Most recently, combined cases of syphilis, gonorrhea, and chlamydia have reached an all-time high in the United States in 2018.1 Certain states and counties disproportionately experience STIs. For example, in Miami-Dade County, Florida, syphilis incidence has nearly tripled over the last 13 years (6.9 cases per 100,000 population in 2005 to 19.5 cases per 100,000 population in 2018).2 Congenital syphilis is also on the rise in Miami-Dade: there were 14 congenital syphilis cases in 2014, 8 cases in 2015, 20 cases in 2016, 32 cases in 2017, and 24 cases in 2018.3 Congenital syphilis is especially concerning, as it can cause miscarriage, stillbirth, and/or other birth complications.4 In the past 4 years, cases of congenital syphilis have nearly tripled (462 cases in 2014 to 1306 cases in 2018).5 Notably, syphilis is a risk factor for transmitting or acquiring HIV and other STIs.6–8 Unlike HIV, antibiotics can easily cure syphilis. However, routine screening for syphilis is not readily available in health care settings. The inability to efficiently screen for syphilis not only impedes treatment of syphilis but also may prevent a missed opportunity for HIV prevention education (if a patient tests positive for syphilis, but negative for HIV).
Previous responses to the syphilis increase were enhanced surveillance, partner services, and collaboration with community-based organizations. However, these approaches were not effective in all communities. Homestead Hospital (HH), 1 of the 11 hospitals of Baptist Health South Florida, the largest not-for-profit health system in South Florida, took an innovative approach in helping diagnose and treat syphilis-infected individuals with the goal of also preventing congenital syphilis and HIV. Homestead Hospital serves patients who are primarily uninsured due to eligibility restrictions, affordability, and/or immigration status. Typically, these uninsured patients use the emergency department (ED) as their routine source of care rather than primary care or urgent care. To concurrently address the syphilis and HIV epidemics, HH incorporated a smart screening algorithm for syphilis into the ED's routine HIV/HCV “opt-out” screening infrastructure. This smart screening algorithm automatically triggers a blood test for high-risk patients (described in Materials and Methods), in addition to the universal HIV/HCV screening. With this technology, preexisting workflows are not affected. Ultimately, implementation of this algorithm allows more patients to get treatment, receive comprehensive prevention services, and, in some cases, avert congenital syphilis. The purpose of this article is to describe the development, implementation, and evaluation of the algorithm.
MATERIALS AND METHODS
The development and implementation of the algorithm occurred in 3 phases: (1) retrospective evaluation of HIV/HCV screening program, (2) creation of smart syphilis screening algorithm based on patient risk, and (3) evaluation of the algorithm's impact on patient identification and linkage to care (LTC). These phases are described hereinafter.
HH's Program Description and Florida Department of Health Collaboration
Homestead Hospital is a 147-bed facility that includes all private rooms, the latest technology and equipment, and a busy ED with 44 private treatment rooms and a trauma room. In 2016, HH achieved Magnet recognition by the American Nurses Credentialing Center as a reflection of its nursing professionalism, teamwork, and superior patient care. Homestead Hospital is the de facto safety net hospital in southern Miami-Dade County. A large number of people living in HH's service area are uninsured or underinsured. Homestead Hospitals' service area is extremely diverse, with 63.9 of its population being Hispanic or Latino, of which 70.5% were white, 23.2% were black or African American, and 1.2% were Asian. The median household income for the service areas is $43,215 for Homestead and $34,545 for Florida City, both lower than Miami-Dade's median household income of $46,338. Approximately 43.1% of individuals in the service area earn less than 200% of the poverty level.9
Homestead Hospital's HIV/HCV screening program supports the Centers for Disease Control and Prevention's (CDC) recommendations for screening and LTC. This program is policy-driven and allows the ED to test a wide array of patients though an opt-out model. A LTC specialist is in-house available for any positive diagnosis and LTC. The FLDOH is an in-kind provider of additional confirmatory testing for HIV, HCV, and syphilis. These additional tests are couriered to the Bureau of Public Health Laboratories – Miami, which includes a Treponema pallidum particle agglutination assay (TP-PA) to determine syphilis infections.
