Linkage to Care Methods and Rates in U.S. Emergency Department-Based HIV Testing Programs – A Systematic Literature Review Brief Report

Abstract

Background

An increasing number of U.S. emergency departments (EDs) have implemented ED-based HIV testing programs since the Centers for Disease Control and Prevention issued revised HIV testing recommendations for clinical settings in 2006. In 2010, the National HIV/AIDS Strategy (NHAS) set an LTC rate goal of 85% within 90 days of HIV diagnosis. Linkage to care (LTC) rates for newly diagnosed HIV-infected patients vary markedly by site, and many are suboptimal. The optimal approach for LTC in the ED setting remains unknown.

Objective

To perform a brief descriptive analysis of the LTC methods practiced in EDs across the United States in order to determine the overall linkage rate of ED-based HIV testing programs.

Methods

We conducted a systematic review of literature related to U.S. ED HIV testing in the adult population using PubMed, Embase, Web of Science, Scopus and Cochrane. There were 333 articles were identified; 31 articles were selected after a multiphasic screening process. We analyzed data from the 31 articles to assess LTC methods and rates. LTC methods that involved physical escort of the newly diagnosed patient to an HIV/infectious disease (ID) clinic or interaction with a specialist health care provider at the ED were operationally defined as ‘intensive” LTC protocol. “Mixed” LTC protocol was defined as a program that employed intensive linkage only part of the coverage hours. All other forms of linkage was defined as “non-intensive” LTC protocol. An LTC rate of ≥85% was used to identify characteristics of ED-based HIV testing program associated with a higher LTC rate.

Results

There were 37 ED-based HIV testing programs in the 31 articles. The overall LTC rate was 74.4%. Regarding type of protocol, 9 (24.3%) employed intensive LTC protocols, 25 (67.6%) non-intensive, 2 (5.4%) mixed, and 1 (2.7%) had unclear protocols. LTC rates for programs with intensive and non-intensive LTC protocols were 80.0% and 72.7%, respectively. Four (44.4%) with intensive protocols and nine (36.0%) with the non-intensive protocols had LTC rates >85%. The linkage staff employed was different between ED programs. Among them, 25 (67.6%) programs used exogenous staff, ten (27.0%) used the ED staff, and 2 had no information. All the programs in the non-intensive group utilized drop-in HIV/ID clinic or medical appointments while seven out of nine of the programs in the intensive group physically escorted the patients to the initial medical intake appointment. There were no significant differences in characteristics of ED-based HIV testing programs between those with ≥85% LTC rate versus those with <85% within the intensive or non-intensive group.

Conclusion

Intensive LTC protocols had a higher LTC rate and a higher proportion of programs that surpassed the >85% NHAS goal as compared to non-intensive methods, suggesting that, when possible, ED-based HIV testing programs should adopt intensive LTC strategies to improve LTC outcomes. However, intensive LTC protocols most often required involvement of multidisciplinary non-ED professionals and external research funding. Our findings provide a foundation for developing best practices for ED-based HIV LTC programs.

INTRODUCTION

Background

The Centers for Disease Control and Prevention (CDC) key strategies to control the HIV epidemic include routinizing HIV testing and linking newly diagnosed patients to care in healthcare settings, including emergency departments (ED).¹ In 2009, the national linkage to care (LTC) rate was estimated at 66%.⁴ In 2010, the National HIV/AIDS Strategy (NHAS) set an LTC rate goal of 85% within 90 days of HIV diagnosis.⁵ While there exists a need for knowledge regarding current trends and optimal strategies for linking newly diagnosed patients to care, the literature on LTC of the ED-based screening program is relatively sparse and comprehensive information regarding LTC methods is lacking. An up-to-date and detailed synthesis of the available literature assessing U.S. ED-based LTC practices is needed. Identification of effective LTC methods through systematic analysis of existing ED-based HIV testing programs will provide valuable guidance for developing best practices for ED-based HIV LTC programs.

