Healthcare News & Insights

How to make big data work for your next project: 4 strategies

Projects that analyze big data give hospitals unique opportunities to explore problems and develop effective, efficient solutions. But before your facility begins its next big data project, try these best practices. 

480631681 (1)Many organizations are looking to trying to analyze and utilize large amounts of information and data to improve care as well as improve the efficiency of their care delivery. Unfortunately, many agencies and facilities struggle with the big data learning curve.

Steven Escaravage and Joachim Roski have a good sense of what makes a big data project successful or not. They’ve worked on big data projects for a number of private healthcare organizations and systems, as well as for the National Institute of Health, the Centers for Disease Control and Prevention, and the Federal Drug and Administration.

They outline some common mistakes and best practices from their experiences with these groups in an article for the Health Affairs Blog.

1. Determine the correct data for your project

According to Escaravage and Roski, many organizations beginning projects default to using data that is easy to collect or already on-hand. They don’t analyze whether that data reflects the central issues the project is meant to fix. This may speed up the project, but it also can lower the value of the results.

Using data you’re already familiar with isn’t bad, but it limits the scope of your project since it doesn’t add any new perspectives for the problem. Brainstorm what other alternative data sources are available for your project. The authors have found that “ when organizations develop a ‘weighted data wish list’ and allocate their resources towards acquiring high-impact data sources … they discover greater returns on their big data investment.”

2. Make sure your initial pilots/data tests have wide applicability

Big data is most effective when it involves and pertains to a wide array of people. If your initial analytic projects are too self-contained, you limit the utility of the results.

The authors give an example of a government health agency they worked with, whose pilots/tests were centered on specific, complex and storage-intensive challenges. Though the project was able to resolve those initial, technical issues, it wasn’t able to provide the kind of results that would lead to larger, organizational challenges like the agency’s leaders wanted.

That means your big data projects will be more effective if they have a wide applicability. That way your facility leaders will be able to understand the information and see how big data can transform their operations.

3. Examine your data’s ‘provenance’ and ‘lineage’

Data mining without looking at where your data came from (its provenance), or how that data has been modified (its lineage), can give your projects misleading results. Be sure your team is scrutinizing new data, their sources and their patterns to make sure the data collections process isn’t “causing a significant effect in the outcome variable of interest.”

4. Focus on the problem, don’t start with the solution

It may seem counterintuitive, but sticking to one rigid goal for your big data project may actually limit its utility. Escaravage and Roski give the example of a federal  health agency that invested in big data, but used conventional approaches to analyze the information. The results didn’t provide many new insights for the agency’s decision-makers.

Instead, the authors recommend giving your big data subject-matter experts direct access to the information and free rein to develop their analytics and discover new or unexpected patterns.

Subscribe Today

Get the latest and greatest healthcare news and insights delivered to your inbox.