Skip to main content

Pitfalls of Analytical Product Development and How to Escape Them

Our health care analysts build data-driven products (dashboards, reports, etc.), and they think through all of the technical implementation steps required to make these products successful. The next step determines the success of the product: pinpointing and avoiding the potential pitfalls that can undermine its usefulness. These pitfalls include:
  • Failure to understand what we really mean by "business intelligence"
  • Poor understanding of the users of our product and their needs
  • Poor data management

Often data sources look like a dangerous cocktail of social determinants of health coupled with genetic, environmental and clinical data with other information thrown in. Finding a meaningful way to manage these data and capitalize on the value of the information can be challenging.

Let’s look at the end user of our analytical products – the provider. The volume, variety and velocity of available information can far exceed any professional’s abilities to process and interpret. For example, our Partnership to Advance Tribal Health (PATH) participating hospitals are bombarded and confused by multiple layers of mandatory reporting and dashboards provided by local area offices, their Medicare Quality Improvement Networks, their Hospital Improvement Innovation Network organization, tribal epidemiology centers, state departments of health and many more organizations.

A recent experience with a PATH hospital illustrates some pitfalls. They shared all of their reports with us and stated that they were perfect and there were no areas for quality improvement plans because they were performing their best. The team was asked to help identify any opportunities in this facility’s reporting. Guess what: we found one pretty easily. The provider had reported an unbelievably small number of ER visits and readmissions, despite their large facility size and occupied bed numbers. Digging a bit deeper into their reports, we found that quite a few patients left their ER without being seen. A financial dashboard reported a different number of readmissions, one that was more appropriate for the scale of this facility. Through some conference calls we helped them identify and focus their quality improvement efforts in areas where their performance showed less than desirable quality. This is just one example of how we can assist our providers to critically process information they receive and identify areas that require their immediate attention.

This experience led me to the idea, though maybe not a new one, of what can we do to assist our customers in this era of “dashboard mania” in which they feel hopeless to navigate. We need to work with them to develop a data navigation strategy to allow them to understand the data that their organization collects themselves, couple it with what they receive from their partners and government agencies, and align it all into something manageable.

In this type of data strategy, identifying trends is key: prepare to rely more heavily on layering data and pair the right data together in order to understand where and how care is being delivered and the nature of any care gaps that exist across networks. This positions providers to target and improve overall care delivery and quality. We need to offer them predictive and prescriptive analytics to transform data into impactful intelligence. Providers need to not only understand what happened, but also predict what will happen in the future and the best strategy to reach desired outcomes and drive quality improvements. And HealthInsight is uniquely positioned to help providers achieve success through data navigation strategy.

Add new comment