Routine HIV Screening in an Urban Community Health Center: Results from a Geographically Focused Implementation Science Program

Abstract

Objective

CDC has recommended routine HIV screening since 2006. However, few community health centers (CHCs) routinely offer HIV screening. Research is needed to understand how to implement routine HIV screening programs, particularly in medically underserved neighborhoods with high rates of HIV infection. A routine HIV screening program was implemented and evaluated in a Philadelphia, Pennsylvania, neighborhood with high rates of HIV infection.

Methods

Implementation science is the study of methods to promote the integration of research findings and evidence into health-care policy and practice. Using an implementation science approach, the results of the program were evaluated by measuring acceptability, adoption, and penetration of routine HIV screening.

Results

A total of 5,878 individuals were screened during the program. HIV screening was highly accepted among clinic patients. In an initial needs assessment of 516 patients, 362 (70.2%) patients reported that they would accept testing if offered. Routine screening policies were adopted clinic-wide. Staff trainings, new electronic medical records that prompted staff members to offer screening and evaluate screening rates, and other continuing quality-improvement policies helped promote screenings. HIV screening offer rates improved from an estimated 5.0% of eligible patients at baseline in March 2012 to an estimated 59.3% of eligible patients in December 2014. However, only 5,878 of 13,827 (42.5%) patients who were offered screening accepted it, culminating in a 25.2% overall screening rate. Seventeen of the 5,878 patients tested positive, for a seropositivity rate of 0.3%.

Conclusion

Routine HIV screening at CHCs in neighborhoods with high rates of HIV infection is feasible. Routine screening is an important tool to improve HIV care continuum outcomes and to address racial and geographic disparities in HIV infection.

Approximately 1.1 million people are living with human immunodeficiency virus (HIV) in the United States.^1,2 HIV testing and treatment are among the most effective prevention interventions. HIV testing is also the critical gateway to receiving HIV treatment and care. Individuals who test positive for HIV and are aware of their serostatus are more likely to reduce their risk-taking behaviors.³ HIV-positive people unaware of their HIV status account for nearly half of new HIV transmissions.^4,5 Early screening, detection, and treatment have positive effects on community health. For example, individuals who are virologically suppressed can reduce transmission to others by 96%.⁶

Since 2006, the Centers for Disease Control and Prevention (CDC) has recommended routine HIV screening in clinical settings.⁷ Despite this recommendation, few clinics routinely offer HIV screening, only 54% of Americans have ever received an HIV test, and 22% of Americans have received an HIV test in the past year.⁸

Members of racial minority groups and individuals from lower socioeconomic backgrounds are least likely to be aware of their HIV infection.^9,10 African American and Hispanic people have eight and three times, respectively, the HIV infection rates of white people;¹¹ are more likely to present for care late in the course of their infection;¹² and have poorer outcomes at every point along the HIV care continuum.^13,14 Moreover, HIV infections cluster geographically. Most new infections in the United States are concentrated in the southeast and in poor, urban neighborhoods that are mostly inhabited by people of color.^15,16 The National HIV/AIDS Strategy calls for directing resources to the most heavily affected parts of the country and reducing HIV-related health disparities.¹⁷ Routinizing HIV screening in neighborhoods with high seroprevalence may be a way to enhance the HIV care continuum in heavily affected neighborhoods and to reduce racial and geographic disparities in HIV infection.¹⁵

Implementation science can be a useful framework in which to explore how to scale public health interventions in real-world settings. The National Institutes of Health defines implementation science as “the study of methods to promote the integration of research findings and evidence into health-care policy and practice.”¹⁸ Implementation science is needed to employ the most effective HIV prevention tools, including HIV testing and treatment in real-world settings and in communities bearing a disproportionate share of the HIV disease burden.^19,20 Implementation science distinguishes implementation outcomes from service and clinical outcomes; frequently used measures to assess implementation include acceptability, adoption, appropriateness, costs, feasibility, fidelity to interventions, and sustainability.^20,21 Implementation science can be particularly useful for evaluating adoption of screening and prevention programming for which screenings (vs. health outcomes alone) are a desirable programmatic outcome.^22,23

We developed a community-based program focused on expanding HIV and hepatitis C virus (HCV) screening and care across three medically underserved ZIP Codes in Philadelphia, Pennsylvania, that have some of the highest rates of HIV and HCV infection in the nation. Our program included a routine HIV screening program at a community health center (CHC), a social marketing campaign to promote and de-stigmatize HIV and HCV screening, door-to-door and street outreach, and rapid HIV and HCV screening on a mobile medical unit. This comprehensive effort aimed to scale HIV screening in both clinical and nonclinical settings in medically underserved neighborhoods, with a focus on linking and retaining patients in HIV care.

