Overuse of Respiratory Viral Panel PCRs on Inpatient Medicine

From the 2018 HVPAA National Conference

Peter Barish (University of California San Francisco), Jahan Fahimi (University of California San Francisco), Michelle Mourad (University of California San Francisco), Ari Hoffman (UCSF)

Background

Viral respiratory illness is a common complaint. Though comprehensive testing is available to identify specific viruses, targeted treatments are limited. Our institutional guidelines for the Respiratory Viral Panel PCR include ICU and immunocompromised status, where results are more likely to change management. Despite this, use remains common on hospital medicine.

Objectives

To study the use of the Respiratory Viral Panel (RVP) PCR in hospitalized patients and to examine potential areas of overuse

To estimate the direct and indirect costs associated with overuse of the RVP PCR

Methods

We obtained computerized electronic health record (EHR) data from patients in the Emergency Department (ED) and on the hospital medicine service at an 800 bed tertiary-care hospital between 11/2016 and 12/2017. We identified all patients who had received respiratory virus testing, which included point-of-care (POCT) influenza, rapid influenza and Respiratory Syncytial Virus (RSV) PCR, and comprehensive RVPs. Patients were categorized by primary encounter location (ED, acute care or ICU), type of testing, and illness severity. Test result time, discharge date, flu status, and length of stay were also abstracted from the EHR. We estimated the proportion of immunocompromised patients through extrapolation of a 60-patient chart review, where immunocompromised status was defined as: the presence of HIV, active chemotherapy, organ transplant, or immunosuppressing medication (e.g. high-dose corticosteroids). Differences in mean length of stay were calculated using parametric tests of significance. A linear regression model was used to adjust for medical complexity using a DRG weight score. We obtained direct cost and charge data from our hospital’s financial database managed by Enterprise Performance Systems Inc.

Results

We identified 2595 total patients with any respiratory viral testing and 1523 patients with an RVP PCR during this time period. The mean time to result of this test was 1.9 days and for much of the year the test was performed only 3 times per week. Of these tests, 12.5% (191) were performed in patients discharged from the ED and overall 26% (395) had results return after the patient was discharged from the hospital. Furthermore, extrapolating from chart review we estimate that only 40% of RVP PCRs were performed according to internal guidelines. The charge associated with each test was $3450, leading to greater than $3 million in excess hospital charges. Relative to patients with other influenza testing alone, those with an RVP PCR experienced a 1.5 day increase in length of stay (6.2d vs 4.7d) and this relationship remained significant after adjustment for medical complexity.

Conclusion

We found that the use of the RVP PCR is common among hospitalized acute care and performed frequently for patients who fall outside the recommended guidelines. There is often a significant delay in result time, leading to a surprising number of patients discharged without an RVP result. Interestingly this test is also associated with an increased length of stay. Based on direct and indirect costs estimates associated with this test, RVP PCR overuse represents a significant potential area for value improvement.

Implications for the Patient

We have shown a clear overuse of the RVP PCR at our institution, and preliminarily, an association with a longer length of stay. This leads to significant excess costs to the medical center, and by extension, to the patient.