Article, Infectious Diseases

A comparison of routine, opt-out HIV screening with the expected yield from physician-directed HIV testing in the ED

a b s t r a c t

Objectives: The Centers for Disease Control and Prevention recommends routine opt-out HIV screening in health care settings. Our goal was to evaluate the feasibility and yield of this strategy in the emergency department (ED) and to compare it to the expected yield of physician-directed testing.

Methods: This is a cross-sectional study in an urban ED during random shifts over 1 year. Patients were ineligible for screening if they were younger than 18 years or older than 64, a prisoner, a victim of sexual assault, in an ED resuscitation room, or had altered mental status. Research associates administered rapid HIV tests and conducted standardized interviews. The patients’ ED physician, blinded to the HIV result, was asked if they would have or- dered a rapid HIV test if it had been available.

Results: Of 7756 ED patients, 3957 (51%) were eligible for HIV screening, and 2811 (71%) of those did not opt out. Routine testing yielded 9 new HIV cases (0.32% of those tested; 95% confidence interval, 0.16%-0.63%). Physician- directed testing would have missed most of these infections: 2 of the 785 patients identified by physicians for testing would have been newly diagnosed with HIV (0.25%; 95% confidence interval, 0.04%-1.0%). Of the 9 new HIV cases, 5 established HIV care, and their median CD4 count was 201 cells/uL (range, 71-429 cells/uL).

Conclusions: Routine opt-out HIV screening was feasible and accepted by a majority of ED patients. The yield of

this strategy only modestly exceeded what may have been observed with physician-directed testing.

(C) 2014


Of the 1.1 million people living with HIV infection in the United States, it is estimated that 1 in 6 or 180 900 people are unaware of their infection [1]. Failure to identify these HIV-infected individuals and link them to preventive and clinical care is an important Public health problem. People who remain unaware of their HIV status are less likely to modify risky behaviors associated with HIV transmission [2]. Likewise, identifying seropositive individuals early and initiating Antiretroviral therapy (ideally when the CD4 cell count is 350-500 cells/mm3) may halt progression to AIDS and further prevent HIV trans- mission, a strategy termed treatment as prevention [3,4]. Ensuring access to HIV therapy for all newly diagnosed patients should decrease HIV-

? Source of support: The STD, HIV, and Tuberculosis Section of the Minnesota State Department of Health supplied the rapid HIV testing kits used in this study.

?? Presented at the Society for Academic Emergency Medicine Annual Meeting, Boston,

May 2011.

* Corresponding author. Department of Emergency Medicine, Hennepin County Medical Center, 701 Park Ave, Minneapolis, MN 55415. Tel.: +1 612 873 3196; fax: +1 612 904 4680.

E-mail address: [email protected] (M.E. Prekker).

related health disparities, a primary goal of the current US National HIV/AIDS Strategy [5].

For these reasons, the Centers for Disease Control and Prevention (CDC) and the US Preventive services Task Force recommend routine (nontargeted) opt-out HIV screening in health care settings where the prevalence of occult HIV infection is greater than 0.1% [6,7]. Routine opt-out screening is meant to remove barriers to HIV testing by informing all patients that an HIV test will be performed unless they de- cline, incorporating consent for HIV testing into the general consent for medical care, and eliminating the requirement for prevention counsel- ing. A streamlined approach to HIV screening, capitalizing on improve- ments in rapid HIV tests, is suited to a high-throughput health care setting like the emergency department (ED). Moreover, ED patients tend to have increased risk for HIV infection as compared to other gen- eral medical settings, and there are often missed opportunities to iden- tify HIV infection in the ED, especially during the year leading up to a diagnosis of HIV [8,9].

Early adopters have successfully implemented routine opt-out HIV screening programs in several academic EDs, yet no consensus has emerged regarding the optimal screening protocol [10-12]. Editorialists have noted the substantial resources required to initiate and sustain

0735-6757/(C) 2014

routine HIV screening and the relatively modest numbers of occult HIV infections found [13,14]. One of the first ED investigations of the com- parative effectiveness of routine rapid HIV screening vs diagnostic test- ing found a significant increase in the number of newly identified HIV infections with the former approach [15]. These findings have yet to be confirmed in a separate ED population.

