Solutions for 3 Critical Data Mistakes in Healthcare Analytics

One thing is for certain, healthcare analytics has changed the landscape of the medical industry. Today, big data is used frequently by organizations and providers to support patients, and rightfully so. It’s hard to ignore its capacity to provide greater efficiencies, decreased costs, reduced errors, and evidence-based decision-making. 

Ultimately, this plethora of data helps you provide better patient care and run your clinics more seamlessly. Unfortunately, critical data mistakes can impede those benefits, especially in wound care. Below, we’ll discuss healthcare analytics mishaps you should be aware of and offer solutions to those challenges. 

Critical Healthcare Analytics Mistakes in Data

1. Failing to Define Your Goals First

As the saying goes, “If you can’t measure it, you can’t manage it.” But what if you don’t even know what “it” is in the first place? This is actually an issue that many clinicians, quality improvement leaders, and healthcare organizations as a whole run into when working with data. Especially when implementing new processes or statistics, it can be hard to know what you need to do.

It’s common for practitioners to examine the data first in an attempt to draw conclusions. However, it’s easy to find yourself stumped halfway through because you’re unsure what you’re genuinely looking for. While the sea of data initially looks intriguing, it suddenly seems overwhelming, causing even the savviest to draw a blank. 

So, why exactly does this happen? Well, this typically happens when you review data without understanding the problem you’re trying to solve. After all, if you’re not sure exactly what you’re looking for, how can you be sure you’ve found it? If you hope to leverage data and make it meaningful, you must define the problem and set a goal. Otherwise, you’ll have data but no direction. 

We’ll share some examples of some of the different types of data you’ll likely see in your role that can be beneficial. But keep in mind that you need an objective to get the most out of it (don’t worry—we’ll discuss more what this might look like in a little bit). 

Data Is All Around Us

Healthcare institutions are presented with tons of data from a wide range of sources. For instance, electronic health records (EHRs) hold an abundance of information about individuals, helping clinicians learn about patient groups and identify trends. A wound-care-specific EHR can be the best resource to gather data on a patient’s medications, diagnoses, and more. 

In addition, imaging data analytics, which research suggests are expected to change the clinical management of wounds, are available. Advancements in ultrasound imaging make it possible to evaluate full-thickness wounds, and perfusion imaging can help clinicians assess functional tissue health status. 

Wearable technology is gaining traction in the wound care space as well. Some can monitor biomarkers like temperature, pH, oxygen, and uric acid. The massive amounts of data that stem from these devices can guide providers on the best treatment approach for an individual patient based on what’s observed. 

Interestingly, artificial intelligence tools can obtain healthcare data and produce even more data. If they’re taught to do so, machine learning algorithms, for example, could recommend wound therapy techniques that improve patient outcomes. Predictive analytics to forecast potential events along a patient’s healing journey can revolutionize the entire industry. 

The Solution? Get Clear on the Problem

These are just a few sources of big data and ideas on how they can assist you. But now it’s time to determine how to maximize its use. The most successful organizations have a clear understanding of how to use healthcare analytics to advance their business. So, here’s what you can do to identify problems.

• Assess common patient satisfaction complaints, such as those provided by the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey. 
• Identify quality issues, such as misdiagnosis, medication errors, readmissions, length of stay, and more. 
• Speak with internal and external stakeholders.
• Review current evidence-based literature to determine whether your organization’s processes still align with new best practices. 

After you confirm the problem(s), you can create some goals, like: 

• Reduce costly penalties associated with readmissions and hospital-acquired conditions.
• Decrease patient’s length of stay without compromising quality of care and patient satisfaction.
• Enhance clinic scheduling, billing, and operations.

Once you’ve established what you’d like to achieve, you can develop data processes that support the end goal. With this clarity, you’ll better understand the type of data you need to obtain, review, and work to improve. 

Organizations that want to thrive in the value-based care world must have clear goals for their healthcare data. 

2. Overloading for the Sake of Overloading

With all of this data available to you, doesn’t it make sense to share it all in a perfectly designed presentation with pie charts, line graphs, and more? Well, not so fast. Truth is, not everything measurable is important and not everything important is measurable. 

The Cons of Sharing Too Much Data

It’s easy to want to share everything we can so that we can dive into all the data we’ve rummaged through. However, this can unintentionally do more harm than good. 

Shifts Discussions Away from the Problems and Goals

There are several complications to simply dumping all collected data onto a slide and going from there. For one, seeing a bunch of data on a screen can overwhelm an audience. You’ll find that discussions on data would be all over the place and drive the focus away from the primary problems and goals you identified and set before. Typically, one to three data points per slide is ideal.

Let’s consider an example. If you and a quality improvement team are meeting monthly to discuss how to decrease the frequency of pressure injuries, should you share data about fall risks, too? Would it be helpful if you explained how many patients developed delirium daily? In this meeting, maybe not. 

It might be more helpful to share data on the following instead. 

• Metrics that show the number of required comprehensive skin assessments divided by the number of comprehensive skin assessments actually being completed daily
• A measure sharing if staff are conducting head-to-toe skin assessments within 8 hours of admission
• The facility’s pressure ulcer incidence rate 

Sharing this type of information may help reveal if and where there are gaps in care. Based on that insight, a team can develop improvement processes and return to their unit to implement and assess them. However, including information irrelevant to that audience would reduce efficiency. 

