Modern medicine has, over the last decade, become an increasingly complex field. While the conditions patients suffer from remain static, our ability to diagnose and treat these conditions has made an exponential leap, in part due to AI and in part due to advances in science, enabled by evolving technologies.
Technology means we no longer need to be separated from our point of care. It can follow us everywhere, an ever-vigilant protector and guardian angel, and its applications are limitless.
We are entering a brave new world. One where the boundaries between human and AI-based care have begun to blur. Where software offers patients more control over their health outcomes and some patients choose to trust this digital voice over the advice of their doctor. It is an emerging quagmire of moral and ethical debates centering around the care of patients and how they are best served. Many doctors are both intimidated and threatened by this encroachment, while others embrace it.
The question we need to be asking is how this all impacts the patient? Where do they lean? Towards reliance on their doctor’s or opting instead for the breadth of knowledge offered by AI.
Confusing tools with solutions
AI isn’t a solution, an independent entity sent from the digital domain to cure all the ills that ail us. Perhaps, fifty years from now, it may be, but for the present, it is merely a tool that doctors must embrace for the sake of their patients. The complexities of modern medicine are beyond the scope of a single individual mind.
AI offers a solution, enabling diagnosis and treatment by expanding the doctor’s range of vision and his knowledge base beyond the scope of his education, experience, and (apologies ahead of time for the ego-bruising) mental capacity.
It’s important to understand that there isn’t a choice to be made here. It isn’t an either-or scenario, either AI or human-based healthcare. In 2021, in healthcare, one cannot exist effectively without the unique talents of the other. Separately they both offer benefits to the patient, together they can achieve god-like status.
There are still impediments to the successful incorporation of AI into modern-day practices. Differential Diagnostic (DX) systems rely on physician input which affects the physician’s ability to focus attention on the patient. Envisage a time, in the next few years, where DX AI listens to the consultation in real-time, absorbing key phrases and expressions from the patient designed to elicit a set of diagnoses that are presented seamlessly in real-time to the physician for consideration.
Emotional and physical queues exhibited by the patient can then be considered by the physician in conjunction with the diagnoses to ensure an accurate evaluation of the patient. AI doesn’t tire, it doesn’t make mistakes and it can access the entire wealth of our medical knowledge in the blink of an eye. DX systems like Isabel from Isabel Healthcare are beginning to truly understand and exploit this relationship for the benefit of the patient.
Overcoming obstacles to broad-scale integration
AI has become a victim of its own success. A myriad of systems all offering different solutions and perspectives, litter the marketplace. This is counterintuitive to developing a cohesive network. There is also the issue of public domain software that belongs in the professional realm. Let me explain.
The Patient Perspective
When you apply for a bank loan, the bank will run software that allows it, based on a number of parameters the bank establishes, to decide if they’re going to take pity on your impoverished self, or if you’re going to be sent packing without a cent. You never see this software, you never have access to it, and you never will. There is an excellent reason for this.
Most members of the public possess little or no knowledge about financial instruments, risk, etc. The same holds true of medicine. It is an incredibly complex and nuanced field, where the way you hold your head when you sit in front of your doctor may alert him to a diagnosis no software is yet able to make. Neither can you (the patient) and that is why public access to professional tools should be limited.
Without a thorough understanding of the subject, the rabbit hole of the internet, ably supported by the myriad of self-diagnostic tools the public now has access to, leads to a flurry of incorrect seld-diagnoses, inappropriate self-administered medications, and actions that inevitably prove more harmful than helpful to the patient.
No software can replace the experience and training of a qualified and caring pair of human eyes and in our opinion, releasing packages that enable the public to play “doctor” is disingenuous to the profession and simply endangers public health.
Patient-based software, however, that seeks to enforce and simply the relationship between the patient and caregiver should be considered a priority. The diversity of new packages launching into the public sphere complicates the development of a single, stable, and broadly adopted product. This matters, as doctors are often forced to adapt to protocols at a practice-based level to integrate with public-facing packages.
The industry would benefit from a more cohesive approach that allows for the development of an industry-based standard, a single API to allow for seamless communication between professional diagnostic, practice-based software, and patient-focused software. Again, in our opinion, the two markets are too diverse to allow for a single solution.
Focus on one or the other for development. You cannot be everything to everyone and building complex packages that attempt this, simply dilutes their efficacy and increases UI complexity. Keep it separate and keep it simple. Your users demand it.
What should a patient be looking for in a software app?
- Your entire patient history (Patient Record) stored in a simple, accessible, and shareable format that is easily accessible to medical personnel and EMTs in the event of an emergency.
- Lucky enough to own a medical monitoring device, smartwatch, or other wearables? Your software package must interface with this tech. If it doesn’t, it isn’t fit for purpose.
- Can your primary care provider (your doctor) communicate and interface with the package, critical, as owning the coolest app on the block is useless unless it has widespread adoption in the industry?
- Do you understand the app and is its user interface simple and intuitive? Not much use having a heart attack and you can’t find the panic button.