In this study, we describe the results of a routine opt-out HIV screen- ing program in the ED of an urban, public hospital with no prior experi- ence with rapid HIV testing or screening. Our primary goal was to assess the feasibility of this strategy. Our secondary goal was to determine if the proportion of patients newly diagnosed with HIV infection is higher among those participating in routine opt-out screening or among the subset of these patients selected by ED physicians to undergo HIV testing (ie, physician-directed testing, which includes diagnostic testing and physician-initiated screening).


The institutional review board of the study institution approved this study with a waiver of written informed consent. The study was funded in part by the STD, HIV, and Tuberculosis Section of the Minnesota Department of Health.

Study design and setting

We used a cross-sectional study design to evaluate a program of rou- tine opt-out HIV screening among adult ED patients in an urban, public hospital. This screening program was not integrated into ED clinical operations; instead, trained research associates evaluated all patients for eligibility, performed point-of-care rapid HIV testing, and interpreted and delivered results to patients in parallel with clinical care. In addition, the patient’s emergency physician, who was blinded to the rapid HIV test result, was asked during the visit if they would have ordered a rapid HIV test for that study patient had such a test been available for clinical use. We compared the performance of routine (nontargeted) opt-out HIV screening to a targeted HIV testing strategy, using physician intention to order a rapid HIV test during the medical evaluation as a surrogate for physician-directed testing.

The study ED is part of an urban, public, safety net hospital and level I trauma center with 420 beds and 98 000 annual ED visits. The institution primarily serves the 1.1 million residents of Hennepin County, Minnesota, which includes the Minneapolis metropolitan area. A substantial propor- tion of patients seen in the ED are racial or Ethnic minorities and unin- sured. The hospital operates a multidisciplinary clinic dedicated to the care of HIV-infected individuals.

The ED supports a year-round research associate program staffed by undergraduate and graduate/professional student volunteers who are trained to screen and enroll patients in research studies. The research associates were trained in this study protocol and assisted 8 study physicians in carrying out routine HIV screening. HIV testing was not available in the ED before this study, except as performed by the clinical laboratory in cases of occupational exposure.

Study population

Patients between 18 and 64 years old who presented to the ED during the study period were potentially eligible to receive an HIV test. Patients were excluded from screening if they (1) spoke a language other than English, Spanish, or Somali; (2) were known to be HIV infected; (3) a prisoner or in police custody; (4) a victim of sexual assault; (5) placed in the ED resuscitation room, which is reserved for critically ill or injured patients; or (6) if an altered sensorium precluded their understanding of the opt-out testing process as judged by a re- search associate and/or study physician. A negative HIV test result at any time in the past or prior enrollment in this study did not exclude patients from participating. Patients could opt out of HIV testing for

any reason. For patients who did not opt out of screening, the attending emergency physician (n = 24) or senior emergency medicine resident (n = 11) involved in their care was asked to participate as a research subject by answering a brief survey to determine if they would have or- dered a rapid HIV test for screening or diagnostic purposes if the routine HIV screening program did not exist and if the ED clinical laboratory of- fered such a test (described in more detail below).

Study protocol

Routine opt-out HIV screening, based on guidelines from the Centers for Disease Control and Prevention [6], was carried out in parallel to ED clinical care between October 26, 2009, and October 19, 2010. All patients presenting to the ED during prespecified 2- or 4-hour blocks of time (336 hours in total) were evaluated for their eligibility to receive a rapid HIV test. These blocks were randomized in a 1:1:1 ratio to occur during day (7 AM to 3 PM), evening (3 PM to 11 PM), or night (11 PM to 7 AM) shifts to ensure a Representative sample of the general ED population was screened. Trained research associates approached eligible patients after they were placed in an ED treatment room to inform them that they would receive an HIV test during their visit unless they declined. Participation in HIV testing was voluntary and did not incur any cost to the patient. A 1-page summary of the Screening procedure was provided to eligible patients in their primary language. This document contained informa- tion on HIV infection, risk factors for acquiring HIV, the opt-out study procedure, and implications of a positive or negative rapid HIV test re- sult. Patients could decline HIV testing at any point during their ED visit by informing a research associate, and any questions about HIV testing were addressed before the test was administered. For patients not opting out, a study physician collected a fingerstick whole blood sample for the Clearview Complete HIV 1/2 assay (Inverness Medical, Princeton, NJ), and research associates interpreted the test result at the point of care under a certificate of waiver from the federal Clinical

Laboratory Improvements Amendments program.

