Readers Write: Little Data

Little Data
By Greg Park

Today’s topic is the methods by which employees and partners obtain information to perform their function, or as I like to call this, the Push vs. Pull method of information dissemination.

Let’s step inside the way-back machine to observe how this was accomplished in 1986.

I was cutting my HIT teeth first as a computer operator and then as “do-everything” guy at a mid-sized hospital in Philadelphia. There were few standards and no one with HIT degrees. You learned from the vendors, by reading technical manuals, and by putting out fires.

My first big hospital project was implementing Shared Medical Systems Spirit platform, which was SMS’s first turn-key mini platform. This led to managing other tertiary clinical applications and methods of creating information.

Each new platform we installed was bundled with standardized reports focused on daily activities, DNFB accounts, and payments by patient type. All reports were QA’d by vendor and staff prior to go-live to certify accuracy. These standard out-of-the-box reports were the lifeblood of staff’s workflows, and believe me, I would hear it if they weren’t delivered by 6:00 a.m.

Reporting exploded as data became richer and tools emerged to create specialized ad-hoc reports. Soon we were creating ER patient flow analysis, profitability by attending physician, nurse staffing by patient acuity, and then linking data between disparate platforms.

Life was good, but we were killing thousands of trees each month. End users were happy because the process of creating these ad-hoc reports was very personal. I would sit with end users to analyze their needs against the data collected. When required, we would add new data fields and workflows to collect that information accurately in our various platforms. Finally we would validate output structures, ensure accuracy, and finally schedule the reports for the desired timeframe.

Each day or week or month, the report would print. As time marched along into the 1990s, we downloaded all reports into our content management platform. Now all reports were readily accessible and audit trails let us know exactly who was (or was not) reviewing their reports.

My formative years were spent in this way understanding need, locating data, and constructing formats to enable user workflows. These were the “Push” years, because it was our responsibility to ensure the report, spreadsheet, or database was created and pushed to users and business partners in a timely way.

Somewhere in the mid-1990s it became clear that something was changing. During a PeopleSoft implementation, I noted that disk space and CPU resources were significantly more robust than the platforms I had encountered. Data could be kept almost indefinitely and sophisticated queries could be run in real time rather than waiting for a day-end process. The writing was on the wall, but the major HIT platforms would take years to catch up. Many hospitals still operate like its 1985, some with basically the same HIT platforms.

The rest of this writing involves relaying the conversations I have had with my customers, so if this is not your experience, please chime in.

Most new EMRs are fashioned with limited standard reports. End users have become the focal point in the process of generating all reports. Generating reports and pushing them to their intended audience has become an anachronism. Let’s call this new method the “Pull” method, because users are expected to pull the data themselves. In this Pull method, users access report writing tools and pull data on their own terms as needed. To the system designers, this must have been intended as empowerment, but for many users it is a speed bump that did not exist before.

This Pull method can be implemented in various ways. In some, the IT department creates general templates for users to enter query parameters. This works fairly well, but does not address those reporting situations where the timing of data generation is critical. The report will look different if run on August 31 as opposed to September 10, and sometimes this is a problem.

Another method is to provide the user with a full query interface. Now end users are playing the role of IT analyst. Maybe this is fair and is a reflection of how IT is part of everyone’s jobs today, but it can be problematic when users select the wrong data or create a query that is particularly taxing to the EMR or its reporting database. In this world, we have a real problem of focusing on the wrong data, or worst yet, not focusing on the data at all.

For the moment (because I know that once HIT is exposed to big data this will change) patient accounting and general finance are your biggest data consumers. From my perspective, they seem particularly annoyed with this new Pull mentality. Many of these users access these new systems exclusively to generate their own reports.

For me, life is always about balance, and I think both the Pull and Push methods have their place. Pull methods are fine in some scenarios, assuming your end users know how to data mine and construct data exports. But when the situation is time sensitive, I want tools that Push information to responsible parties. This means end users are immediately notified the information exists, that it is readily accessible, and that they are expected to review it immediately.

I know this topic is not at the top of your mind considering MU, ICD-10, and all of your other requirements, but think about it during some downtime and consider whether you are doing your best to get the right information to the right people at the right time.

Greg Park is director of enterprise solutions for Dbtech of Edison, NJ.