Increases the Potential for Errors

Having access to a goldmine of data is exciting, but let’s not forget that data management is also complex. If you’re trying to share all the data possible, you may end up with some inaccuracies. 

In the world of big data, you’ll likely come across unstructured data. In fact, research suggests that 80% of electronic healthcare data are unstructured. So, if you have a mix of structured and unstructured data, you’ll want to ensure it’s standardized before you present it to others. This is easy to overlook if you’re trying to share a lot of data. 

Healthcare analytics are powerful, but poor data or ineffective management can result in: 

• Distorted assumptions that skew quality of care and healthcare delivery. 
• A negative financial impact, as reimbursements are often tied to quality. 
• Unidentified public health trends. 
• Negative downstream effects. 

The Solution? Select the Right KPIs for the Dashboard

If you’re trying to figure out what data to present, ensure you’re measuring the proper Key Performance Indicators (KPIs). KPIs are quantifiable measures that determine an organization’s performance against a set of targets, objectives, or industry peers.

Not sure what some of your KPIs should be? Consider looking at quality measures from: 

• The Joint Commission
• The Centers for Medicaid and Medicare Services (CMS), or
• The Agency for Healthcare Research and Quality (AHRQ) 

CMS has approved quality measures specific to wound care that could affect reimbursements. One such measure is the Merit-Based Incentive Payment System (MIPS), which wound care practitioners and organizations should pay close attention to and ensure they’re compliant with. 

Finally, create a dashboard. Some business intelligence tools utilize data visualization to provide actionable feedback on performance, capture adherence to research-based practices, display resource utilization, and more. Ensure the information presented on the dashboard aligns with the objectives set and is tailored to the audience. Decision-makers should be able to examine the data points and tell a story.

3. Absence of Clearly Defined Definitions

If Department A captures and shares data one way, and Department B captures and shares data differently, how do you bring the information together to make it meaningful? This is an issue that organizations occasionally run into, as departments and specialties often operate in silos. This can also happen across the industry as a whole, as is seen in wound care, with different facilities and different practitioners treating patients.

In wound care, this can be particularly problematic, as there are scenarios where subjectivity can occur. Even the National Institute of Health shared that ” each clinician will have widely differing and distinct opinions on wound therapy depending on prior experiences.” But this doesn’t just begin with wound therapy approaches. These differing opinions, or subjectivity, occur during wound assessments, too. 

For example, depending on someone’s background, whether they are a specialist in wound treatment or in another specialty like emergency medicine or even general practice, and what tools they use to measure, they may interpret a wound measurement or degree of tissue injury in another way than someone else. This level of subjectivity makes meaningful data allocation quite difficult. And if healthcare analytics aims to improve patient care, data standardization is needed. 

What is Data Standardization?

Data standardization includes “methods, protocols, terminologies, and specifications for the collection, exchange, storage, and retrieval of information.” In other words, it organizes data in a common format so it’s easily accessible. 

When data standards aren’t established in an organization (or between organizations or across industries), it creates more than just an internal issue. Lack of standardization can stop the share of information between healthcare organizations and payers for insurance reimbursement. There’s a potential for healthcare providers and pharmacies to have issues sharing data about prescriptions or for healthcare facilities to have difficulty accepting data from clinical laboratories.

Furthermore, it halts the reutilization of healthcare data to meet quality reporting and patient safety requirements. Over time, these data standardization roadblocks will make it difficult for companies to leverage predictive analytics to drive clinical and operational decision-making. As a result, it’s crucial for departments and stakeholders to agree on how to store data so it can be used collaboratively. 

The Solution? Define Data Points and Decrease Subjectivity  

If you want to get on the right track, it helps to take a look at what standardizing healthcare data involves. It’s much more than just defining “what,” it also includes establishing the following. For these purposes, we’ll just talk about things your organization can do to standardize your own data, though these best practices also work across industries and facilities.

• Terminologies: Agree on the medical terms that all providers at the organization will adopt. It’s possible for a condition or treatment to be referred to by numerous codes or abbreviations. 
• Knowledge representation: Deciding how to design, organize, and present data in a manner that allows systems, like machine learning, to interpret and use it. 
• Definition of data elements: Confirming how the data will be collected and exchanged. 
• Data interchange formats: Determine the formats you’ll use for exchanging data, such as Health Level Seven International (HL7) or Fast Healthcare Interoperability Resources (FHIR). 

Another important element you’ll want to focus on is attempting to minimize subjectivity and maximize objective data points. To do so, look to technology. 

AI-powered wound imaging platforms can utilize artificial intelligence and mobile wound imaging to measure wounds without a paper or physical ruler. Some offer features that can highlight the risk of amputation or predict treatment plans. These can help practitioners better evaluate and treat the wounds you see.  

Get the Most Out of Healthcare Analytics 

Chronic and non-healing wounds can impair patient outcomes and quality of care, making them a fiscal and physical burden. Thankfully, technology, along with extensive healthcare data, can be used to not only promote the healing of chronic wounds but potentially prevent them altogether. However, this heavily relies on reliable data collection and analysis. 

Now that you’re privy to some of the most common data mistakes in healthcare analytics and know how to solve them, your organization is better prepared for what’s coming in wound care. The industry is quickly transitioning from descriptive (the past and now) to predictive (the future) analytics, and it’s important to get on board as the change occurs.