We performed a systematic review of studies that reported LTC data from U.S. adult ED HIV testing programs. The study aimed to (1) summarize and compare types of LTC protocols and (2) determine whether specific LTC methods were more effective than others.

METHODS

Study selection and Data Abstraction

We conducted a comprehensive search on Pubmed, Embase, Web of Science, Scopus and Cochrane databases for studies published as of June 2013, using search terms related to HIV, linkage to care, and emergency department (Appendix 1). Only full text articles and conference abstracts were selected. Studies that reported data in commentaries or editorials, contained pediatric data, had no LTC data, or had overlapped data from the same study site were excluded. The study selection process and the result are presented in Appendix 1.

Data Items and Definitions

Each program’s LTC data from the final selected articles were independently extracted both quantitatively and qualitatively and verified through discussion and consensus with the senior author. The key data included are outlined in Table 1.

Table 1.

Description of Included Studies.

First Author, Year	Journal	City	Study Period	ED Volume	Academic ED
Kelen et al. 19967	Ann Emerg Med	Baltimore, MD	Jun – Aug 1992	50,000 visits	Yes
Kelen et al. 19998 Subset A	Ann Emerg Med	Baltimore, MD	1993 – 1995	50,000 visits	Yes
Kelen et al. 19998 Subset B	Ann Emerg Med	Baltimore, MD	1993 – 1995	50,000 visits	Yes
Glick et al. 20049	AIDS Educ Prev	Chicago, IL	Jan 2001 – Jan 2002	44,000 visits	Yes
Kendrick et al. 200410	AIDS Res Let	Chicago, IL	Not available	120,000 visits	Yes
Lyons et al. 200511	Ann Emerg Med	Cincinnati, OH	1998 – 2002	80,000 visits	Yes
MMWR June 200712 Subset LA	MMWR Morb Mortal Wkly Rep	Los Angeles, CA	Jan – Dec 2005	47,736 visits	Yes
MMWR June 200712 Subset NY	MMWR Morb Mortal Wkly Rep	New York, NY	Jan – Dec 2005	72,948 visits	No
Brown et al. 200713	J Acquir Immune Defic Syndr	Washington, DC	Sep – Dec 2006	not provided	Yes
Silva et al. 200714	Ann Emerg Med	Chicago, IL	Apr 2003 – Aug 2004	44,000 visits	Yes
Lyss et al. 200715	J Acquir Immune Defic Syndr	Chicago, IL	Jan 2003 – Apr 2004	not provided	Yes
Haukoos et al. 200716	Acad Emerg Med	Denver, CO	Oct 2004 – Mar 2007	55,000 visits	Yes
Mehta et al. 200717	Acad Emerg Med	Boston, MA	Nov 2003 – May 2004	77,000 visits	Yes
White et al. 200918	Ann Emerg Med	Oakland, CA	Apr 2005 – Dec 2006	75,000 visits	Yes
Christopoulos et al. 201019	J Acquir Immune Defic Syndr	New York, NY	Jan 2006 – Dec 2007	69,398 visits	Yes
Haukoos et al. 201020	JAMA	Denver, CO	Apr 2007 – Apr 2009	55,000 visits	Yes
Torres et al. 201121 Site B	Ann Emerg Med	Midwest Region	Jun 2007 – Apr 2008	87,626 visits	Yes
Torres et al. 201121 Site C	Ann Emerg Med	South Region	Jun 2007 – Apr 2008	27,913 visits	Yes
Torres et al. 201121 Site D	Ann Emerg Med	Midwest Region	Jun 2007 – Apr 2008	89,615 visits	Yes
Torres et al. 201121 Site E	Ann Emerg Med	West Region	Jun 2007 – Apr 2008	52,528 visits	Yes
Torres et al. 201121 Site F	Ann Emerg Med	Northeast Region	Jun 2007 – Apr 2008	52,542 visits	Yes
Sattin et al. 201122	Ann Emerg Med	Augusta, GA	Mar 2008 – Aug 2009	77,000 visits	Yes
Schrantz et al. 201123	Ann Emerg Med	Chicago, IL	Jan 2007 – Nov 2008	not provided	Yes
Christopoulos et al. 201124	AIDS Patient Care STDS	San Francisco, CA	Feb 2007 – Nov 2008	55,000 visits	Yes
Hoxhaj et al. 201125	Ann Emerg Med	Houston, TX	Oct 2008 – Apr 2009	96,000 visits	Yes
Wheatley et al. 201126	J Urban Health	Atlanta, GA	May 2008 – Mar 2010	>100,000 visits	Yes
White et al. 201127	Ann Emerg Med	Oakland, CA	Feb 2007 – Jan 2008	75,000 visits	Yes
Wilbur et al. 201128	Ann Emerg Med	Indianapolis, IN	Mar – Apr 2008	100,000 visits	Yes
Martin et al. 201229	Acad Emerg Med - Abstract	North Carolina	Jun 2008 – Sep 2011	not provided	Yes
Rothman et al. 201230	Acad Emerg Med	Baltimore, MD	Nov 2005 – Oct 2009	60,000 at Johns Hopkins ED and 50,000 at Johns Hopkins Bayview ED 60,000 visits	Yes
Humphrey et al. 201231	AIDS Patient Care STDS	New Orleans, LA	2008		Yes
Bernstein et al. 201232	Acad Emerg Med	Boston, MA	Nov 2004 – May 2008	not provided	Yes
Sclesinger et al. 201233	Ann Emerg Med - Abstract	Los Angeles, CA	Mar 2011	180,000 patients	Yes
Calderon et al. 201334	Sex Transm Dis	New York, NY	Oct 2005 – Aug 2012	not provided	Yes
Christopoulos et al. 201335	J Acquir Immune Defic Syndr	San Francisco, CA	Nov 2008 – Apr 2009	not provided	Yes
Ruffner et al. 201336	Acad Emerg Med - Abstract	Cincinnati, OH	2002 – 2010	not provided	Yes
Haukoos et al. 201337	Ann Emerg Med	Denver, CO	Jun 2010 – Sep 2010	55,000 visits	Yes

