10 ways that AI is shaking up healthcare - Part 1
As an MD, I admit that we don't see this revolution in our daily lives, yet these technologies are advancing medicine at incredible speeds.
AI systems are expected to be implemented in 90% of U.S hospitals and globally in 60% of hospitals and insurance companies combined, and thus will deliver easily accessible, cheaper and quality care to 70% of patients. Furthermore, the AI market for healthcare applications is expected to achieve rapid adoption globally, with a CAGR of 42% by 2021. Frost & Sullivan even project that by that time these systems will be generating $6.7 billion in global revenue in the field of healthcare alone.
Let’s look at 10 different ways Artificial Intelligence is bringing about impactful change.
Mining Electronic Health Records
The Electronic Health Record is a single location where all patient data originating from different sources are gathered together into one repository. Imagine a situation where a policy-maker at the ministerial level has to make an important decision on allocating resources for that particular year. How is he/she going to achieve that? In collaboration with public health professionals, data scientists and informaticians, thousands, if not millions, of anonymized patient records will be analyzed against specific clinical coding standards to provide insights in a couple of hours within an interactive dashboard backed by Artificial Intelligence. This is impossible in an environment where paper clinical documentation reigns. This process usually still takes months and is riddled with scientific inaccuracies.
“But EHR mining does not only have application in the context of high-level decision-making. In fact, it offers multiple opportunities to directly improve patients’ healthcare experience.“
But EHR mining does not only have application in the context of high-level decision-making. In fact, it offers multiple opportunities to directly improve patients’ healthcare experience. For example, how about using artificial intelligence in patient recruitment for clinical trials? EHR Mining enables patient matching. That is, if a Health System collaborates with a Patient Recruitment System (PRS), then the PRS could easily reach out to specific patients to offer them the chance to participate in an available clinical trial.
Advanced consultations and chatbots
It’s snowing outside and you feel as if your nose is about to explode and you can barely breathe. You struggle to decide whether to make your way to the doctor or just stay inside and get better with the usual concoction of medications. You decide to be responsible and call your favorite GP, who informs you that they won’t be able to make it today as their car has broken down. The most support they can offer at that moment is general advice and an appointment another day.
This won’t happen anymore. You just take your smartphone out of your pocket, start an interaction with your favorite, reliable chatbot – a special app you can interact with in the form of a conversation. This chatbot has been scientifically validated by a team of qualified and experienced doctors. Outcome: you get useful advice and, in the case of an emergency, a video-call appointment within a couple of minutes.
Interest in diagnostic chatbots is growing. Check out Anna’s article on Healthcare and Chatbots and, if you like, test it out with something as simple as ‘I have a headache’. Be ready to be marvel as I did!
The transition from paper-based clinical documentation to electronic clinical documents has not been the easiest. In fact, many physicians and other HCPs still see data entry as a challenging part of their daily jobs. This pain point, however, has brought about some strategic partnerships, such as the recent one between Nuance Communications and Epic – one of the world’s leading patient electronic record providers. The cooperation has resulted in the integration of the Artificial Intelligence capabilities of Nuance's AI computer-assisted physician documentation tool into the Epic NoteReader module for clinical documentation improvement. By analyzing relevant patient notes using deep learning and natural language processing technologies, the Nuance CAPD tool can spotlight certain clinical indicators in an electronic medical record and alert doctors when data is missing or needs clarification.
“An emergency room radiologist may look at as many 200 cases a day, and some of these, such as lower body CT angiography, can include as many as 3,000 images per case.“
All kinds of healthcare professionals do repetitive and tedious jobs that are distinct to their specialty and are a necessary part of their day-to-day work. AI can also help with specialty- or role-specific activities. Radiologists, for instance, see countless radiological images. In fact, an emergency room radiologist may look at as many 200 cases a day, and some of these, such as lower body CT angiography, can include as many as 3,000 images per case. This can bring about eye fatigue, and with radiologists being a scarce resource in many countries, it is crucial to facilitate this image viewing process. Given this pressing need, IBM Research, through their project Medical Sieve, has created an image-guided informatics system filtering the important clinical information physicians need to know about the patient for diagnosis and treatment planning.
Enhanced patient care pathways
Patients today have a number of opportunities to improve their care experience, and one of the most crucial elements is getting connected to the right provider in the right health system and at the right time. This is known as a Patient Access Solution, and Kyruus and PokitDok seem to be the leaders in this field. In simpler language, patient access has to do with booking an appointment, figuring out who is the right doctor for you, and organising a blood test, but also much more. There are now startups offering a comprehensive solution which takes your doctor visits to the next level through the nifty combination of AI and doctors. Forward is one example. It takes care of every step of your primary care experience: from having your vitals analyzed by a futuristic-looking scanner to being seen by a doctor with a large touch-screen.
Rising to the challenge: better public health with Artificial Intelligence – Part 1
Public Health, that is, population-based medicine, has a pivotal role in the development of Artificial Intelligence for health and well-being.
Personalized treatment regimens and clinical decision support
The potential for personalized treatment regimens to bring about effective treatment with long-lasting positive effects is impressive, to say the least. Furthermore, supported by Clinical Decision Support, clinicians have the opportunity to provide specific treatment with less anxiety about treatment-related issues such as adverse effects of drugs. There are a number of technology companies taking the lead in this area.
“The potential for personalized treatment regimens to bring about effective treatment with long-lasting positive effects is impressive, to say the least.“
IBM Watson for Oncology is collaborating with oncologists to provide clinicians with evidence-based treatment options. The beauty of this is the analysis of the meaning and context of both structured and unstructured data within clinical notes and reports. This documentation can prove to be critical in selecting a personalized treatment pathway, especially given the inclusion of attributes originating from the patient’s file, external research and data.
IBM is not the only company working on this; in fact, Google DeepMind, amongst their many projects, is collaborating with the University College of London Hospital on the particularly difficult task of treating cancers of the head and neck. These affect over 11,000 people a year in the UK alone. The anatomical complexity of this region of the body requires well-planned radiotherapy to ensure that no healthy structures are damaged and thus isolate the cancerous tissue to be treated from healthy tissue. DeepMind’s Research project is focused on using Machine Learning to speed up the segmentation process while maintaining accuracy. This process usually takes up to four hours for head and neck cancers, and the project believes that they could potentially free up clinicians’ time to focus on patient care, education and research. The excitement about this project is that this could be transposed to a number of cancers and scale up to the whole national health system and thousands of other patients.
This article is Part 1 of a 2-article series on the Present and Future of Artificial Intelligence within Healthcare. Check out the second piece!