Ditto AI
Blog

3 examples of AI detecting disease with advanced health informatics

Posted by Ditto on Nov 27, 2019 8:15:00 AM

Picture this: A 50-year-old man has an angina attack and requires emergency medical assistance. A few years prior, the man was diagnosed with an allergy to Nitroglycerine, a common drug used to treat angina.

Emergency response teams arrive, but they don’t have access to the man’s patient records and so treat him according to protocol.

They give him Nitroglycerine. He reacts badly, and now emergency staff must deal with his allergic reaction, as well as his condition.

If the medical staff had access to his patient records at the scene – and if an AI could quickly scan his records and uncover his previous diagnosis – the man could have received tailored care without error.

This is just one example where AI and advanced health informatics can benefit the healthcare industry.

But what about using AI for disease detection?

 

1. Diagnosing cancer as early as possible

More than 90 percent of women diagnosed with breast cancer at the earliest stage will survive for at least five years, compared to 15 percent for women diagnosed with the most advanced stage of the disease.

Early cancer diagnosis is challenging. Oftentimes, we react to physical symptoms and seek treatment, by which point treatment might be ineffective. Using AI and health informatics, however, means we can adopt a proactive approach.

Download our guide to responsible AI for your business

AI can use big data to study and analyse the prerequisites to cancer. It can then begin to detect early onset symptoms, helping doctors deploy preventative treatment and save lives.

 

2. AI and urinalysis: giving power to the patients

Healthy.io is a start up that uses AI and health informatics to give urinalysis power to the patient.

A smartphone camera, a mobile app and a home-based urinalysis kit means patients can self-test for diseases including UTIs, antenatal and chronic kidney disease, and share results with healthcare providers if there’s an issue.

Not only does this reduce pressure on medical staff – who conduct more than 42 million urinalysis tests annually in the UK – but it also means patients can benefit from early diagnosis and the privilege of home privacy.

 

3. Acute kidney injury

Acute kidney injury (AKI) affects one in five hospitalised patients in the UK, causing at least 40,000 deaths each year. When diagnosed on time, however, 25 percent of AKI deaths are preventable.

Google DeepMind has developed an AI that gives clinicians a 48-hour head start in treating the disease. Paired with an app called Streams, the technology can send patient test results direct to clinician phones to alert them about individuals who are in danger.

According to a report by Deloitte, the app uses advanced health informatics to provide real-time access to the most relevant clinical information, such as blood test results and medical history. The report also estimates that the workload of nurses using Streams can be cut by two hours per day.

 

AI and early diagnosis: it’s saving lives

Diagnosing a disease as early as possible is one of the most important factors to conquering it. But without physical symptoms, early diagnosis is a difficult thing.

With AI we can study patterns in patient timelines and predict outcomes; it means staff get real-time access to healthcare records when they need it; and it means we can diagnose diseases earlier and deploy proactive treatment, rather than reactive treatment.

The real question, however, is: How does an AI technology reach its diagnosis? For that, we need to make AI explainable. Only then can it become accountable.

Download our guide to responsible AI for your business

Topics AI at work Machine Learning