The pandemic has been a catalyst for increased artificial intelligence (AI) adoption within the healthcare industry. In fact, AI helped to reveal patterns that led to the identification of the novel coronavirus, now known as COVID-19. A cluster of unusual pneumonia cases in Wuhan, China were detected using AI, specifically natural language processing and machine learning (ML). These cases were tracked, located, and reported to various healthcare organizations and would soon mark the start of today’s public health crisis.
Now AI is being used for screening, faster diagnosis, and even more use cases to come. Despite the tremendous opportunities that come with AI adoption, there are barriers. Specifically, barriers that pertain to data management and that must be overcome by healthcare organizations.
Challenge #1. Siloed, Multiple Types of Data Streaming at Different Velocities
Healthcare organizations manage mass quantities and types of data that are usually stored in disparate systems spread across each organization. According to Qlik’s Ritu Jain, “Depending on the source and type of data used to train the model, results [from data] can be skewed. To get the most accurate insights from AI and ML, you need a single, unified repository to store all relevant data.” This is where data lakes can serve as a central hub for mass amounts of data, helping to make AI adoption achievable via accessible, yet secure, data management.
Challenge #2. Raw, Unrefined Data, Without Consistent Metadata
In order for machine learning models to be successful, they must be fed with relevant, up-to-date, and analytics-ready data. Where a data lake serves as the single source for untransformed and raw information, “data with no tagging, or common descriptions explaining what it means, can’t be used for ML,” explained Jain. “It lacks the markers on what it is supposed to teach the model. Further, standardizing, formatting, and refining raw data to get it in a consumption-ready state can be time-consuming and code-intensive, requiring specialized skills.” This is where a management solution that accelerates the delivery of actionable data is crucial.
To overcome the challenges associated with AI adoption and implementation healthcare organizations need the right infrastructure and management. With this is in place, mass amounts of data can be stored and transformed into analytics-ready information so that AI and ML initiatives can uncover insights and yield powerful results.
The adoption of AI and ML will continue to grow as the world works to combat the pandemic and other health crises to come. With barriers including mass amounts of data and the need to turn raw data to information that’s analytics-ready, adoption will be challenged. The key is for those in the healthcare industry to invest in the right infrastructure and management. In doing so, data is made accessible and usable to solve some of the world’s biggest problems including today’s public health crisis.
Read about the third challenge to AI adoption here.