Q&A: Addressing expected mortality
As a benefit of membership in the ACDIS CDI Leadership Council, Associate Editorial Director Linnea Archibald sends out “missed connections” emails to Council members with questions from other members on a regular basis. Anyone with experience related to one of the questions was invited to respond and Archibald connected them with the question-asker.
In order to share the information more broadly, the answers are periodically compiled and shared as Q&A-type articles in the monthly Leadership Council newsletter, the CDI Leadership Insider (CLI). This Q&A was originally shared in the June 2022 edition of the CLI and answers have been de-identified for the purposes of this article.
Q: Our system is noticing higher mortality numbers and CDI is working with quality to try to address the issue. In our current CDI software, we only have the severity of illness (SOI)/risk of mortality (ROM) available and not a concurrent expected value. When running reports for the past month’s mortalities, we noticed many patients whose SOI and ROM are both at 4 but whose expected value is much lower than expected. Are other leaders seeing this issue? How are they addressing it? How are other leaders assessing their mortalities? Are they using SOI/ROM?
Council member answer #1: We opened a pre-bill mortality queue in our CDI software. These cases are reviewed post-coding. The final coding is downloaded onto an Excel template in a CSV file for entry into our software that contains a risk calculator. We arrive at an initial score and view the risk contributions suggested to us by the software based on the model group the case falls into.
We then review the coding and documentation with a fine-tooth comb, looking to see if the present on admission (POA) status was assigned correctly, whether all the POA conditions were captured in the documentation and coding, etc. Depending on those answers, we either issue a query or send a note back to coding. Intense physician education is being done with physician groups as well.
We then capture the work on an Excel spreadsheet noting whether it was a coding or documentation issue and noting both the initial score and the improved score.
Council member answer #2: We use the Vizient® ROM calculator. Vizient drivers are based on what was POA for the patient. The SOI/ROM is a good indicator but does not always capture the entire picture. With some DRGs, even though they are maxed out at a 4/4, they still have a low mortality score. Heavy hitters such as sepsis, COVID, cerebral vascular accident, etc., need other conditions to be captured. It can be small things too, such has whether the patient came from a nursing home or was admitted through the ED, their age—those are the things that often get missed. In other instances, the POA status of a condition does not get coded as POA Y.
Council member answer #3: Historically, we only did a second level prebill review on mortalities that had an SOI/ROM less than 4/4. When I took over the program, I started reviewing mortality accounts with several of our quality leaders and noticed we had the same issue as already mentioned: our patients had an SOI/ROM of 4/4, but they had a low expected mortality.
Upon further review, I noticed we were not capturing many of the risk-adjusted diagnoses that would have impacted our expected mortality. We are currently revamping our process to include a second-level review on all mortalities regardless of SOI/ROM, working with a calculator to help identify risk adjustment opportunities, and we recently attended a quality boot camp to further educate ourselves on risk adjustment diagnoses, patient safety indicators, mortalities, etc.
Council member answer #4: We are a Vizient shop at the moment. Our initial findings were the same as others described: capturing appropriate POA status, admit source, picking up BIPAP, picking up appropriate risk adjustment diagnoses, DNR, and making sure all pertinent diagnoses made it to the hospital claim. We have a post-coding, prebill process in our technology. CDI reviews targeted cases with an SOI/ROM score less than 4/4, inpatient to hospice with scores less than 4/4, and all sepsis deaths. We want to move to a process in which we review all deaths once we are fully staffed.
During the review, CDI uses the Vizient risk adjustment calculator to verify all diagnoses that impact risk adjustment are captured. Sometimes a query is warranted and at other times the coding can be updated based on current documentation. We have a couple of physician advisors who review these cases with CDI, and at times, they will suggest queries and/or reach out to the provider of record. Physician education is key.