First Author, Year	Type of HIV Test	No. of Positive	No (%) Linked	Received Results in the ED	Description of Linkage	Type of Linkage
Kelen et al. 19967	ELISA	24	23 (95.8)	No	HIV/ID clinic referral through appointments	Non intensive
Kelen et al. 19998 Subset A	Standard EIA	63	25 (39.7)	No	HIV/ID clinic referral through appointments	Non intensive
Kelen et al. 19998 Subset B	Rapid HIV test (SUDS)	15	6 (40.0)	Mixed†	HIV/ID clinic referral through appointments	Non intensive
Glick et al. 20049	ELISA	13	9 (69.2)	No	Positive patients physically escorted to clinic or introduced to case manager	Intensive
Kendrick et al. 200410	Rapid HIV test (SUDS)	46	36 (78.3)	Yes	Not available	NA
Lyons et al. 200511	Standard EIA	39	31 (79.5)	No	Positive patients physically escorted HIV clinic	Intensive
MMWR June 200712 Subset LA	Rapid HIV test (OraQuick®Advance)	13	11 (84.6)	Yes	Medical Appointment	Non intensive
MMWR June 200712 Subset NY	Rapid HIV test (OraQuick®Advance)	19	15 (78.9)	Yes	Medical Appointment	Non intensive
Brown et al. 200713	Rapid HIV test (OraQuick® Rapid)	9	8 (88.9)	Yes	HIV/ID clinic referral through appointments	Non intensive
Silva et al. 200714	Rapid HIV test (OraQuick®Rapid)	8	3 (37.5)	Yes	HIV/ID clinic referral through appointments	Non intensive
Lyss et al. 200715	Rapid HIV test (OraQuick® Rapid)	83	65 (78.3)	Yes	HIV/ID clinic referral through appointments	Non intensive
Haukoos et al. 200716	Rapid HIV test (OraQuick® Rapid or OraQuick®Advance)	15	12 (80.0)	Yes	HIV/ID clinic referral through appointments	Non intensive
Mehta et al. 200717	EIA	7	4 (57.1)	No	Positive patients physically escorted to the HIV clinic for their intake appointment	Intensive
White et al. 200918	Rapid HIV test (OraQuick®Advance)	101	90 (89.1)	Yes	HIV/ID clinic referral through appointments	Non intensive
Christopoulos et al. 201019	Rapid HIV test (OraQuick®Advance)	24	17 (70.8)	Yes	HIV/ID clinic referral through appointments	Non intensive
Haukoos et al. 201020	Rapid HIV test (Uni-Gold™ Recombigen)	19	19 (100)	Yes	HIV/ID clinic referral through appointments	Non intensive
Torres et al. 201121 Site B	ELISA	25	23 (92.0)	No	Positive patients physically escorted to the ID center for their intake appointment	Intensive
Torres et al. 201121 Site C	Rapid oral	10	10 (100)	Yes	Telephonic counseling from ID fellow and follow up HIV/ID clinic appointment	Non intensive
Torres et al. 201121 Site D	Rapid finger stick	70	35 (50.0)	Yes	HIV/ID clinic referral through appointments	Non intensive
Torres et al. 201121 Site E	Rapid Venipuncture	50	50 (100)	Yes	HIV/ID clinic referral through appointments	Non intensive