After testing was completed and before the end of the ED encounter, the supervising ED physician, blinded to the rapid HIV test result, was asked if they would have ordered a rapid HIV test for diagnostic or screening purposes had such a test been available in the ED clinical lab- oratory at the time of the study. The theoretical decision to order an HIV test was at the discretion of the ED physician. No protocols, guidelines, or decision aids existed at the study institution regarding targeted test- ing of ED patients for HIV infection.

A research associate notified patients of a nonreactive HIV test result in person during their index ED visit, and no posttest counseling was provided. We considered a reactive rapid HIV test to be a preliminary positive result. All patients with reactive rapid HIV tests underwent ve- nipuncture by an ED nurse for conventional serum HIV Antibody testing and reflexive confirmatory Western blot. Indeterminate rapid HIV tests were repeated with a second rapid test in the ED. When a reactive rapid HIV test was encountered, members of the research team informed the patient of their preliminary positive status, performed brief posttest counseling, and gave the patient written HIV educational material along with a follow-up appointment in 7 days or less with an HIV coun- selor in the outpatient HIV clinic at the study institution. The supervis- ing ED physician was notified of any preliminary positive result, and a communicable disease report was sent to the Minnesota Department of Health in accordance with state law.

All eligible patients, including those who opted-out of HIV screening, were invited to participate in a brief, standardized interview with a re- search associate. Patient demographics, HIV risk factors, and prior expe- rience with HIV testing were collected during the interview.

Study outcomes

The primary outcome was the proportion of patients with newly di- agnosed HIV infection among all patients screened during the study

period. A new diagnosis of HIV infection was defined as a reactive rapid HIV test in the ED followed by a positive confirmatory Western blot in a patient not known to already be HIV infected based on personal history and a review of the medical record. Secondary outcomes were 3-fold:

No reason given N = 433

Recently tested N = 302 No risk for HIV infection N = 103 Not pertinent to ED visit N = 95 Already participated in study N = 85 Ready for discharge N = 63

Didn N = 52

Didn status N = 13

ED patient encounters during study period N = 7,756

Eligible for opt-out HIV screening N = 3,957


New diagnosis

N = 2

Age <18 or >64 years N = 1,750 Missed during clinical care N = 934 Altered sensorium N = 670 Placed in resuscitation room N = 159 Known HIV infection N = 151

Prisoner N = 57

Sexual assault N = 27 No exclusion reason given N = 51

New diagnosis

N = 7

Previously diagnosed* N = 1

False positive

N = 1

Screened for HIV infection

N = 2,811

Physician would have ordered HIV test

N = 785

Physician would not have ordered HIV test N = 1,514

Positive rapid HIV test N = 3

Positive rapid HIV test N = 8

Positive rapid HIV test N = 3

Physician not available to answer N = 512

Opted Out

Previously diagnosed* N = 1

Previously diagnosed* N = 2

(1) the theoretical yield of physician-directed HIV testing as compared to the observed yield of routine opt-out screening; (2) the acceptance of routine, opt-out HIV screening among eligible patients; and

(3) patient-reported satisfaction with the screening protocol.

Sample size assessment and data analysis

Based on prior research, we estimated that 1.8% of the general ED population at the study hospital has known HIV infection (unpublished data). Assuming an approximately 2% overall prevalence of HIV infected patients (diagnosed and undiagnosed), we calculated a priori that a sample size of 3012 patients not opting out of testing would provide 95% confidence intervals (CIs) of 0.5% around the overall HIV prevalence estimate. The data are summarized with median and interquartile range for continuous variables and with proportions for categorical variables. The unit of analysis was the patient encounter. All data were entered into an electronic database (Excel; Microsoft, Redmond, WA), and de- scriptive statistical analyses were carried out using Stata version 12 (Stata Corp, College Station, TX).


