Tomorrow’s Medicine
Tomorrow’s medicine will be unrecognisable to the clinicians of today.
And the clinicians of tomorrow will look back at us as we might look back at those in the time of Osler; doing their best, certainly, but ignorant and outgunned.
So what will Tomorrow’s Medicine look like?
The short answer is - “who knows?”
I have no specialised knowledge or skill here, other than being a clinician who has found some time to start thinking about these things.
But I would like to know who does.
So I present to you my thoughts in mostly rambling1 question form.
Some science fiction on my part, lots of questions, with a few predictions thrown in.
Let’s begin:
Thoughts of the Future:
# Pre-Hospital
= Wearable technology as surveillance for functional decline, clinical deterioration, biochemical derangement
= Automatic scheduling of Telemedicine consult
= Wearable technology automatically detects critically deranged physiology and initiates call to emergency services
= Bystander notified by proximity alarm
- High quality camera phone can assesses temperature heart rate, respiratory rate, colour + perfusion to automatically triage urgency
- Would there even be operators? Or just a set regime of triage questions, integration of vital signs, GPS data integrated with voice recognition?
= A drone is dispatched to the GPS coordinates with essential equipment - defibrillator
= City wide traffic changes are coordinated to allow easy passage of emergency services
= Patient Identification prompts Electronic Health Record to display immediate information: Allergies, Meds, PMHx, recent trends (from wearable monitor), Advanced Care Plans
# Arrival to Hospital
= Triage information is coordinated with ED capacity information, hospital capacity, traffic and other data to guide future arrivals to ED versus bypass in the case of ED overload
= Slap on a patch - provides vital signs, ECG, and intermittent transdermal serological sampling with wireless transmission of results
# Resuscitation
= A - Video laryngoscopy with mobile, flexible blade* and smart endotracheal tube combining automatic flexible fibre optic function and dynamic cuff pressure monitoring (and correction).
*I imagine a Laryngoscopy Blade that can automatically move soft tissue to the appropriate location with prevention of tissue damage to allow best view.
= B - Continuous monitoring and modification of ventilator settings based on prior probability information from EHR, continuous data from ETT/Circuit, alertness from a BIS monitor, compliance monitoring.
= C - Continuous Cardiac USS* monitoring coordinated with invasive arterial monitoring and automatic titration of vasoactives.
*I imagine a thick gel pad slapped over the praecordium within which are dozens of oscillating probe heads - the combined imaging information is automatically synthesised into standard and 3D TTE images with automatic analysis of standard measurements
*Then of course the TOE version of the same - with automatic advance and withdrawal functions linked in to the ETT securement device and multiple probe heads that synthesise and analyse an image
= D - Continuous EEG, TOF and arousal monitoring with automatic titration of sedatives and paralysis that includes information from ventilator.
= E - Temperature controlled smart* mattresses (robust and reusable of course) linked with wearable combined temperature, colour and doppler peripheral perfusion sensors.
*I imagine a mattress that senses the pressure (literally landscapes of high force over small area) and automatically adjusts it’s topography to prevent (minimise) pressure areas on patients
# Decision Making
= Rapid radiological reporting by non-humans
= Rapid histopathological reporting by non-humans
= Near point-of-care identification of pathogens (within the limits of colony doubling times etc)
= Rapid genomic analysis of pathogens for targeted antimicrobial therapy in order to reduce ineffectiveness and ensure appropriate antimicrobial stewardship
= Then continuously monitored inhibitory, bactericidal and peak concentrations of antimicrobials
= Auto-anticoagulated intravascular devices for continuous-intermittent sampling for point of care analysis
= Big Data-assisted decision making for both clinicians and families - population level statistics for survival, morbidity based on presenting issue, clinical course, invasive and imaging data, microbial information etc
= Nanorobots and sensors for targeted drug delivery
= Gene editing for personalised medication strategies
= AI assisted polyp identification
# Surgery and Procedures
= Robotic surgery becomes more autonomous with decreasing Surgeon oversight
= Augmented reality laparoscopy with Heads Up Display style overlay of anatomy and surgical planning
= Automatic inhibition of cut/coagulate based on image analysis of anatomy
= Within-theatre point of care 3D printing of autografted biological structures for surgical implantation (negating the need for transplantation)
= Automated and targeted surgery with nanorobots
= Augmented reality glasses for vein location, surface anatomy, vital signs during procedures
= USS output to augmented reality glasses
= Automatic ‘make safe’ of eg Seldinger needle based on real time imaging data to prevent errors
# History Taking
= Patients converse with Chatbots based on automatic prompts then natural language processing algorithms synthesise and organise into the standard format: PC, HPC, ROS, PMHx, PSHx, Meds, Allergies, FHx, SocHx, Travel Hx etc etc
= Allows the narrative to be structured automatically, then summarised
- Then clinician clarification of key points, important negatives and positives
= Would almost certainly put medical scribes out of business and save hours of time
# Physical Examination
= Live diagnostic probabilities based on
- PMHx, age, gender, ethnicity, habitus, colour, perfusion, vital signs, work of breathing and any bedside investigations - all of which could be achieved by smart thermal cameras in each examination room or patient cubicle that are linked to EMR and the summarised narrative history
- Probabilities are then adjusted by clinical findings during physical examination
- And live integration (and analysis) of POCUS to diagnostic probabilities
= All of which gives recommendations on reasonable, least invasive and cheapest ‘next step’ investigations.
