By Lisa English, Hayes Management Consulting
Twitter: @HayesManagement
Twitter: @lisasoutrunning
Yes, the terms “big data” and “analytics” are buzzwords, but they clearly highlight a shift toward data-driven decision-making with a real measurable impact on outcomes in many different industries. Savvy digital marketers now mine your digital breadcrumb trail to offer you more of what you like and attempt to discern what you need before you are aware of it yourself. This not only drives sales, but also actually helps consumers – if they aren’t “creeped out” by the “Big Brother is watching” when you post on social media then immediately see ads pick up on a word from your post.
In the healthcare industry, the move to population health is just one obvious application for sophisticated analytics. As we appropriately engage our best and brightest in solving the core healthcare issues of our society, we find the key questions that analytics can help answer: What treatments drive positive health outcomes for patients? How can we curb wasteful, ineffective healthcare? Ultimately, why does the US spend more on healthcare than all developed nations while getting only mediocre healthcare outcomes when looking at our population as a whole?
We need to find answers to these big picture questions, but meanwhile, many others in the healthcare industry have problems to solve in their own, smaller domains. They struggle with how to drive changes today while waiting for the statistics PhD students to turn their attention to the types of operational improvements those in the trenches deal with every day. Analytics provides a platform to monitor multiple areas of risks or interest, support continuous risk assessment and monitoring, and complement limited compliance resources – day-to-day issues organizations face every day.
While it’s clear analytics can play a significant role in solving these tactical issues, there is still much to learn. This post is the first of a two-part miniseries that will explain a side of healthcare analytics that you may not have known before. Here are eight things about healthcare analytics that may surprise you and some tips to get you started, taken from “aha!” moments our customers have experienced in their own analytics journey.
1. Few healthcare organizations have analytics you can get your hands on
Healthcare organizations with large, expensive enterprise wide analytics are often focused on expansive initiatives like population health. We can all agree that improving the health of our patients is the primary goal, but we still need to leverage data for other operational goals like improving revenue flow and minimizing risk.
We often see compliance leadership left hungering for analytics that helps them to focus on their key risk areas. The information you need exists, but it isn’t helpful if you can’t get your hands on it. Impediments may include IT group bandwidth and lack of expertise with complicated analytics tools. Actionable analytics requires iterative refinement of available data with subject matter experts who can pull out the “learnings” that are real and worthwhile.
2. Analytics helps, even in areas you think you’ve got under control
Effective organizations know that to solve business problems you have to build operational solutions. You no doubt spend much of your energy reviewing root causes of problems and fixing them properly.
This is certainly true for effective healthcare compliance groups. You can never audit your way out of improper billing. You need to diagnose and fix the problem at the core. But how do you know you’ve been successful?
One large healthcare organization had a rude surprise when they implemented risk-tracking analytics. Kwashiorkor (diagnosis code E40) has been a target in the OIG annual work plan for several years. Kwashiorkor is a form of malnutrition that most often affects children in developing regions of the world where there is famine or a limited food supply and is rarely the proper code for patients in the US. Overuse of this specific code (versus the code for other more general forms of protein deficiency) was driving up Medicare DRG reimbursement for patients due to overstating their level of “complications and comorbidities.”
The compliance officer was aware of this risk and had instituted training on proper usage as well as operational checks to prevent inappropriate use of this code. She was confident that a claim could not leave her organization with this diagnosis code.
On implementing a KPI dashboard of curated analytics risks, the first things that jumped out as she viewed her data dashboard were several instances of the code being used over the past year. Her subsequent investigation uncovered a gap in the process related to interfaces with an external system. Properly designed analytics should monitor a broad array of issues on an ongoing basis, including known problems as well as current and emerging risks.
3. Reports are not enough
Every quarter, hospitals receive PEPPER reports from the CMS comparing them to other US hospitals of the same type. The report highlights hospital metrics where the hospital is above the 80th percentile or under the 20th percentile in several key measures. Hospitals flagging high constitute a risk for overbilling, and may trigger CMS audits to verify if the billing levels are appropriate.
It is likely that some group in your organization is reviewing these reports to understand what is driving your organization to be well out of the norm and is tasked to either fix problems or be prepared to defend your billing in the case of an OIG audit. One hospital was above the 80th percentile for a cardiac related DRG code. Their Clinical Documentation Improvement group was monitoring the reports and had been working with and educating their cardiology physicians.
Their compliance officer obtained a compliance dashboard that included some PEPPER related metrics. Wondering if it might be useful to the CDI group who monitored metrics, she showed them the dashboard. As they drilled into the details underlying this PEPPER metric, they immediately realized that the providers driving their outlier status were not the cardiologists.
Actionable analytics requires providing the ability to drill into the details in order to correctly diagnose and respond to the insights gained. Detailed PEPPER analytics based on the prior month’s data were immensely helpful in diagnosing the problem, and allowed for timely monitoring of the effectiveness of any mitigation efforts. When compared to static reports on two-year-old data, there really is no comparison.
4. Look at things sideways
Some problems are just hard to see via traditional operational reports. You can often find that looking at measurement metrics from a different angle can give you fast actionable insight.
Recently we worked with a group of customers to measure critical care time. One indicator that critical care time may be overbilled is a provider who seems to be billing an unfeasibly large number of critical care hours in one 24-hour work day. This may flag your providers as outliers for Medicare audits.
It is not practical to monitor this via traditional claim scrubbing or random-pull auditing methods. However, by looking at your data organized by provider and service date and by applying estimated minutes to each billed code (per coding guidelines), you can quickly identify providers who may have risky or borderline billing patterns in this area. Generally, organizations find a few providers who flag as a risk and they can easily follow up with a detailed audit for the provider and the specific dates of services.
This article was originally published on Hayes Management Consulting and is republished here with permission. The second post of this miniseries can be seen here.