Among 7756 ED visits during random blocks of time that routine HIV screening was offered over 1 year, 3957 patients were eligible to partic- ipate, and 2811 patients (71% of those eligible) did not opt out and were tested; 14 (0.5%) had a positive rapid HIV test. One of these patients had a false-positive rapid HIV test based on confirmatory testing, and 4 others had previously been diagnosed with HIV infection and were re- ceiving outpatient care (but failed to disclose this when assessed for study eligibility). The remaining 9 patients (0.32% of all patients tested; 95% CI, 0.16%-0.63%) were newly diagnosed with HIV infection (Figure). Patient characteristics including demographic data, prior HIV test- ing, and select HIV risk factors are given in Table 1. One patient with a confirmed new HIV diagnosis presented to the ED for a problem poten- tially related to HIV infection (axillary Cutaneous abscess), whereas the other 8 patients presented for a variety of reasons: musculoskeletal low back pain (n = 2), inguinal hernia, shoulder pain, minor closed head in- jury, minor injuries from a motor vehicle crash, seizure in a patient with preexisting epilepsy, and alcohol intoxication (this patient was approached after regaining sobriety and did not opt out of screening). Following the index ED visit, 5 of 9 patients with newly identified HIV infection were successfully linked to outpatient HIV care; the median

CD4 cell count in that group was 201 cells/uL (range, 71-429 cells/uL). The potential yield of an alternate HIV testing strategy (ie, physician-

directed testing) was explored by asking the supervising ED physician at the end of the visit if they would have ordered a rapid HIV test if this test was available in the ED clinical laboratory. The response rate was 82% (ie, the treating physician was identified and available to an- swer this question in 2229 of 2811 visits). Physicians intended to test 785 study patients (34%) for HIV during their index ED visit; the ratio- nale for testing was “screening” in 613 patients (78%) (ie, targeted test- ing based on the patient’s risk profile) and “diagnostic” in 172 patients (22%) (ie, testing to determine if HIV infection was contributing to the patient’s presenting problem). Patient demographics and the distribu- tion of select HIV risk factors did not differ substantially between those patients who physicians would and would not have tested (Table 2). Physician-directed testing would have identified 2 of 9 new HIV diagnoses found during routine opt-out screening. The prevalence of new HIV infection potentially identified with physician-directed test- ing was 0.25% (95% CI, 0.04%-1.0%) or 2 in 785 patients.

Approximately one-third of ED patients eligible for HIV screening during the study period opted out for the reasons listed in Figure. A

Figure. Accrual of study patients and the subsequent results of HIV testing. ?”Previously diagnosed” refers to ED patients who did not disclose their HIV-seropositive status to research associates during assessment for eligibility to participate in ED HIV screening and, therefore, underwent rapid testing instead of being excluded from the study.

prior negative HIV test was cited as the reason for opting out by 26% of patients. In fact, two-thirds of patients who did not opt out of testing re- ported having a negative HIV test result in the past, including one-third (n = 605) who tested negative in the past 12 months. The distribution of age and sex did not differ between the patients who underwent HIV testing and those who opted out (mean +- SD age, 39 +- 12 years and 51% male in both groups). A greater proportion of eligible ED patients opted out of HIV testing during the evening shift (3 PM to 11 PM, 623 of 2004 patients [31%] opted out), as compared to the day shift (7 AM to 3

PM, 382 (27%) of 1396 patients opted out) and the night shift (11 PM to

7 AM, 141 [25%] of 557 patients opted out) (P = .008 by ?2 test).

Approximately 90% of ED patients who participated in routine opt- out HIV screening reported at the end of their visit that they were either “satisfied” or “very satisfied” with their experience (89%, or 2507 of 2811 patients). Of the 9 patients with newly diagnosed HIV infection, 1 patient each was “very satisfied” and “satisfied” with their testing experience; the remaining 7 patients did not answer.


Routine opt-out HIV screening using a rapid test is feasible in a high- volume ED and is able to identify patients previously unaware of their HIV infection. The nontargeted screening protocol was acceptable to pa- tients, as evidenced by a relatively low proportion of eligible patients opting out (29%), and among those not opting out, nearly all patients were satisfied with their HIV screening experience. Although the abso- lute yield was modest (9 new HIV diagnoses made among 2811 patients

Table 1

Characteristics of ED patients tested for HIV infection

Negative HIV

Positive HIV test: confirmed

Table 2

Select patient characteristics stratified by ED physician intent to order HIV testing

If a rapid HIV test were available, would


test,a n = 2798

new diagnosis, n = 9

you have ordered this test?

