By Syed Haider, senior architect, X by 2
In the health insurance space, a world of resources and effort has been expended in recent years on the premise that the more data, the better, at times without regard for how to use the data. Unfortunately, many insurers who make that journey too often find they are little better off than when they started.
The more-data-the-better focus usually results in acquiring the kinds of solution provider tools that can help health insurers aggregate and normalize years, and sometimes decades, worth of data. Once in a repository, the data is married with front-end tools that give business decision makers and consumers easier and more effective access to it.
It all works well on paper (pun intended), but often not in reality. That’s because the health insurers have either shortcut, mismanaged or even skipped the key to success in any information improvement effort — addressing their data credibility issues before they start their data improvement journey.
After all, this is the hard part that the main solution providers and even consultants cannot help you with. Why? Because of the simple fact that the people who understand the current state of any insurer’s data — good, bad, and ugly — work for the insurer, and are often buried in the deepest recesses of the company.
They are the ones who have survived most of the root problems, such as poor data quality, including legacy systems that have multiple revisions layered on them, the fragmentation of business rules over time, proprietary data stores with poor edits, and many more. All of these causes produce the effect of a slow ride down the data degradation road, to the point that the consumers of the data — inside and outside the insurer — just don’t believe it anymore.
No amount of fancy tool-filtering or slick user interfaces will fix this. It has to be addressed holistically and collaboratively from within the insurer.
There are several ways to approach this, but the best and most direct way is old-fashioned — creating a dedicated team of business and IT subject matter experts who have the domain knowledge and authority to build the new data governance model for the company.
More often than not, that’s the rub for most insurers, and why so many of these data transformation efforts never seem to deliver to their full potential. However, that doesn’t mean it’s not the right thing to do, as difficult as it is.
First and foremost, insurers must decide how important their data journey is to them, and then they must dedicate resources and investments accordingly. If, as with most insurers, data transformation is one of the highest priorities, then there must be an agreement to fully dedicate the appropriate IT and business resources to the endeavor. That’s a difficult thing to do in the best of times, but it’s absolutely critical to the long-term success of the transformational effort.
Once that is settled, the next step is to create the new data governance model for the company, which is generally where the real fun begins.
The task is a basic, but bedeviling one: identify the operational business definitions for the company. Among other things, that means being able to address and answer fundamental questions like: How is the data defined? What data is more important for the long term? What data is less useful? What data does the insurer need that it does not currently have? How will the insurer go about acquiring the desired data? And of course, the million-dollar (or billion, depending on the insurer’s size), how will data be made actionable?
Answering these questions requires the data governance group to reach out to all parts of the organization to solicit input and time, and to continually extol the benefits to the organization of this process. People need to understand why it’s so important.
However, many companies stumble at this point due to the time pressures of daily operations, the reluctance of some people to give up what they know about the current data and even cultural issues around cross-departmental and divisional collaboration. It is also critically important that the data governance group has executive support and the authority to make the hard decisions about what data stays and what data goes.
Too often with insurers, these efforts get bogged down over what a particular data element really means, and to avoid any confrontation, the data element gets accepted into the data model with multiple definitions. This should be heresy to any data governance group worth its salt. Allowing such things is not data governance; it’s simply setting the insurer up for suboptimal results from its data transformation effort.
People will naturally have different opinions about how to best define something, but the necessary effort has to be put into resolving these differences in a way that optimizes the outcomes. One example of how to resolve this is to invite the data tribal elders — those subject matter experts that have lived and breathed the data since time immemorial — into the process, making them accountable for educating the group on what they know and what they think should be done.
Insurers must also deal with the hard data issues that cannot be rehabilitated. This is typically data that has been modified several times from its original form, and the source system that produced it no longer exists in the insurer. It can also involve dealing with the people who support and consume that data, and having difficult conversations about putting the data out of its misery and moving forward with new formats and supporting systems.
To paraphrase a famous Italian Renaissance political philosopher, when it comes to the quest for data credibility and quality, the end justifies the means. The potential benefits to any health insurer are so great — process effectiveness, customer insight and intimacy, the veracity of information, foundation for real business intelligence and the golden ring of truly actionable data — that it is worth a little organizational discomfort to get there.
The alternative to not making the effort is something that most health insurers should avoid at all costs, if necessary.
About The Author
Syed Haider is a senior architect for X by 2, a technology consultancy focused on the practice of architecture for the insurance and healthcare industries based in Metro Detroit.