Impacts of Sexually Transmitted Disease/HIV Outreach Sites on the Effectiveness of Detecting New Infections in Baltimore City, 2015 to 2018

Abstract

Background 

Effectiveness of sexually transmitted disease (STD)/HIV mobile van outreach programs has been shown to assist in identifying new cases of syphilis and HIV among high-risk populations. However, specific types of mobile outreach testing and their varying abilities to assist in STD/HIV case identification have not been thoroughly explored.

Methods 

We analyzed data from the Baltimore City Health Department syphilis and HIV testing intake interviews and subsequent laboratory results for clients at any of the 6 different Baltimore City Health Department mobile outreach testing setting types between 2015 and 2018. A total of 8098 syphilis and 7946 HIV testing encounters were included. Cross-tabulation and proportion estimates were used to obtain the volume of new syphilis and HIV diagnoses across the different mobile outreach testing setting types. Multivariable firth logistic regressions for rare events were used to obtain odds ratios for new syphilis diagnoses across the different mobile outreach setting categories.

Results 

Testing at street corners yielded the highest proportion of new syphilis cases (0.83%), whereas at community centers, it produced the highest proportion of new HIV cases (0.48%). The adjusted odds for new syphilis were 2.14 among street corners compared with community centers.

Conclusions 

The findings indicated that street corners were the most successful at detecting new syphilis cases, whereas community centers were the most successful at detecting new HIV cases. However, because of the small number of new cases detected and low uptake per testing event between 2015 and 2018, STD/HIV mobile testing does not seem to be effective overall.

In Brief

Analyses of clients who used sexually transmitted disease/HIV mobile outreach services in Baltimore, MD, found that street corners for syphilis and community centers for HIV were most effective at detecting new cases.

Mobile outreach testing has demonstrated effectiveness in identifying new and early sexually transmitted disease (STD)/HIV infections in urban settings among high-risk individuals.¹ A prior study indicated that mobile outreach was more successful than clinic-based STD/HIV testing at recruiting population of color, individuals with lower educational attainment, individuals with binge drinking behavior, first-time STD/HIV testers, and substance use individuals.² Despite the apparent benefits of mobile outreach testing for STD/HIV, limited research on the varying effectiveness in case identification of different types of mobile outreach settings is available. Prior literature has often grouped all settings of mobile outreach testing together to determine effectiveness and feasibility.^1,3,4 However, effectiveness may also depend on the variations in settings at which outreach is performed, not only the mode (i.e., mobile vs. fixed) being used.

Better understanding of specific mobile outreach setting types can facilitate outreach planning for health departments and community-based organizations. The data offer an opportunity to compare the varying outputs from the different mobile outreach testing setting types: community centers, community events, drug treatment/mental health centers, schools/colleges, shelters/assistance programs, and street corners. Effectiveness, for this analysis, will be defined as the number and proportion of new syphilis and HIV cases identified at each setting.

MATERIALS AND METHODS

Sexually transmitted disease/HIV community outreach involves STD/HIV education, counseling, and testing in the community, especially in areas with high burden of disease throughout Baltimore City. We have grouped the mobile outreach settings into 6 different setting types: community centers, community events, drug treatment/mental health centers, schools/colleges, shelters/assistance programs, and street corners. Community centers include churches, recreational centers, and senior centers. Community events consist of community-sponsored events, government-sponsored events, festivals, fairs, clubs, and entertainment events. Finally, street corners consist of street-based outreach settings. Outreach sites were chosen based on geospatial analysis of STD/HIV burden in Baltimore City and requests for STD/HIV testing by the community.

All clients who visited any of the outreach sites were administered an intake interview questionnaire before STD/HIV testing. Sexually transmitted disease/HIV mobile van testing at all outreach sites was free of charge, and clients who chose to get STD/HIV tests were provided with small monetary or gift incentives. Intake interviews consisted of demographic, sexual history, and behavioral history questions. Serologic testing was performed to diagnose syphilis and HIV.

Syphilis and HIV are reportable diseases.⁵ Positive syphilis and HIV laboratory results obtained during the outreach visit were cross-checked with the Maryland Department of Health statewide PRISM sexually transmitted infection surveillance database to determine if an infection of HIV or syphilis was newly or previously diagnosed.

