To link, or not to link: that is the question for physician documentation
Since the adoption of the ICD-10 code set, clinical documentation integrity specialists (CDISs) have experienced an ongoing dilemma of capturing complication codes with regard to adverse drug events versus ‘known’ or ‘inherent’ side effects.
At our organization, CDISs ask our physicians to add linking verbiage to their documentation for accuracy, which in turn, has a negative impact on their publicly reported scores with complications, and other data. For example, hypokalemia, due to ‘diuretic,’ has a T501x5A code added. Unfortunately, CDISs are caught in the middle. Our clinical documentation integrity (CDI) department is driven by accuracy and truth, however, publicly reported data from ICD-10 codes may indirectly harm our organization and physicians with reportable scores by attributing side effects as something ‘bad’ or ‘unintended consequence.’
The accuracy of documenting linkage has an indirect consequence to represent negatively our organization and physicians and the excellent care they provide when these links are made. In turn, it is not our intention to complicate matters for our coding professionals to suggest they not follow their guidelines to capture these diagnoses when linked.
With surgical procedures, the coding guidelines address ‘inherency’ and require the documentation of the surgeon to designate something as a complication or inherent to procedure. Coding guidelines also direct the coder to query if the documentation is vague. Should the same criteria apply when it comes to known side effects of medication, where the desired effect may also include known side effects, but the benefit supersedes the harm? Should we require the physician to deem an ‘adverse drug effect’ as a known side effect when most occasions, it is common occurrence? Should we ask the physician to ‘exclude’ the linking verbiage? Should they enter ‘as expected’ or should they simply avoid addressing these issues?
Coding guidelines specify that to justify a coding a complication, it must be clinically evaluated, diagnostically tested, and therapeutically treated; the complication must also result in an extended length of stay in the hospital necessitating increased resources related to care. The condition should not be part of routine care or the routine outcome of an expected procedure. When it comes to surgery, coders should clarify that there was a complication prior to assigning a complication code. The physician must agree and must document that the condition is a complication.
Should the same criteria apply when it comes to known side effects of medication, where desired effect may also include known side effects, but the benefit supersedes the harm? Should we require the physician to deem an ‘adverse drug effect’ as a known side effect when most occasions, it is a common occurrence? Should we ask the physician to ‘exclude’ the linking verbiage? Should they enter ‘as expected’ or should they simply avoid addressing these issues?
CDISs are caught in between, because we seek truth and accuracy. However, as clinicians, we do not necessarily consider side effects with the same measurement as ‘catastrophic’ events.
It is my belief that side effects should have an algorithm or require further clarification as to whether they are inherent or actual severe consequences that resulted in unexpected, untoward events that caused harm and increased both therapeutic measurements and length of stay (LOS). Most patients who receive diuretics or steroids in the inpatient setting require monitoring and potassium supplements, or insulin for that matter. It is unfortunate that during hospitalization, this may require capturing an adverse drug event code in the process.
The same issue is true of addressing ‘leukocytosis’ likely due to steroids while acknowledging that another disease process is going on that may not be infectious. It is through further assessment and lab testing that differential diagnoses may be ruled out. Should we discourage the physician to address these other issues?
I wonder what other CDI/Coders are experiencing, and if anyone has suggestions?
It is my hope that the utilization of ICD-10 codes for adverse drug events would be reported as they were intended, and that as clinicians, we can differentiate what is considered an ‘adverse’ event versus an inherent side effect, which would be monitored in an inpatient hospital setting along with excellent care.
Program Note:
Listen to Sharon Savinsky report on this topic today on Talk Ten Tuesdays, 10 a.m. EST.