Identifying Areas for Improvement in the HIV Screening Process of a High-Prevalence Emergency Department

Abstract

Since 1993, the Centers for Disease Control recommendations for HIV testing were extended to include persons obtaining care in the emergency department (ED). Situated in Newark, New Jersey, the University Hospital (UH) ED serves a community with a greater than 2% HIV prevalence, and a recent study showed a UH ED HIV seroprevalence of 6.5%, of which 33% were unknown diagnoses. Electronic records for patients seen in the UH ED from October 1st, 2014, to February 28th, 2015, were obtained. Information was collected on demographics, ED diagnosis, triage time, and HIV testing. Random sampling of 500 patients was performed to identify those eligible for screening. Univariate and multivariate analysis was done to assess screening characteristics. Only 9% (8.8–9.3%) of patients eligible for screening were screened in the ED. Sixteen percent (15.7–16.6%) of those in the age group18–25 and 12% (11.6–12.3%) of those in the age group 26–35 were screened, whereas 8% (7.8–8.2%) of those in the age group 35–45 were screened. 19.6% (19–20.1%) of eligible patients in fast track were screened versus 1.7% (1.6–1.8%) in the main ED. Eighty-five percent of patients screened were triaged between 6 a.m. and 8 p.m. with 90% of all screening tests done by the HIV counseling, testing, and referral services. Due to the high prevalence of HIV, urban EDs play an integral public health role in the early identification and linkage to care of patients with HIV. By evaluating our current screening process, we identified opportunities to improve our screening process and reduce missed opportunities for diagnosis.

Introduction

Since 1993, the Centers for Disease Control (CDC) recommendations for routine HIV counseling and testing were extended to include persons obtaining healthcare in the emergency department (ED). In 2003, those recommendations were reinforced with the CDC revisions advising opt-out screening for patients in all healthcare settings and that persons at high risk for HIV be screened at least annually. They also advised lowering the barrier to screening by recommending that separate written consent for HIV testing should not be required and that prevention counseling should not be required as part of HIV screening programs.¹

Those recommendations were reinforced in 2013 when the United States Preventive Services Task Force (USPSTF) revised their previous recommendations and recommended routine testing for patients of age 15–65 years in any healthcare setting. While previously the USPTF had made no recommendations regarding routine screening, changes in the HIV treatment guidelines and our improved ability to halt the HIV progression to AIDS with early detection led the USPTF to make this recommendation with an A grade, indicating that there is a high certainty that the net benefit is substantial. In addition, the USPTF recommendations included a recommended screening interval of 1 year for patients, “known to be at risk of HIV infection, those who are actively engaged in high-risk behaviors, and those who live or receive medical care in a high-prevalence setting.” High prevalence was defined as an HIV seroprevalence of >1% in the community.² Those recommendations were further supported by the American College of Emergency Physicians policy statement accepting the important role the ED plays in HIV screening.³

Current CDC estimates are that about 1.2 million people have HIV and that 20% of them are unaware of their diagnosis.⁴ Situated in Newark, New Jersey, the University Hospital (UH) ED serves a high-prevalence HIV community. Data from December 2013 show that Essex County had 1240 cases of HIV per 100,000 people (1.2%).⁵ The city of Newark itself had ∼5767 people living with a diagnosis of HIV in a population of only 278,500 resulting in a prevalence of more than 2%.⁵

In the city of Newark, UH serves as the safety net hospital, providing care to underserved populations who are often at increased risk of HIV. While the overall prevalence in Newark is 2%, the prevalence in high-risk groups treated by UH is higher.⁶ In a published study of 2008 looking at de-identified blood specimens from the UH ED, they discovered ∼6.5% of all patients were HIV positive and that one-third was unaware of the diagnosis.⁷

An additional study reviewing new HIV infections seen in the UH Infectious Diseases practice indicated that 36% of newly infected individuals had a missed opportunity for screening with 45% of those missed opportunities occurring in the UH ED.⁸ This is consistent with data from other studies in the United States indicating that many newly diagnosed patients with HIV have missed opportunities for screening and potentially, as a result, present with late-stage disease.^9
–11 Routine HIV testing in the ED is seen as one way to significantly reduce late diagnosis and delays in treatment.¹²

Currently, UH has programs in place to provide counseling, testing, and referral services (CTRS) in the ED, 7 days a week, and linkage of care processes to ensure that newly diagnosed patients are linked to care with an HIV provider within 1 day. However, despite these services, patients with HIV are still being missed. What was not in place was a standardized process for the CTRS to screen or assess patients as well as electronic decision support systems that could help prompt providers to initiate testing. The goal of this study was to assess the current gap in HIV screening in our ED to develop interventions to enhance our rate of screening and reduce missed opportunities for diagnosis and linkage to care.