Age, median (IQR), y 39 (27-49) 43 (32-46)

Male sex, n (%) 1412 (50) 7 (78)

Race, n (%)

No Yes

Variable n = 1514 Diagnostic,

n = 172

Screening, n = 613

African American 1118 (40) 5 (56)


917 (33)

4 (44)

New HIV diagnoses, n





268 (10)


Age, median (IQR), y

40 (28-50)

33 (24-48)

39 (28-49)

Native American

234 (8)


Male gender, n (%)

764 (50)

78 (45)

318 (52)


200 (7)


Race, n (%)

None given

61 (2)


African American

591 (39)

83 (48)

250 (41)

Birthplace, n (%)


511 (34)

40 (23)

197 (32)

United States

2327 (83)

9 (100)


157 (10)

19 (11)

71 (12)


147 (5)


Native American

112 (7)

13 (8)

44 (7)

Central or South America

62 (2)



115 (8)

13 (8)

35 (6)


44 (2)


None given

28 (2)

4 (2)

16 (3)

Other/none given

218 (8)


Select risk factors for HIV infection, n (%)

HIV testing exposure, by

>=2 sexual partners in last year

319 (21)

55 (32)

166 (27)


Any history of intravenous drug use

111 (7)

19 (11)

74 (12)

Prior HIV test, n (%)

1883 (67)

7 (78)

Blood transfusion before 1985

95 (6)

14 (8)

37 (6)

Timing of prior HIV test, n (%)

Any history of sex with HIV+ person

67 (4)

11 (7)

33 (5)

b 1 mo

153 (8)

1 (14)

or IV drug user

2-6 mo

255 (14)

1 (14)

Any history of exchanging sex for

52 (3)

12 (7)

24 (4)

6-12 mo

192 (10)

3 (43)


1-3 y

580 (31)

1 (14)

or drugs

3-5 y

361 (19)

1 (14)

Men who have sex with mena

49 (6)

7 (9)

24 (8)

Select risk factors for HIV infection, n (%)

>=2 sexual partners in the past year

Any history of intravenous

677 (24)

255 (9)


1 (11)

An important goal of ED-based HIV screening is to link newly diag-

nosed patients with specialized, outpatient care for HIV infection. The cur- rent study did so in only 5 (56%) of 9 patients despite a referral by the

drug use

Blood transfusion before 1985 Any history of sex with HIV+

179 (6)

147 (5)

1 (11)


treating ED physician to the HIV clinic at the study institution and written material with a date and time for this appointment before discharge. Pro-

person or IV drug user

grams reporting successful linkage in upwards of 75% of newly diagnosed

Any history of exchanging sex

107 (4)


ED patients used a linkage-to-care specialist or social worker to facilitate

for money or drugs

the counseling and follow-up process [15,16]; however, the contribution

N 5 y

195 (10)



147 (8)


a Denominator for the proportion is the number of males in each group.

Men who have sex with menb 102 (7) 3 (43) Abbreviations: IQR, interquartile range; IV, intravenous.

a Negative HIV screening includes those with a negative rapid HIV test (n = 2797) and a false positive rapid HIV test (n = 1).

b Denominator for the proportion is the number of males in each group.

screened), this study describes 1 possible way to implement the CDC HIV screening recommendations and adds to a growing body of litera- ture in this area.

To date, 6 other studies have reported on results of routine opt-out HIV screening in the ED. This literature, including the current study, amounts to a total of 68881 ED patients screened for HIV across several US regions (Table 3) [10-12,15-17]. There is a wide range of screening offer rates (31%-89%) and screening acceptance rates (24%-91%) across these studies. It is possible that this variability is due to differences in where and in whom screening was carried out as well as how the screening process was incorporated into the ED workflow (Table 3). In the current study, HIV screening shared a physical space with, but was performed in parallel to, ED clinical care (ie, patients were approached after they had been roomed in the ED and those not opting out were tested at the bedside). This structure could have influenced the relative- ly high acceptance rate (71%) of this study, as compared to a program based in the ED triage area (31%) [11] or another initiated by nonclinical registration staff (24%) [15].

