The Breadth and Depth of Clinical Documentation

  • Dr. Terrance Govender
  • September 06, 2018

In my last post, "What Should My CMI Be?", I discussed how, despite misleading conclusions that can arise from using CMI alone as a metric to measure CDI performance in the short term, many programs face questions when CMI declines in the context of improved documentation performance. I also demonstrated how ClinIntell’s clients are immediately able to gauge true CDI output/performance from a single metric: Documentation Score, which is independent of actual CMI and controls for fluctuations in your patient mix over time. In this post, I will discuss how a true improvement in CDI should include the breadth and depth of documentation, going beyond merely optimizing DRG assignment accuracy.

In the early stages of CDI, optimizing the DRG and achieving the shift to accurately capture patient severity of illness was at the forefront of our scope and activities. As CDI has matured and adapted to the evolving and increasingly demanding healthcare environment, the scope has rightfully moved beyond that, with a belief that the entire breadth and depth of the patient acuity can be captured through accurate and complaint clinical documentation. Reasons underlying this progression from the initial scope include the value-based reimbursement models (e.g., bundled payments), quality scores (e.g., VBP, Truven measures), and population driven reimbursement (Medicare Advantage).

While this is a good mission to adopt, as it speaks to capturing the full depth of patient severity, there are still many challenges that exist to pursuing the capture of these diagnoses. Some of the challenges are based on CDI productivity measures that are centered around optimizing the DRG and, hence, that is where most of the resources and time are spent.

Furthermore, CDSs are aware of query fatigue that may exist, especially amongst poorly engaged physicians. In addition, some of the sickest patients have numerous conditions with lacking documentation and physicians may only tolerate one or two queries on the same patients. Nevertheless, the benefits of capturing the full extent of patient severity outweigh these challenges and if CDI is to be perceived as a profession that has truly evolved over the years, then these issues should be addressed.

A benefit to adopting a clinical diagnosis approach to enhancing overall CDI performance is that the performance metrics can resonate with both the CDI function and physicians. Physicians can relate to underperforming on a specific diagnosis by sharing data on their observed versus expected rates and how those rates compare with their peers. In contrast, sharing metrics such as CC/MCC capture, DRG shift, or CMI trends will probably not engage physicians in the same way.

Advanced Analytics and Clinical Condition Performance

ClinIntell’s advanced analytics provides clients with observed versus expected performance trends on specific high-impact, under-documented clinical conditions. Such types of performance data are often familiar to physicians, as they are similar to readmissions, mortality indices, etc. With the adoption/implementation of standardized definitions, the answering of CDS queries, and a heightened awareness around under-reported diagnoses for their patient mix, physicians are able to “close the gap” over time and sustain the improvements.

The advantage of ClinIntell’s advanced clinical condition analytics over traditional benchmarking is that they can predict the prevalence of a diagnosis specific to a patient population, even if a diagnosis is woefully under-documented by physicians across the nation. In addition, the expected prevalence of the diagnosis (green line below) fluctuates from quarter to quarter due to fluctuations in the patient mix; another key variable that ClinIntell takes into consideration that benchmarking does not.

Drug Dependence chart

Thus far, ClinIntell’s clients have increased the capture of Drug Dependence (1% to 9%) through focused education and reporting, in spite of the low national performance on this diagnosis, with a clear majority of IM/FM physicians (362 hospitals) with a minimum of 3,000 Medicare discharges only reporting this diagnosis at 0-1% of stays.

Coded Rate Comparison for Drug Dependence

1All Internal Medicine physicians at U.S. hospitals (n = 568) with at least 3,000 Medicare discharges, sourced from the 2016 Medicare Standard Analytics Files (SAF) database, which includes Medicare fee-for-service claims.

The approach of addressing the breadth and depth of clinical documentation, and not just the DRG shift, is especially valuable in specialties with high-acuity patients, like pulmonary disease/critical care. It is easier to attain with “W MCC” DRG in these patient populations because they often have multiple MCCs and only one non-excluded MCC is required for the highest severity DRG. As a result, such patients may be deprioritized by the CDI function if physicians already documented MCCs on their own. While the analytics often reveal that they may not have a large CMI gap to close (DRG already optimized), clinical condition performance metrics will reveal significant gaps in the reporting of other high-severity diagnoses such as malnutrition, which may influence other quality and risk adjustment metrics.

ClinIntell can calculate the clinical condition gaps at the health system, hospital, specialty, and individual physician levels, as the analytics are based on claims-level data. This helps monitor and guide your efforts towards improving the capture of high-severity diagnoses through a myriad of initiatives that range from education to workflow process improvement. Furthermore, a Quality Department equipped with a list of industry popular “risk factor” diagnoses will be able to get immediate insight into under-reporting of specific risk factor clinical conditions by specific physicians and/or groups, allowing for a more proactive versus reactive approach to documentation improvement.

The case for fully capturing the breadth and depth of clinical documentation is an easy one to make, but only through advanced analytics can a health system or standalone hospital develop and execute a focused 80/20 approach required for success. If you engage physicians with this focused approach, then percentage increases in CMI will inevitably take care of itself.

“You’ve got forever; and it’s a mile wide and an inch deep and full of alligators.”

- Jim Thompson


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