Furthermore, a data reconciliation agreement between the DOH-Miami-Dade and HH allows for the sharing of HIV-positive clients' information for surveillance and LTC. In the hospital, an LTC specialist provides results of the tests to the patients. The DOH certifies these LTC specialists to provide counseling and referrals. In addition, the DOH-Miami-Dade provides a dedicated disease intervention specialist (DIS) to HH. A DIS is dispatched to the hospital and/or patient home for every new diagnosis to provide comprehensive counseling, link patients to care or prevention services (including preexposure prophylaxis), and facilitate partner notification.
Retrospective Evaluation of the Screening Program
Over the past couple of years, clinicians and screening program staff members anecdotally saw an increase in syphilis infections. Thus, they conducted a retrospective evaluation of patient records to confirm their observations and worked with the FLDOH to validate their findings. From May 2016 (the day that the screening program started) through October 2017, 18,000 patient records who were screened through the screening program were reviewed. Of these records, 183 patients were diagnosed as having HIV, with a 1% seropositivity. When tracing the patient's medical history, of the 183 patients with HIV, 44 (24%) were screened for syphilis prior or concurrent. Thirty-nine (21%) had a reactive syphilis diagnosis. Notably, 13 (33%) of the rapid plasma reagin (RPR) reactive (syphilis cases) were found to be reactive before their HIV diagnosis. This 33% was one of the driving forces to develop a systematic screening algorithm to target these at-risk patients.
Smart Syphilis Screening Algorithm Creation
From October 2017 to April 2018, information technology developers, clinicians, researchers, and screening program staff worked together to build and pilot test the smart screening algorithm. We used the already implemented HIV/HCV routine screening infrastructure to develop a rule to autopopulate a syphilis treponemal (enzyme immunoassay, or EIA) test. To help capture the at-risk patients, a broad category of symptoms associated with STIs was chosen for the algorithm. The smart algorithm generates a syphilis EIA test order if any of the following broad categories of risk are identified:
- reason for medical visit including symptoms of syphilis, for example, vaginal bleeding, penile lesions, and groin pain;
- positive pregnancy test results, such as pregnancy test blood/qualitative and pregnancy test on urine;
- positive STI result (e.g., gonorrhea and chlamydia) on date of service; and/or
- historical positive STI result.
If the syphilis TP Screen (EIA) is positive, a quantitative RPR (confirmatory) test is ordered. A nonreactive RPR will be sent out to the State Laboratory for additional confirmatory testing with TP-PA. A nonreactive TP-PA could be determined as a false positive (unlikely of syphilis infection). The FLDOH confirms whether a reactive RPR (syphilis) is a previous or current infection. A DIS is then assigned to the patient, and the patient is linked to care.
In addition, there is a relinkage alert system for all patients who test positive through the HIV/HCV/syphilis screening. If a patient is determined to be lost to care, the LTC specialist will document on the patient's medical record that the patient “Needs to be Connected w/(HIV/HCV/syphilis) care.” As soon as the patient returns to the hospital, the LTC specialist will get an e-mail with the patient's information and the reason the relinkage alert was completed (e.g., “Needs to be connected with Syphilis care,” “Unable to contact,” and/or “Missing information”). This mechanism helps our LTC specialist follow up with patients to understand and address any barriers that impeded patients from originally following up with medical care for their HIV, HCV, and/or syphilis.
Smart Syphilis Screening Algorithm Evaluation
To evaluate the implementation of the smart syphilis screening algorithm, we tracked the numbers of patients screened for syphilis, number of patients who tested positive for syphilis, and number of patients who were linked to care. We also reviewed medical records to document patients' demographic characteristics to understand which patient population was being affected by the syphilis epidemic.