Despite differences in implementation, the prevalence of confirmed newly identified HIV infection was remarkably similar across studies using routine opt-out methodology, ranging from 0.2% to 0.6% of pa- tients tested (Table 3). This prevalence exceeds the CDC’s threshold to perform routine HIV screening in a health care setting (0.1%) as well as the threshold associated with cost effectiveness (0.05%) according to 1 study [18].

of a dedicated navigator was less robust at another institution [17].

Routine screening, synonymous with universal screening, aims to test all ED patients regardless of HIV risk profile or chief complaint to maximize the number of tests performed, thereby identifying the larg- est absolute number of HIV infected patients unaware of their status. The implicit tradeoff is that many ED patients will be tested who are at Very low risk for HIV infection. It has been suggested that screening in a resource-limited setting like the ED could be more efficiently and ef- fectively done using a risk-based (targeted) strategy [13,14]. In the cur- rent study, however, ED physicians would have tested only a third of patients who participated in screening but would have missed 7 of 9 occult HIV infections picked up with routine HIV screening. Similar find- ings were reported in an observational study comparing routine opt-in HIV screening with diagnostic testing [19]. In terms of the absolute yield of new HIV diagnoses, diagnostic HIV testing performed somewhat bet- ter as compared to routine opt-out screening in a quasiexperimental prospective study [15]. Importantly, both of these articles differ from the current study in that they used existing ED personnel to incorporate routine HIV screening into ED operations and both report a lower screening offer rate than the current study. Future multicenter compar- ative effectiveness research should follow up on the pragmatic benefits and drawbacks of non-risk-based (routine) vs risk-based (targeted) HIV screening [20].

If and when ED patients with a prior negative HIV test should be

rescreened is controversial. The CDC and US Preventive Services Task Force guidelines both cite insufficient evidence to inform an optimal screening interval. The current study speaks of the impact that the pres- ence and timing of a recent negative HIV test may have on future accep- tance of ED screening. A recent negative HIV test was the most common reason for opting out, given by 1 in 4 patients in the current study (Figure). Most ED patients (67%) screened for HIV reported having

Table 3

Existing literature on routine, opt-out HIV screening in the ED

Structure Process Outcome

Study Population and setting Integrated into ED


HIV test Screening offer rate

Screening acceptance rate

New HIV diagnosis rate

Follow-up Median CD4


Current study

Geren et al [16]

Sattin et al [12]

White et al [11]

Hoxhaj et al [17]

Haukoos et al [15]

Brown et al [10]

7756 patients, age N

18 y (40% African American, 10% Hispanic). Urban, public safety-net hospital and trauma center.

71556 patients, age N 18 y (48% Hispanic). Urban, public safety-net hospital ED in Phoenix, Arizona. 20612 patients,

ages 13-64 y (half are African American). Public, academic ED in Augusta, Georgia

serving 6 counties, a mix of urban and rural areas.

26757 patients, age N 15 y (half are African American, one-third Hispanic). Urban, teaching hospital and trauma center in Oakland, California.

24686 patients, age N 18 (44% Hispanic, one-third African American). Urban, academic hospital ED in Houston, Texas.

31,472 patients, age N

16 (one-third Hispanic,

14% African American). Urban, public safety net hospital and

trauma center in Denver, Colorado.

13240 patients, ages 13-64 (half are African

American, 4% Hispanic). Urban, teaching hospital ED in Washington, DC.

No. Trained research assistants performed testing and delivered results.

Yes. Triage nurses identified eligible patients who were undergoing venipuncture and tests performed in hospital lab.

No. Experienced HIV counselors performed counseling, testing, and result delivery.

Yes. Registration staff identified patients for screening and referred them to an HIV tester in triage area.

Yes. Staff nurses identified eligible patients who were undergoing venipuncture and tests performed

in hospital laboratory.

Yes. Registration staff identified patients for screening; nurses or techs draw blood that is sent

to hospital laboratory for testing.

No. Trained research assistants performed testing and delivered results.

