Skip to content Skip to navigation

Getting the Diagnosis Correct: What's the Impact of HCIT? (Part 2)

May 24, 2012
by Joe Bormel
| Reprints

As noted in Part I of this blog, I learned that as a physician, I had a misconception of the importance and interdependence of complete clinical documentation and coding integrity.  This was made readily apparent during the ACDIS conference and seminar two weeks ago in San Diego.

In essence, when the diagnoses are incomplete or non-specific, bad things happen from documentation subject to interpretation.  Competent doctors may appear to have higher mortality than their actual level because imprecise documentation has made the risk adjustment blind to their truly sicker patients. 

The expected lengths of stay for their patients may similarly paint them as unable to get their patients discharged in a reasonable amount of time, because the expectations are based on inadequate characterizations of their patients’ diagnoses.  Therefore, these physicians can appear to provide relatively poor quality care and to be inefficient, even if they are the most effective and perhaps most caring docs with the best results.  And that's just the direct impact on those doctors.

As a consequence, their organizations will rank poorly in terms of case mix, mortality, and a range of other measures that grow every year as pay-for-performance matures.  Also, their ability to collaborate and coordinate care with other providers will suffer.

Coding is hard enough when the diagnoses accurately capture the patient's situation, but when the diagnoses are non-specific, absent, or wrong, it adds work to everyone's plate.  That’s why Clinical Documentation Improvement (CDI) programs have become such important tools for getting the diagnoses correct. 

As we role out EHRs with the meaningful use measures of up-to-date problem lists, let's elevate the role and responsibilities of those HIM professionals who are trained and credentialed to help the community of providers get the problem lists right.  Their skills are focused on using an important body of knowledge, little known outside the HIM community until recently, to do just that.  This is upstream, concurrent with care and EMR use, and critical to quality measurement, clinical decision support, and at least half of the measures in MU Stage 1 – and likely in the majority of subsequent stages in healthcare modernization.

Specific Recommendations:

1.  Get Yourself Educated:  If you are a physician serving as physician advisor, a C-suite executive (especially CIO and CFO), or an HIM director with less than one FTE per hundred facility beds, you need to attend a conference like the one described here.  Unless you are aware and can comfortably speak to the necessity of a strong CDI program in your facility from any of those roles, any further recommendations are pointless.  This isn’t a topic you can learn by reading about it in a book, blog, or 45 minute inservice.  No one can, for a variety of reasons.

2.  Educate Others:  This is a big topic; in short, it’s an on-going, multi-faceted program.  It ranges from Quarterly staff meetings, one-on-one interactions, and a strong “query” program to communicate and educate in the context of patient care.

3.  Analytics:  There are CDI-specific operational analytics, in addition to traditional metrics like case mix index with appropriate, comparative benchmarks.  And, of course, documentation integrity impacts all other performance measures, including quality and safety.  You need to be collecting, validating, disseminating and trending CDI metrics, as you do with other management information.

4.  Develop CDI workflow for EHR deployment and optimization:  Where and when diagnoses are collected shouldn’t be left to chance, or merely reflect how it was done on paper.  There’s an additional, critical nuance here.  From a CDI perspective, the difference between a comorbidity and a complication can be a matter of when it was documented.  As a result, the admission history and physician exam note, especially the past medical history section, takes on a dramatically heightened importance.   Similarly, the progress notes and discharge documentation carries diagnoses, and their characteristics under CDI requires several hours to adequately elaborate.  Finally, CDI includes related query practices.  All modern EHRs have a variety of messaging systems, including but not limited to inboxes.  Several hours were spent at the CDI conference sharing experiences with how to integrate this component in the workflows involving EHRs.

Although emerging technologies, like speech recognition, computer assisted documentation with coding, and enhanced decision support will play critical roles, our practice of medicine and healthcare delivery will need to evolve and progress as well.

What do you think?



Dr. Bormel,
I like your recommendations. They are a common sense approach to what we need to accomplish.

Could you please tell me your position on investing in CDI before a hospital implements its CAC strategy? It seems to me that of importance is that CAC will be much more effective if it can actually find the codes required to meet the expectations defined in your blog.

Doc Benjamin

Thanks for your comment and great insight.

The simple answer is Yes, a strong CDI program is almost always a prerequisite to Computer-Assisted Coding. As you suggest, the quality of input to a CAC will be an important determinate of it's effectiveness. The qualifier "almost always" reflects that organizational governance, history, and culture issues often drive decisions.

CDI and CAC solutions have related but different goals. The former, CDI is best described as follows:

"Clinical Documentation Integrity:
Assurance that diagnoses and treatments described by the treating physicians accurately reflect the patients’ severity of illness using officially sanctioned terminology (ICD-9-CM) and are appropriately captured and reported by the treating facility."

CAC is more of a downstream tactic to address coding accuracy and coder productivity. It addresses a set of very but different problems. When the patient's record is substantively incomplete or inaccurate, as described in this post and as occurs commonly, CDI is the single, most important focus an organization can have.

Thanks again for your comment and question.

Joe Bormel,

As always, you've written an excellent, beautifully articulated blogpost. I think, beyond the specifics of why, how, and when to implement a CDI program, your blog points up the need for physicians to become truly engaged in processes that until recently, few wanted to be truly engaged in. Let's face it: the average practicing physician is focused on treating the patient in front of her/him, and views clinical documentation as, at best, a necessary evil, a process that might help him/her look back at the patient's record later on and help the next physician as well; but beyond that, physicians see documentation primarily in the context of the need to get paid.

That's why I think the transition to ICD-10 will be transformative in more ways than one. The potential to really look at patient populations and their situations, at far more granular levels, through analytics, will be one of the benefits coming out of the increased granularity of the ICD-10 system. But to use any system well takes time, and physicians and physician leaders absolutely need to be a part of the conversation, and fully engaged, to make sure they and their colleagues--and ultimately, of course, their patients--get the best out of all of this.

As always, you've hit the nail on the head with your observations, Joe!

Best, Mark

Dr. Joe,
Your blog points out the serious financial and quality perception risks inherent to transitioning from human-based abstracting and coding to MU's EHR-based quality measure reporting, problem list generation, health information exchange and financial system feeds. As we evolve out of the paper world, we seldom think of the human factor that is central to how we view coding and abstracting. Human HIM professionals can and do exercise judgment and arrive at intuitively reasonable conclusions based on what is stated in the record, implied by the record and absent from the record. For example, a patient may have a known condition other than the condition which constitutes the reason for visit or hospitalization. In the not infrequent situation that the other condition is not explicitly codified into the EHR, in our current world we can and do rely on the human coder to recognize the fuller context and ensure proper coding which includes the other condition. This is the way we've been doing it forever and we don't even think about it. We neglect to remember how binary and literal computers are; computers and EHRs are generally not yet at the point where, like IBM's Watson, maybe, they can make the intuitive leaps and judgment calls that trained humans make, and hence the faulty outcomes that your blog post so well describes.
As you suggest, the solution needs to very deliberately reinvent the process and not to simply assume that computers have yet achieved proficiency at data abstration anywhere near the level of competency of HIM professionals.

These were a great couple of posts, Dr. Bormel. The age-old problem with clinical documentation improvement efforts is trying to convince physicians that there is something in it for them; I think you've demonstrated very capably here that the stakes are high for docs.

Most physicians provide wonderful healthcare, but sadly it doesn't always appear that way on paper. That's where a good CDI program led by an engaged physician advisor can really make an impact.

Brian Murphy, Director ACDIS