For the syphilis analysis, zip codes were coded into 3 categories of low, moderate, and high based on a separate 2015 Baltimore City Health Department (BCHD) geospatial analysis of early syphilis rates by zip codes.⁶

The number of cases for each setting type was cross-tabulated, and Pearson χ² and Fisher exact tests were conducted to assess statistically significant associations (using P values) between new HIV/syphilis diagnoses and setting type. Subsequently, bivariable and multivariable firth logistic regression models for rare events were used to obtain unadjusted and adjusted odds ratios of new syphilis diagnoses for each of the outreach setting types compared with community centers, used as the reference setting type. There were too few new HIV cases to warrant accurate regression analysis; therefore, only syphilis cases were used. Proportion estimates using the Clopper-Pearson exact method were also calculated with 95% confidence intervals and SEs to assess the likelihood of new syphilis and HIV diagnoses in each mobile outreach setting type.

RESULTS

Statistical analyses used cross-sectional retrospective outreach data from July 1, 2015, to June 30, 2018, at 236 mobile van outreach sites. There were 8098 syphilis testing encounters (423 under community centers, 719 under community events, 371 under drug treatment/mental health centers, 250 under schools/colleges, 2379 under shelters/assistance programs, and 3956 under street corners) across 5337 unique clients and 7946 HIV testing encounters (417 under community centers, 725 under community events, 364 under drug treatment/mental health centers, 261 under schools/colleges, 2278 under shelters/assistance programs, and 3901 under street corners) across 5342 unique clients for the period of analysis. Inclusion criteria consisted of all clients who sought syphilis and HIV testing at any of the BCHD 236 mobile outreach sites and were 18 years or older during this period.

For syphilis, across the observed period, the community centers category had an average uptake of 4.8 encounters per event (88 total events), community events had 10.7 encounters per event (67 events), drug treatment/mental health centers had 6.3 encounters per event (59 events), schools/colleges had 7.4 encounters per event (34 events), shelter/assistance programs had 7.3 encounters per event (324 events), and street corners had 7.0 encounters per event (560 events).

For HIV, the community centers category had an average uptake of 4.7 encounters per event (89 total events), community events had 11.0 encounters per event (66 events), drug treatment/mental health centers had 6.3 encounters per event (58 events), schools/colleges had 7.6 encounters per event (35 events), shelter/assistance programs had 7.0 encounters per event (327 events), and street corners had 7.1 encounters per event (552 events).

Study population characteristics for clients tested for syphilis are shown in Table 1. Among clients tested for syphilis, 3177 (59.5%) were male. When demographic characteristics were stratified by testing setting categories, mean age for shelters/assistance programs was the highest, followed by community centers, drug treatment/mental health centers, street corners, community events, and lastly schools/colleges. For race, most clients were Black, ranging from 71.8% to 91.5% at each setting. Most male clients seen for syphilis testing were reported as non–men who have sex with men, ranging from 88.9% to 97.7%. Clients mainly reported being stably housed at the time of the testing encounter, ranging from 64.7% to 88.2%. Of 8098 syphilis testing encounters overall, there were 3828 clients (47%) who were seen for mobile outreach testing only once, whereas all other clients had multiple encounters throughout the 3-year period of analysis, with some clients being encountered as many as 13 times.

