New technologies offering new ways to solve complicated and costly problems are entering the health care sector . Artificial intelligence (AI) is one of them. To succeed with AI, you need to think in an analytical manner, to build an analytical organisation and, as a part of that, you need an analytics platform.
Artificial intelligence makes it possible for machines to learn from experience, adjust to new input and perform humanlike tasks. Most AI examples that we hear about today, from chess-playing computers to self-driving cars, rely heavily on deep learning and natural language processing. Using these technologies, computers are trained to accomplish specific tasks by processing large amounts of data and recognising patterns.
AI in health care uses algorithms and software to approximate human cognition in the analysis of complex medical data. The primary aim is to analyse relationships between prevention or treatment techniques and patient outcomes.
Why do you need an analytics platform
Despite many investments in both innovative and legacy AI environments, most AI projects outside of the high-tech industry are stuck in the prototype phase. Rarely do they deliver the expected business value. The primary challenge is to operationalise AI applications and embed them into enterprise business processes. You need an analytics platform to succeed.
An analytics platform is a software foundation engineered to generate insights from your data in any computing environment. Built on a strategy of using analytical insights to drive business actions, this platform supports every phase of the analytics life cycle – from data, to discovery, to deployment.
With an analytics platform you can develop managed, governed AI applications that are scalable and have an integrated security model. In addition, these applications can be evolved using the support of an end-to-end analytics life cycle.
SAS customers who use the SAS® Platform to develop and deploy AI applications achieve crucial benefits:
- Availability of AI techniques such as deep learning and natural language processing.
- The ability to move AI applications from prototype, development and test all the way to production. In short, support for industrialisation of AI applications within the enterprise.
- Participation in new innovative ecosystems by leveraging open source through SAS APIs to augment and add capabilities such as data transformation, distributed machine learning and automated deployment.
How can artificial intelligence help health care
By using AI in an advanced analytic predictive model based on new data and historical behaviour, health care professionals receive diagnostic suggestions based on patterns from both statistical and real-time data.
Many organisations right now are devoting copious resources to develop algorithms based on AI. As I see it, we will see the real value from AI when we succeed in having the algorithms brought into a production environment. This means that the algorithm will bring value to one or many hospitals, and not only to the person who developed the algorithm. That is where many health care organisations are struggling, because they do not have the correct platform for this process.
Many organisations still struggle with AI. Find out what practical issues they face when implementing AI, including people and platforms, in the recent SAS study The Enterprise AI Promise: Path to Value.
If health care organisations decide to start their analytical journey, they need to understand the importance of having a visualisation of the data set so they understand where to start. Algorithms for more predictive outcomes can be added when an organisation knows what to focus on.
Increase the speed and quality of diagnoses
When patients are hospitalised, they expect to be diagnosed fast and treated for their exact disease. That is not the case at all hospitals today. The challenge is to improve the basis for increasing the speed and quality ofpatient diagnoses so hospitals can achieve the benefits that lead to patient advantages.
Detect cancer at an early stage using AI
It is a common wish that more cancer patients survive their diagnoses. If patients are diagnosed with cancer at an earlier stage, more patients can be cured and more lives might be saved.
New techniques for prediction can take up the challenge. AI makes it possible for machines to learn from experience, adjust to new input and perform humanlike tasks. For example by combining data from biochemical blood tests collected over the past 10 years with a technique from the analytical platform, we it would be possible to find out if an AI system could be trained to recognise patterns in new data for early cancer prediction.
Examples from the real world
Fighting hospital-acquired infections with AI
Karolinska University Hospital in Stockholm and Sygehus Lillebælt in Denmark are working together in developing AI algorithms to predict which admitted patients are in risk of getting a hospital-acquired infection. The vision is to develop models that predict the infection an average of five days earlier than doctors could.
The predictive models will be based on different data types: structured data from lab results, medications and diagnoses, and unstructured data from medical records describing the treatment of patients and description of X-rays.
Once preliminary tests are approved, the AI algorithms will be implemented at all hospitals in the Stockholm County Council and the Region of Southern Denmark. These algorithms could be the difference between life and death for hospital patients.
Soon, we will start doing AI on how to make a correct diagnosis in the shortest possible time to ensure correct treatment.