Will AI be the doctor’s liberator?

Apr 17, 2018 - reading time 6 mins

Jesse Hirsh is an internet professional with 25 years’ experience in research, writing, and public speaking. His key interest is helping people understand what the future holds for them. He regularly appears on television, radio, and social media like Twitter, Medium, Instagram, and Tumblr. He has a TEDx talk on the future of authority and once hosted an interfaith television show called 3D Dialogue. He completed a Master’s Degree at Ryerson University on the rise and increasing role of algorithms in the governance of our society and is currently investigating the impact of artificial intelligence and cognitive computing on all sectors and industries.

By Jesse Hirsh Click here to read less

As a tool, AI offers tremendous potential, but as a concept, it requires a new kind of human agency and responsibility.

These are early days in the era of artificial intelligence (AI) and it can be difficult to understand the impact of this technology amid all the hyperbole. For health care professionals (HCPs) this poses a cultural and professional challenge. The embrace of AI could result in a more patient-centric health system with improved health outcomes and longer lives, or it could lead to an inhuman and automated system that treats health as a commodity rather than a social value.

Which path is travelled will be determined by our literacy – put simply, our ability to understand and use AI-based technologies appropriately.

HCPs must therefore be curious, but also committed to finding out how these tools can bend to their vision of what healthcare should be, rather than the other way around. Too often when new tools emerge, out of ignorance or blind faith, users are expected to bend towards those tools, when those tools should bend towards the users.

Intelligent assistants and patient data

Some physicians will spend upwards of 70% of their time managing data rather than managing patients. Yet what if a lot of this data management could be automated?

Voice recognition can be used to automate the recording of notes, which allows us to envisage a future where conversations between a patient and their doctor are automatically recorded and transcribed. The system could highlight words that require human attention and prompt for tests or follow up when the conversation or context reflects the need. Once the visit is complete, the HCP can review the recorded exchange for both accuracy and to authorize the actions desired.

AI depends on data. The more data is available, the greater the potential of that AI. Presently, electronic medical record (EMR) systems do not have the ability to import data generated by the patient. However, given the rise of wearable technologies, the patient is increasingly generating more data than their physician is in a position to collect. An AI-driven EMR would be able to both import this data and make sense of it, allowing the EMR to more accurately reflect the overall health of the patient.

Perhaps the patient should have more direct contact with their health record? Bright.md is an example of an automated intelligent assistant that helps screen patients. Programmed to interact with the patient and ask questions about their need to see an HCP, it helps manage scarce resources by collecting important information from the patient while also assessing who they should see and what resources they may require.

AI will require human supervision for the foreseeable future, but it can adequately perform mundane tasks that free up humans to focus on the more fundamentally human stuff.

Certainly, when it comes to data collection there’s much that can and is being automated, not to mention that the patient can and should be contributing data directly.

AI as augmented intelligence

In oncology, however, AI is already performing more advanced tasks, albeit with human assistance. Having been trained by oncologists, machine learning systems are moving towards being able to diagnose specific kinds of cancers based on test results. This could then help chart a customized treatment plan for fighting that cancer and improving outcomes.

These systems can process all of the available medical literature, which is of course beyond human beings. An AI can not only scan and index all of the literature but continuously integrate new research as it emerges. This gives such a system abilities that a human professional will never have.

Yet what gives AI its true potential is how it can work with a human HCP and help augment their knowledge so that the best decision can be made.

This works in oncology because of the wide range of medicines and methods that can be used to combat cancer. A human can’t keep track of the various options and near infinite potential combinations. Yet AI can sift through all that data to find the connections and overlooked gems.

As this technology continues to empower and offer HCPs and patients more effective means of treating illnesses, the next step is to use AI to improve preventative health and assist people in avoiding illness. Or, in the case of chronic conditions, avoiding flare-ups and unnecessary complications.

The potential of prevention

This comes back to the issue of data. We presently have vast oceans of health information that we do not use. And, as with the example of keeping up with medical literature, we could be using systems to process this in way that leads to better patient outcomes.

Personalized AI will lead to more personalized health, which then helps patients manage health so that illness can be managed better or avoided entirely. Think of it as the personal assistant that, combined with relevant health data and accessing the latest medical literature, can help guide users through their day and lives with a focus on maintaining and improving their health.

This will have a profound impact on aging and geriatric care. Technology is already extending our lifespan and will continue to do so as our health system is able to keep us living longer and healthier. But aging is not easy and the sorts of health issues developed over time require teams of HCPs, each with varied disciplines and expertise.

AI can help coordinate and consolidate a healthcare team by ensuring everyone is on the same page – the patient’s.

AI can and should also ensure that as people age, they’re better able to maintain their independence, mobility and dignity. This is where the promise of the smart home has potential, but only if it is on the terms of the resident and respects their security and privacy.

It is essential that, as we add automation to our healthcare systems, we focus on the needs of the patient and the importance of patient-centric care. There must be a strong emphasis on consent, privacy and security, but it also involves empowerment, agency and the opportunities technology affords patients, in particular, the ability to have responsive and effective healthcare.

In present terms, we’ve split our health systems in two. This split is not between private and public healthcare, but rather between literate and illiterate healthcare. If the HCP has a high level of tech literacy, they’re able to provide a higher level of care to their patients than if they have a lower level of literacy. This also plays out for the patient (and their family). The higher their literacy, the greater their ability to use technology to improve their care and their overall health.

As we move forward, HCPs should not fear technology, but rather embrace the challenge and opportunity to use it to help improve the care they’re able to offer their patients.