Methods

Electronic records for patients seen in the UH ED from October 1st, 2014, to February 28th, 2015, were obtained. Information for each visit was electronically collected on age, gender, ED diagnosis, ED acute diagnosis ICD9 codes, care location within the ED, triage date and time, disposition, and if they were tested for HIV during that encounter. Excluded patients included those younger than 18 years of age, patients whose charts were entered in error, and patients who walked out before being triaged.

Patients who were tested for HIV in the ED and patients with HIV as part of their ICD9 diagnosis were electronically identified. Random sampling of 500 patient medical records without known HIV, encompassing 1374 visits, was conducted to determine the percentage of patients with HIV not documented in the ED diagnosis, those screened in the past 12 months, and those patients who were eligible for screening. Patients eligible for screening were defined as patients without a diagnosis of HIV who did not have documented testing in the prior 12 months (Fig. 1).

FIG. 1.

All patients with emergency department (ED) visits from October 2014 to February 2015 and older than 18 years of age, who made it through triage, were included in this analysis. This included 27,300 visits from more than 19,500 unique patients. Patients who were screened for HIV or had known HIV by diagnosis code were separated out electronically. A random sample of 500 patients, encompassing 1374 visits, who were not screened in the ED were then manually reviewed to determine the estimated number of eligible visits for screening.

Univariate and multivariate analysis was done to assess the percentage of eligible patients screened by age, gender, care location, ED diagnosis code, disposition, and time of day. This project was approved by the Rutgers Biomedical and Health Sciences Institutional Review Board.

Results

There were a total of 27,312 eligible visits in the ED between October 1st, 2014, and February 28th, 2015. Major care, fast track, and trauma were responsible for 91% of total ED patients seen. Overall, the ED population was 49% male and 51% female. However, major care had a slight majority of males, and fast track had a slight majority of females.

Only 9% (95% CI 8.8–9.3%) of eligible patients were screened in the ED. Overall, 7.7% (95% CI 7.5–8%) of eligible males and 10.2% (95% CI 10–10.5%) of eligible females were screened. Sixteen percent (95% CI 15.7–16.6%) of those in the age group 18–25 and 12% (95% CI 11.6–12.3%) of those in the age group 26–35 were screened, whereas only 8% (95% CI 7.8–8.2%) of those in the age group 35–45 were screened, with a roughly 2% higher female screening rate in each age group (Fig. 2).

FIG. 2.

Eligible patients who were screened for HIV were grouped by age and gender. Younger patients and female patients were more likely to be screened than older male patients. However, the majority of new diagnoses seen in the infectious diseases practice were males with an average age of 37.

Discharged patients (12%, 95% CI 11.6–12.3%) and patients who left against medical advice (8.4%, 95% CI 8.2–8.6%) were significantly more likely to be screened than those transferred to observation (3.8%, 95% CI 3.7–3.9%) or admitted (3%, 95% CI 2.9–3.1%) (Fig. 3).

FIG. 3.

Eligible patients who were screened for HIV were grouped by disposition. Discharged patients and patients who left against medical advice were significantly more likely to be screened than those transferred to observation or admitted.

Twenty percent of eligible patients in fast track were screened versus 1.7% (1.6–1.8%) in major care. Eighty-five percent of all patients screened in fast track were triaged between 6:00 a.m. and 8:00 p.m., compared with 71% of all screening in major care being done at that time. Peak screening occurred on patients triaged from 9:00 a.m. to 3:00 p.m. (Fig. 4). On weekends, screening was less likely in patients triaged after 3:00 p.m. than when compared with weekdays (Fig. 5).

FIG. 4.