TABLE 1 - Characteristics for Syphilis and HIV Mobile Van Testing

Syphilis (n = 5337 Clients)	Site Categories
	Community Centers (n = 264)		Community Events (n = 567)		Drug Treatment/Mental Health Centers (n = 232)		Schools/Colleges (n = 212)		Shelters/Assistance Programs (n = 1273)		Street Corners (n = 2789)
	Mean (%)	SD (n)	Mean (%)	SD (n)	Mean (%)	SD (n)	Mean (%)	SD (n)	Mean (%)	SD (n)	Mean (%)	SD (n)
Age, y	45.0	13.8	36.5	13.8	44.8	11.4	24.9	11.3	48.0	11.5	44.2	13.1
Sex
Female	47.3	125	52.6	298	42.2	98	59.0	125	28.0	356	41.2	1150
Male	51.9	137	47.1	267	57.8	134	41.0	87	71.9	915	58.7	1637
Unknown	0.8	2	0.4	2					0.2	2	0.1	2
Race
Black/African American	85.6	226	71.8	407	74.6	173	91.5	194	79.3	1009	84.2	2349
White	6.8	18	18.9	107	24.6	57	2.4	5	17.7	225	12.6	351
Other	6.4	17	9.2	52	0.7	2	6.1	13	2.6	33	2.9	80
Unknown	1.1	3	0.2	1					0.5	6	0.3	9
Education level
HS/GED/equivalent	54.2	143	45.7	259	65.1	151	11.3	24	62.1	790	63.6	1774
Less than HS	18.6	49	9.2	52	20.7	48	0.5	1	20.5	261	17.2	479
More than HS	26.9	71	42.9	243	13.8	32	87.3	185	16.3	207	17.1	477
Unknown	0.4	1	2.3	13	0.4	1	0.9	2	1.2	15	2.1	59
Screening zip code by prevalence
Low			0.7	4	49.1	114	80.7	171			0.4	11
Moderate	80.3	212	42.0	238	43.5	101	10.9	23	73.1	930	62.2	1734
High	19.7	52	57.3	325	7.3	17	8.5	18	26.9	343	37.4	1044
MSM status
No	97.0	256	88.9	504	97.8	227	96.2	204	96.4	1227	97.7	2726
Yes	3.0	8	11.1	63	2.2	5	3.8	8	3.6	46	2.3	63
Housing status
Literally homeless	2.3	6	2.3	13	6.9	16	0.9	2	23.1	294	5.0	140
Stably housed	86.7	229	88.2	500	76.7	178			64.7	823	82.8	2308
Unstably housed	0.8	2	0.5	3	1.3	3	86.3	183	2.1	27	1.2	32
Unknown	10.2	27	9.0	51	15.1	35	12.7	27	10.1	129	11.1	309
HIV (n = 5342 Clients)	Community Centers (n = 270)		Community Events (n = 577)		Drug Treatment/Mental Health Centers (n = 235)		Schools/Colleges (n = 222)		Shelters/Assistance Programs (n = 1244)		Street Corners (n = 2794)
	Mean (%)	SD (n)	Mean (%)	SD (n)	Mean (%)	SD (n)	Mean (%)	SD (n)	Mean (%)	SD (n)	Mean (%)	SD (n)
Age, y	44.9	13.8	36.8	13.7	45.0	11.6	24.6	10.8	47.9	11.5	44.0	13.2
Sex
Female	46.3	125	53.0	306	42.1	99	59.9	133	28.1	349	41.2	1151
Male	53.0	143	46.6	269	57.9	136	40.1	89	71.9	894	58.7	1641
Unknown	0.7	2	0.4	2					0.1	1	0.1	2
Race
Black/African
American	85.6	231	72.1	416	74.5	175	91.4	203	78.6	978	84.0	2346
White	7.0	19	19.2	111	24.3	57	2.3	5	18.6	231	12.7	356
Other	6.3	17	8.5	49	1.3	3	6.3	14	2.4	30	3.0	83
Unknown	0.7	3	0.2	1					0.4	5	0.3	9
Education level
HS/GED/equivalent	54.8	148	46.3	267	63.8	150	11.3	25	62.9	782	63.7	1781
Less than HS	20.0	54	9.0	52	20.9	49	0.5	1	19.6	244	17.0	476
More than HS	24.8	67	42.1	243	14.9	35	87.4	194	16.4	204	17.3	482
Unknown	0.4	1	2.6	15	0.4	1	0.9	2	1.1	14	2.0	55
MSM status
No	96.3	260	89.8	518	98.3	231	96.4	214	97.0	1206	97.9	2736
Yes	3.7	10	10.2	59	1.7	4	3.6	8	3.0	38	2.1	58
Housing status
Literally homeless	2.6	7	2.3	13	6.8	16	0.9	2	23.3	290	5.3	147
Stably housed	86.3	233	87.4	504	77.0	181	86.0	191	64.6	804	82.6	2307
Unstably housed	0.7	2	0.5	3	0.9	2			2.4	30	1.2	33
Unknown	10.4	28	9.9	57	15.3	36	13.1	29	9.7	120	11.0	307

GED indicates General Equivalency Diploma; HS, high school.

Study population characteristics for clients tested for HIV are shown in Table 1. Among the mobile outreach clients tested for HIV, there were 3172 (59.4%) men. The mean age of clients was highest at shelters/assistance programs, then drug treatment/mental health centers, community centers, street corners, community events, and schools/colleges. Most clients who were tested for HIV were Black, ranging from 72.1% to 91.4% across outreach setting types. Education-level attainment for clients tested for HIV across mobile outreach category types was similar to that among clients tested for syphilis. Most male clients tested for HIV also reported no MSM status, ranging from 89.8% to 98.3% across setting categories. A majority of clients tested for HIV also reported being stably housed, ranging from 64.6% to 87.4% across outreach setting category. Of the 7946 HIV testing encounters, 3879 clients were seen at mobile outreach once, whereas the remainder of clients were seen multiple times, with some clients with as many as 12 encounters.

Overall, there were a total of 42 new syphilis and 15 new HIV cases identified through BCHD's mobile outreach testing efforts between July 2015 and June 2018.

Of 3956 syphilis testing encounters conducted at street corners, 33 (0.83%) were confirmed new cases. This is the highest proportion of new syphilis cases across the 6 mobile outreach setting categories (Table 2).

