How to select the right health data for your use case: 3 key considerations

Recently, I had the pleasure of sitting on a panel during the 40th Annual Conference  hosted by the National Association of Health Data Organizations (NAHDO). While my fellow panelists and the moderator explored different health data use cases, three themes surfaced time and again in our discussion and other conference presentations.

If your organization is undertaking a data project in the New Year, here are three crucial considerations before you dig in.
 

Consideration 1: Where will you get the data?

Data sets can come from public or private sources, each with pros and cons.

Public sources like All-Payer Claims Databases (APCDs) provide comprehensive, publicly available data — ideal for projects analyzing trends across all payers at a summary level.

Private sources, such as your electronic health record or claims data, are better suited for custom improvement projects.

Other projects, such as those requiring benchmarks and comparisons but also actionable insights on specific patients, may benefit from data assembled from multiple sources. This can be challenging depending on whether the data is already collected and accessible.

At Solventum, we rely heavily on claims data for running data enhancement fields and researching improvement opportunities. This approach works well because the data is already collected and doesn’t require additional lift. However, adding other data points could unlock deeper analysis.
 

Consideration 2: How granular will you get?

Too much data can be as daunting as too little. Think about what data will help you achieve your purpose and meet your project timeline.

Granular data points can bring insight but take time to interpret.

Some projects span years, with subsequent phases extending even longer.

When designing your project, draw on data that is present and easy to access. Consider aggregating clinically similar procedures, patients or people to identify patterns at a higher level. While individual occurrences matter for care management, aggregate patterns and variation within cohorts can reveal opportunities in claiming, process or strategy.
 

Consideration 3: How will you maintain data consistency?

Consistency is critical for meaningful comparisons. You cannot compare services, quality, costs, or outcomes if the instances being compared are disparate.

For example, “Facility A” performing high-acuity procedures on patients with severe illness cannot be compared solely on complication rates to “Facility B,” which handles lower-acuity cases and transfers complex patients elsewhere. The complexity of service and illness burden can skew results — Facility A may have unavoidable complications even when minimizing poor outcomes.
 

Why it matters

Data-based projects vary in design, but the ultimate goal is to uncover opportunities for improvement, whether that means enhancing quality, reducing costs or identifying care gaps. By carefully considering where your data comes from, how granular it needs to be, and how to maintain consistency, you’ll set your project up for meaningful, actionable insights.

For organizations focused on reducing complications and improving outcomes in ambulatory care, methodologies like Ambulatory Potentially Preventable Complications (AM-PPCs) can help identify performance gaps and guide quality improvement efforts.

Shannon Garrison, MBA, MJ, is a health policy manager, clinical and economic research, at Solventum.