Eligible patients who were screened for HIV were grouped by triage time and location of care. The majority of patient screening in fast track occurred from 8 a.m. to 5 p.m. (87% of all testing) with screening rates peaking at 25.2% between 10 and 11 a.m. While major care had a significantly lower rate of screening, they had a much more even distribution of screening throughout the day. Overall, the majority of patients were screened in fast track during hours when the HIV counseling, testing, and referral service (CTRS) was available. Screening dropped off markedly overnight when the CTRS was unavailable.

FIG. 5.

Eligible patients who were screened for HIV were grouped by triage time and day of the week (Weekday, Monday–Friday). The emergency department has dedicated HIV screeners available from 8:30 a.m. to 10:00 p.m. with two screeners from 11:30 a.m. to 7:30 p.m. on weekdays and one screener from 9:00 a.m. to 5:00 p.m. on weekends. The majority of patients on both weekdays and weekends were screened when the HIV counseling, testing, and referral service (CTRS) was available. Screening dropped off earlier on the weekends (5:00–10:00 p.m.) when the CTRS had shorter hours.

Ninety percent of patients were screened through a rapid HIV test, while only 10% of patients were screened using laboratory-based methods. Individuals with a diagnosis suggestive of increased risk of HIV infection were still screened at rates less than 50% (Table 1).

Table 1.

Selected Risk Factors for HIV and Percentage of Patients Screened

ICD9	EMR description	Number screened	Percentage screened (%)
599	URIN TRACT INFECTION NOS	111	12
788.1	DYSURIA	62	18
292	DRUG WITHDRAWL	3	3
V72.41	PREGNANCY TEST NEGATIVE	13	10
V74.5	SCREEN FOR VENEREAL DIS	178	41
V01.6	VENEREAL DIS CONTACT	55	36
V65.45	CONSL OT SEX TRNSMT DIS	18	33
V73.98	SCRN UNSPCF CHLMYD DIS	130	28
625.9	FEM GENITAL SYMPTOMS NOS	44	23

EMR, electronic medical record.

Of the 1711 patients screened, six were positive (0.3%). Three of the six positives were identified in fast track by point of care rapid testing provided by the CTRS, while the other three were identified in the main ED through serum testing. All of the patients testing HIV positive were triaged between 7:30 a.m. and 6:30 p.m.

Discussion

Our study showed that despite ready availability of HIV testing, a very small proportion of patients receiving care in the ED were tested. Assessing HIV screening trends in the ED is an important first step in the development of practical interventions to identify HIV-positive individuals. With a renewed emphasis on treatment as prevention, identification through screening is the critical first step in enrolling people in the cascade of care. Within the United States, Newark, New Jersey, is disproportionally affected due to a socioeconomic and historic factors, and explicit research within this community is needed to help end the epidemic.⁶

Despite evidence that rapid HIV testing in the ED is an effective way to find previously undiagnosed patients and link them to care,¹³ there remains a substantial number of patients in high-prevalence communities not offered HIV screening.¹⁴ ED screening remains a fundamental part of effective HIV screening programs, as for many people with HIV, the ED is their only interaction with the healthcare system.^15,16 Individuals within these communities may even expect to be screened,¹⁷ or even more worrisome, falsely assume that they have been tested for HIV in the ED when they have not.¹⁸

Given the resources the HIV CTRS dedicated to testing and the presumed provider knowledge of HIV prevalence in Newark, the 9% overall screening rate of eligible individuals was shocking. Newark suffers from a common issue in healthcare of limited resources and, even with a dedicated CTRS program operating at maximal rate, it cannot screen all eligible individuals. The CTRS was available on weekdays from 8:30 a.m. to 10:00 p.m. with two screeners available from 11:30 a.m. to 8:30 p.m. with one screener available on the weekends from 9:00 a.m. to 5:00 p.m.

This resource constraint is best demonstrated by our data indicating that 85% of patients screened were triaged between 6:00 a.m. and 8:00 p.m. and more than 90% were screened with the rapid diagnostics provided by the CTRS. This was also reflected in the weekend data. Previous studies have demonstrated that routine rapid HIV testing was accomplished more frequently by dedicated HIV counselors than by ED staff.¹⁹ However, these studies found that counselors increased screening rates by being more successful in convincing patients to agree to screening. This is a time-intensive process and can limit the absolute number of people they can screen.

