113 Impact of Scribes on Billed Relative Value Units in an Academic Emergency Department

Abstract

To assess the impact of scribes on relative value units (RVUs) in both adult and pediatric emergency departments (EDs). A prospective observational study was developed in a tertiary care academic ED that contains both an adult and pediatric ED. Completed patient charts were coded by an external billing and coding company, then returned and mapped by ICD-9 diagnostic codes to allow for comparison by diagnostic groups. All ED patients from February 1, 2015 through September 30, 2015 were included. There were no exclusions. Scribes were hired and trained between February and May. After completing individualized training by a staff member with more than 10 years experience in implementing scribe programs, scribes were scheduled to provide 1-to-1 support to a provider (attending, resident, nurse practitioner, or physician assistant) during the duration of that provider’s shift. Comparisons were made between scribed and non-scribed patient visits. There were 49,389 patient visits during the study period, of which 7,865 (15.9%) were scribed (ie, seen by a provider working with a scribe). The study population included 39,926 adults (80.84%) and 9,463 pediatric patients (19.16%). Amongst the adult patient cohort, the charting generated for scribed patient visits produced 0.20 additional RVUs per patient compared to non-scribed visits (p<0.001). More specifically, scribes generated additional RVUs in ESI level 2 (0.27 higher RVU/patient, p<0.001) and level 3 (0.13 higher RVU/patient, p<0.001) adult patients. Scribed charts had higher RVUs in several diagnostic categories, including chest pain (0.36 higher RVU/patient, p<0.001), ear nose and throat diagnoses (0.43 higher RVU/patient, p=0.040), leg fractures (0.60 higher RVU/patient, p=0.027), heart (0.40 higher RVU/patient, p<0.001), psychiatry (0.29 higher RVU/patient, p=0.002), and respiratory diagnostic categories (0.31 higher RVU/patient, p<0.001). Only vision-related diagnoses led to less RVUs per patient when scribed (0.52 less RVU/patient, p=0.027). For adult encounters, a multivariable model revealed scribes positively impacted RVUs even after adjusting for patient sex, acuity, month and daily census. Scribed encounters generated 0.08 less RVUs per pediatric patient compared to non-scribed encounters (p=0.007). ESI category had no effect on RVUs for pediatric patients. Categories where scribed charts had higher RVUs than charts without a scribe included joint complaints (0.67 higher RVU/patient, p=0.034) and vomiting/diarrhea (0.30 higher RVU/patient, p=0.047). Categories where scribed charts had less RVUs per patient were in ear nose and throat diagnoses (0.18 less RVU/patient, p=0.025) and headache (0.53 less RVU/patient, p=0.023). Scribes had a positive impact on RVUs in adult patients, but not the pediatric population. Amongst adult patients, scribes led to higher RVUs in ESI level 2 and 3 adult patients but not ESI 4 and 5 patients, perhaps suggesting an inability of scribes to improve revenue capture on lower acuity patients.