Torres et al. 201121 Site F	Rapid Venipuncture	34	27 (79.4)	Yes	HIV/ID clinic referral through appointments	Non intensive
Sattin et al. 201122	Rapid HIV test (OraQuick® Rapid)	35	26 (74.3)	Yes	Positive patients physically escorted ID/HIV clinic	Intensive
Schrantz et al. 201123	ELISA	28	25 (89.3)	No	ID clinic social worker attempted to reach out and make appointments for patients who tested positive	Non intensive
Christopoulos et al. 201124	Rapid HIV test (Uni-Gold™ Recombigen)	65	46 (70.8)	Yes	The linkage team met patients in person if notified during business hours and otherwise contacted patients for providing HIV clinic appointment	Mixed
Hoxhaj et al. 201125	Chemiluminescence testing (ADVIA, Centaur®)	80	34 (42.5)	Mixed†	HIV/ID clinic referral through appointments	Non intensive
Wheatley et al. 201126	Rapid HIV test (OraQuick® Rapid)	126	94* (74.6)	Yes	Positive patients provided appointment at either the local health department or hospital-affiliated ID program	Non intensive
White et al. 201127	Rapid HIV test (OraQuick®Advance)	29	22 (75.9)	Yes	HIV/ID clinic referral through appointments	Non intensive
Wilbur et al. 201128	Rapid HIV test (OraQuick®Advance)	5	4 (80.0)	Yes	HIV/ID clinic referral through appointments	Non intensive
Martin et al. 201229	3rd Generation ELISA	22	22 (100)	Mixed†	Positive patients physically escorted to the HIV clinic for their intake appointment	Intensive
Rothman et al. 201230	Rapid HIV test (OraQuick®Advance)	100	83 (83.0)	Yes	HIV/ID clinic referral through appointments	Non intensive
Humphrey et al. 201231	Rapid HIV test	99	59 (59.6)	NA	HIV/ID clinic referral through appointments	Non intensive
Bernstein et al. 201232	ELISA	13	12 (92.3)	No	HIV/ID clinic referral through appointments	Non intensive
Sclesinger et al. 201233	Rapid HIV test	25	22 (88.0)	Yes	In person encounter by ID specialist in the ED	Intensive
Calderon et al. 201334	Rapid HIV test (OraQuick®Advance)	295	227 (76.9)	Yes	Positive patients physically escorted to the HIV clinic for their intake appointment	Intensive
Christopoulos et al. 201335	Rapid HIV test (Uni-Gold™ Recombigen)	23	23 (100)	Mixed†	Linkage team performed clinic intake and medical assessment and nurse practitioner provided primary care till hand-off to specialist	Intensive
Ruffner et al. 201336	Not Available	170	119 (70.0)	NA	Linkage procedures included in-person and telephone follow-up, attempts at repeat contact for missed appointments, and accompanying patients to initial appointments	Mixed
Haukoos et al. 201337	Rapid HIV test (OraQuick®Advance)	14	14 (100)	Yes	HIV/ID clinic referral through appointments	Non intensive