Females and younger individuals were found to be screened at statistically significantly higher rates. With adults aged 35–44 screened at 8%, those aged 45–54 were screened at 6.5%, and those aged 55–65 at 3.6%. The trend of screening females at higher rates is similar to that seen in Arizona where an electronic medical record (EMR) was utilized to improve HIV screening. Interestingly, the screening rate by age was dramatically different in Arizona where while those aged >65 had the lowest rates of screening, those aged 18–24 had the next lowest rates, and those aged 45–64 were screened at almost 40%.²⁰ Most worrisome about the declining HIV screening with age seen in our study is that the average age of a new diagnosis presenting to our infectious diseases practice was 37 years,⁸ and a study from NYC suggested that adults older than 50 years of age may represent as much as 20% of new diagnosis.²¹

Discharged patients and patients leaving against medical advice had higher rates of screening than those sent to observation or admitted to the hospital. The disparity in discharged patients may be explained by the fact that the majority of screened individuals were screened in fast track and not in the main ED. In addition, it is also possible that those individuals are more likely to have been healthier and, therefore, more accessible to the HIV CTRS staff. While the CTRS was free to roam the ED and initiate screening, informal discussions with them indicated that they found screening people in fast track more efficient.

In the main ED, if the CTRS was not roaming and screening independently, then testing was physician directed. Studies have demonstrated that physician-directed testing misses a high number of HIV infections.²² In addition, UH is an academic medical center and major care is staffed by both residents and attending physicians. Residents have demonstrated limited knowledge of HIV screening guidelines and limited knowledge of linkage to care programs. Common barriers to resident testing were noted to be competing priorities and forgetting to order the test.²³ However, even surveys of ED physicians have indicated that there are numerous competing factors affecting their ability to provide HIV testing.²⁴

Most alarming in our analysis was the finding that individuals with another diagnosis suggestive of increased risk of HIV infections were screened at rates less than 50% (Table 1). Lamentably, this is not different than seen in other studies, including in primary care settings²⁵ and actually better than that seen in another academic ED.²⁶

Numerous ways to improve HIV screening rates have been demonstrated across the country, including utilizing decision support systems with electronic alerts²⁷ or hard stop alerts,²⁴ using the electronic medical record to preorder tests,²⁸ kiosk systems²⁹ or videos³⁰ for patient education, and utilizing patient care technicians for rapid testing.³¹

One intervention currently under consideration is combining CTRS with nurse-led triage initiated routine opt-out testing for those in the main ED. Routine opt-out HIV screening has demonstrated the highest rates of HIV screening ranging from 4% to 71% in eligible patients^22,32
–34 and has been demonstrated to identify more cases of HIV than targeted screening.³⁵ In addition, this reduces the burden on the physicians and residents.

One consideration is that the primary issue with opt-out screening has been the high rate of individuals electing to opt-out in some studies. One of the advantages of the current counselor-based system is higher rates of consenting to HIV testing. A potential combination approach of opt-out screening with CTRS or other services available to those who decline screening may help dramatically improve overall screening rates. Finally, once any new plan is implemented, there remains the difficulty in maintaining a system over the long term.³⁶

One limitation of this study is that due to the large number of patients a random sample was done. The data are presented in the Results section with confidence intervals and statistical significance held even when using conservative assumptions. However, the usual caveats to data acquired through sampling apply.

The major limitation of this study is that we were unable to track the rate of people who were offered and declined screening. While experience and informal discussions with CTRS indicated a low rate of refusal, an earlier study done in the UH ED demonstrated agreement rates of 48% for screening.³⁷ Other studies have demonstrated patient refusal rates as high as 88.5%³⁰ with females and older people as being more likely to refuse.³⁸ This limitation is particularly concerning, because while most patients refuse testing because they believe they are at low risk, a study from Washington D.C. indicates that the rate of HIV-positive individuals who declined testing was 2.74 times higher in those who were tested.³⁹ Due to the high prevalence of HIV in the Newark community, the UH ED plays an integral public health role in the early identification and linkage to care of patients with HIV. With less than 10% of all ED visits resulting in proper HIV screening, our current screening processes have resulted in a substantial number of missed opportunities for screening. Using a protocol-driven or decision support process to ensure that all patients are offered screening is the likely next step to reduce the missed opportunities for diagnosis.

