With the steady improvement of medical informatization and the rapid development of advanced technologies, the traditional healthcare system has been in a state of profound transformation for more than 10 years. Big data, algorithms and computing power are creating countless breakthroughs as deep learning demonstrates its potential and value across a range of healthcare applications.
In the late 2000s Fortune Global 500 healthcare companies ramped up AI deployment in the industry, from in-hospital diagnosis and treatment to drug supply chain and out-of-hospital scenarios. AI has since become a key industry component globally and is believed to have the potential to solve major healthcare challenges while overcoming resource shortages, reducing costs and improving patients’ clinical experience.
I. Global Healthcare Industry Market Overview
The healthcare industry has experienced rapid development over the past decades, owing to increasing public health awareness, a global economy that continues to grow, and the aging of populations. Deloitte predicts global healthcare spending will hit US$8.7 trillion by 2020, a compound annual growth rate (CAGR) of 4.3 percent.
According to data from the Organisation for Economic Co-operation and Development (OECD), US healthcare expenditures accounted for 16.8 percent of the nation’s GDP in 2015, while the corresponding figure for China was 5.4 percent, suggesting China has relatively more room to grow regarding healthcare spending.
II. AI Technologies in the Healthcare Industry
A wide range of AI technologies including computer vision, natural language processing (NLP), data mining and robotics have been implemented in specialized healthcare scenarios such as healthcare and pharmaceutical logistics, chronic disease management, in-hospital diagnosis and treatment, health insurance, etc. Computer vision and robotics are the most commonly used mature technologies in the healthcare system.
Computer vision is applied to assist medical image processing and robotic surgery and to support facial recognition and document reading in pharmacies and hospitals.
Data mining is mostly used for healthcare data management and medical payments.
NLP is highly valuable in scenarios that need text processing, e.g. electronic medical record systems.
Robotics can significantly improve healthcare automation and are widely applied for pharmacy customer service, healthcare guidance, pharmacy automation, surgery, etc.
Deep learning is commonly used for intelligent medical imaging in medical diagnosis.
III. AI Application Scenarios in Fortune Global 500 Healthcare Companies
In-hospital diagnosis and treatment: This involves general disease diagnosis and treatment scenarios that relate to all patients. With abundant valuable healthcare data extracted by machine learning, robotics and other AI-related technologies, the diagnosis and treatment process can become more efficient.
Drug supply chain: AI is primarily applied in pharma R&D and sale processes. As an innovative tool, AI can perform highly efficient compound screening during drug discovery. During drug logistics anddistribution, machine learning technologies can effectively predict and forecast for the business cycle.
Out-of-hospital scenarios: People who live with chronic diseases or under sub-optimal health status can enable real-time monitoring of their physiological parameters via wearable devices, with data uploaded to the cloud for further analysis. This enables abnormalities to be more quickly detected so clinicians can make earlier interventions. Medical underwriting is another important scenario where AI is being applied in the healthcare industry.
Other scenarios: Other popular AI applications outside the scope of our Fortune Global 500 healthcare companies coverage include drug production and hospital management.
IV. AI in Fortune Global 500 Healthcare Companies: Technology, Scenario and Application
V. Representative AI Application Cases in Fortune Global 500 Healthcare Companies
HCA Healthcare’s SpoT Algorithm for Sepsis Identification: Leveraging machine learning technology and informed by data from millions of hospitalizations, US-based HCA created the SpoT algorithm to quickly detect sepsis. By monitoring patients’ physical parameters such as temperature, pulse, respiratory rate, white blood cell count and lactate level, SpoT can help clinicians and nurses identify suspected sepsis early. The algorithm operates with 100 percent sensitivity, meaning all true sepsis positives can be identified,
Abbott’s Maya Virtual Assistant: Abbott India became the first pharmaceutical company in the country to deploy an AI Virtual Assistant for its sales force in early 2018. The pilot project involved some 3,000 sales employees engaging with the personal assistant and support BOT Maya to get them ready for the day. Maya communicates with users in a simple natural language via voice or text, and can assist them with fetching information, handling FAQs, completing administrative tasks, receiving reports, and tapping into an enterprise knowledge base like SalesForce or Tableau.
Novartis’s Research Platform for Pathology: Pathologists and data scientists from Novartis partnered with tech startup PathAI to train an AI platform that is able to assist in digital pathology tasks. Users can examine pathology images with “machine eyes” to capture hidden, subtle or complex but informative patterns that are difficult for human pathologists to discern. To train a model to distinguish cell types, PathAI broke down each training slide into about 10,000 smaller sub-images which a team of consulting pathologists labeled.
Medtronic’s Guardian Connect Continuous Glucose Monitoring (CGM) System: Medtronic has launched the Guardian Connect CGM system, which uses an inserted miniature sensor to conduct 24/7 monitoring of glucose levels in the interstitial fluid under the skin. Data can be directly sent to a smartphone via a wireless transmitter. Users can view the latest glucose readings, examine historical trends, track daily activities that may affect glucose levels, and receive low or high threshold glucose value alerts.
VI. AI Trends and Challenges in Fortune Global 500 Healthcare Companies
1. Patients receive smarter management for chronic diseases outside the hospital.
2. Process management in hospitals becomes more efficient.
3. Medical robots become more widely deployed with more applications.
4. AI penetrates the entire drug R&D process.
1. Concerns remain regarding the disclosure of confidential or private personal healthcare data.
2. Information silos are an a issue with healthcare institutions.
3. It is difficult to perform quality control on medical image data labeling and annotation.
4. The accuracy of computer-aided diagnosis must continue to improve.
Source: Synced China
Localization: Tingting Cao | Editor: Michael Sarazen