- Most importantly perhaps is this one single thing. Does your app simplify and improve your health relationship with your care provider? Are appointments simple to make, do you receive updates about cancellations, openings, upcoming vaccinations, etc? Does your pharmacist ping you reminders for repeat scripts and can you access related services and support groups relevant to any conditions you suffer from?
- Are you able, in a pandemic society, to wait in your car until the app goes off with that message you’ve been waiting for? The doctor will see you now.
- Does your app support multilingual interaction? This single aspect may save your life.
The Professional Perspective
The problem IT faces in medicine is IT. It permeates every corner of the profession, bits of randomized and unassociated code that perform various functions critical to the daily running of a professional practice. Think of a pack of Lego blocks of different colors and sizes, stuck together in a random haphazard fashion to form a psychedelic house of data. Pull it apart and you’re left with exactly that. Unintellible noise that cannot be interpreted.
Not for the benefit of the profession nor the patient.
The medical field is huge and data sets can differ dramatically from one discipline to another, but developing compartmentalized solutions merely perpetuates data isolation. More colored blocks that sort of fit, but don’t contribute towards a stable, interperable data structure.
Approach any doctor in any profession with a new data solution and expect eye-rolling and sighs. Rather than simplifying the profession, IT has complicated it. Add more fields and checkboxes to the doctor’s life and the doctor checks out. There are movements underfoot in the industry to address this, but they have their work cut out. Doctors are or should be, the developer’s first port of call. For broad adoption, software should address two key issues.
- Simplify the doctors life. Key to this is not adding new processes to worloads.
- Improving patient outcomes and communication
Integrating systems isn’t the responsibility of practice managers. It falls to the IT profession to remove these hurdles and provide point-of-care software that can seamlessly communicate with patient portals and federal and state-based government platforms. Arguably, the architecture of these government platforms should be based on emerging medical data standards, not the other way around, as is currently the trend.
Inverting the pyramid in this way is counterintuitive and leads to the development of products that are far from suited for the actual day-to-day functioning of practices and are most certainly not patient-care based.
The patients and the caregivers form the base of the IT Medical pyramid, building architecture that obeys any other structure is counterintuitive and doomed to failure.
Give the doctors what they are crying out for. Seamless tools built to operate intuitively and that take advantage of the latest technologies to achieve this. Save them time, improve their patient interactions and outcomes and do it unobtrusively. Integrate these tools with patient portals and forget about addressing external, non-medical entities for now.
Focus on the correct pyramid, the industry, on the patients and their care providers. Create bespoke, silent and effective tools that address these issues and doctors will no longer see you as a threat, but in your real role.
That of an indispensable partner.
Why does any of this matter?
It’s a fair question. Medicine has survided and thrived till now, pretty much on its own and will continue to advance, albeit far more slowly, without the intervention of AI. Patients however, you and I, will pay the price.
It’s probably worthwhile contextualizing the term “AI” first. Any software that makes decisions based on inputs it receives is an example of artificial intelligence (AI). We are unaware of the programming inthe background that has enabled the software to make seemingly random choices. It’s been taught by its programmers how to make associations and link things together, giving it the appearance of being intelligent. It isn’t.
More modern software is designed to evolve and grow on its own, it can teach itself and “learn” from the input it receives, but only in the context of its programming. We are a few generations and movies away from machines working out that the best way to eradicate disease permanently is to eradicate humans. For the foreseeable future, computers form an integral part of medicine, and for good reason.
Let’s check in on Bob. He’s on holiday abroad and has a nasty run-in with a jellyfish. Bob is rushed into a hospital and a foreign doctor who cannot understand Bob evaluates him and prescribes medication. The doctor is able to immediately access Bob’s entire medical history via the app on Bob’s phone. Multilingual support overcomes the language barrier and allows the doctor to pick up two medications Bob uses every day. The doctor is able to adjust his prescribed medication to prevent Bob suffering cardiac arrest from a drug interaction.
He is also able to see that Bob will require evaluative hospitalization as he is in an at-risk category for developing potentially life-threatening anaphylaxis in response to the jellyfish sting. Bob’s app absorbs the new data and treatment provided and his doctor in the US is immediately notified. The foreign doctor also notices an abnormally high count on Bob’s bloodwork that isn’t related to the jellyfish. On his return home, Bob’s doctor is able to admit Bob for further testing and administer early preventative lifesaving treatment by removing a tumor.
Without the app (yes, and the jellyfish, we know) Bob could have died twice, once from a drug interaction, and in the second instance, from an undiagnosed disease which his doctor was alerted to. In the same way, Bob’s smartwatch can save him from an undiagnosed heart attack by interacting with his Patient App and providing early warning to his GP.
The health benefits offered by technology to the patient are endless. Improvements in access to care at the appropriate moment increase exponentially. Preemptive care is massively important to both patients and the healthcare infrastructure of any country. and it is for these reasons and others that IT and AI must be accepted and effectively integrated into modern healthcare as a matter of priority.
Technology means we no longer need to be separated from our point of care. It can follow us everywhere, an ever-vigilant protector and guardian angel, and its applications are limitless. It must be embraced.