Footnotes

Author Disclosure Statement

Dr. Swaminathan served on an advisory board and receives grant support from Gilead Services. Dr. Zucker is supported by an NIH T32 Training Grant.

References

Branson

, Handsfield

, Lampe

, Janssen

, Taylor

, et al. Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings. MMWR Recomm Rep, 2006 Sep 22; 55(RR-14):1–17; quiz CE1-4. PubMed PMID: 16988643.

Moyer

. Screening for HIV: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med [Internet]. 2013; 159(1):51–60. Available at: http://www.ncbi.nlm.nih.gov/pubmed/23698354.

HIV testing and screening in the emergency department. Policy statement. Ann Emerg Med, 2014; 64:563.

HIV in the United States | Statistics Overview | Statistics Center | HIV/AIDS | CDC. Available at: http://www.cdc.gov/hiv/statistics/overview/ataglance.html (Last accessed February 17, 2016 ).

NJDOH—HIV/AIDS, STD and TB Services—County and Municipal Statistics. Available at: http://www.state.nj.us/health/aids/repa/aidsdata.shtml (Last accessed February 17, 2016 ).

El-Sadr

, Mayer

, Hodder

. AIDS in America—forgotten but not gone. N Engl J Med, 2010 Mar 18; 362(11):967–970. PubMed PMID: 20147707; NIHMSID: NIHMS720631; PubMed Central PMCID: PMC4568988.

Mohammed

, Martin

, Sadashige

, Jaker

, Paul

. An anonymous unlinked sero-prevalence survey of HIVHCV in an urban emergency department. J Clin Virol, 2013; 58 Suppl 1:e19–e23.

Derose

, Zucker

, Cennimo

, Swaminathan

. 376 Rutgers—New Jersey Medical School HIV screening compliance—Are we missing opportunities?. Open Forum Infect Dis, 2015; 2(suppl 1):S149.

Lascar

, Freer

, Phiri

. Tailoring HIV testing in a setting of late HIV diagnosis: Is the tide turning?. Br J Hosp Med October 2015; 76:592–595.

10.

Nakao

, Wiener

, Newman

, Sharp

, Egan

. Falling through the cracks? Missed opportunities for earlier HIV diagnosis in a New York City Hospital. Int J STD AIDS, 2014; 25:887–893.

11.

Chin

, Hicks

, Samsa

, McKellar

. Diagnosing HIV infection in primary care settings: Missed opportunities. AIDS Patient Care STDS, 2013; 27:392–397.

12.

Hempling

, Zielicka-Hardy

, Ellis

, Majewska

, Fida

. Routine HIV testing in the emergency department: Feasible and acceptable?. Int J STD AIDS, 2015:pii: 0956462415613727.

13.

Richey

, Carpenter

, Barbeau

, Hadi

. Rapid HIV testing in a New Orleans emergency department is effective in identifying new HIV diagnoses and in linking patients to care. J La State Med Soc, 2014; 166:28–33.

14.

Hsieh

Y-H

, Kelen

, Beck

, et al. Evaluation of hidden HIV infections in an urban ED with a rapid HIV screening program. Am J Emerg Med, 2015; 34:180–184.

15.

Signer

, Peterson

, Hsieh

Y-H

, et al. Scaling up HIV testing in an academic emergency department: An integrated testing model with rapid fourth-generation and point-of-care testing. Public Health Rep, 131 Suppl:82–89.

16.

Lin

, Dietz

, Rodriguez

, et al. Routine HIV screening in two health-care settings—New York City and New Orleans, 2011–2013. MMWR Morb Mortal Wkly Rep, 2014; 63:537–541.

17.

McAfee

, Tung

, Espinosa-Silva

, et al. A survey of a small sample of emergency department and admitted patients asking whether they expect to be tested for HIV routinely. J Int Assoc Provid AIDS Care, 2013; 12:247–252.

18.

Khakoo

, Lindsell

, Hart

, Ruffner

, Wayne

, Lyons

. Patient perception of whether an HIV test was provided during the emergency department encounter. J Int Assoc Provid AIDS Care, 2014; 13:506–510.

19.

Walensky

, Reichmann

, Arbelaez

, et al. Counselor- versus provider-based HIV screening in the emergency department: Results from the universal screening for HIV infection in the emergency room (USHER) randomized controlled trial. Ann Emerg Med, 2011; 58(1 Suppl 1):S126–S132.e1–e4.