This article presents results from the clinical implementation science program to routinize HIV screening at a CHC in an urban neighborhood with high rates of HIV infection. We present results from the clinical program focused on the acceptability, adoption, and penetration of routine HIV screening in an urban CHC.

METHODS

This program was executed at a CHC that provides primary care, behavioral health, pediatric, family planning, and dental services. The CHC serves a patient population with limited access to health care in a southwest Philadelphia neighborhood with 2%–3% HIV seropositivity. The clinical care delivery model uses medical assistants for patient intake, vital signs, and triage processes, while nurse practitioners deliver most clinical care.

At baseline, in March 2012, the CHC used a risk-based HIV screening model in which patients were offered a rapid HIV test upon request or if their provider believed their risk behaviors warranted screening. The health center then developed a routine HIV screening program in which medical assistants routinely offered a rapid HIV test to all patients ≥13 years of age who had not received a test in the last 12 months. We assessed the acceptability of routine screening with patients, adoption of routine HIV screening policies, and penetration of a routine HIV testing offer by medical assistants.

Acceptability of HIV screening

The definition of acceptability commonly cited in the implementation science literature was used to guide the program: “The perception among patients that a given treatment, service, or practice is agreeable or satisfactory.”²² To measure routine HIV screening acceptability prior to the program's commencement, a baseline needs assessment was conducted in March 2012 among a sample of 516 health center patients. This survey was offered to all patients presenting for care for five consecutive days at the clinic, during April 2012, and 516 of 573 (90.0%) patients responded. The survey explored willingness to take an HIV test, self-perceived HIV risk, HIV testing history, and demographics, and took approximately three minutes to complete.

Adoption of routine screening policies and procedures

In implementation science terms, adoption is frequently defined as the intention, decision, or action to employ an innovation or evidence-based practice.²² Evaluation included assessments of adoption of the policies and health infrastructure necessary to implement a routine HIV screening program. Several important policies related to implementing routine HIV screening were adopted. Not all policy changes happened concurrently, and many evolved over time in accordance with lessons learned.

First, health center leadership adopted policies endorsing a transition from risk-based screening to a routine offer, including modifying the informed consent process to eliminate written informed consent requirements. Second, all medical assistants, nurse practitioners, front desk staff, and other personnel were trained to create a clinic-wide culture promoting HIV screening. Scripts, which were developed to guide conversations between providers and patients, included information about routine screening and the high rates of HIV infection in the surrounding community. Clinical staff members were trained to handle sensitive topics, including discussing HIV risk behaviors, delivering HIV diagnoses, and understanding the importance of HIV care and treatment.

Third, electronic medical records (EMRs) were modified to include prompts about screening during the medical assistant intake process and reasons patients might decline screening (Figure 1). The EMR was modified to collect more information as the program evolved. Collecting data disaggregated by the medical assistant allowed us to develop continuing quality improvement (CQI) efforts to evaluate the HIV screening offer and patient acceptance rates in real time.

Figure 1.

Emergency medical record (EMR) intake form at an urban health center in Philadelphia, Pennsylvania, 2012–2014

HIV = human immunodeficiency virus

Other CQI efforts were also adopted as the program evolved. To jump-start screening, a financial incentive structure to reward medical assistants who improved their HIV screening offer rates was employed. The medical assistant with the highest testing rate was awarded a $50 gift card. This incentive structure evolved in accordance with evolving programmatic goals. In year two, monthly progress reports assessed testing offer and patient acceptance rates by medical assistant. In year three, a wall of fame was created to publicly display monthly testing offer and acceptance rates by medical assistant.

Penetration of routine HIV screening

We defined penetration as the integration of a practice within a service setting.²¹ We evaluated program penetration by measuring HIV screening offer and patient acceptance rates over time by a medical assistant. We determined offer rates by dividing the number of HIV tests offered by the total number of eligible patients. The EMR tracked HIV screening offer, decline, and acceptance rates. We defined the offer rate as the total number of eligible patients who were offered screening by a medical assistant. Patients were considered eligible if they were ≥13 years of age. We defined the acceptance rate as the total number of patients offered who accepted screening. The overall screening rate was the number of eligible patients who underwent testing. We also measured the reasons patients declined screening. Crude odds ratios (ORs) and 95% confidence intervals were calculated using bivariable logistic regression analysis.