TABLE 2 - Syphilis and HIV New Cases

Site Categories	No. New Cases	Proportion	SE	Exact (95% CI)
Syphilis
Community centers (n = 423)	1	0.0024	0.0024	0.00006–0.0131
Community events (n = 719)	1	0.0014	0.0014	0.00004–0.0077
Drug treatment/mental health centers (n = 371)	0	0	N/A	N/A
Schools/colleges (n = 250)	0	0	N/A	N/A
Shelters/assistance programs (n = 2,379)	7	0.0029	0.0011	0.00118–0.006058
Street corners (n = 3956)	33	0.0083	0.0014	0.00575–0.01169
HIV
Community centers (n = 417)	2	0.0048	0.0034,	0.00058–0.0172172
Community events (n = 725)	0	0	N/A	N/A
Drug treatment/mental health centers (n = 364)	1	0.0027	0.027	0.000069–0.01521
Schools/colleges (n = 261)	0	0	N/A	N/A
Shelters/assistance programs (n = 2,278)	4	0.0018	0.0009	0.00048–0.00449
Street corners (n = 3901)	8	0.0021	0.0007	0.00089–0.004037

Among the HIV cases identified, street corners yielded the largest number of new HIV cases with 8 cases (Table 2). The proportion of new HIV cases was highest among community centers (0.48%).

Correlates of Mobile Outreach Testing Setting Types and New Syphilis Diagnoses

Through multivariable firth logistic regression for rare events, new syphilis case diagnosis was associated with certain mobile van outreach types, although results were not statistically significant (Table 3). Street corners had odds of 2.14 (P = 0.436) of detecting new syphilis infection compared with community centers, after adjusting for sex and race. We were not able to run the firth logistic regression for rare events for HIV, as only 15 new HIV cases were detected during the 3-year period of analysis.

TABLE 3 - Correlates of Outreach Types

Site Categories	Unadjusted OR (95% CI)	P	Adjusted OR (95% CI)	P
Community center	Reference	N/A	1.0	N/A
Community event	0.59 (0.06–5.67)	0.646	0.43 (0.04–4.17)	0.466
Drug treatment center/mental health	0.38 (0.02–9.33)	0.553	0.29 (0.01–7.28)	0.454
School/college	0.56 (0.02–13.85)	0.725	0.58 (0.02–14.21)	0.736
Shelter/assistance program	0.89 (0.13–4.63)	0.898	0.79 (0.13–4.63)	0.793
Street corner	2.40 (0.47–12.39)	0.294	2.14 (0.41–11.12)	0.463

DISCUSSION

We discovered that not all mobile outreach types demonstrated the same effectiveness in detecting new syphilis and HIV cases. For syphilis outreach testing, street corners were the most successful among the 6 mobile outreach setting types at identifying new syphilis, whereas for HIV outreach testing, community centers were the most successful at identifying new cases of HIV. Multivariable firth logistic regression for rare events confirmed further the importance of street corner testing compared with community centers for syphilis; although not statistically significant, clients who were seen at street corner mobile outreach testing settings were more than 2-fold likely than those at community centers to be diagnosed as a new syphilis case after controlling for sex and race.

The small number of new cases limited the power of evaluating the outcome of new syphilis or HIV case finding across the various mobile outreach testing setting categories. Future studies should be conducted to further analyze the ability to identify new syphilis and HIV, especially early syphilis or new HIV seroconversion, as these specific diagnoses are essential triggers for public health actions and can produce the most health impact. Although not statistically significant in this analysis, the findings offer a valuable guide for planning and targeting areas, time, and resources for syphilis and HIV mobile outreach testing in Baltimore City.

Overall, whether because of incentives or easy access, there was a large number of clients who returned to BCHD mobile outreach testing; some clients returned for outreach testing services as many as 13 times. However, when outreach categories were stratified by outreach events, average uptake per event was low. In addition, although street corners and community centers obtained the highest proportions of new syphilis cases, proportions are still quite small. Based on prior studies, greater cost-effectiveness was experienced through testing at clinics in fixed-settings compared with mobile outreach testing models, and rapid testing protocols were shown to be more cost-effective compared with standard blood-based testing protocol.^7,8 Although the analysis has limited power, because of the small proportion of newly identified syphilis and HIV cases, the low uptake per outreach event in this analysis, and BCHD's STD/HIV traditional blood-based testing method, the STD/HIV mobile outreach observed during the established period does not seem effective overall. To further assess effectiveness though, additional effectiveness studies should be conducted to look at new STD/HIV among specific high-risk groups.