First Author, Year	Linkage Staff	Follow-up Phone Call	Duration to Linkage	Extramural Funding
Kelen et al. 19967	Exogenous	Yes	NA	Emergency Medicine Foundation
Kelen et al. 19998 Subset A	Exogenous	Yes	NA	HIV test kits were donated by Murex
Kelen et al. 19998 Subset B	Exogenous	Yes	NA	HIV test kits were donated by Murex
Glick et al. 20049	Exogenous	Yes	NA	Receiving funding from unspecified source
Kendrick et al. 200410	Exogenous	NA	NA	CDC
Lyons et al. 200511	Exogenous	Yes	NA	Ohio Department of Health and the Cincinnati Health Network (Ryan White Title III funds).
MMWR June 200712 Subset LA	Exogenous	NA	NA	Not available
MMWR June 200712 Subset NY	Exogenous	NA	NA	Not available
Brown et al. 200713	Exogenous	Yes	NA	Department of Health, District of Columbia, and Gilead Sciences
Silva et al. 200714	Exogenous	Yes	NA	CDC
Lyss et al. 200715	Exogenous	Yes	Within 4 months	CDC
Haukoos et al. 200716	Indigenous	Yes	NA	Colorado Department of Public Health and Environment and CDC
Mehta et al. 200717	Exogenous	Yes	0,0,17,128 days	CDC
White et al. 200918	Indigenous	Yes	Median 14 d	CDC and NIH
Christopoulos et al. 201019	Exogenous	Yes	NA	NIH
Haukoos et al. 201020	Exogenous	NA	NA	CDC, AHRQ, Colorado Center for AIDS Research
Torres et al. 201121 Site B	Exogenous	Yes	NA	CDC and HRSA
Torres et al. 201121 Site C	Exogenous	No	NA	Foundation grant, local health department
Torres et al. 201121 Site D	Exogenous	No	NA	CDC and HRSA
Torres et al. 201121 Site E	Indigenous	Yes	NA	No, hospital and laboratory budget
Torres et al. 201121 Site F	Indigenous	No	NA	CDC and HRSA
Sattin et al. 201122	Exogenous	No	NA	CDC
Schrantz et al. 201123	Indigenous	Yes	NA	Bristol-Myers Squibb Virology Fellows Research Training Program
Christopoulos et al. 201124	Indigenous	Yes	90% by 90 days	NIH
Hoxhaj et al. 201125	Indigenous	Yes	NA	CDC and Houston Department of Health and Human Services
Wheatley et al. 201126	Exogenous	NA	NA	CDC
White et al. 201127	Indigenous	Yes	Median 7 days for opt-in and 14 days for opt-out	CDC
Wilbur et al. 201128	Exogenous	NA	NA	Marion County Ryan White Part A.
Martin et al. 201229	Indigenous	No	NA	Not available
Rothman et al. 201230	Exogenous	Yes	1 Year	Maryland Department of Health and Mental Hygiene, Baltimore City Health Department, The Gilead Foundation
Humphrey et al. 201231	NA	NA	6 months	Not available
Bernstein et al. 201232	Exogenous	Yes	NA	NIH
Sclesinger et al. 201233	Exogenous	NA	NA	Not available
Calderon et al. 201334	Exogenous	No	NA	Public Health Solutions of New York City
Christopoulos et al. 201335	Indigenous	Yes	Median-3.5days	NIH
Ruffner et al. 201336	NA	Yes	NA	Not available
Haukoos et al. 201337	Exogenous	NA	NA	Colorado Department of Public Health and Environment, NIH