20.

McGuire

, Moore

. Using a configurable EMR and decision support tools to promote process integration for routine HIV screening in the emergency department. J Am Med Inform Assoc, 2016; 23:396–401.

21.

Ellman

, Sexton

, Warshafsky

, Sobieszczyk

, Morrison

EAB

. A forgotten population: Older adults with newly diagnosed HIV. AIDS Patient Care STDS, 2014; 28:530–536.

22.

Prekker

, Gary

, Patel

, et al. A comparison of routine, opt-out HIV screening with the expected yield from physician-directed HIV testing in the ED. Am J Emerg Med, 2015; 33:506–511.

23.

Bares

, Steinbeck

, Bence

, et al. Knowledge, attitudes, and ordering patterns for routine HIV screening among resident physicians at an urban medical center. J Int Assoc Provid AIDS Care, 2014:pii: 2325957414554006.

24.

Schnall

, Sperling

, Liu

, Green

, Clark

, Vawdrey

. The effect of an electronic “hard-stop” alert on HIV testing rates in the emergency department. Stud Health Technol Inform, 2013; 192:432–436.

25.

Joore

, Reukers

DFM

, Donker

, et al. Missed opportunities to offer HIV tests to high-risk groups during general practitioners' STI-related consultations: An observational study. BMJ Open, 2016; 6:e009194.

26.

Klein

, Martin

IBK

, Quinlivan

, Gay

, Leone

. Missed opportunities for concurrent HIV-STD testing in an academic emergency department. Public Health Rep, 129 Suppl 1:12–20.

27.

Schnall

, Liu

, Sperling

, Green

, Clark

, Vawdrey

. An electronic alert for HIV screening in the emergency department increases screening but not the diagnosis of HIV. Appl Clin Inform, 2014; 5:299–312.

28.

Medford-Davis

, Yang

, Pasalar

, et al. Unintended adverse consequences of electronic health record introduction to a mature universal HIV screening program. AIDS Care, 2016; 28:566–573.

29.

Rothman

, Gauvey-Kern

, Woodfield

, et al. Streamlining HIV testing in the emergency department-leveraging kiosks to provide true universal screening: A usability study. Telemed J E Health, 2014; 20:122–127.

30.

Aronson

, Marsch

, Rajan

, Koken

, Bania

. Computer-based video to increase HIV testing among emergency department patients who decline. AIDS Behav, 2015; 19:516–522.

31.

Egan

, Nakao

, VanLeer

, Pati

, Sharp

, Wiener

. Increased rates of rapid point-of-care HIV testing using patient care technicians to perform tests in the ED. Am J Emerg Med, 2014; 32:651–654.

32.

Haukoos

, Hopkins

, Conroy

, et al. Routine opt-out rapid HIV screening and detection of HIV infection in emergency department patients. JAMA, 2010; 304:284–292.

33.

Montoy

JCC

, Dow

, Kaplan

. Patient choice in opt-in, active choice, and opt-out HIV screening: Randomized clinical trial. BMJ, 2016; 532:h6895.

34.

Haukoos

, Hopkins

, Bucossi

. Routine opt-out HIV screening: More evidence in support of alternative approaches?. Sex Transm Dis, 2014; 41:403–406.

35.

Lyons

, Lindsell

, Ruffner

, et al. Randomized comparison of universal and targeted HIV screening in the emergency department. J Acquir Immune Defic Syndr, 2013; 64:315–323.

36.

Rayment

, Rae

, Ghooloo

, et al. Routine HIV testing in the emergency department: Tough lessons in sustainability. HIV Med, 2013; 14 Suppl 3:6–9.

37.

Martin

, Salaru

, Mohammed

, Coombs

, Paul

, Cadoff

. Finding those at risk: Acute HIV infection in Newark, NJ. J Clin Virol, 2013; 58 Suppl 1:e24–e28.

38.

Setse

, Maxwell

. Correlates of HIV testing refusal among emergency department patients in the opt-out testing era. AIDS Behav, 2014; 18:966–971.

39.

Czarnogorski

, Brown

, Lee

, et al. The prevalence of undiagnosed HIV infection in those who decline HIV screening in an urban emergency department. AIDS Res Treat, 2011; 2011:879065.