CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

Conclusion: These data are concordant with recent literature suggesting a higher prevalence of HCV infection in non-BC white males. That 40% of HCV testing occurred within the non-BC is encouraging, considering testing in this group is likely driven by direct ascertainment of risk. However, this also presents a barrier to implementing more targeted screening practices as risk-factors for HCV testing are often unstructured data and not yet searchable with current technology. Importantly, primary care providers are obtaining RNA tests and referring HCV Ab+ patients to care at high rates. Gaps appear in transitioning into specialty care. Work is currently underway to better identify risk associated with testing, and educate providers on HCV linkage and retention using best practices form BC initiatives. 544 STRATEGIES FOR UTILIZING EHR TO IMPROVE BIRTH COHORT TESTING IN THE HOSPITAL SETTING Lorlette Moir , Shuchin Shukla, Uriel Felsen, Jeffrey Weiss, Kimberly Cartmill, Jonathan M. Schwartz, Paul Meissner, Kimberly Yu, Yuming Ning, Alain H. Litwin Montefiore Med Cntr, Bronx, NY, USA Background: Hepatitis C is the leading cause of liver disease in the United States with a 2% prevalence nationally. In 2012 the Centers for Disease Control and Prevention (CDC) released updated recommendations for Hepatitis C testing and screening of persons born between the years of 1945-1965. In 2014 New York State (NYS) expanded on the CDC recommendations by introducing the New York State Hepatitis C Testing Law which mandates that all residents in this birth cohort should be offered Hepatitis C screening within the inpatient or outpatient setting. Accordingly, Montefiore Medical Center implemented birth cohort prompts into Electronic Health Record (EHR) platforms. We describe the impact of EHR prompts on identification of unique Hepatitis C positive patients in the hospital setting. Methods: Montefiore Medical Center initiated a birth cohort screening prompt across three inpatient settings in March 2015 to encourage providers to test eligible patients. Providers are prompted to test for HCV if the patient had not been previously tested for HCV Antibody. To assess the effect of this prompt, we compared the proportions of patients who received first Hepatitis C tests and the average number per month of newly identified positives before and after the implementation of the EHR prompt (January 2014 through February 2015 and March 2015 through May 2016). Two sample t-tests were used to assess the statistical significance of these observed differences. Results: From January 2014 through February 2015, an average of 5% of birth cohort patients received HCV testing. Following the implementation of the EHR prompt in March 2015 through May 2016, this increased to 29% (P<0.0001). Additionally, the average number of newly identified positives among patients who had not previously been tested rose from 12.2 per month in the pre-implementation period to 32.7 per month in the post-implementation period (P<0.0001). Conclusion: The Hepatitis C screening EHR prompts have had a demonstrable positive effect on the proportion of patients within the birth cohort who are tested, and on the number of Hepatitis C positive patients identified. Given these positive results, EHR implementation of birth cohort strategies should be standardized within inpatient settings. 545 IMPLEMENTATION OF AN EHR PROMPT REVEALS LOW ADHERENCE TO HCV TESTING RECOMMENDATIONS Alexander Geboy 1 , Whitney Nichols 1 , Chinyere Ukaegbu 1 , Stephen Fernandez 2 , Peter Basch 3 , Dawn Fishbein 1 1 MedStar Hlth Rsr Inst, Washington, DC, USA, 2 MedStar Hlth Rsr Inst, Hyattsville, MD, USA, 3 MedStar Hlth, Washington, DC, USA Background: The prevalence of hepatitis C virus (HCV) among the Birth Cohort (BC) born during 1945-1965 is five times higher than adults born in other years. Though there is a productive discussion about effective linkage to care strategies for this population, healthcare systems are failing to adequately screen eligible patients. Identifying system-wide gaps in adherence to federal screening recommendations is paramount to uncovering the full burden of disease and planning a course toward HCV Elimination. Methods: Beginning July 2015, MedStar Health (MSH) activated a clinical decision support (CDSS) Electronic Health Record (EHR) prompt. Eligible BC patients were neither previously HCV tested nor positive. The prompt was triggered at