Q&A: ‘Unable to determine’ query responses and KPIs

CDI Strategies - Volume 16, Issue 52

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 July 2022 edition of the CLI and answers have been de-identified for the purposes of this article.

Q: At our organization, the provider response of “unable to determine” is considered a disagree. What goals or key performance indicators (KPI) are facilities/systems using to measure provider responses? For example, we use agree rate of 80% or greater, disagree rate of 20% or less, which includes “unable to determine.” Is anyone reporting “unable to determine” responses independent of disagree responses? If so, what is the goal for these responses?

Council member answer #1: We track and trend the following types of responses:

  • Agree
  • Disagree
  • Clinically insignificant
  • Clinically undetermined
  • No response
  • Remove

The report we share with our stakeholders shows our overall response rate, agree rate, and disagree rate. We don’t share the other four query response metrics (insignificant, undetermined, no response, and remove) with stakeholders, nor do we have targets for those metrics. We track and trend those query response types, and approach physician groups or individuals as needed for education based on the metrics.

Council member answer #2: We track very similar metrics. The “unable to determine” in our CDI software also goes to a “disagree.” Because of the following regarding present on admission (POA) status, however, we will mark our queries appropriately (which may be an “agreed” if the provider is clinically unable to determine):

CMS indicator

CMS POA meaning

POA equivalent





Clinically unable to determine

Counts as POA = YES






Counts as POA = NO

If we have a CDI specialist with a higher-than-average response of “disagreed,” the manager or team leader will complete an audit to determine if the query was valid, provided all clinical indicators, and is non-leading. Sometimes providers will just say “unable to determine” and we know that should not count against the CDI professional when the query was appropriate. I know this isn’t an easy solution, but it is how we currently handle the process.

Council member answer #3: We’ve recently removed the option of “unable to determine” from the templated response. We left the option of “other, please specify” in an effort to reduce the number of quick replies of “unable to determine.” Part of our escalation process is for “unable to determine” responses be sent to the attending for second opinion as well.

We report out agree, disagree, and no response to the chiefs of service.

Council member answer #4: Our “unable to determine” responses are included in our response rate as a response; however, it’s not reported out separately on our KPIs. We only report the agree rate of the queries that are responded to.

There are times when “unable to determine” is an appropriate response, but we found that often providers answer that way because they don’t understand what is being asked, don’t know how to respond to it, or they’re just trying to close out the query pop up.

We had a lot of “unable to determine” responses about two years ago so we came up with a process to reduce this response type.

Whenever we get an “unable to determine” response, and it has a financial impact (providers don’t know this but we had to limit the number somehow), the CDI specialist has a follow-up face-to-face conversation with the provider. This allows us to educate the provider on the ask and why we need them to respond/the value of the clarification. It also allows the provider to educate the CDI specialist on what they are thinking and/or why the query cannot be answered. Our goal was to drive down the “unable to determine” responses, but moreover it is about educating and engaging our providers and growing the CDI team.

The hospitalists are the attending on 65-85% of the inpatients, so I meet with the hospitalist medical directors at each hospital quarterly around their CDI metrics. I also use this time to review the patient-level detail for any “unable to determine” responses we received during that quarter with potential financial impact. They actually enjoy going over the cases and we have had some great conversations educating them and some great takeaways for the CDI team too. 

Over the past two years we have driven our financial impact “unable to determine” responses down from over 60 per quarter to an average of nine.

Council member answer #5: We break out our “unable to determine” responses on our reporting. We also report alternative diagnoses and uncodeable responses. Our goal is to have an 85% agreement rate (we are currently in the 90s). We report this information to administration and present it at our utilization management committees. If one of our hospitalist groups “unable to determine” rate goes up, our physician advisors meet with the directors of the group.  We also escalate many of our “unable to determine” responses to the physician advisors.

Found in Categories: 
Ask ACDIS, Clinical & Coding