As the program evolved, we used data-driven CQI measures, including disaggregating data trends by medical assistants, prominently posting signs in the clinic to promote screening, and implementing an incentive scheme, to encourage medical assistants to offer and conduct HIV screening.

RESULTS

Acceptability

Of the 516 needs assessment patients, 403 (78.1%) were female, 410 (79.5%) were African American, and 232 (45.0%) reported annual household income <$10,000. Three hundred sixty-two respondents (70.2%) reported that they would accept an HIV test if offered during the course of routine clinical care, but most respondents (n=316, 61.2%) believed that they were not at risk for contracting HIV. Among the 405 respondents (78.5%) who believed they had been previously tested for HIV, 123 (29.6%) had not been tested in the past year (Table 1).

Table 1.

Characteristics of respondents to an HIV needs assessment survey at an urban health center, Philadelphia, Pennsylvania, 2012–2014

HIV = human immunodeficiency virus

GED = general educational development

Adoption and penetration

Primary process outcomes for adoption included the routine screening policy and creation of an EMR that prompted screening and allowed for measuring testing offer, consent, and decline rates. The EMR was -modified over time to reflect lessons learned and ways to enhance offer and acceptance rates. For example, when we learned that patients frequently believed that they had been tested for HIV in the last year, but had not been tested, the EMR text flashed red to guide provider discussions with patients about screening (Figure 1).

Those who self-identified as American Indian/Alaska Native, white, and Asian were significantly more likely to decline testing than those who self-identified as black/African American (ORs=3.12, 2.23, and 1.76, respectively), and females were 1.34 times more likely than males to decline testing. Individuals aged 19–35 years were least likely to decline testing compared with those aged 13–18 years (OR=1.19) and >35 years (OR range: 1.72–9.50) (Table 2).

Table 2.

Demographic characteristics of patients who were offered, accepted, and declined HIV testing at an urban health center in Philadelphia, Pennsylvania, 2012–2014

HIV = human immunodeficiency virus

OR = odds ratio

CI = confidence interval

Ref. = reference group

From January 2012 to August 2014, a total of 5,878 HIV tests were conducted (Figure 2). Seventeen of the 5,878 patients tested positive for a seropositivity rate of 0.3%, and 13 of the 17 patients diagnosed with HIV were previously unaware of their infection. The overall HIV testing offer rate was 59.3% (13,827 of 23,317 eligible patients), but ranged from 24.6% (234 of 952 eligible patients) to 81.3% (591 of 727 eligible patients) during the course of the program. Among 13,827 patients offered screening, 5,878 (42.5%) accepted screening, but this acceptance rate also ranged from 29.8% (176 of 591 tests offered) to 50.8% (253 of 498 tests offered) during the course of the program. The overall screening rate was 25.2% (5,878 tests completed of 23,317 eligible patients). Offer and acceptance rates varied by month (Figure 3) and even more widely by medical assistant. Two medical assistants nearly always offered screening, two medical assistants offered screening about 50% of the time, and two medical assistants almost never offered screening. The most common reason cited for declined screening was that patients thought that they had recently been tested or perceived that they were at low risk for contracting HIV (Figure 4). Factors unrelated to the screening program, such as leadership changes and staff shortages, also influenced offer rates (Figure 3).

Figure 2.

Policy changes and HIV screening offer rates by month at an urban health center in Philadelphia, Pennsylvania, 2012–2014

^aThe incentive program provided a financial incentive structure to reward medical assistants who improved their HIV screening offer rates; the medical assistant with the highest testing rate was awarded a $50 gift card.

^bPublicly displayed monthly testing offer and acceptance rates by medical assistant

^cStandard protocols for carrying out HIV screening using blood draw

^dTraining on standing orders for HIV screening using blood draw for medical assistants

HIV = human immunodeficiency virus

EMR = electronic medical record

Figure 3.

Number of HIV screenings offered and conducted by month at an urban health center in Philadelphia, Pennsylvania, 2012–2014

HIV = human immunodeficiency virus

Figure 4.

