Artificial intelligence (AI) in healthcare is no longer the imaginings of the movie screen – it has arrived and is becoming more widely used in medical practices.

We’re not quite at the level of sophistication and comfort where we can rely on AI on its own though. As a developing technology, there is still much more to be learned and tested, but there is another way to use this technology without giving it complete autonomy: Augmented intelligence.

Free download: Augmented intelligence in healthcare, a Q & A

What is augmented intelligence?

To understand augmented intelligence you must first have a good grasp on what artificial intelligence means. This is otherwise known as “machine intelligence” and involves the development of computer systems that can perform tasks which usually require human intelligence. AI covers machine learning, robotics and natural language processing. For example, what if a computer program could read an X-ray and determine which bone is broken and the type of break?

In most applications within healthcare, we’re not there yet in terms of allowing AI to “make decisions”, “talk” to patients or have the final word on any kind of diagnosis. It’s not just about technology development, but about patient preferences. In one survey, while 42% of patients were comfortable with their doctors using AI data for healthcare decisions, 62% stated that they want communication via phone call. People understandably want to make sure that they’re still getting a personalized, human touch.

Augmented Intelligence is an answer that puts artificial intelligence in an assisting role to the human operator. Its emphasis is to enhance human intelligence rather than replace it. So in our example of AI that could read an x-ray, it might point its findings to an ER doctor or orthopedic surgeon for them to examine and have the final word.

Augmented intelligence

Why use augmented intelligence?

Choosing augmented intelligence gives you the benefits of artificial intelligence, while reducing or avoiding some of the pitfalls. For patients, it can often be reassuring – everyone wants to know that they’re being examined and listened to by a qualified human.

In terms of AI in healthcare, there are a few current limitations, for example:

  • Ethical questions about the boundaries between the physician’s and the machine’s role in patient care.
  • Ethical concerns about ensuring patients make informed consent and that patient autonomy is intact. For example, it must be clear to the patient the role that AI will play in their care, whether that is during surgery, diagnosis or at any other point.
  • AI doesn’t provide that “human” touch and that’s what people want from healthcare. If you contrast this to banking or other less “emotional” fields, a lot of people are happy with a high technology, low touch environment.
  • Machine learning “learns” from what happens around it. This means that in some cases, it has been shown to pick up on prejudices that human operators may have. For example, it may not provide equally accurate predictions of outcomes across race, gender or socioeconomic status.
  • Machine learning isn’t perfect. As it learns from what it sees, there will be some scenarios it just hasn’t seen yet that an experienced doctor may have. For example, consider reading an MRI or CT scan – there are many different variables and of course, rare cases that may come up only once every few years.

Augmented intelligence, where the AI is used as an extra tool for the human operator, allows busy clinicians to take a proactive approach to care in many cases. This could also be helpful for patients in terms of giving them preventive care before they get to a point where they need treatment.

As an example, AI might monitor factors such as blood pressure, weight and blood sugar readings, all known risk factors for cardiovascular disease. Poor readings could be flagged for the patient’s physician so they learn of them earlier than they otherwise might if they only see the patient for routine check-ups.

Taking an augmented intelligence approach also means that the human is there to intervene and double-check results. For example, what if AI didn’t flag anything on a CT scan, but the eyes of the experienced physician can spot an abnormality?

Dr. Anushka Patchava is an expert in the field of augmented intelligence in medical care and an advisor to the United Nations. Here’s how she positions augmented intelligence:

“We don’t want to completely replace doctors and healthcare workers; we need to find synergistic partnerships, where it’s neither high touch/low touch or low tech/high touch, it’s tech and touch working in combination to deliver the best outcomes for both physician and patient.”

How is augmented intelligence applied in healthcare?

Augmented intelligence can help in areas of healthcare where human intelligence can benefit from assistance from AI. Here are a few examples of how it is being applied currently:

  1. Workflow optimization. This can be in the form of procuring results more quickly for physicians and improving referral services. “We should free our experts to undertake expert work by removing as many non-essential tasks for them as possible,” noted Angie Craig, assistant director of operations and performance, Leeds Teaching Hospitals Trust, National Health Service, United Kingdom. “That’s where artificial intelligence comes in.”
  2. Improved communication flow and work prioritization. For example, a study of Medic Bleep users revealed that they saved significant amounts of time per shift through a time and motion study using the communication app.
  3. Genomics is an area undergoing significant development with AI. “The use of artificial intelligence to make sequencing and analysing DNA faster, cheaper, more accurate and more available could have significant implications on the way we deliver healthcare.”
  4. Diagnostics. A Lancet review comparing studies of deep learning performance in detecting disease against the performance of healthcare professionals found that AI has made significant improvements over the last few years. Diagnostic performance between the humans and the AI was approximately equivalent.
    This does have some limitations, as indicated earlier. For example, AI is being used in breast screening, but a certain number of those screens fall into an ambiguous area that requires further analysis by human operators.
  5. Disease diagnosis and drug design. For example Bayer is working with tech companies to develop software to help diagnose disease and develop drugs that will treat them. The AI looks at information such as symptoms, test results, disease cause, medical images and doctor’s reports.
  6. Patient communication and customer service. Healthcare chatbots are becoming increasingly popular, with applications such as streamlining admissions, improving at-home care and assisting with follow-up. 
Augmented intelligence

There are plenty of other areas that are undergoing development within healthcare fields. It is hoped that augmented intelligence will streamline healthcare and reduce costs. It has the potential to enhance and improve the performance of doctors and other healthcare workers, while allowing them to focus on the most important areas that require their expertise.

Free download: Get our Q & A on augmented intelligence in healthcare here

Final thoughts

Augmented intelligence is taking its place among the technologies that are the future of healthcare. For private practices, augmented intelligence can prove to be a solution that gives them the best of both worlds, while mitigating the limitations of each.

For example, augmented intelligence can help to improve workflows and productivity, while assisting the human operator with their tasks. Taking an augmented position means that the human is still heavily involved, respecting patient wishes for personalized care, while enjoying the benefits of expediency that AI can bring.

Healthcare providers should look to ways AI can enhance their current operations. It may prove to be the perfect assistant.