Harjinder Sandhu, PhD is the Founder and CEO of Saykara, a company working to combat the epidemic of physician burnout that has surfaced from increasingly burdensome documentation requirements and time spent on electronic health record (EHR) data entry.
You initially began your career in academia as a computer science professor and then transitioned to entrepreneurship. Could you discuss what inspired you to pivot to being an entrepreneur?
Having spent years in academia teaching and writing research papers about the transformative power of computing, I was inspired to apply that knowledge to a real-world scenario, to something that could impact industry in a substantive way. Originally, my plan was to spend a year or two helping get a company off the ground, then return to academia. However, after going through the ensuing dot com bust, the challenge and thrill of building something that makes a real difference in people’s lives was too great, and I never went back to my post as a professor.
Could you discuss the genesis story behind Saykara?
I founded Saykara in 2015 with a goal of applying the untapped potential of speech recognition, machine learning and natural language processing to solve one of the most significant issues facing healthcare provider organizations today, which is the cumbersome and time-intensive nature of clinical documentation.
Although current-day speech recognition solutions are very capable of capturing physicians’ verbatim dictation, using them to complete the requisite clinical documentation still entails a considerable amount of time and effort, with physicians remaining tethered to a computer, mouse and keyboard — editing content as they go and navigating the complexities of the electronic health record (EHR) system. Therein lies the bigger problem.
Over the past five years, Saykara has developed and launched a platform that uses artificial intelligence (AI) to interpret conversations between physicians and patients and automatically construct the clinical notes, orders and referrals, etc. needed for the medical record. And we do this through a voice-enabled assistant named Kara that physicians access via a mobile app.
I’ve personally been involved with speech recognition, machine learning and natural language processing technologies for the past 20 years, and Saykara is actually my third health tech startup. My first startup was a company called MedRemote, which I co-founded in 2000 with my friend, Kulmeet Singh, and eventually sold to Nuance Communications, the global leader in speech recognition technologies. In 2011, the two of us co-founded another company called Twistle, which is still around today, and although I continue to serve on its Board of Directors, my full-time focus since 2015 has been my role as CEO of Saykara.
How big of an issue is clinical documentation and how much time does this process normally take for physicians?
How big of an issue is it? One word: HUGE.
Physicians have been tasked with performing an ever-increasing amount of clinical documentation stemming primarily from insurance billing requirements, public reporting and regulatory mandates. Physicians are not only losing time with patients, that time is being impeded by having to perform data entry to the EHR.
An often-referenced study published in the Annals of Medicine observed that physicians spend nearly twice as much time doing administrative work as seeing patients. Specifically, the study found that for every hour physicians spend on direct clinical face time with patients, they spend nearly two hours on EHR and desk work. And for many physicians, this carries over into an additional one to two hours each night, something commonly referred to as ‘pajama time.’
Numerous other studies and surveys have observed that between 50% and 71% of physicians are experiencing burnout, which has worsened during the COVID-19 pandemic. The Medscape 2020 Physician Burnout & Suicide Report revealed that the top contributors to burnout include “too many bureaucratic tasks (e.g., charting, paperwork),” “spending too many hours at work,” and “increasing computerization of practice (EHRs).”
What is Saykara doing with machine learning and natural language processing technologies that is unique?
What we’re doing in the context of conversational AI is completely unique. We can understand conversations that naturally occur between physicians and patients and generate structured data from those conversations in a way no other company does. As well, the ubiquitous qualities of our platform means it can readily adapt to any area of medical specialty, and we’re able to tailor the user experience to each individual physician’s style and preferences.
We currently have a semi-autonomous hybrid model that pairs our AI with a human-in-the-loop reviewer, which helps strengthen our knowledgebase and ensures that we return accurate results to our users from Day 1.
Whereas the speech recognition component of our system captures what is being said, the natural language understanding component interprets what is being said. For instance, if a patient comes in for knee pain, the system first has to understand that knee pain is a symptom. It then has to build a story around knee pain based either on what the patient is saying or the physician is summarizing. When did it happen? What caused it? How severe is it? The system next has to take that story in its raw data format and compose a comprehensive, high-quality clinical note, which is what the physician sees.
Our system is also able to filter out the “noise,” and by that I mean it understands the kinds of information a physician needs to gather based on the reason for a patient visit, then it accurately predicts and anticipates what is relevant and what is not. For example, if a patient goes to see a rheumatologist for arthritis and starts talking about the anxiety they’ve been experiencing over buying a new house, the rheumatologist will not likely want or need to document those details. Our system can learn this and focus solely on what’s relevant to the rheumatologist’s clinical note. On the other hand, if a patient arrives at a family practitioner’s office and starts talking about the anxiety they’ve been experiencing over buying a new house, the family practitioner will very likely want and need to document those details.
Could you discuss the process for how a physician would use Saykara while interacting with a patient?