Whole blood, fingerstick (Clearview Complete Rapid)

Whole blood (Architect HIV Ag/Ab combo)

Oral swab (OraQuick Rapid)

Oral swab (OraQuick Rapid)

Whole blood (ADVIA Centaur)

Whole blood, venipuncture (Uni-Gold)

Oral swab (OraQuick Rapid)

51% (n = 3957) 71% (n = 2811) 0.3% (n = 9) 5/9 linked to

ambulatory HIV care

78% (n = 55500) 39% (n = 27952) 0.3%a (n = 78) 64/78 linked to

ambulatory HIV care

45% (n = 9343) 91% (n = 8504) 0.4% (n = 35) 28/35 received

confirmatory results

76% (n = 20280) 31% (n = 6273) 0.4% (n = 21) 16/21 linked to

ambulatory HIV care

N/A 57% (n = 14093) 0.6% (n = 80) 60/80 received HIV

test results, and 34/80 linked to ambulatory HIV care

89% (n = 28043) 24% (n = 6762) 0.2% (n = 15) All 15 patients

linked to care

31% (n = 4187) 60% (n = 2486) 0.4% (n = 9) 13 patients (50%)

lost to follow-up








a The new HIV diagnosis rate in the study of Geren et al [16] includes both Acute infections (presence of HIV-1 p24 antigen, n = 18) and chronic infections (presence of HIV antibody, n = 60).

been tested previously (one-third of those in the past year), along with 7 of 9 newly diagnosed HIV-positive patients (5 of those reported a neg- ative test in the past year) (Table 1). Based on these data, ED patients appear willing to be screened for HIV more than once, and some much more frequently than that. Until more data are available, it may be pru- dent to omit a specific screening interval from the exclusion criteria of future studies.

Of special interest to others considering an ED-based routine HIV screening program, the process of implementing the current protocol in our ED led to important secondary benefits for the study institution. This project was conceived by a group of resident physicians in a com- bined emergency medicine and internal medicine training program with buy-in from a core group of faculty from emergency medicine and infectious diseases, at a time when the institution did not support rapid HIV testing for diagnosis or screening outside of occupational ex- posure. Anecdotally, as the study progressed with robust dissemination of rapid test results to ED patients and a very low false-positive rate,

much of the local bias on the part of clinicians or administrators about the stigma associated with HIV testing was dispelled. The program strengthened partnerships between departments inside and outside of the institution including with local public health agencies. As the ED HIV screening project concluded, rapid HIV testing was made available institution wide through the clinical laboratory.

This study should be interpreted in light of several limitations. First, ED physicians were not able to order a rapid HIV test for diagnostic or screening purposes during the study period, and their intent to test, as assessed by a face-to-face brief survey, was used as a surrogate for actual physician-directed HIV testing. Supervising ED physicians were not available to respond to this survey for 18% of study patients. However, our results are strengthened by complete physician-directed testing data for all newly diagnosed HIV-positive study patients, and all physi- cians were blinded to the result of rapid HIV screening when they responded to the survey. The intent to order a rapid HIV test may over- estimate actual physician test-ordering behavior, and therefore, our

comparison with routine opt-out screening may be conservative. Sec- ond, it is possible that individuals visited the ED and were screened more than once during the study period. We intentionally did not quan- tify repeated testing, as there was no exclusion for time interval since previous negative HIV test. The unit of analysis in this descriptive study is the ED visit, not the individual patient. Third, routine (ie, uni- versal) HIV screening was the study goal, but we recognize that 934 pa- tients (12% of the total ED census during the study period) were not assessed for eligibility due to conflicts with clinical care, imminent ED discharge at the beginning of a screening period, or deterioration in pa- tient condition. Finally, the study population was adult ED patients pre- senting to a single, urban, academic, safety net hospital, and the results may not be generalizable to other settings or populations.

In summary, implementation of a routine opt-out HIV screening pro- gram in an urban, academic ED without prior HIV testing experience was feasible and accepted by a majority of patients. The yield of new HIV diagnoses with a routine, opt-out strategy (0.32% of all those screened) was only modestly greater than what may have been ob- served with physician-directed testing. On the other hand, the absolute number of occult HIV infections identified was 4 times higher with rou- tine opt-out screening as compared to what may be observed with physician-directed testing. Although the decision to implement routine or targeted HIV screening in an ED must take into account local HIV prevalence and available institutional resources, we observed important and durable secondary benefits of beginning a routine opt-out rapid HIV screening program at our institution.


