Switched-on healthcare – cognitive computing

Watson’s reach already extends beyond the US – other countries that are now undertaking Watson-based projects include Thailand, the Netherlands, Australia, Argentina and South Korea. In 2015, IBM announced a partnership with Mubadala Development Company, owned by the Abu Dhabi Government, to bring Watson to the MENA region. The two companies have teamed up to create a joint venture, Cognit, which will create a network of entrepreneurs, start-ups and public-sector organisations to apply Watson in innovative ways in a number of vertical markets.

For the UAE in particular, the collaboration promises to act as a spur to economic growth. The country is focusing increasingly on expanding its knowledge sector, and the opportunities Watson offers to businesses and public sector organisations to develop data-driven applications will provide a big boost. The IT market in MENA is already growing fast: at the end of 2015, technology research firm Gartner predicted that IT spend in the Middle East would reach $213 billion in 2016.

Information whiz

Much of Cognit’s early focus will be on healthcare, an area in which the technology has already demonstrated its potential. Take Watson for Oncology, an application designed to support cancer physicians in making diagnostic and treatment decisions, and currently being used at the Memorial Sloan-Kettering (MSK) Cancer Center in the US. Watson for Oncology has a store of structured and unstructured clinical data, including thousands of historical patient cases, 204 medical textbooks, papers from 304 medical journals and 12 million pages of text. The historical patient cases are crucial, because they enable Watson to understand how MSK doctors have treated patients in the past.

Oncologists who want to consult Watson about a particular patient can enter the patient’s medical record (which includes information about family history and current medications) as well as details of test results and scans. Watson is then able to analyse that information and to check its database to suggest the likelihood of the patient having a particular illness, as well as appropriate treatment. Both the diagnostic suggestions and treatment options are scored, so that oncologists can make an informed decision about which are likely to work best.

“Watson isn’t meant to replace the oncologist; it’s meant to empower them with information very quickly so that they can come to a good conclusion for the patient.” says Dale Potter, healthcare transformation partner in the IBM Watson Group. Each suggestion from Watson is accompanied by references to the medical literature, so that the oncologists can read the supporting evidence. Watson learns from each new piece of patient information, so the more it is used, the more refined its diagnostic capabilities and treatment suggestions become.

Spreading expertise

Potter sees several benefits in bringing Watson to MENA. One is that it will provide much needed clinical decision support in countries where there are fewer experienced oncologists: “If you look at North America or western Europe, there is no such thing at present as a shortage of oncologists, because people who go through the training to be an oncologist gravitate towards those countries, compared with Bahrain or Yemen or other markets where it’s not particularly attractive for expat physicians to move their family and work there. So this will allow a limited number of oncologists to be much more confident in their treatment decisions.”

In particular, says Potter, Watson will help level out the inequalities between the experience of cancer patients living in urban areas and those in rural areas: “In many countries in MENA, there’s a discrepancy between the success of treatments of people who live in urban centres versus those who live in rural centres, so one of the benefits that we want to explore in this region is whether we can democratise expertise [beyond] the big centres. If we can somehow push this expertise out to where people live, it’s a more comfortable experience for the patients because they’re close to their families and their loved ones, and they can be assured that the care they receive is evidence based and the same as they would receive in an urban centre.”

Another benefit relates to the way in which diseases such as cancer manifest themselves differently among different ethnic groups. “In the MENA region, and specifically the UAE, the median breast cancer age for women in particular is shockingly young,” says Potter. “We don’t know why, so there are some opportunities to continue to train Watson to take these ethnic and biological differences into account as it makes decisions.”

Finally, says Potter, less research is carried out in MENA than in other parts of the world. Watson could help change this: “Using this tool and making decisions based on standardised evidence could heighten the interest and need for research coming out of MENA.”

The use of Watson in MENA could also provide benefits to other markets, adds Potter. The expertise of physicians in MENA in treating young women with breast cancer, for example, could be applied elsewhere: “The physicians in the West are less experienced than people in MENA, because here they deal with it very frequently. If they can make advances in this region, then those advances could be pushed out to Western Europe, to Australia, Canada, the US, because they don’t see it as much.”

Initially, 80–90% of Watson’s application in MENA will be aimed at clinicians: because the key statistic on treatment success in cancer is the five-year survival rate, it will be at least five years, says Potter, before enough local clinical data will have been gathered to enable research to be carried out on Watson’s impact. One of its local uses, however, will be as a training tool: “You can put a real patient clinical condition into the system, and allow residents to explore and read the literature, and it guides them towards the relevant literature. Some of the institutions I’m talking to are considering using this as a testing tool to give a clinical case to a learner, a resident or a fellow, and then have them see the Watson recommendation and decide if they agree with Watson or not.”

Oncology represents an ideal focus for the technology, Potter believes. Despite the big advances in cancer research in recent years, there is still a substantial hit-and-miss element in treatment. Potter participated in a seven-year study in Ontario of breast cancer patients, which found that 44% of first-line treatments were reversed within 60 days, because they weren’t the appropriate treatment for the particular patient. Watson will help oncologists provide treatments that are much more personalised to the patient’s particular characteristics: “Watson takes into account the attributes of that patient: it’s not just that this person is a 44-year old person with breast cancer, stage 2b, it takes into account every single relevant attribute – the fact, for example, that they have a poor cardiac function and ejection fraction of 45, which would mean their cardiac performance is low. That’s very relevant in the treatment decision.”

Potter emphasises, however, that Watson can only give advice: it doesn’t replace a clinician’s own expertise. So, for example, it doesn’t have any ability to take emotional factors into account. Imagine, says Potter, a woman who has early stage breast cancer and asks if she can postpone treatment for 60 days to attend her daughter’s wedding. While Watson would automatically say ‘no’, an oncologist might be happy, after weighing all factors, to agree to this request – just as they would be prepared to consider alternative treatments for a patient who wanted to minimise hair loss or preferred a gentle treatment to a more aggressive one.

Unlimited potential

Watson won’t only be used in oncology, however. The MENA region has a growing problem with diabetes (mostly type 2, resulting from an increasingly obese population); 37 million people, nearly 10% of the population, have the condition, and Potter believes that Watson can help provide doctors with the knowledge they need to improve diagnosis and treatment: “A common factor is that doctors aren’t equipped with the information to say, ‘Here are 150 patients, and here are the ten you really need to pay attention to who are at high risk for diabetes.’ If they had some easy, quick way of knowing that, they could provide interventions to minimise the risk of diabetes not just here, but all over the world.”

In the US, Watson is being used as a tool in precision medicine, in which genetic information about an individual is used to help diagnose or treat a disease. So could the technology be used in this way in MENA? “There are more substantive battles today that need to be fought before we get to that level of sophistication, and that will have a higher impact than focusing on precision medicine,” says Potter. But he thinks this will change in time: “Very little genomic testing is done in the region. And I think probably in the next 18 to 24 months, we’ll have a different conversation.” Potter is hopeful that, with a greater emphasis on genomic testing, more genetics experts will be attracted to the region, creating opportunities for the development of precision medicine.

It’s still early days for Cognit, but the future looks bright. The experience of MSK and other clinics in the US suggests that Watson offers huge potential to transform healthcare – as well as providing a big boost to the region’s economy.