Advanced analytics is coming to healthcare in a big way. Is your organization ready?

According to a recent Forrester survey1, 90 percent of healthcare organizations have implemented analytics and/or plan to do so in the next 12 months. As the use of advanced analytics technique like machine learning, artificial intelligence and predictive analytics becomes more widespread across the industry, are you ready to make the most of it in your organization? This piece considers some of the key considerations to bear in mind as you get started.

Big data has come to healthcare in a big way. Over the past several years hospitals have adopted a variety of digital systems like electronic health records and PACS imaging systems that are generating large amounts of structured and unstructured data. In many cases the penetration of these systems into the market is over 90 percent2. The cost of these implementations can be massive, with both Mayo Clinic and Partners Healthcare estimating that their Epic EHR implementation costs at close to $1 billion3.

One question that inevitably arises is how to extract the maximum value from these investments? A key vector for value creation lies in extracting operational, financial, and clinical insights from the data that these systems capture. To achieve it and build the foundations for future success, it is important to know how to use the data to generate new revenue streams, improve operational efficiencies, or improve clinical outcomes.

However, executing on an analytics strategy to capitalize on this data is often not straightforward. It’s telling that in a recent survey, 94 percent of hospitals4 said they're not capturing the information necessary for population health analytics. We’ll see in this article that there are several common challenges to implementing a sustainable advanced analytics program in healthcare, but that none of them are insurmountable.  

Advanced analytics shows value across clinical, operational, and financial use cases

There are many use cases for advanced analytics in healthcare. Precision medicine, the application of genomic data to define tailored care plans, leverages analytics across massive data sets to improve cancer treatment outcomes. Hospital operations teams are using advanced analytics to forecast patient admission rates and streamline patient flows through the hospital to decrease length of stay and increase revenue. Cybersecurity teams are analyzing real-time data flows in and out of hospital data centers to detect and protect against data breaches. 

There are many more examples that touch nearly every function in a healthcare organization. Despite the breadth of use cases, the challenges with implementing advanced analytics are often similar. Below we’ll discuss three of these challenges in more detail: data strategy, use case selection, and culture.

Strategy: Overcoming the hurdles to enable advanced analytics

One of the key challenges with building a data strategy to support advanced analytics is that the work often happens retrospectively, after the key systems that are generating data have been implemented. A good practice is to make data strategy a part of your enterprise’s process for implementing new digital systems.

For many organizations, getting a grasp of all the potential data available for analysis can feel overwhelming. Where are all the data silos located and who owns them? What structured and unstructured data is available? How can this data be made available consistently and reliably for analytics? Mapping out all the relevant data sources and establishing data governance processes is the first step. If multiple sources generate the same data point, which is most accurate? What are the common errors and omissions?

The next stage is integrating the data into a centralized environment like a data lake with a suitable analytics toolkit. It’s important to make sure your data and analytics environments are flexible. Over time you may find that your team needs to use new frameworks, and if they are successful then the list of data sources you’ll want to incorporate will surely grow to improve upon existing models or build new models.

Underpinning all this must be a robust data security strategy that accommodates integration of clinical data sources into a unified analytics environment while providing requisite access controls and cybersecurity protections.

Use Cases: Knowing what you need to get the best answers

While the mechanics described above are important, use case selection can make or break an analytics program, especially in the early days. Experienced analytics leaders will work closely with line of business and clinical leaders to vet a mixture of short- and long-term analytics use cases that can be achieved with the data available (or data that can be acquired) and with a clear path to implementation. Analytics teams must work closely with stakeholders and end users to balance the technical capabilities of the analytical approach with the real-world realities of healthcare.

Use cases are evaluated against a well-defined and mutually established set of key performance indicators (KPIs) to determine whether the project was a success. Over time, a better-integrated data infrastructure will emerge as more and more relevant use cases drive integration with business and clinical workflows. 

Culture: Bringing the organization with you

Finally, there is the issue of skills and culture. Introducing analytics as part of the clinical workflow can be unsettling for staff that are used to doing their jobs in a certain way. It is therefore essential to have buy in for analytics initiatives at an executive level, to provide you with the funding and direction needed to get projects off the ground. At the same time, ensure staff that will be required to integrate analytics insights into their daily work are educated on how to do this and shown the value these insights can bring to them and their patients.

The benefits of introducing advanced analytics into an organization can be far reaching. While a complex undertaking, advanced analytics can further clinical, operational, and financial aims. The good news is that whatever data you already have available, and whatever reporting you’re already doing, this can act as your foundation for more integrated data collection and analytics. Getting there is a journey but one you can start today, with the help of your IT team.

To find out much more about how to make your data work harder for you, your staff and your patients, read Intel’s latest white paper.