With the world’s largest population, China generates an unprecedented amount of medical data. It’s an ideal harvest ground for the development and application of cognitive solutions that extract insights from data to improve medical management.
The landscape of Chinese healthcare has undergone huge transformation over the past 20 years. The digitization of medical records resulted in a surplus of structured and unstructured data. And although algorithms have been developed to analyze and understand most structured data, current systems have a tough time making sense of unstructured information, such as clinical narratives, medical images, medical reports, operation notes, and discharge summaries from the roughly 7 billion hospital visits each year in China.¹
The sector needs a faster approach to help researchers and doctors cope with the data explosion. At IBM Research – China, researchers developed a set of cognitive healthcare tools, which use machine learning, natural language understanding, and analytical reasoning to aid in diagnosis and improve care giving.
Patient Similarity Analytics and Risk Prediction Optimization
Cardiovascular diseases are one of the top health concerns among Chinese. Right now, there are about 290 million cardiovascular patients living with the illness – that’s one in every five people in China. It is becoming a pressing issue for China’s public wellness.²
That’s exactly why IBM Research – China dove into this domain starting in 2015. The intention is to work with top cardiovascular experts and doctors in China to help build a highly intelligent cognitive model to identify key factors that may predict stroke onset. Together with top hospitals in Beijing, we constructed thousands of patient features from an even bigger set of patient data and applied machine learning to discover factors which may cause or prevent stroke. For instance, living with a spouse or partner was found to be a significant protective condition to help patients avoid stroke.
Combining data-driven research with cognitive technologies helps clinicians make new discoveries and expand their understanding. For example, patients with atrial fibrillation (AF) are often prescribed a certain type of anticoagulant to help reduce the risk of stroke occurrences. However, for some patients, it will have the opposite effect and increase the chances of stroke or hemorrhage. Our research-based cognitive dashboard can learn from a vast amount of desensitized patient data and divide a population of AF patients with conventional high risk for stroke into multiple subgroups. Then, it could be used to identify a truly low risk patient group unsuitable for the medication aforementioned and an additional group who would be extremely responsive to the treatment.
Cognitive advisor for diabetes
Another effort from IBM Research –China is centered around cognitive decision making for healthcare providers looking after type II diabetes patients. A cognitive advisor can learn from up-to-date, public information — including diagnosis guidelines and research papers — to produce information for human experts to review and manage. The goal is ultimately to pair this with real-world patient data, and generate reliable insights for practitioners in any given clinical setting.
On a daily basis, practitioners need to see a long line of patients and help manage their changing conditions. The daily management of diabetes involves many factors, from insulin sensitivity to carbohydrate intake, making every diabetes patient starkly different. A cognitive advisor would potentially help practitioners throughout the day, assisting in areas such as evidence support and risk analysis so they can provide each patient more personalized care.
The healthcare industry is infused with big data and I couldn’t think of a more challenging yet exciting industry to apply our innovations. IBM is working with partners across the industry to help clinicians enhance patient care. By leveraging the power of augmented intelligence, we are dedicated to enabling domain experts to sail beyond traditional understanding of chronic illnesses and improve care to patients in China and beyond.
 Source: China Daily official figures