Hospital-associated venous thromboembolism prophylaxis use by risk assessment at a large integrated health care network in Northern California.

Abstract

Hospital-acquired venous thromboembolism (HA VTE) is a preventable complication in hospitalized patients. We aimed to examine the use of pharmacologic prophylaxis (pPPX) and compare two risk assessment methods for HA VTE: a retrospective electronic Padua Score (ePaduaKP) and admitting clinician's choice of risk within the admission orderset (low, moderate, or high). We retrospectively analyzed prophylaxis orders for adult medical admissions (2013-2019) at Kaiser Permanente Northern California, excluding surgical and ICU patients. ePaduaKP was calculated for all admissions. For a subset of these admissions, clinician-assigned HA VTE risk was extracted. Descriptive pPPX utilization rates between ePaduaKP and clinician-assigned risk as well as concordance between ePaduaKP and clinician-assigned risk. Among 849,059 encounters, 82.2% were classified as low risk by ePaduaKP, with 42.3% receiving pPPX. In the subset with clinician-assigned risk (608,512 encounters), low and high ePaduaKP encounters were classified as moderate risk in 87.5% and 92.0% of encounters, respectively. Overall, 56.7% of encounters with moderate clinician-assigned risk received pPPX, compared to 7.2% of encounters with low clinician-assigned risk. pPPX use occurred in a large portion of low ePaduaKP risk encounters. Clinicians frequently assigned moderate risk to encounters at admission irrespective of their ePaduaKP risk when retrospectively examined. We hypothesize that the current orderset design may have negatively influenced clinician-assigned risk choice as well as pPPX utilization. Future work should explore optimizing pPPX for high-risk patients only.