primary care visits only upon clicking the “View All Protocols” (VAP) button. It contained seven discrete, actionable options, each traceable and monitored to determine system-wide adherence to HCV BC testing recommendations. A qualitative analysis is presented. Results: Between 7/1/2015 and 6/30/2016, 77,575 patients were identified as eligible. Testing occurred at 133 primary care sites by 470 providers across MSH. Providers clicked the VAP button for 29,668 (38%) eligible patients seen, accessed the HCV CDSS prompt for 21,675 patients (28% of total denominator; 73% of clicked VAPs), and took an action within the prompt for 20,528 (26% of total; 95% of prompts accessed). Of these: 6,768 patients (9% of total denominator; 33% of prompt actions) were HCV tested, 4,426 patients (5%; 22%) were not screened [1807 (41%) declined, 39 (1%) had a history of HCV positivity, 1912 (43%) previously screened negative, 349 (8%) deferred, and for 319 (7%) it was not indicated]; there were 9,334 actions (45%) that were unaccountable, these were likely printing an HCV handout. There were 1,356 additional tests conducted outside of the CDSS prompt, for a total of 8,124 tests. Conclusion: Adherence to BC recommendations was low at approximately 11% (8,124/75,305). It is concerning that 62% of providers did not access the VAP. Next steps will provide targeted education to PCPs and a health maintenance dashboard in a new EHR; consideration will be given to implementing standing orders. Barriers to HCoC initiation are evident, and exemplify the observation that only 50% of those infected are ever tested. Creating and implementing new best practices with supporting policy changes are essential if Elimination of HCV is to be a realistic possibility. 546 DEVELOPMENT OF AN EMR-BASED ALGORITHM TO PLACE PATIENTS IN THE HCV CARE CASCADE Jason Zucker , Justin Aaron, Henry W. Evans, Peter Gordon, Michael T. Yin Columbia Univ, New York, NY, USA Background: Disease specific care cascades have become important public health and organizational tools to characterize gaps in care, drive process improvement, and target resources. Their construction is often laborious and to be useful for ongoing process improvement they need to be maintained. Hepatitis C virus (HCV) is the most common blood-borne infection in the United States and the leading cause of cirrhosis, hepatocellular carcinoma, and liver transplant. The advent of highly effective antiviral agents has the potential to end the HCV epidemic if effective and efficient care engagement could be realized. We sought to create an algorithm to electronically derive an individual’s location in the HCV care cascade from data available in the electronic medical record (EMR) from a single hospital. Methods: We included all new institutional diagnosis, defined as patients with a first HCV antibody positive (Ab) test from 2013 to 2015 with positive or no confirmatory testing. Patients with HCV Ab+ tests prior to 2013 and those with negative confirmatory testing were excluded. To create the cascade we identified 5 milestones including: positive HCV Ab, HCV RNA testing, linkage to care defined as an outpatient or inpatient visit with an infectious diseases or gastroenterology provider, prescribed treatment, and sustained virologic response (SVR). An algorithmwas developed to categorize patients into each stage of care. To evaluate accuracy we created a reference standard to replicate a clinician’s review of the chart. A single researcher without access to the algorithm performed the reference standard review on a random sample of 129 patients. Results: The algorithm identified 1225 patients with a new institutional diagnosis of HCV infection. The algorithm identified 711 (58%) patients with a detectable HCV RNA, 330 (27%) patients linked to care, 100 (8%) patients prescribed treatment, and 39 (3.2%) with SVR (Figure). The algorithm correctly categorized 117 of 129 (90%) patients compared to 126 of 129 (98%) for the reference standard. 6 of 12 (50%) errors identified were related to physician documentation of outside hospital records. Conclusion: Using commonly available data from an EMR, our algorithm has a high accuracy for placing individuals in the HCV care cascade, and identified significant gaps at each step of the care cascade at our institution. An electronic care cascade provides a method to readily measure and monitor performance in HCV treatment and care over time.

Poster and Themed Discussion Abstracts

CROI 2017 229