HIV- Human immunodeficiency virus; ELISA-Enzyme linked immunosorbent assay; EIA- Enzyme immunoassay; SUDS-Single use diagnostic systems MMWR June 2007²⁰ Subset Oakland was not included as we were able to infer that the data from this site would be included under White et al. 2009²⁶ Torres et al. 2011²⁹ Site A was not included, as it did not have any linkage data in the article.

^*Number linked to care was obtained from rounding off the percentage linked from total number of positives.

^†Mixed meant that some patients received test results before leaving the ED while others didn’t.

Indigenous linkage staff meant that the ED clinical providers or staff who were employed as part of regular ED operations performed linkage to care; exogenous linkage staff meant that the personnel who were employed specifically for HIV testing program and not as part of regular ED operations performed linkage to care

Following rigorous examination of the linkage modalities, each study was categorized based on the type of LTC protocol used. We operationally defined LTC protocols as “Intensive”, “Non-intensive” or “Mixed“. Only “direct linkage”, operationally defined as LTC as a result of an active or passive referral process from an ED testing program,⁶ was analyzed in our study. “Intensive” LTC protocol was defined as when a patient who had an HIV positive test was linked to care (1) by a healthcare worker who physically escorted the patient to a specialty clinic, or (2) when the patient received in-person interaction from an HIV specialist [e.g. Infectious Disease (ID) specialist, ID fellow, HIV specialty nurse] while in the ED. “Mixed” LTC protocol was defined as a program that employed intensive LTC protocol only part of the coverage hours which varied by the time of the day or day of the week. All other forms of linkage protocols were defined as “non-intensive” LTC protocol. “Unclear protocol” was when the LTC protocol was not described in the article.

RESULTS

A total of 37 testing programs from 31 publications (28 journal articles and three conference abstracts) which met inclusion criteria were included (Table 1).^7–37 All programs except one were reported from academic EDs in urban settings and 27 (72.9%) reported receiving extramural funding (Table 1). The 37 programs accounted for 204,783 HIV tests of which 1,816 were HIV infected (0.9% seropositivity). Most programs (59.5%) used point-of-care (POC) tests. Overall, nine (24.3%) utilized an intensive LTC protocol, 25 (67.6%) non-intensive, two (5.4%) mixed, and one (2.7%) unclear. The overall LTC rate was 74.4%. The programs that employed intensive LTC protocols had an average LTC rate of 80.0% and the non-intensive protocol ones had 72.7%. Four (44.4%) programs with intensive protocols and nine (36.0%) programs with the non-intensive protocols had LTC rates >85%.

We observed substantial heterogeneity with regard to the exact methods of linkage modalities and key program data reporting as summarized and detailed on Table 1. The composition of the linkage staff also varied: 25 (67.6%) programs used exogenous staff, ten (27.0%) used the ED staff, and 2 (5.4%) had no information. Twenty-two (59.5%) programs made attempts to call patients to ensure LTC, six (16.2%) did not, and the remaining nine (24.3%) did not report the information. Only 8 (21.6%) programs reported duration from diagnosis to LTC which varied from zero to 365 days and only 4 reported an LTC rate of >85% within 90 days of diagnosis. There were no significant differences in ED volume, type of HIV test, linkage staff, use of follow-up call, and receipt of extramural funding between those with ≥85% LTC rate versus those with <85% within the intensive or non-intensive group.

