Artificial intelligence in healthcare

Artificial intelligence (AI) enables computer systems to perform tasks that typically require human intelligence, learning from experience and adjusting to new inputs to mimic human cognitive functions.

While the banking and housing sectors have been quick to realise the benefits and adopt AI, healthcare has traditionally been hesitant.

Aside from being more efficient and cost-effective, these technologies have the potential to transform many aspects of patient care, as well as streamline administrative processes within provider and pharmaceutical organisations.

While there has been an ambitious discussion of whether AI doctors will eventually replace their human counterparts, 6B don’t think this is desirable or viable. Instead, we see AI working in harmony with medical professionals, helping them make better decisions, and even replacing human decisions in certain functional areas like radiology.

Scaling AI in healthcare

AI devices currently being used in the healthcare sector fall into two main categories:

Machine learning (ML) techniques that analyse structured patient data like imaging to determine the probability of disease outcomes.
Natural language processing (NLP) methods that extract information from unstructured data like clinical notes to transform texts into machine-readable data.

For organisations that operate in the sector, AI is fast becoming a top area of growth and investment. From apps that help patients to manage care themselves to man-made pancreases that release insulin in response to changing blood glucose levels, the industry is waking up to the benefits of using artificial intelligence in conjunction with human intelligence.

The scale to many solutions is small, but the increasing adoption across public and private providers indicates that momentum is accelerating.

But what is needed for large scale adoption?

The logical starting point would be to address largely routine and repetitive administrative tasks that absorb significant time from doctors and nurses. This would help optimise operations and pave the way for widespread adoption.

Next, we would expect AI solutions to help ease the transition from hospital-based to home-based care, enabling the remote monitoring of patients with alert systems in place. But this will require a deep cultural change as the workflow of clinicians would be drastically altered.

Finally, the integration of AI solutions in clinical decision making based on lessons learned from clinical trial studies, playing an important role in how patient care is investigated and delivered.

The impact of large scale adoption will change how healthcare is delivered altogether. It will allow for enhanced patient care at every level and will help fulfil an ever growing employment gap in the sector, rather than replacing jobs completely.

Streamlining administrative processes

To fundamentally improve patient care, it’s of the utmost importance that healthcare providers find a way to free up doctors and nurses so they can spend more time interacting with patients.

This means reducing administrative tasks that could otherwise be automated.

Robotic Process Automation (RPA) is relatively inexpensive and performs repetitive tasks far quicker than any human could, to a higher degree of accuracy, and it’s easier to program than other AI technologies.

This technology is allowing medical professionals to focus purely on delivering the best quality patient care, with tasks like updating patient records and billing history now being performed to a higher standard by bots.

But digitalisation has brought about its own problems – namely user burnout. To combat this, electronic healthcare record developers are using AI to create more intuitive interfaces and automate routine processes that consume so much time.

Developers are exploring the possibility of recording clinical interactions with patients and then leveraging AI to sift through this information for future retrieval, which would save valuable time in the documentation process. These technologies will also help users to prioritise their workflow and operate more efficiently based on patient priority.

Early detection and diagnosis

Detecting an illness early can be the difference between life and death; it opens the door for future care and treatment that can cure or lengthen survival time. It’s why we have periodic health screenings for diseases like cancer.

The proliferation of consumer wearables and other medical devices combined with AI is enabling the healthcare industry to monitor patient risk more effectively.

With the majority of the UK adult population having access to a smartphone, a growing proportion of health related data is being generated on the go. From mobile apps that track step count to devices which monitor heart rate, these tools provide invaluable insight into the health of the population.

AI will play a key role in assessing this data trove and extrapolating actionable insights that can save lives. Although patients are likely to be wary of sharing data on such a large scale, they are more likely to trust the advice of a doctor than a multinational corporation like Google or Facebook.

By collecting data on a granular and continuous level rather than simply when a patient is under observation at a hospital, we can begin to mitigate a growing threat to populations around the world: antibiotic resistance. Electronic health record data can help identify patterns and create faster and more accurate alerts for patients at risk due to overuse before they even begin to show symptoms.

Universalising healthcare

The use of AI not only has the power to help fill employment gaps in the UK, but it can also help fill critical shortages of trained physicians in the underserved and developing regions of the world.

In a world where access to healthcare provision is so unfairly skewed in favour of countries in the western world, technology can help mitigate the severe deficit of qualified clinicians by fulfilling diagnostic duties typically allocated to humans. The role of an ultrasound technician can be filled by AI to detect signs of pregnancy, vascular diseases and more, providing local doctors and nurses with the vital support they need.

These life-saving tools can be deployed via an app to low resource areas in the world, evening the playing field for patients and reducing the need for a trained physician to be on site.

However, it’s important to note that algorithm developers will need to ensure that the data they use is compatible with diverse communities, as a one-size-fits-all approach will be counterproductive and potentially harmful to the people they are trying to help.

Research advancements

The road from research lab to patient is long and costly in the world of drug development. More often than not, billions of pounds and several years are lost to drugs that don’t even make it to market. The early weeks of the coronavirus pandemic and the uncertainty surrounding a viable vaccine put this at the forefront of everyone’s mind.

Drug and research discovery is an AI application very much in its infancy, but by directing the latest advances in AI to streamline this famously lengthy process, we can hope to slash the time to market for new drugs and the eye watering sums associated with them.

Machine learning algorithms can help optimise the clinical trial phase, offering advanced predictive analysis that researchers can use to select candidates, drawing on data that compares their medical history to a specific target population.

This will also help to improve the overall safety of clinical trials, as researchers can keep a closer eye on participants and monitor whether they are responding adversely to treatment.