Reasons for decline given by patients declining HIV screening at an urban health center in Philadelphia, Pennsylvania, 2012–2014

^aOf 13,827 HIV tests offered, 7,949 offers were declined. All subjects gave one principal reason for declining.

HIV = human immunodeficiency virus

Among the 17 individuals who tested positive for HIV, eight received care in-house and nine received care at other HIV subspecialty care centers. At the time of publication, all patients had been linked to care, 12 individuals had been prescribed antiretroviral therapy, and seven individuals had suppressed viral ribonucleic acid. Of the 17 individuals who were diagnosed during the program, six presented late in the course of their HIV infection with a CD4+ count <350 cells per cubic millimeter.

LESSONS LEARNED

Using an implementation science approach, we successfully adopted routine screening policies at an inner-city health center in a neighborhood with high rates of HIV infection, limited access to health services, and high rates of poverty. Integrating HIV testing into existing workflows, adopting system-wide policy changes, using EMRs to track trends and promote CQI efforts, and educating the entire clinical staff were important components of the program. These four pillars of routine HIV screening helped inform program -implementation^24,25 and a needs assessment exploring feasibility of this intervention among the clinic's patient population.

This approach culminated in a more than 10-fold increase in HIV testing rates compared with baseline, a nearly 60% testing offer rate, and nearly 6,000 patients tested, many of whom might otherwise not have undergone screening. Although a universal offer rate—defined as offering an HIV test to every eligible patient—was not achieved, these findings suggest that implementing policies and programs to promote routine HIV screening efforts is a feasible and effective way to screen and diagnose people living with HIV, particularly in neighborhoods with high rates of infection.

Notably, HIV testing rates increased in tandem with each of our data-driven CQI efforts. For example, at the project's outset, the primary reason patients declined testing was because they were not expecting screening. Designing posters and placing them prominently in waiting areas and examination rooms resulted in an immediate decline in the number of patients reporting that they did not expect a test. Offer rates also improved when we developed and introduced scripts to train providers on how to routinely offer screening.

Moreover, while testing uptake was initially low, it jumped rapidly when an incentive scheme was introduced to reward medical assistants who achieved the highest offer rates. The incentive program was an important tool to jump-start a new screening program, but there were diminishing returns to the incentive program over time, even as the program evolved to adopt new goals and screening targets.

HIV testing rate changes often corresponded with major events unrelated to the program, such as leadership changes and human resource constraints. For example, HIV testing rates decreased when two medical assistants took leaves of absence. HIV testing rates also declined when the clinic was keenly focused on enrolling and insuring new patients during the Affordable Care Act's open enrollment period. Lastly, the overall offer and acceptance rates declined briefly when the clinic transitioned from a rapid testing model. In the original model, the clinic relied on grant funds for buying rapid tests that used fingerpricks and delivered results within minutes. In the new model, the CHC adopted standing orders for a blood draw for routine screening, sent blood work to external laboratories for analysis, and billed for services. Given that other routine screening models have successfully achieved high levels of testing with laboratory-based models, this decline is likely not a forecast of future screening rates,²⁴ but a reflection of the learning curve associated with adopting a new screening model. Testing offer rates bounced back after the brief decline associated with the transition and will likely continue to increase as CQI efforts related to laboratory-based screening continue.

The biggest challenge to increasing screening rates was the high percentage of patients who declined screening. Higher test decline rates among women and those aged >35 years warrant further exploration and intervention. Although many patients answered that they had been screened in both the acceptability survey and during our screening offers, it is likely that many had never been previously screened. This finding is supported by other HIV screening studies in Philadelphia that found patients often mistakenly believed they had been screened because a health provider drew blood.^26,27 These findings, robust across several local studies, suggest that more efforts are needed to educate patients and providers about Pennsylvania's revised HIV testing law, which eliminates written informed consent requirements for HIV testing but still requires documentation of consent.^28,29 Many patients still need to be educated about the importance of undergoing screening because they live in neighborhoods with high rates of infection, which can increase their HIV transmission risk.^17,30,31 Patients also need to be informed that unless they specifically provided consent to test, they likely have not undergone HIV screening.

Although the clinic provides HIV specialty care, nine of the 17 patients who tested positive for HIV ultimately received HIV care elsewhere, namely at larger HIV subspecialty care centers. Nevertheless, the model used at this CHC, which screened and diagnosed patients in a medically underserved community and then referred them to other subspecialty care centers, highlights important opportunities to dramatically expand HIV screening programs.

