Guest post: Pre- and post-query CDI metrics

By Howard Rodenberg, MD, MPH, CCDS

Along with my CDI work, I’m still pulling a few shifts each month as an ED doc. Having been in “The Pit” for nearly three decades, I’m intimately familiar with the Laws of Emergency Medicine, such as:

• The chances of serious illness decrease logarithmically as the number of individual complaints rise. (The Group Corollary is that as the number of persons in the exam room rises, the chances of anyone having a serious illness precipitously falls.)
• You don’t change the mortality rate. It’s still one per person.
• You can’t fake diaphoresis.
• If it’s sticky and it’s not yours, don’t touch it.
• When in doubt, examine the patient.

One other is that vital signs are like gold—incredibly valuable in assessing the patient’s condition and providing hard numbers as evidence of the care required and provided. But “hard numbers,” like pulse and respiratory rate, only give you part of the story. Every clinician has seen patients where the numbers (vital signs) look good but the person doesn’t, and the opposite is sometimes also true. The picture is fleshed out by the physician’s education, experience, and intuition—things that are not easy to quantify but no less critical to successful care.

Similarly, some things in CDI are easy to measure. It’s not difficult to calculate a facility’s case mix index (CMI), and it’s also not difficult to determine the financial impact of queries issued by CDI staff. Like a patient’s vital signs, these metrics are golden, as pieces of “hard evidence” about your CDI program’s performance. However, all measures have their limits, and just as with vital signs, we need to use another set of “soft” information to fully evaluate the efficacy of the CDI effort.

In a prior blog post, we reviewed why the CMI may be a poor indicator of the effect of a CDI program, even when the program’s queries demonstrate significant financial return. But, queries are not all a CDI program does. Much of our effort is devoted to provider education, either directly through interactions with physicians or indirectly by using the queries as teaching tools.

We tend to evaluate our teaching efforts in terms of what might be called “process objectives” such as the number of meetings we attend, the volume of presentations we give, and the numbers of tress cut down in the distribution of CDI Pocket Guides. But, if we’re going to try to fully demonstrate the value of our CDI work, it would be helpful to be able to quantify the effect of these knowledge-based efforts as well with some sort of objective measure.

I’ve been looking for a way to do so, and purely out of selfish motives. It’s pretty easy for a CDI professional to demonstrate their worth to an organization. Take the financial impact of the queries they generate, subtract their salary, and…voila…you have a positive return on investment that secures continued employment.

It’s not as easy with a physician advisor, however.

I don’t generate the queries, and while I may occasionally follow-up with a recalcitrant physician, the main thing I do is work on provider education and develop institutional definitions, policies, and procedures for the CDI team. Quantifying “leadership” is inherently elusive. So, is there any way I can objectively demonstrate value for what I do in an increasingly cost-conscious healthcare world? (And, because having discovered this thing called “daylight” and this meal called “lunch “after years of working ED nights, I really don’t want to go back into the dark. Plus, I’m a lot less pale.)

This process of worrying about my own job security has, I think, led me to devise a method which I think can help demonstrate the value of the education piece of a CDI program. The easiest way to do so is to look at the CMI before queries are issued and answered, against a pre-you baseline. This version of the CMI reflects what your efforts have accomplished in terms of education alone, while the final CMI shows the baseline plus the effect of queries. A steady rise in your baseline CMI before queries should serve as objective evidence of the educational value of your program.

Alas, it’s often not that easy to find this number. For instance, I can easily find the final CMI for any month or year. But the final CMI includes those records in which the individual case’s relative weight (RW) has been affected by queries, so it can be hard to figure out what it was before those queries were issued.

This is where some math comes in. At our facility, we report how many queries have financial impact, and what the CMI was of that group of queries before and after the query process. It turns out I can use those numbers to calculate what the CMI would have been if no queries were issued.

Here’s an example, using whole numbers to make the math easier. Just as in our calculations regarding CMI in a prior post, this also uses the concept of RW as “points.” Let’s say your program’s CMI for a month with 1,000 admissions was 1.0, and that you know you issued 100 queries with financial impact that month within the 1,000 charts. The CMI of those 100 charts before your queries was 1.0, and after queries were answered the CMI for that group of charts rose to 2.0. To calculate the CMI before the queries went out:

• Total 1,000 records with CMI 1.0 = 1,000 RW points
• 100 records (within the 100) after queries at CMI 2.0 = 200 RW points
• Remaining 900 records have 800 RW points (these are charts that have not changed by CDI review and queries)
• 100 records before queries at CMI 1.0 = 100 RW points
• CMI before queries = 800 RW points from 900 unchanged charts plus 100 RW points from queried records before queries = 900 RW points/1,000 records = CMI 0.9

Hopefully your data system will be able to show you the CMI both before and after queries, and you won’t need to solve these mathematical puzzles. But it is reassuring to know there’s potentially a way through this conundrum.

Just like vitals, however, the numbers are not the whole story. In my blog post about CMI, we talked about needing to know the context of a statistic to make sense of its meaning. We’ve also talked about the natural history of CDI programs, and that CDI numbers often follow a sigmoid-shaped curve. The same is true here. Your CMI before queries—your “education number”—can only be interpreted in the context of your CDI program as a whole. The addition of new physicians or services to your hospital may also cause CMI to rise irrespective of the scope and depth of your provider education. The age of your program matters as well. If you have a brand new program and you can track this number from the start, you have a wonderful opportunity to demonstrate impact. If your program is more mature, the pre-query CMI is much harder to interpret, especially as most CDI metrics plateau as programs age.

Objectively demonstrating the value of CDI educational efforts doesn’t guarantee job security. But it does help put a number to those intangibles of a CDI effort, and allows you to fully describe your program’s effect on the institution’s bottom line. And maybe it keeps me in the sunlight. That would be nice.

Editor’s note: Rodenberg is the adult physician advisor for CDI at Baptist Health in Jacksonville, Florida. Contact him at howard.rodenberg@bmcjax.com. Advice given is general. Readers should consult professional counsel for specific legal, ethical, clinical, or coding questions. Opinions expressed are that of the author and do not represent HCPro or ACDIS.