  1. Centers for Disease Control and Prevention (CDC). Monitoring selected national HIV prevention and care objectives by using HIV surveillance data–United States and 6 dependent areas.; 2011. [Accessed April 18, 2014].
  2. Marks G, Crepaz N, Janssen RS. Estimating sexual transmission of HIV from persons aware and unaware that they are infected with the virus in the USA. AIDS 2006; 20(10):1447-50.
  3. Department of Health and Human Services. Guidelines for the use of antiretroviral agents in HIV-1 infected adults and adolescents. ContentFiles/AdultandAdolescentGL.pdf. [Accessed April 18, 2014].
  4. Cohen MS, Chen YQ, McCauley M, Gamble T, Hosseinipour MC, Kumarasamy N, et al. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med 2011; 365(6):493-505.
  5. White House Office of National AIDS Policy. National HIV/AIDS Strategy for the United States.; 2010. [Accessed April 18, 2014].
  6. Branson BM, Handsfield HH, Lampe MA, Janssen RS, Taylor AW, Lyss SB, et al. Re- vised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings. MMWR Recomm Rep 2006;55(RR-14):1-17.
  7. Moyer VA. Screening for HIV: US Preventive Services Task Force recommendation statement. Ann Intern Med 2013;159(1):51-60.
  8. Jenkins TC, Gardner EM, Thrun MW, Cohn DL, Burman WJ. Risk-based human immu- nodeficiency virus (HIV) testing fails to detect the majority of HIV-infected persons in medical care settings. Sex Transm Dis 2006;33(5):329-33.
  9. Liddicoat RV, Horton NJ, Urban R, Maier E, Christiansen D, Samet JH. Assessing missed opportunities for HIV testing in medical settings. J Gen Intern Med 2004;19(4):349-56.
  10. Brown J, Shesser R, Simon G, Bahn M, Czarnogorski M, Kuo I, et al. Routine HIV screening in the emergency department using the new US Centers for Disease Con- trol and Prevention Guidelines: results from a high-prevalence area. J Acquir Im- mune Defic Syndr 2007;46(4):395-401.
  11. White DAE, Scribner AN, Vahidnia F, Dideum PJ, Gordon DM, Frazee BW, et al. HIV screening in an urban emergency department: comparison of screening using an opt-in versus an opt-out approach. Ann Emerg Med 2011;58(S):S89-95.
  12. Sattin RW, Wilde JA, Freeman AE, Miller KM, Dias JK. Rapid HIV testing in a south- eastern emergency department serving a semiurban-semirural adolescent and adult population. Ann Emerg Med 2011;58(S):S60-4.
  13. Haukoos JS. The impact of nontargeted HIV screening in emergency departments and the ongoing need for targeted strategies. Arch Intern Med 2012;172(1):20-2.
  14. Moran GJ, Talan DA. Process and models for HIV screening in the emergency depart- ment: can and should we do this? Ann Emerg Med 2011;58(S):S172-3.
  15. Haukoos JS, Hopkins E, Conroy AA, Silverman M, Byyny RL, Eisert S, et al. Routine opt-out rapid HIV screening and detection of HIV infection in emergency depart- ment patients. JAMA 2010;304(3):284-92.
  16. Geren KI, Lovecchio F, Knight J, Fromm R, Moore E, Tomlinson C, et al. Identification of acute HIV infection using fourth-generation testing in an opt-out emergency department screening program. Ann Emerg Med 2014;64(5):537-46.
  17. Hoxhaj S, Davila JA, Modi P, Kachalia N, Malone K, Ruggerio MC, et al. Using nonrapid HIV technology for routine, opt-out HIV screening in a high-volume urban emergen- cy department. Ann Emerg Med 2011;58(S):S79-84.
  18. Sanders GD, Bayoumi AM, Sundaram V, Bilir SP, Neukermans CP, Rydzak CE, et al. Cost-effectiveness of screening for HIV in the era of highly active antiretro- viral therapy. N Engl J Med 2005;352(6):570-85.
  19. White DAE, Scribner AN, Schulden JD, Branson BM, Heffelfinger JD. Results of a rapid HIV screening and diagnostic testing program in an urban emergency department. Ann Emerg Med 2009;54(1):56-64.
  20. Haukoos JS, Hopkins E, Bender B, Sasson C, Al-Tayyib AA, Thrun MW, et al. Compar- ison of enhanced targeted rapid HIV screening using the Denver HIV Risk Score to non-targeted rapid HIV screening in the emergency department. Ann Emerg Med 2013;61(3):353-61.