused experiences. In this article you’ll learn how AI can reduce wait times, restore empathy to patient-doctor encounters, speed up test results, detect dangerous diseases earlier, and improve both the consistency and quality of care, to keep you and future generations happy and healthy.
A Greater Focus on the Patient
It’s easy to get the impression that your healthcare team lacks the capacity to properly care for you and in a way, it’s true. Physicians now spend an estimated 40% of their time on paperwork. In fact, paperwork is one of the leading causes of physician burnout according to a recent Medscape study. From a medical-legal perspective, “If you didn’t document it, you didn’t do it”.
AI-powered speech-to-text transcription is alleviating this burden. Thanks to recent advances in speech recognition, a physician’s spoken words can now be accurately transcribed to text. Documentation is generated on-the-spot simply by “thinking out loud” into a mobile headset or app. This really does save time and money. In a recent study, 84% of physicians reported they could better focus on patients when using AI-assisted documentation, while medical transcription costs decreased by 80%.
You may soon see AI assistants similar to Amazon’s Alexa listening in on doctor-patient visits. Of course, such technologies must obey stringent medical privacy regulations such as HIPAA. But with advancements in edge computing, which lets AI operate on-device without sending information over the cloud, the barriers to adoption are dropping.
From a patient’s point of view, AI will free your doctor to focus more completely on exploring, understanding, and addressing your needs.
Your Doctor is Better-Prepared
Gone will be the days when a physician enters the room while hurriedly flipping through an unorganized mass of paperwork. When doctors barely have time to read your medical history, you might wonder about the quality of their medical decision-making.
How can the situation be improved? Important information is often unstructured and scattered across many pages of medical documentation. But imagine instead if the key details relevant to your medical case were neatly summarized for your doctor on a tablet or mobile device. AI can already achieve this. Image recognition and natural language processing (NLP) can extract the contents of test results, charts, or hand-written doctor’s notes from a variety of sources. It can then aggregate and streamline this medical information into a simplified, easy-to-digest electronic format. For example, the output might be a succinct summary of the patient’s case, including results of key diagnostic tests organized into tables or charts that help illustrate the patient’s history.
Another challenge is the sheer volume of records a physician has to sift through to locate a key piece of information (or discover its absence). AI can learn clinician treatment patterns to recognize which information is most important in a given patient encounter, and surface this information in real time. Furthermore, natural-language question-and-answer systems can help physicians query medical documentation more efficiently. Doctors can verbally ask a question about a patient’s health history and receive a plain-english response, such as a summary of the patient’s recent encounter information.
By providing your doctor with faster, easier access to the information most relevant to treating you, AI will help maximize the benefit you get during the short time available for your appointment.
Faster Test Results
Alice arrives at the ER with a searing pain in her abdomen. After an MRI scan, she waits an agonizing 3 hours to receive the results. Only then does she learn that the images were unclear and the scans must be repeated. In the meantime, her pain worsens.
What’s going on? Today, medical scans enter a queue to be inspected sequentially by a radiologist. Hours may pass before the radiologist reaches a particular scan. Until then, this critical piece of medical data is not actionable. Meanwhile, Alice might have a life-threatening condition needing immediate intervention, such as severe appendicitis.
Instead imagine an AI assistant that pre-assesses a scan ahead of a radiologist. This AI could determine that Alice’s scans need repeating before she even leaves the scanner. It could also identify that Alice may need urgent treatment, and move her case to the top of the radiologist’s inspection queue (Figure 1). Once a radiologist begins their inspection, an AI assistant could accelerate their workflow by directing attention to regions of interest within a medical image (Figure 2). And since AI can already match or surpass the accuracy of radiologists, it may eventually eliminate the need for human inspection altogether. Imagine receiving your test results within seconds instead of hours!
Figure 1: Artificial intelligence could provide smarter triaging of medical cases by performing pre-assessments on test data such as medical scans, and assigning greater priority to urgent cases. The animation above illustrates the reprioritization of a medical image inspection queue by an AI. Image source: Aidoc Inc.
Figure 2: AI can direct a physician’s attention towards easy-to-miss anomalies; for example, by indicating regions of interest within a medical scan. Image source: Aidoc Inc.
AI diagnostic assistants must continue to gain the trust of physicians. Transparency in how AI arrives at its decisions is an important factor. However, it is only a matter of time – AI will continue gaining traction as an assistive tool working alongside physicians to provide you with better quality assurance, smarter triaging, and faster turn-around times.
Dangerous Diseases are Caught Sooner
Today over 25% of lung cancers and 15% of breast cancers are missed by routine chest X-rays. Such diseases are treatable if caught early. Unfortunately, cases often go undetected until it’s too late.
AI can catch hard-to-spot anomalies in medical data that a doctor might miss. For example, an always-vigilant AI could detect an abnormality in an X-ray, such as a suspicious nodule, even if it isn’t specifically being examined for. Studies have shown that pairing physicians with AI can reduce the danger of false negatives, in which patients are erroneously discharged. AI never gets tired, and it continually learns and adapts to improve its performance – all without having to return to med school! With the help of AI, ailments will be found and diagnosed earlier, leading to better outcomes for the patient and reduced cost of care.
Consistent High-Quality Care
Have you or someone you know experienced the stress of going from doctor to doctor, trying to diagnose a tricky ailment or find a treatment that works? In some cases it can take years to get answers.
