If you work in the healthcare industry, you can’t swing a cat these days without hitting someone proselytizing about the importance of data. Big data, small data, data silos, data warehouses, data lakes, data-driven organizations. It’s all about data.
The processes and sources that generate raw data can sometimes be opaque and hidden from end-users behind layers of cleaning and standardization. Did the data come from a survey? An EMR? Who’s authorized to enter data into the system? What does the data entry form look like? Is there a drop-down menu or does the software allow users to enter free text answers? All of these questions can illuminate potential potholes for stakeholders seeking to leverage data. Putting blind faith in “data” without understanding where it came from ignores the messy, chaotic reality of the world; and in healthcare, it ignores the hectic day-to-day operations and goings-on within modern hospitals.
Let’s take electronic medical records (EMRs) as an example. Most hospitals have customized EMRs. If a patient goes to two different hospitals with the same principal complaint, we’d have every reason to believe that the data for those two visits would not look exactly the same. In one instance, a physician might pick from a pre-filtered list of common diagnosis codes; in the other, the physician could search for a more specific and descriptive ICD code. Perhaps one physician enters “abdominal pain” as the complaint, while another is short on time and simply shortens it to “abd pain”. And the more hospital IT systems involved, the more complex the problem of reconciling data becomes.
Hospital IQ processes data from many different data sources, even within a single customer; it would be easy to simply take in the raw data, generate some reports, and call it a day. But we gain valuable insight into the nuances of our clients’ data by engaging with stakeholders, especially clinical informaticists and other front-line staff, at multiple points in the deployment:
- Discovery – Discussion with stakeholders helps Hospital IQ understand the goals of a particular client. Some may just want to replace existing reporting; others might hope to leverage data for operational initiatives such as improving the management of observation patients or decreasing length of stay.
- Data extraction – Existing hospital reports may run off of several different data sources. Systems analysts and clinical informaticists are familiar with the systems and know not only which sources can be trusted, but also which fields within the data to be wary of. Conversations with stakeholders can help tease out, for example, if physicians don’t enter a preliminary or working DRG for admitted patients.
- Data validation – With their knowledge of the hospital and its operations, staff stakeholders often know where to poke and prod the data to see if it holds up. They know what values certain metrics should have and will know if something smells fishy.
- Configuration – Every hospital has different internal policies that need to be reflected in any reporting. By engaging stakeholders, Hospital IQ can ensure the platform is configured in order to ensure the smoothest user experience.
Without context, data can easily be misinterpreted or misrepresented. Combining local knowledge with centralized analytics, therefore, is a critical step to ensure that hospitals get the most out of their data.