Healthcare Analytics, Population Health Management, Healthcare Big Data

Quality & Governance News

Cultivating a Big Data Mindset for Coders, Clinical Documentation

Clinical documentation improvement is the key to better ICD-10 coding, but it's also vital for preparing healthcare organizations for a big data environment.

- For the past few years, clinical documentation improvement efforts have been mostly associated with the ICD-10 transition – the often-delayed switch to an upgraded classification and coding system that requires enhanced detail and specificity in every patient’s health record.

Clinical documentation improvement and big data analytics

Leading up to the conversion, most providers worried about if the modernized code set would negatively impact their revenue cycles, and some opponents of the new requirement even forecasted imminent financial doom for the majority of the industry. 

Thankfully, the naysayers have been largely wrong about the monetary impacts of ICD-10, but that doesn’t mean that the process has been free of trouble – or that the code set has yet produced the wide-ranging analytics and quality improvement benefits touted by its supporters.

Even organizations that have developed a long-term view of the value of their big data analytics assets are still struggling with deeply ingrained process problems that made everyday coding for reimbursement a difficult proposition. 

At the University of New Mexico Hospitals (UNMH), Executive Director of HIM Catherine Porto, RHIA, isn’t just responsible for training coders to be speedy and accurate when working with ICD-10. 

She is also trying to raise the level of information management across the board in order to prepare the organization for a future where the quality of clinical documentation is more closely tied to reimbursement than ever before.

Supported by 3M coding technology and four decades of experience in the health information management field, Porto has a clear vision of how to achieve these goals, and will also be sharing her expertise with her fellow HIM professionals as the 2017 Speaker-Elect of the AHIMA House of Delegates.

But competing priorities in different care settings, lingering bad habits, EHR usability issues, and the constant challenge of managing multiple approaches to clinical care have made it difficult to generate a commitment to consistency, specificity, and accuracy in every member of data integrity continuum.

To overcome these obstacles, healthcare organizations like UNMH must develop tailored methods of education and encouragement that address differences in opinions and personal preferences while generating buy-in across all departments and job titles.  

Tensions can run high when providers are faced with queries about the content of their clinical notes, or when CDI and coding staff butt heads over details for billing.  No professional wants to be told they have to make changes or corrections to their work, especially when the revisions take extra time and effort, but sometimes a few tweaks are unavoidable.

“We need to reinforce that these aren’t extra tasks we’re asking them to do just because we want to,” Porto said to HealthITAnalytics.com during the 2016 AHIMA Convention in Baltimore.  “They need to do these things because this is how you code correctly in an environment where quality matters.”    

Organizations that start looking at data integrity as an opportunity to enhance quality instead of just a necessity for payment will lead the healthcare system towards achieving its reform goals, she asserted, but making that leap while retaining high levels of productivity and payment under the current system is admittedly a big ask for everyone involved.

“Under the Evaluation and Management (E/M) billing system we’re using, physicians have become so focused on the elements to get a specific level for their billing that they don’t really pick up the specificity or the diagnosis, and they don’t tie the two together with their assessment and plan," Porto said.  "They just learn what the billers want and give them the bare minimum, because they don't always have time for all these other things.  So the chart doesn’t become a clinical document anymore, and it doesn’t tell the patient’s story well.”

Coders, meanwhile, might be so focused on getting bills out the door on time that they may not be capturing all of the relevant details that are actually present in the chart.

“Under ICD-10, we’ve got coders that are working very hard on getting their productivity numbers up, and so they’re zipping through these things really fast.  There’s no point in having speed if you’re not billing properly, though.”

“One particular area that’s really difficult is post-op respiratory failure.  The coder sees ‘respiratory failure’ in the chart, so they just code ‘respiratory failure.’ Well, if the diagnosis occurs during the post-op period, it’s going to count as a patient safety indicator (PSI), so you have to record that it’s within that timeframe for quality reporting and payment purposes,” she pointed out. 

“Coders have to look at more than just the term in front of them.  They have to see it in context and take the time to read deeper into the circumstances surrounding that term if they want to tell the most accurate story.”


The Role of Healthcare Data Governance in Big Data Analytics


Storytelling is a recurring theme in the clinical documentation improvement world, and for good reason.  A patient’s chart often centers around a narrative that will, ideally, capture every detail of their progression from wellness to illness and back again. 

But sometimes treating patients as characters in a story can backfire, Porto said, if the narrator makes the mistake of focusing only on a single defining trait without recording how the patient's condition develops and changes over time.