Our 25.2% overall screening rate was lower than the 56% screening rate in a model program recently cited in Morbidity and Mortality Weekly Report.²⁴ However, that program used a laboratory-based model in which blood was drawn rather than a rapid testing model.^24,25 Laboratory models have been hypothesized to enhance sustainability and billing more than rapid testing models. It is worth noting, however, that offer rates in our program generally surpassed rates in another six-site study employing rapid screening technology, but this program experienced higher patient test decline rates.³² Given that 77.5% of patients in our representative needs assessment reported an interest in screening, further efforts to educate clinic staff members about how to offer screening in a resource- and time-constrained environment might enhance overall HIV screening penetration. Even if our clinic's test decline rates remained static, HIV screening could be increased considerably if our offer rates approached those cited in the New York model. Transitioning to a laboratory-based model may also offer opportunities to further enhance screening.

Routine screening is a critical component of CDC's high-impact HIV prevention strategy²⁶ and is an important tool for diagnosing more individuals earlier in the course of their HIV infection.^32,33 Moreover, provider and patient perceptions about HIV risks are often inaccurate,²⁷ and risk-based screening is an insensitive criterion for effectively diagnosing people who are unaware of their infection.^34,35 Routine screening was implemented across an entire urban CHC, overcoming many of the commonly cited obstacles to routine HIV screening, such as stigma, concerns about confidentiality, lack of knowledge about informed consent laws, and financial barriers.^36–38

Limitations

Our analysis was subject to several limitations. Although all patients testing positive were immediately linked to care, focusing on patients' HIV treatment outcomes was beyond the scope of this analysis. Additionally, this population was overwhelmingly female, which reflects the large proportion of patients who presented for prenatal and pediatric services. Local HIV infection rates are often higher among men than women. As such, other urban clinics with fewer prenatal and pediatric patients might have even higher rates of HIV infection than the rates identified in our program.

Our seropositivity rate of 0.3% was lower than the seropositivity rate in the surrounding ZIP Codes, which may be attributable to several factors. First, we employed a universal vs. risk-based offer, which means we offered screening even to those who were at low risk for HIV. Second, the highest-risk individuals often did not present for clinical care. Third, many of the patients were women presenting for prenatal care services or children presenting for pediatric well visits, and these populations generally have lower HIV seropositivity rates than other subpopulations. CDC guidelines endorse clinical screening as a highly cost-effective intervention if HIV diagnosis rates are >0.1%,³² which suggests that this program was indeed cost-effective.

Lastly, at the time of publication, in an effort to enhance sustainability, our program was transitioning away from using rapid HIV test kits and toward using standing orders for universal blood draws to test all eligible patients. This model includes onsite phlebotomy services, off-site laboratory testing, and billing for both phlebotomy and laboratory services. Although standing orders for universal blood draws for HIV screening can enhance HIV screening volume, provider reimbursement, and program sustainability,^24,25 analyses related to this latest programmatic change were beyond the scope of this evaluation.

CONCLUSION

We now have a robust set of tools to help prevent HIV transmission and treat HIV, including HIV testing, linkage to and retention in HIV treatment and care, pre- and postexposure prophylaxis, and needle exchange. Although these tools have the potential to dramatically affect the acquired immunodeficiency syndrome epidemic in the United States, they must be expanded outside the controlled environment of research programs. Using an implementation science framework that focuses on acceptability, adoption, and penetration of routine HIV testing in a real-world, inner-city setting, HIV screening increased in an urban community with high rates of HIV infection, wide racial disparities in infection, and limited access to testing and treatment services. This approach to optimizing routine screening facilitated development of a program to use patient data for CQI efforts. This approach evolved in response to lessons learned and new challenges allowed for improvement of this program consistently over time and facilitated relatively high rates of HIV screening.

These findings suggest that busy urban community clinics often have competing priorities for personnel time, among other challenges, but do provide critical opportunities to enhance HIV screening and diagnosis. Expanding similar routine screening programs in CHCs in neighborhoods with high rates of infection across the country could have overwhelmingly positive effects on diagnosing, linking, and retaining HIV-positive patients in care. Expanding routine HIV screening presents an important public health opportunity to comply with CDC routine screening guidelines and reduce geographic and racial disparities in HIV infection.