Traditional Methods Versus The ClinIntell Approach For Identifying CDI Opportunity

  • Dr. Terrance Govender
  • February 07, 2018

Identifying opportunities for improvement of severity capture through accurate documentation and coding at your organization is crucial for several reasons. Some of the key points that come to mind are:

  • Measuring performance on risk adjustment models
  • Setting goals for the CDI program AND physicians
  • Identifying high versus low opportunity specialties and physicians
  • Monitoring annual and monthly performance
  • Developing a focused education program
  • CDS staffing
  • Determining the scope of your program

Historically, the industry has used benchmarking to measure and compare physician performance in CDI. We have found that benchmarking can, frequently, be very unreliable. We here at ClinIntell believe that one needs to consider the patient mix unique to each facility when identifying opportunity. Just because a facility is similar to another in terms of size, service lines, and specialties, doesn’t mean that their patient mix is the same.

Benchmarking assumes that your patient mix is static, when really, there are fluctuations month in and month out. In addition, if you consider the pool of facilities you are benchmarking against, and if all facilities within that pool are “low performers”, you will still have high and low performing outliers, which means that you could still be benchmarking against the best low performer. A static patient mix does not control for factors that are unique to your facility and can have a significant impact on expected rates for certain diagnoses at your facility.

This same benchmarking concept can be applied at a more granular level when dealing with your performance on specific diagnoses. Consider a diagnosis that is historically and woefully under-documented across the industry, like drug dependence and chronic respiratory failure. Even the top performers in a benchmarking model will be under-documenting these diagnoses for their actual patient mix, and your performance on these under-documented conditions will not be flagged as an opportunity to improve on.

An Alternative?

One option that would consider your unique patient mix when identifying CDI opportunities for your facility would be to do a chart audit of every single patient that walks through the door and gets admitted over a significant period of time. Not only is this unfeasible (since it is time consuming), but here are some other reasons why this is not a good option:

  • The manpower required to audit every single inpatient chart is too costly
  • The results are auditor-dependent
  • Opportunities can only be identified based on information or clues already provided by the physician – no information means no query or opportunity can be identified

Because we are all familiar with the DRG system and what is required to shift the DRG to a more accurate, heavier weighted DRG, it makes sense that there may be under-reporting of many chronic conditions in your patient mix once the DRG shift has occurred. Many diagnoses are not only classified as MCCs or CCs, but in addition, there may be HCC value to the accurate capture of these clinical conditions.

ClinIntell’s Solution

Faced with the challenges mentioned above, ClinIntell’s advanced analytics were developed to estimate the true severity of your unique patient mix based on your claims data and then quantify the negative impact on CMI from under-reported patient severity, as well as the degree of under-coding of specific diagnoses. And our model controls not only for the patient mix specific to your facility, but also for the unique setting; two major variables that benchmarking simply cannot factor.

Over the years, we have calibrated and validated our algorithms – now at their highest level of refinement – through feedback of coded data associated with physicians who have improved their documentation practices on specific diagnoses. This allows us to identify those under-documented conditions for your facility, even if the most appropriate facilities to benchmark against are poor performers on those diagnoses. The advantages of such a data analytic approach include identifying opportunities for specialties, physicians, and diagnoses per physician, based on their specific patient mix.

Our model also allows for performance monitoring down to the physician level, with a root cause report that empowers you to focus your efforts and have meaningful conversations with specific providers. Since our model predicts the prevalence of high-impact conditions in the patient mix attributed to each physician, we can provide you with an “observed versus expected” performance over time, not only for diagnoses that allow for accurate DRG assignment, but also for diagnoses that are of HCC value.

The ClinIntell model, which looks at your unique patient mix and controls for characteristics specific to your facility, is the next best thing to performing a chart review on every single patient that gets admitted over a significant period. It gives you insight into your patient mix that you wouldn’t ordinarily have through conventional methods.

To view your facility’s opportunity to improve severity reporting, as well as the highest opportunity docs and diagnoses based on your facility’s 2016 publicly available Medicare data, go to and request a complementary analysis.

“Innovation is the new competitive advantage.”

-Julie Sweet


Dr. Terrance Govender
VP of Medical Affairs, ClinIntell, Inc.