The problem is that no single physician can be an expert in all things. A disease that is easily diagnosed by a specialist might be missed by a general practitioner. But imagine if healthcare teams had an AI assistant possessing the knowledge of the world’s top specialists and experts? One that would advise on next-best-steps, to help frontline teams decide which treatment to prescribe or which diagnostic test to run.
Just like Uber’s self-driving cars can pool and share their collective knowledge to improve each other’s performance, AI-powered healthcare apps will integrate knowledge from all practitioners across an entire health network. By learning from past patient encounters – what steps were taken and their outcomes – AI assistants will help practitioners make the best-possible decision given the available evidence.
Healthcare providers have already started connecting AI to electronic health records (EHR) systems to provide clinical decision support. The result will be higher-quality care that is more consistent across physicians. Eventually such systems will also help democratize the quality of care worldwide by making collective healthcare knowledge available in remote or resource-limited regions.
Treatment is Tailored to You
The one-size-fits-all approach doesn’t work in medicine. A drug that works well for one person might be ineffective for another, or cause potentially-harmful side effects. But determining the effectiveness of a treatment today is still largely through trial and error.
The goal of Personalized Medicine is to leverage patient data to remove this guesswork. AI-based clinical support systems can aid this goal by fine-tuning their recommendations based on your specific medical profile. This might include information available about your genome, immunome, or proteome, in addition to your medical history.
Leveraging AI for personalized medicine will help your healthcare team provide you with the right treatment, the first time. Reducing uncertainty around the course of treatment will save lives, especially in high-stakes situations such as cancer, where finding an effective drug early is crucial.
Seamless Referrals
If you’ve ever had to see a specialist or get a medical scan, you might recall waiting days or weeks just to get an appointment date. Why the delay? Booking time on diagnostic equipment or with a specialist is a time-consuming process for physicians and their already-busy staff. It may require coordination between multiple parties, pre-clearance with insurers, and the creation of briefing reports.
AI can automate much of this work behind the scenes, such as navigating calendars and booking systems to find open timeslots, and pre-populating medical briefings with the relevant information. AI can also combine voice synthesis with speech recognition to assist scheduling over the phone. An example from the consumer world is Google’s Duplex. It can phone a restaurant to place a reservation on your behalf, while using speech disfluencies such as “hmm”s and “uh”s to sound more natural.
Altogether, you can expect a much quicker turnaround on referrals and appointment bookings down the road.
Appointments are Easier to Get
Ever feel frustrated by the length of time it takes to see a doctor? Physicians need to be available at the right place and at the right time to adequately meet the needs of patients.
AI can optimize a physician's schedule using data about the number, type, and location of requested medical appointments. For example, it might identify that demands for cardiac care at a particular hospital are highest on a Tuesday. It might also help determine the best duration for an appointment. Patients scheduled for 15-minute appointments sometimes spend a full hour with physicians, increasing wait times for other patients and hurting productivity. To mitigate this, AI can recommend questions that scheduling staff can ask patients to more accurately assess the appropriate length of their visit.
Patient no-shows are another source of inefficiency. AI can pre-screen patients to determine who might be at risk of missing an appointment (due to factors such as age or past absences). It can then issue a reminder or confirmation request via phone or text, which the busy office staff might otherwise lack capacity for.
By optimizing scheduling and reducing inefficiencies, AI will help ensure your doctor is available when you need them.
Clinics are Open Longer
Ever been forced to take time off work because your doctor isn’t available in the evening? Wonder why clinics often close early? One reason is that doctors often need the evening hours for paperwork. This might involve painstakingly re-reading their patient encounter notes to identify which billing codes to send to insurers so that their practice can be reimbursed for expenses.
Instead of your doctor toiling away on this mundane task, AI-based natural language processing (NLP) can read through their electronic notes to auto-extract important information and populate relevant medical forms. For example, an AI could identify the correct billing codes and pre-fill much of the paperwork needed for reimbursement. It could similarly automate other labour-intensive duties including writing patient referrals, hospital admittance forms, or diagnostic reports. In this workflow, the AI assistant pre-populates the paperwork, and the physician reviews and verifies it before submission.
With far less paperwork to burden them, physicians will have more hours available to spend with patients.
Closing Remarks
Artificial intelligence will help physicians work more efficiently, accurately, and with greater focus on the patient. It will restore empathy to healthcare and ultimately save lives. So when can we expect AI to enter our doctor’s office? The transformation has already begun. In fact, a hallmark of well-designed AI products is that you don’t even realize AI is there. Seamless integration into existing doctor-patient workflows and user-focused product design are both keys to the successful adoption of AI.
Some challenges remain. Developers must continue gaining physician trust by providing transparency in the AI’s decision-making process. Complex healthcare data security regulations are also a hurdle, albeit one that can be mitigated through a privacy-by-design attitude towards AI application development.
There is also enormous potential for AI to understand health on a global level based on real-life data. For example - what is the true impact of processed foods on cardiovascular health? How much sodium can we safely consume? Are certain pharmaceuticals as effective as companies claim? The health insights AI could gain by crunching medical data from millions if not billions of people would be far more meaningful than the small “artificial” studies often conducted on small population samples by organization that might have ulterior motives. But we as a society have to come up with answers to issues such as data privacy. The challenges aren’t purely technological. Are populations willing to grant access to their medical data if it could lead to better care?
In the meantime, the coming years will see more AI-based tools working alongside physicians to provide you with healthcare that is faster, more effective, more accessible, and arguably more human. Smile! The future is bright.