“We have a hospitalist group, for example, that has a line for patient identification at the top of the page,” she explained.  “It’s just something to trigger the physician’s memory of their patient.  It might say something like, ’96-year-old hip fracture,’ because physicians can’t keep all the names straight, so that’s how they think of them and how they discuss the patient with their colleages or residents.”

“Now, that may be what this patient came into the hospital with, but that might not be her only diagnosis.  She might develop pneumonia, so now she’s also a pneumonia patient.  Or she may have diabetes and hypertension, so she’s not just an orthopedic case.  She’s much more complex than that, and if you keep thinking about her as a hip fracture first and foremost, you might miss other aspects of care that she needs to get well.”

That way of thinking may be a quick and easy way to keep patients straight during busy rounds, but it can negatively influence the quality of documentation, she said, because retaining “hip fracture” as the patient’s primary diagnosis across her entire stay may not be the most accurate way to bill for services rendered.

“I think we’ll see a bigger role for coders in trying to change that mentality, because they will learn not to keep taking that initial diagnosis forward, and then they’ll push the physicians to stop doing that, too,” Porto predicted.  “Everyone needs to look further into the data.  What happened today to the patient that didn’t happen yesterday?  What do we need to add to the record so that it's as accurate and comprehensive as possible?  Physicians and coders all need to capture that, but they’re not used to doing it.”

CDI technology and computer assisted coding (CAC) applications can help organizations drill down into the data, report on quality and outcomes more accurately, and prevent the need for contentious conversations about who missed what the first time around.

“If you don’t have software that’s going to help you determine that this patient may have experienced a hospital acquired condition (HAC) or that this development is a patient safety indicator, then you might be in some trouble,” Porto said. 

“If you don’t have a software tool that can flag those problem areas for you and tell your coders to slow down and look at the chart again, you won’t be able to get them into that data analytics mindset, and that focus on quality reporting, which most of them don’t currently have.  That’s another skill they’re going to have to learn.”


The Difference Between Big Data and Smart Data in Healthcare


Nurturing that mindset is important for coders and HIM professionals who are responsible for developing a pool of big data that can eventually be used for more advanced analytics, and it’s equally vital for physicians if they want to succeed in a patient-centered environment where transparency and quality are key.

“The potential effect on their quality scores is scaring them to death,” says Porto.  “They’re very competitive, but it’s also a hard sell for a lot of providers who don’t want to be asked to do more and more in the electronic health record.  We have to go back and think about what tools we can use in the EHR to make it easier for them to give us more information.”

As with many recent efforts to reform the healthcare system, securing physician buy-in is essential. “The key to generating buy-in is having physician experts translating for you,” Porto said. “Clinicians and HIM aren’t always on the same page, and they sometimes feel as if they're in opposition to each other.  It can certainly cause a few ruffled feathers.” 

“I had one educator that grabbed a chart and just started highlighting all the areas in the document that could have been more specific – it sent one provider off the deep end,” she recalled.  “Because this physician was great at documentation and had beautiful notes.  She could have been more detailed in certain areas, sure, but it wasn’t intended to a rebuke of her work.  She thought we were ‘grading her paper,’ though, and it really touched a nerve.” 

“So I sat down with her talked about what that felt like to her, and asked her how we needed to change our approach, because that wasn’t what we were trying to do.  What we see as an opportunity for improvement may seem like a criticism to someone else, so we had to learn how to present things in a way that didn’t alienate physicians.”

Porto also spearheaded a physician champion program that fosters education and dialogue between physicians and coders.

“We took about twelve of our most troublesome PSIs and HACs and recruited a physician champion for each area,” she explained.  “It’s educational on both sides, and if we can get a good clinician partner who’s really interested, that’s a big help.”

Facilitating productive discussions between physicians, coders, and CDI staff is just one of the many tasks piled on the shoulders of any HIM director, but Porto believes the HIM department should be even more involved in projects like EHR optimization.

“HIM should be involved in designing workflows and how to use structured and unstructured data in the EHR,” she stated.  “It doesn’t make sense to have everything in a structured field, of course, but there are a lot of things that should be in that format.”

More structured data may help to reduce some of the confusion over creating clinical documentation that includes the data necessary to code accurately, she added, and it would allow for easier reporting across a variety of big data analytics needs.

“I don’t want people to have to dig around in the charts hunting for bits of information if we have an audit or we want to do analytics,” she said.  “I want to be able to run a report and give it to them.  Standardized data would help us do that, but that comes back to designing the right interfaces and making the workflows easy for physicians.  That’s a big project for everyone, and it’s going to take time.” 

“We’ve come a long way towards improving data integrity in healthcare and preparing for a big data analytics environment where quality and value are really paramount, and it’s exciting to see the progress, but there’s still a lot we still have to do.”

Continue to site...