What can I eat in pregnancy? App aims to answer with help from IBM’s Watson | Innovation
Daffi is pregnant with her third child. On holiday in Thailand recently she wondered whether it was safe for her to eat prawns. She asked her new pregnancy app, which reassured her that she could go ahead and pop the prawn in her mouth.
Daffi's partner is Jonathan Lipnik, who developed Nutrino with his friend Yaron Hadad. They aim to cut through the plethora of nutrition information to provide customised advice for an individual. “Our nutrition insights platform is built on three pillars – machine learning, optimisation and big data,” says Hadad.
Hadad's interest in nutrition started when he tried lots of different diets as a teenager to help cure his migraine and sinusitis. For him, becoming vegan was the answer, but in his quest for a solution he found that dietary information is overwhelming and often contradictory or misleading.
“Nutrino is designed to compress the broad knowledge that exists in nutrition, and turn my three years of experimentation and research into a few minutes for other people.” The first product appeared two years ago and a new version for pregnant women has just been announced.
The Nutrino app, powered by IBM's supercomputer Watson, claims to guide women through pregnancy. For $15 (£10) for the duration of pregnancy, the app gives personalised meal recommendations and nutritional support by combining Nutrino's nutrition database and Watson's natural language capabilities.
Watson is a computer system that can answer questions posed in natural language. Its cognitive computer system understands natural language and isn't programmed but learns. Watson analyses unstructured data like news articles using natural language processing to understand grammar and context. If you ask it a complex question, it evaluates all the possible meanings and determines what is being asked. Then Watson presents answers based on the supporting evidence and quality of information it finds in the text.
A Nutrino user needs to input the stage of their pregnancy, health goals, eating habits and preferences. They choose from a list of questions about nutrition in pregnancy. The app then uses Watson to search the data in Nutrino's platform and, using the individual's profile, customises the answer. In time, the answers become more tailored to the user.
A pregnant woman may, for example, want to know how much coffee she can safely drink when pregnant. She can type or ask and Nutrino will tell her: “There is conflicting data on how much caffeine is safe to consume during pregnancy. Most of the data does not suggest an increased risk of adverse pregnancy when consuming less than 300mg per day of caffeine. In order to stay on the safe side, many nutrition authorities recommend limiting caffeine consumption to under 200mg per day.” There are references that you can click on to find out more.
This information is very similar to that available free on the NHS choices website in the UK. The trick to avoiding information overload in pregnancy is to stick to a couple of tried and tested, reliable sources of information. GP Dr Carmel Mond recommends NHS choices and patient.info as good starting points for any medical query.
GP Dr Sharmili Bose thinks access to up-to-date dietary advice in pregnancy is very useful. “Guidance changes all the time. For instance, pregnant women used to be told to avoid peanuts, especially if there was a history of allergic conditions, but current evidence suggests they're safe so the advice has changed. I'm not sure whether you need customised advice but it's certainly good to have one reliable go-to source of information.”
Hadad says that the new app is especially useful for expectant women who use wearable devices to monitor their exercise, sleep and other activities. “If the device detects that a woman isn't sleeping long enough or awakens too often, it might suggest to her, unbidden, to compliment (sic) her diet late in the day with foods that contain the natural sleep-inducing hormone melatonin, such as almonds or raspberries.”
The problem with this is that there's no reliable evidence linking duration of sleep with outcome of pregnancy. Most pregnant women wake up during the night, often to do a wee. Suggesting that this is detrimental to the developing baby seems alarmist and ill founded. And although a wide range of foods, including almonds, may boost melatonin production, there is no evidence that extra afternoon top ups will impact on sleep.
Industry analyst Sam Jacobs says that this technology platform can connect pregnant women to the internet of things. “Remote devices connected to powerful processors can give people real time, customised health advice. I predict that this is the start of a wave of applications that will change the way we live.”
Nutrino is likely to appeal to women who already track their diet and exercise. The fitbit and Apple Watch generation may prefer to get their information about pregnancy by talking to their wrist rather than chatting to their mums. But even Watson may struggle to provide the common sense and personal experience that complements scientific knowledge. Your mum may come up with some random, non-evidence-based stuff, but then so does Nutrino.