When you hear the word “analytics,” what’s the first thing that comes to mind?
If not, you’re actually in the minority. Despite the prevalence of analytics in nearly every field and industry today, the term itself continues to be commonly (and incorrectly) defined as “pre-packaged software that provides a nice display of information you already have.”
You know – a dashboard.
Of course, that’s a tremendous underselling of what analytics actually is. But it also goes a long way toward explaining why the buying process for analytics software can be challenging, especially if the purchasing decision requires consensus. Does everyone in the group agree on what you’re actually buying? More importantly, what are you actually buying?
A good place to start is to get everyone on the same page. Analytics is the systematic computational analysis of data or statistics (if you Google it). It’s both the action and the outcome of applying computer science, mathematics, and artificial intelligence to understand and extract actionable information from data.
Analytics is a powerful form of technology that can truly drive your business forward. If you help others better understand what it is and what it can do, they’ll see the value too – and you’ll be positioned to prove the ongoing value of your investment successfully.
Leveraging the right type of analytics at the right time – and demonstrating ROI
Let’s say you’re considering an incremental investment in analytics capabilities. How do you decide whether or not it’s worth it? How do you demonstrate ongoing value and build a compelling ROI narrative?
First, distinguish between the type of analytics you’re employing for different purposes. Analytics can be:
- Descriptive – uncovering what happened and helping you understand why.
- Predictive – leveraging the understanding of what happened to forecast what’s most likely (or unlikely) to occur in similar instances in the future.
- Prescriptive – defining actionable paths that, if followed, create the greatest likelihood of a desired outcome (or the smallest likelihood of an undesired outcome), and making recommendations based on those insights.
Realistically, you’ll need to employ some combination of all three – so part of the value of your investment lies in developing the foundation of these capabilities.
The bulk of your ROI, however, will come from making analytics actionable in a meaningful way.
Making analytics actionable
Let’s start with an example based on core KPIs. For instance, let’s use HFMA’s MAP Keys – things you should absolutely be measuring and monitoring, like aged AR as a percentage of billed AR.1
A dashboard might be able to display this metric and even tell you your fastest and slowest payers. While this is great information, it’s something you could likely get by asking the right person within your organization. And it’s not inherently actionable.
What makes analytics actionable is when it accomplishes something beyond, in effect, asking that right person for the basic information you need. This is achieved through automated analysis traveling upstream and downstream in the data to pinpoint what drove those days in AR and what the full, long-term impact was (descriptive analytics).
Analytics also identifies the right points of intervention, so you can target where you expend resources to address the underlying causes. The truth is, there are often lots of ways to solve a problem, and some are better than others. It all depends on the combination of the relevant elements at any given time.
Think of it like a Rubik’s Cube. The steps to solve the cube most efficiently will depend on how the individual tiles are arranged at the time you begin attempting to solve it. You might need to start from the beginning, and follow the rudimentary rotations that lay the foundation for solving subsequent layers of the cube. Or those layers might already be solved, in which case starting over would actually be counterproductive (and would lead to unnecessary rework).
Knowing the steps you could potentially take to solve the puzzle is actionable information, but it doesn’t become meaningful until you add the insight of which steps are most likely necessary or of greatest value in that instance (predictive analytics) as well as a more granular recommendation on which of those specific steps to follow (prescriptive analytics).
Hey, that’s a great segue to the next section.
Making analytics actionable and meaningful
So let’s say you pinpoint every single thing that could possibly be improved in your organization – down to the cent and down to the minute.
Then what? Hire more people? Ask everyone to work more hours? Lock yourself in the boardroom and refuse to come out until things are better?
Not very realistic or productive options – and unlikely to have any positive, meaningful impact on your operations.
The right analytics software, however, can turn this potentially overwhelming wealth of opportunities for improvement into a triaged, continually self-refining system. Software automation can accomplish the legwork of understanding which of all these dozens or hundreds (or thousands) of things actually deserve your attention. An analytics engine can identify ways to not have to bring as many of these things to your attention – figuring out single fixes that have significant aggregate benefit because they address and prevent similar instances of a common problem.
At the day-to-day level, analytics shows you where to focus your training and can help you prioritize staff time and other resources. It enables you to understand the details of your revenue cycle and AR days, and see the historical and near real-time differences between individual actors: which payers have the longest turnaround times for claims, for example, and which coders have the highest denial or rejection rate. Analytics empowers you to catch and address problems in your reimbursements before they become entrenched and erode a significant slice of your revenue.
Returning to the example of MAP Keys (or KPIs like them), let’s distinguish between actionable and actionable and meaningful:
- Understanding the makeup of your denial volume is important – by payer, plan, procedure, and so forth.
- This insight is inherently actionable, as you can see which teams or which types of claims or which payers need special attention.
- But it only becomes meaningful when it is further analyzed to understand long-term financial impact, impact on patient care/patient satisfaction, and overall value of reworking/resubmitting.
- And the most meaningful outcome would be to pinpoint the causes of the denials and implement the most efficient upstream change to documentation, coding, eligibility verification, etc.
- Days in AR, a very important metric, but also one that can be tough to parse.
- Let’s say, for a given payer, you typically get paid within 35 days. Is that 30 days with 5 days to get it reconciled in the system? Or is it 20 days plus 15 days to get it into the system?
- Once you have this information and other similar insights into AR performance, how evident are the right solutions? What are the real sources of the delay(s), and how can they be mitigated?
- Write-offs, bad debt, and unrealized revenue.
- Those all sound fun.
- Not really, but how do you go about managing them down when you’ve already been trying to address them for years (in some cases, decades)?
- You find new solutions. “Meaningful” in this case means the identification of novel approaches or previously unknown pockets of opportunity.
Analytics becomes meaningful when it identifies previously hidden or unknown correlations and connections across your data. Finding these patterns enables you to target your effort.
As a real-world example, it’s the difference between rewarding collection activity based on simplistic metrics (number of touches, dollar amount owed) – and targeting it based on true probability of collection and total overall cost to collect.
Laying the groundwork for meaningful, actionable analytics
The capabilities you need to support and maximize the value of these forms of analytics include:
- Data aggregation and normalization – the cleaner and more comprehensive your data is, the more meaningful and accurate the insight that can be gleaned from it;
- Interoperability and direct tie-in with workflow software – the more seamless the integration with your other systems, the more efficiently actionable the insights become; and
- Access to relevant third-party data for model validation – this is particularly important for predictive and prescriptive analytics, because the models used in these forms of analytics may need to look to de-identified third-party data to understand rare instances (or to get a head start on understanding new types of instances when you add new service lines, begin to serve different patient populations, or operate under new reimbursement models).
Keep all of this in mind the next time you’re discussing analytics, especially if the colleagues you’re discussing it with are a little fuzzy on what analytics actually means. Helping them better understand will help your entire organization, because you’ll make the right investment and reap all of the benefits that come with it.
You’ll be able to spotlight what happened, why it happened, and ways to prevent (or recreate) it in the future. The proof of your success will become evident as that future becomes the present, and then the past. Descriptive analytics will enable you to demonstrate the impact of changes you implement. Predictive analytics can be leveraged to show the granular or aggregate likelihood of undesirable outcomes decreasing over time (and what the associated long-term impact is likely to be). And prescriptive analytics will enable an ongoing, continuously self-refining roadmap of how you can further improve processes, even as you continue to make progress toward your goals.