During the course of an office visit or procedure, or what’s referred to as a “clinical encounter,” the physician simply opens our iOS app (most typically on an iPhone), taps on the patient’s name (a daily list is imported from the physician’s EHR or scheduling app) then simply talks naturally, interacting with the patient in a completely normal and customary manner. Our AI assistant, named Kara, listens ambiently to the entirety of the conversation and when the clinical encounter is complete, the physician taps the app to end the recording. Our platform then interprets and transforms salient content required for notes, orders, referrals, etc., and the structured and narrative data is populated directly to the patient’s chart in the EHR system, ready for the physician’s review and sign-off.
Saykara offers a sophisticated and flexible platform that allows physicians to use the solution in multiple different modes. Could you discuss some of the different modes?
What I just described is the ‘ambient mode.’ Physicians also have the option to use our AI assistant in the ‘recap mode,’ which also allows them to simply talk naturally — no special code words or commands are required. However, instead of having Kara listen to the entirety of the conversation with a patient, the physician provides one or more brief reflective summarizations, either during and/or following the encounter. The process thereafter is the same, as is the resultant content.
What market segments and industries is Saykara currently focused on serving, and how do you license your platform?
At the uppermost level, Saykara serves the healthcare industry. Our customers include independent physician practices as well as hospitals, health systems, ambulatory surgery centers and urgent care centers. We have a software-as-a-service subscription model where customers pay a monthly fee to use our platform.
How has the feedback been from practicing physicians?
The feedback has been overwhelmingly positive. This is truly life changing for physicians and they use those very words. Our AI assistant completely eliminates the ‘pajama time’ I described earlier and it reduces the amount of time physicians spend interacting with the EHR and preparing clinical documentation by an average of 70%. Most importantly, it gives physicians more face time with their patients. Physicians are able to give patients their full attention without distraction.
We have dozens of testimonials regarding the impact our solution has had on productivity, work-life balance and the physician-patient relationship. For example, a family practitioner said, “Now, with Saykara, I can see a patient in 15 minutes and I chart for 30 seconds. It’s just been a game changer.” A pediatrician said, “My spouse can tell you that I’m a much happier person when I come home knowing that I don’t have this weight of charts over my head.” An orthopedic surgeon said, “I can give my full attention to my patients, and I even have time during visits to talk with them about their vacations, grandkids or restaurants they recommend.” A vascular surgeon said, “It understands the context and intent of what is said, then eloquently and accurately creates a note in our EHR. No more nights and weekends proofreading and signing charts.” And the CEO of a multi-specialty physician group said, “Using the Saykara mobile AI assistant allows our providers to focus on their patients and create personalized, meaningful and collaborative interactions.”
What are your views on the future of AI in healthcare?
The AI revolution in healthcare is still in its infancy. That said, the AI capabilities applicable to the healthcare industry are truly vast and have tremendous potential for good. You’ll find AI applications for the clinical, operational, financial, technical, administrative and security realms of healthcare, with solutions for IoT device monitoring, chronic condition management, prescribing of medications, decision support, disease detection, care planning, workflow automation, revenue cycle optimization, supply chain management, clinical trials, and the list goes on. I would say that overall, these applications are still very early in their lifecycles and we will see many more capabilities as they mature.
What we’ve seen with AI technology in other industries is that it reaches an inflection point where it begins evolving at a much more accelerated rate and becomes able to do more and more in shorter and shorter spans of time, with tremendous accuracy. I think we are almost at that inflection point for many of the applications within healthcare. There is undoubtedly a great deal that nobody has even contemplated at this point. Over the next three to five years, I anticipate we will see incredible AI disruption in healthcare, in a good way.
What kinds of emerging technologies are you most excited about?
The technologies I’m most excited about are AI systems that can interact with human beings through conversation, which I believe will have an impact across every industry. When we reach the point that computing devices can understand our conversations, this will open up a whole new world of possibilities that will transcend every aspect of our lives.
Right now, you have to learn how to interact with a computing device on its terms, not on your terms. The parameters are pre-defined and you need to figure out how to use it best to meet your needs. What’s on the horizon in terms of the AI revolution is that you’ll be able to interact with a computing device on your terms, in the same way you naturally interact with other human beings, through language, with the expectation that the device will respond accordingly. That’s what we’re focused on at Saykara, and it’s a trend across many other industries as well. I think we will quickly look back and wonder how we ever lived without these systems, much like we look back and wonder how we ever lived without our smartphones.
Thank you for the great interview, I fully agree with you that AI is evolving at a much more accelerated rate. Readers who wish to learn more should visit Saykara.
- Humayun Sheikh, CEO of Fetch.ai – Interview Series
- Research Team Aims to Create Explainable AI for Nuclear Nonproliferation and Nuclear Security
- Pieter VanIperen, Founder & Managing Partner of PWV Consultants – Cybersecurity Interviews
- Advance in Microchips Brings Us Closer to AI Edge Computing
- AI Researchers Create Video Game Playing Model That Can Remember Past Events