DISCUSSION

This descriptive review included 37 U.S. adult ED HIV testing programs that performed over 200,000 HIV tests and identified approximately 1,800 HIV infected individuals. The overall LTC rate of 74% is comparable to the 76% found in a systematic review and meta-analysis by Marks et al on LTC in HIV testing programs in ED and urgent care center settings on studies conducted mostly prior to the 2006 CDC HIV testing recommendations.³⁸ Our data showed that ED-based LTC rates are yet to reach the national goal of 85% while some types of LTC protocols might possess some promising results toward reaching this goal. These findings provide baseline comprehensive data on ED-based LTC rates that can be used to evaluate future progress with continued adoption of the 2006 CDC recommendations.

Programs with intensive LTC protocols had an average LTC rate of 80% while the non-intensive LTC protocols had an average of 72.7%. In fact, 44.4% of the programs that utilized intensive LTC protocols surpassed the national goal of 85% (36.0% in the non-intensive group did). The findings from this study suggest that intensive LTC protocols might be a better practice model for ED-based programs. Intensive LTC protocols may be more successful in linking patients to care because they provide an opportunity for a relatively immediate, in-person, guided transition of the patient from the testing staff to an HIV specialty or primary care provider. This model may help to earn patients’ trust in the provider and healthcare system by providing support and a direct introduction to comprehensive HIV care.³⁹

However, one of the major limitations of implementing an intensive LTC protocol in an ED is that it requires multidisciplinary collaboration (ED staff, subspecialty ID/HIV physicians, and public health professionals), resources, and involvement of multiple agencies for best LTC outcomes. For those reasons, EDs with limited or no extramural funding and resources may not be able to implement the intensive protocol. It is therefore imperative to collaborate with public health or community-based resources to enforce success in such resource-limited settings. From our study, there were no notable differences (e.g. POC assay) between those with ≥85% LTC rate versus those with <85% within the non-intensive group. The difference in LTC outcomes may have been due to how they implemented their protocol by having better follow-up call processes, helping the patient make the appointment, and/or collaborations with the health department and community-based organizations to help follow up with patients. The next step to facilitate implementation and improvement of LTC services is to develop detailed recommendations for LTC best practices for ED settings with different levels of resources and capacity.

The definition of LTC by each program varied and might not truly indicate a successful linkage. Using indicators such as CD4 counts and viral loads may be more indicative of a successful LTC but these data are often not accessible to the ED testing programs. In addition, the measure of subsequent second or third visits is more indicative of retention to care, which is beyond the scope of ED testing programs.

While our analyses were rigorously conducted, several limitations exist. Many studies from our literature search did not provide comprehensive LTC information. This limited the analyses that could be performed and the strength of inferences that could be drawn. Future researchers should report comprehensive testing and LTC data to minimize this limitation. In addition, only eight programs reported the duration to successful linkage. The NHAS goal of 85% linkage within 90 days can only be assessed if this information is reported. Publication bias likely exists due to unbalanced publication between programs with high LTC rate and those with inadequate LTC rate. In addition, programs with research resources and publication expectations from funders are more likely to publish. The data included in this literature review are mostly generated from academic EDs. Thus, the practices in community EDs are unknown. Though our operational classification as intensive and non-intensive, simplifies the diverse LTC modalities into few broad categories, it may not reflect all the true characteristics of the various linkage protocols in practice currently. Lastly, despite the rigorous literature review, we may have missed some relevant publications. The search was completed for studies published by June 2013. Therefore, we might have missed more recent data on current practices.

CONCLUSION

The average LTC rate of an ED-based HIV testing programs published prior to July 2013 was still below NHAS’s goal of 85%. However, approximately 45% of the programs that utilized intensive protocols surpassed the 85% NHAS goal, trended more favorably than the 36% in those that utilized non-intensive protocols. This suggests that, when possible, ED-based HIV testing programs should adopt intensive LTC strategies to improve LTC outcomes. Standardization for reporting HIV LTC methods and results would improve future syntheses of trends and best practices. Future research should utilize in-depth qualitative and implementation science methodologies to gather data that will inform LTC best practices. Continued efforts to improve LTC modalities and practices in EDs are crucial to the health of ED patients newly diagnosed with HIV and in controlling the HIV epidemic.