Priya Kulasagaran finds out more about how machines and algorithms are changing the face of healthcare
Thanks to popular culture, Artificial Intelligence (AI) tends to bring up images of dystopian futures where humans are enslaved by their robot overlords. The reality however is that AI can be harnessed to improve our lives, and this is already proving to be the case in healthcare.
“We shouldn’t be afraid of AI,” says Philips’ Healthcare Informatics and Population Health Management (Asean & Pacific) business leader Fernando J Erazo. “It will never replace humans, but instead elevate the roles of humans. A clinician should never have to spend time information-finding, toggling screens and consolidating case notes. If we can do that with machines, clinicians can come in and really concentrate on making the best therapy choices.”
AI is more about using algorithms to make sense of data rather than creating self-aware robots. In the world of healthcare, this is particular pertinent as a single patient will easily generate tons of data — a full-body computed tomography (CT) scan for instance, can generate hundreds of images. The task at hand then is connecting medical professionals with all this information, and helping them make sense of it.
“AI in itself is not new, but our currently ability to improve the user interface is; the user experience is what will make the difference in the near future. With the power of computing and the Cloud, we can really present the data to healthcare teams in a sensible way. So, (machine) automation can help with the way images are processed and displayed, but you have to combine that with the workflow of hospitals, how the clinicians need to use the software — if no one knows what to do with the information, you get nothing,” explains Erazo.
Personalising healthcare through machines
Harking to its roots of developing medical imaging such as x-rays, Philips is now moving into the data and informatics space in healthcare. An example of this its Intellispace platform, a clinical suite of solutions across disciplines such as oncology, radiology and cardiology which uses AI.
“The underlying concept with the different Intellispace products is to get all the relevant information to the healthcare team to enable earlier and faster decision-making; this will help them deliver better and more personalised care,” says Erazo. “The portal has over 80 clinical applications, so its usage varies, and it’s also being used in academic and research settings.”
A deceptively simple application of Intellispace can be found in oncology, namely its tumour tracking function. A cancer patient is likely to have a multidisciplinary team handling his or her care — physicians, surgeons, oncologists, palliative care professionals — who all need to weigh in on how best to proceed with treatment.
“And it’s very important for this team to reach an agreement on a simple thing: is the tumour shrinking?” says Erazo. “If a patient is undergoing treatment like chemotherapy or radiation, they can’t just extend the patients therapy unless they know it’s working. So how the tumour tracking helps is that it gives a quantification; the machine algorithm can explain that the tumour has shrunk by X%. We can now calculate this across several cancer types, as long as tumour is solid.”
Achieving such objective measures can also help in another area of medicine: neurology. With mental health poised to be next huge healthcare burden, there is a need for medical professionals to be able to access more accurate data on the brain.
“Here, with clinical teams looking at brain images, healthcare teams will be having the same debate over whether a certain area of the brain has shrunk or not. By having quantifiable data, doctors can be much more confident in doing a ‘first time right’ diagnosis. For patients, and their family members, this means they can get treatment earlier instead of waiting months for the results of different tests. That’s what data can do — give you a more precise diagnosis earlier and faster,” says Erazo.
Unlocking the power of data
One of the most remarkable features of modern medicine is that we are now living longer. The question is whether we are living out those years in good health. It is no secret that non-communicable diseases are major health burden worldwide; in Malaysia alone, heart disease is among the leading causes of death among adults.
As doctors have yet to discover a foolproof cure for chronic diseases such as diabetes and cardiovascular diseases, the next best step is helping patients manage these conditions. While health education is an important aspect at mitigating these diseases, a big problem healthcare professionals face is trying to identify possible health conditions before they worsen.
“Where the system gets broken is when patients are reactive,” says Erazo. “I do predict that in 50 years, our grandchildren will look back and say the way we treat heart disease is almost barbaric. Unfortunately, what we do today is wait for the condition to worsen, and then intervene in an aggressive way, like with a cardiac bypass, or very invasive surgery. But what if we can detect issues earlier, and use less invasive therapy to deal with problem?”
With data, this is an easily achievable possibility. From image-guided therapy, and software to track the information gathered from a patient, doctors are now able to catch patients much earlier instead of waiting for them to show up at hospital on the verge of a heart attack.
“In cardiology, you need the total picture of care, otherwise you literally cannot see the consequence. So, screenings are important of course; your first encounter with a cardiology practice is an electrocardiogram (ECG), then you may go for a second one a year later. What is important is not have all the data just sitting there, because you need to see a trend — that trend is going to help me find cases before patients develop a condition that is too acute,” says Erazo.
In this respect, Erazo stresses that there needs to be a greater push for healthcare systems to develop secure IT infrastructures that enable public and private hospitals to freely exchange a patient’s medical records. “There is a need for greater advocacy for a one-patient-one-record system so the data doesn’t get lost — then we can look at low cost interventions earlier on,” he says.