The Future of Healthcare: AI and Digital Transformation
As a part of a ‘Talking Tech’ speaker series hosted by Digital Nova Scotia, Brilliant labs contributed to thoughts on what the future of healthcare might look like as it pertains to Artificial Intelligence and digital transformation.
The panel was hosted by Derek Leung of the Digital NS Marketing team and hosted such individuals as Mark Opauszky, Gail Tomblin-Murphy, and Emily Johnston in addition to Will Collins. Each of these individuals was uniquely qualified to carry the conversation, revolving around how dramatically things have changed with technology and how to keep up, on the merits of their experience in the tech field.
The subject of the series was pertaining to the overlap between digital tools and healthcare and was intended to provide the audience with venues for exploration on the topic. Different disciplines of computer science and healthcare are individually siloed and have minimal overlap but the potential for implementation is immense. Given that all of Canada has a consolidated healthcare system on a province by province basis many of the digital tools ease the burden of administration. While it might seem easier to have a consolidated healthcare system it also raises the potential for systemic overload in which an overabundance of disorganized information yields poorer outcomes if not properly managed. The same could also be true of private networks which would excessively cut costs to accommodate shareholders or private interest groups. In short there are two ways things can go wrong: (1) administrative or cost overruns which fall in excess on the citizen or (2) an overtly lean system in which poor quality resources and training do not address the need. Striking a balance between cost and quality is the focus of the Nova Scotia Health Innovation Hub.
Benefits of the existing e-health platforms implemented during the pandemic
Talking Tech speaker Gail Tomblin-Murphy brought up the usefulness of the existing e-health platforms implemented during the pandemic. The ability to call nursing or health staff prior to making the decision to go to the emergency room has the implicit benefit of not overloading the emergency room or other such hospital waiting rooms.
Furthermore, where doctors visits can be done virtually the cost saving measures of infrastructure are significant. This is especially true of remove communities in which they may not have access to healthcare facilities and may have to travel several hours to access reasonable care. Computerized medical records that are secure and routinely audited have the potential for reducing these inefficiencies. Additionally, where it is possible to interpret medical data results a ‘meta-analysis’ can be conducted in which the results of several studies can be compiled and disseminated.
Practical implications of technologies in healthcare
Now let’s turn to the practical implications of such technologies by assessing the potential stated by Mark of Sparrow Bioacoustics. A cardiac incident, like a heart attack, can be anticipated prior to it occurring by looking for irregularities in rhythm. Prior to a cardiac event there are a number of medication and lifestyle interventions that can be conducted and prevent future issues.
For instance, perhaps the individual has high blood pressure that would benefit from some form of medication but it is unclear which variety would work best for them. With a series of simple bioacoustic and genetic tests it is possible to determine which medications might best tend to the health of the patient using AI. Furthermore, continuous monitoring of the patient’s physical activity is another means of preparing for a cardiac event. In the event that our hypothetical patient becomes overwrought with exertion and keels over a medical team could be dispatched at a whim. Most defibrillators, which restart the heart when it fails, also contain a degree of computer intelligence that allows it to correctly time restarting the cardiac cycle. If we could combine a defibrillator with a means of administering medication prior to an individual having a cardiac event the issue could be mitigated. Such a device would rely on artificial intelligence to ensure the patient was monitored prior to medical intervention.
Digital resources in the healthcare crisis: the overprescription of pain management drugs
Lastly but perhaps most importantly the management of pain is another use case combining digital resources and healthcare in the case of Pain Coach. As some of our audience may be aware a healthcare crisis is ongoing due to the overprescription of pain management drugs namely oxycontin. This was the result of excessive lobbying and prescription advocating on behalf of Purdue Pharma which subsequently led to the opioid crisis.
The reality is that pain can be extraordinarily difficult to manage following surgeries or precipitous healthcare events. Emergence of a research field known as ‘Biofeedback’, in which patients modulate their experience of pain with attentional focus, may hold some of the answers to blunting the overconsumption of opioids following surgery. This is easier said than done but mental resilience in dealing with pain and continuous reporting will ultimately alleviate some patients suffering. While this is something of a stretch in implementation the need is very real and affects our communities on a daily basis.
What could it mean to have Artificial Intelligence (AI) rolled out in healthcare?
If we were to have a computer program capable of making medical decisions it might take the brunt of decision making off of healthcare professionals. There is the potential for Healthcare-AI to be either a force for good or a force for disruption and this depends largely on how the technology is rolled out.
For instance, when patients go to the doctor and receive a ‘biopsy’, in which a sample of tissue is taken for analysis, the sample must be processed and imaged to see if it contains cancerous, infected or diseased cells. This process of biopsy collection, fixation, staining and imaging is quite laborious and requires specialized personnel. An example might include a biopsy for skin cancer which would require a trained medical professional to determine if it is ‘abnormal’. Computer vision is capable of doing such image analysis and providing medical recommendations with a great deal of accuracy. This would limit the need for human intervention to making a decision call rather than cycling through hundreds upon hundreds of images.
Potential for misuse of Healthcare-AI
The potential for misuse of Healthcare-AI is also immense as well. Should the medical data of patients be compromised and taken by hacker groups it makes those individuals subject to exploitation. Many hospital networks have been the subject of ‘lockout events’ in which passwords are changed and no longer have access to a breadth of patient medical information. Subsequently this will mean that patients cannot receive their medications or upcoming appointments for surgeries if ‘locked out’. Many non-state hacker groups will charge a ransom fee for access to the network again but it is not guaranteed that the information will be intact or not sold off on a separate platform. Imagine a scenario in which an individual that requires administration of insulin for treatment of diabetes but cannot get access to the medication due to a hacker ‘lockout event’. The results could be devastating.
In closing, what can you and the Brilliant Labs community do about taking part in digital health?
We at BL encourage students and learners to not just use technology but to create it as well. Through a process of ‘participatory design’ we might take apart an existent technology and conceive of ways in which it can be modified to suit our needs. Perhaps something like a health technology could be simple like changing colors to reflect our moods. In fact many heart rate monitors can be run off of platforms like the bboard, microbit, arduino or ras pi allowing students to tinker with how they might personalize it. The digital healthcare future is coming and we had best be prepared.