= With decision tree logic based on positive and negative tests results.
# Nursing Care
= This is an important one.
= Could we reduce the rate at which comfortably sleeping elderly sick people are woken in the middle of the night to have vital signs taken?
= Could smart, thermal cameras monitor HR, RR, Temp, Perfusion; synthesise values and trends; and alert a clinician to potential deterioration for closer monitoring?
= Could this reduce nursing loads?
= Could wearable, dose adjustable patches or smart infusions do the same for medication rounds?
= Could we simply see humanoid robots assisting with medication rounds, lifting and bed changes?
# Research and Guidelines
= The days of lying to medical students to coerce them into unpaid roles as slaves of spreadsheets for their research supervisor must nearly be over…
= Continuous, automatic analysis of massive data sets by AIs will nearly preclude human input to confidently answer clinically relevant questions
= How will the ethics of trial enrolment change if we know that the AI will give definite answers on lifesaving therapeutics if entire populations are enrolled? Would it be ethical to opt out? Could you opt out?
= Guidelines could be personalised to genomic data - of both patient and microbe - to guide optimal care; how would we ensure privacy?
# Insurance
= This one could get icky.
= Could we see ‘computer says no’ scenarios in which care is refused based on cost/benefit analyses?
= Fraud detection would almost certainly benefit healthcare systems in terms of dollars saved.
= But the most brutal actuary will be an AI - personalised insurance premiums and coverage plans.
= Since wearable technology will be linked to preventive care, insurance premiums could be adjusted on a day-by-day or minute-by-minute basis - a ‘Black Mirror’ episode in the making.
= Would one be penalised (in premiums) for removing their wearable health monitor?
= This one gets dystopian fast - particularly if you live in a particular country…
# Medicolegal Matters
= Could human clinicians then be held to the standards of a near omniscient AI in settings of negligence and malpractice claims?
= Blockchain technology will, allegedly, provide tamper-proof and decentralised storage of medical records. But could it ensure privacy? Are our current standard of privacy a reasonable basis from which to judge?
= Legal research and analysis of medical data would be accomplished by AI - the courts would almost certainly move quicker.
# Geographical Constraints
= Could cheap cameras and Starlink’ed software improve care in low income countries or rural communities? - almost certainly yes.
= It seems like issues of resource rationalisation would greatly be improved.
= Medical tourism would almost certainly (continue to) explode - one could search the best doctor, best unit, best hospital globally for their particular condition or desired procedure.
# Ethics
= Patient privacy and data security will be obvious concerns.
= But what about algorithm bias - what happens if we see that particular subsets of the population are penalised with higher insurance premiums based on indelible factors - because we programmed the AI to “reduce costs”?
= How can equity of access to best medical care be maintained when these technologies will arrive quickly and almost certainly in metropolitan centres first?
= Could one company monopolise an algorithm that provides better care?
Conclusion:
Again, I have no specialised knowledge here. But I am convinced that Medicine will drastically change in response to the explosion of new technologies.
I think this could range from simple, probably obvious and harmless starts -
- like having a computer listen to a narrative history of three days of unwellness with productive cough to automatically produce a crisply documented subjective history;
- through to ethically ambiguous uses of synthetic biology, genome editing and big data that could verge on eugenic...
This is all food for thought.
I would love to hear what you all think.
And if you happen to know of those working in any of these little niches, I would love to check out what they are working on.
Thanks,
Luke.
P.S. You will be unsurprised to learn that I also asked ChatGPT2 about all this as well - it gave me good points on ‘Nanorobots’ and ‘Blockchain’ but wasn’t quite up to speed with POCUS... Just to let you know.
I have obviously left out massive chunks of the Healthcare ecosystem here. I have tried to touch on the Hospital-based Medical and Nursing aspects because that is what I am most familiar with; but consider my enormous omissions of Community Care, General Practice, Pharmacy, Physiotherapy, Occupational Therapy, Speech Therapy, Healthcare Administration, Medical Education, Government and Private Industry - all of whom would have their own drastically different futures based on technological development.
(And this is by no means a complete list of omissions…)
And I never really went into any medical or surgical specialties, just stuck with generic 'very sick people' in hospital. So the possibilities here are vast.
Prompt: "You are a Medical Futurist. You consider how the practice of Medicine will change with the advent and improvement of technologies such as Artificial Intelligence, Machine Learning, Big Data, Synthetic Biology, Nanotechnology, Genomics and Medical Devices.
Question: How will medicine change in the future in the following domains?
- Community Care
- Pre Hospital Care
- Electronic Health Records
- Emergency Care and Resuscitation
- Clinical Decision Making
- Medical Imaging
- Surgery
- Medical Procedures
- History Taking
- Physical Examination
- Patient Monitoring
- Wearable Devices
- Nursing Care
- Research
- Health Insurance
- Medicolegal Matters
- Ethical Conundrums"