6 Critical Points to Consider When Evaluating Healthcare Data

Background

Healthcare payers, providers, and other value-based care organizations face challenging questions as they work to improve treatment performance and cost. How healthy are patient cohorts and panels in a particular market? What treatments are the most effective for the most prevalent conditions in a population? How widespread are preventative screenings and vaccinations?

Answering these questions requires access to the most significant tool in a healthcare decision maker’s arsenal: claims data. But that data can be incredibly difficult to utilize for effective decision-making thanks to two issues: its volume, and its quality.

A joint Dell EMC and IDC study recently estimated the total volume of global healthcare data to exceed 2,314 exabytes. That’s more than 23 quadrillion terabytes and should give some sense of the scale and challenge that healthcare data analysts face: it’s not easy to make data-driven healthcare decisions when you’re drowning in exabytes. Further, so much of that data is of poor quality.

Numerous studies posted by the National Institute of Health highlight the data quality challenges healthcare analysts face, and yet pressures to make decisions on this data are only increasing thanks to the rise of value-based care. Spurred by patients, governments, and investors, payers seeking data-driven methods to reduce healthcare costs are constantly seeking ways to improve healthcare data quality in order to make the most informed care decisions.

There are three recommended criteria you should use to evaluate healthcare data quality: breadth, depth, and trustworthiness. We’ll examine these elements below, and share how they can be used to make your data more actionable.

Criteria to Evaluate Healthcare Data Quality

When CareJourney’s members evaluate healthcare data quality, we help them look at three variables:

Breadth
Does the data set cover a significantly sized population to be of statistical significance? Does it address a full spectrum of population demographics, provider specialties, or care episodes? In many (although not all) cases, the broader your data set, the better macro insights it will provide.

The gold standard for healthcare data breadth is the inclusion of Medicare and Medicaid data alongside commercial data. Including information from the largest healthcare payers in the U.S. is a must to increase your healthcare data quality.

Depth
The size of your data set is one thing, but its depth is where data analytics can begin to add value by improving decision-making. Do you have longitudinal treatment data for every cancer patient in a particular practice? Do you have a list of all costs by billing code for your providers in a given timeframe? The more detail you have about care delivery over time, the better your analytics will be.

There is no “ideal” standard for data depth. Rather, you are looking to ensure you have the detail required to make value-driven decisions. An example might be data that profiles physicians and labels them according to the most common treatments they perform, versus the specialty they once labeled themselves as in an open data set.

Trustworthiness
The importance of basing care decisions that affect millions on “good” data cannot be overstated, and decision-makers will rest easier knowing that the data they use to do so has been validated by clinical expertise. These validations should include clinical and provider reviews of claims coding to ensure that the data represent real-world scenarios and should be prepared by subject matter experts who have deep experience in the healthcare industry and can help correct for outliers, population demographics, and common assumptions that do not reflect reality.

Data sets that are large enough to reduce the margin of error, have good detail into episodic care delivery at the provider level, and have a level of standardization that can hold up to scrutiny will be of the highest quality. While there is no single best practice to evaluate data breadth, depth and trustworthiness, as long as you take these three factors into consideration and prioritize those that are most important to your organization, you will be on the right path.

How to Make Healthcare Data Actionable

Once you’ve ensured you’re working with high-quality healthcare data, the next step is to make that data actionable in order to support value-based care decisions. To do so, CareJourney focuses on the three areas below.

Taxonomy
The more structured and codified your healthcare data becomes, the easier it will be to understand, make decisions and come to consensus. An ideal healthcare taxonomy will organize provider billing into a phenotype that can be used to build a real-world picture of how a given specialist delivers care. This taxonomy can then be used to make decisions in areas like plan support, market focus, and care coverage.

For example, if the majority of orthopedists who work on foot and ankle issues provide a certain treatment plan for a broken foot, and data shows that more generalized doctors often deviate from that norm and deliver less than ideal care, a payer may identify the need to refer more patients to specialists in order to improve outcomes. But the proper taxonomy (built from a deep data set) must first be in place in order to codify what “typical” care looks like.

Episode Grouping
Care episodes often involve a number of visits to a variety of providers over a significant period of time (and may even cross networks). Grouping these episodes is extremely challenging, but doing so allows payers and providers to understand how prevalent a particular treatment or condition is among a certain population, and how likely it will be to occur.

Providers in cities, for example, may be more likely to care for a younger, healthier population that will not have as much need for cardiac treatments, physical therapy centers, or diabetes medications that are more commonly required in areas with aging populations. Episode data helps make care allocation decisions.

Note that accessing episode grouping data from large, clinically vetted data sets is an excellent way to improve data trustworthiness, depth, and breath. CareJourney leverages an episode grouper for this reason.

Appropriateness
One of the most important derived data points in value-based care is appropriateness: how a provider delivers care for a particular condition. Significant clinical research has been done to identify the most cost-effective care plan for many common medical issues based on data that show what treatments have the best chance of success given the patient’s demographics, level of health, medical history, and the treatment’s cost.

For example, before progressing to shoulder surgery, did the provider offer a steroid injection? Was physical therapy performed? The more specific data a payer has on the care that was delivered, the better decisions it can make about the most effective treatments and coverage decisions.

Additionally, some providers may see better results with particular patients than others. Care data can be used to identify and reward efficient providers with higher payments, or to encourage less efficient providers to adopt new treatment plans. CareJourney has developed a Provider Performance Index based on appropriateness measures to help payers understand which providers deliver the most appropriate care.

Treatments, legislative changes to Medicare, patient symptoms (such as COVID), and payer priorities are constantly evolving, making data-driven healthcare decision-making one of the most challenging activities in the industry.

Many payers, providers, and value-based care organizations choose to work with a healthcare informatics partner due to the complexity of the analysis involved and the fast-moving pace of the industry. Request a demo to learn how CareJourney can help.