Figure 1 outlines the number of ED patients tested through the screening program and the syphilis smart screening algorithm, tested reactive, and linked to care. In summary, from April 2018 to August 2019, the algorithm triggered 4806 syphilis TP screens (EIA), of which 122 syphilis TP screen (EIA) tests were reactive (2.5% seropositivity). Of the reactive tests, based on reactive reflex confirmatory tests (RPR and TP-PA), with determination by DOH, 59 patients required treatment and/or additional follow-up from their DIS. All of the 59 patients with reactive syphilis tests requiring additional follow-up and/or treatment are considered to be linked to care. Table 1 describes the characteristics of patients who screened reactive for syphilis TP screen (EIA). In summary, of those 122 patients with reactive syphilis TP screen (EIA) results, 32 were pregnant. The most common trigger (reason for medical visit) for the screening algorithm was abnormal vaginal bleeding. The DOH determined our program successfully averted 9 cases of congenital syphilis. On average, these 59 patients were 30.44 years old. Most of which were non-Hispanic/black women, and approximately 25% were coinfected with another STIs.
Our program identified syphilis cases that were previously diagnosed and treated elsewhere in the health care system. In calculation of the 122 reactive syphilis TP screens (EIA) in the hospital, 63 with reactive syphilis TP screen (EIA) test were determined to be previously infected, previously treated, and required no additional follow-up by the DOH's DIS. It is important to note that a reactive syphilis TP screen (EIA) test requires a reflex test to confirm their syphilis infection. Patients with a reactive syphilis TP screen (EIA) are documented as syphilis cases (reported to DOH); however, the reflex syphilis confirmation test (RPR and TP-PA) determines if the case is considered previously infected, newly diagnosed, or a false positive (based on biological factors). With the close collaboration with DOH, we are able to confirm if a patient requires additional follow-up and/or treatment. Furthermore, based on documentation retrieved from the DOH data reconciliation and the patients' medical records, it was determined that 32 of the 59 patients with reactive syphilis test, which required follow-up and/or treatment, were not considered newly diagnosed and had previous documentation of reactive syphilis test in the past. However, based on their reactive TP screen (EIA) and reactive confirmatory test (RPR and/or TP-PA), Of these cases, DOH follow-up on these patients through DIS outreach as their status is unknown to DOH because of being unable to contact or being uncompliant with previous treatment. Therefore, 27 of the 59 patients with reactive syphilis tests are considered newly diagnosed syphilis infection (no history of syphilis infection at HH or elsewhere in the health care system) and require follow-up treatment. Of which only 20 of the newly diagnosed syphilis infections had documentation of treatment that complies with guidelines. Because of limited data, currently, it is unknown if 7 of the newly diagnosed syphilis infections have successfully complied with the DOH DIS outreach and follow-up including receiving appropriate treatment.
We were able to develop and implement a smart screening algorithm for syphilis into our current ED-based HIV/HCV screening program. With such implementation, we were able to seamlessly integrate the syphilis screening into our existing workflow. The syphilis test is a system-generated order based on specific criteria, so there is no staff involvement with the ordering process. The syphilis linkage workflow was easily incorporated into the DOH-Miami-Dade DIS' role. In fact, it simplified the DIS' workload because they are able to meet the patient at the hospital as opposed to searching for them in the community. The Laboratory did not experience any workflow issues with processing the additional syphilis tests because of the relative low volume. The smart algorithm was created from data from patients who had previously presented at the ED and later tested positive for syphilis. From this analysis, we chose the most frequently reported symptoms, other factors that could indicate a higher risk for syphilis infection, and recommendations from the US Preventive Services Task Force (i.e., screening all pregnant women for syphilis)10 to create the algorithm. Implementation of the algorithm was successful, and we were able to identify and link to care 59 patients who were confirmed positive for syphilis. In communities with high HIV and syphilis infection rates such as Miami-Dade County, inclusion of a smart syphilis screening algorithm to an HIV/HCV routine screening program can be a useful tool in HIV/HCV/syphilis education, prevention, and treatment efforts. We found this to be especially true for the prevention of congenital syphilis: our program averted 9 cases (Miami-Dade reported a total of 24 cases of congenital syphilis in 2018).3 Pregnant women in our service area including uninsured and undocumented are very likely to lack appropriate prenatal care. One in 23 infants in Miami-Dade county was born to women receiving late or no prenatal care.11 Therefore, by integrating their pregnancy status into our screening program, we ensure that they are screened for syphilis and can help prevent congenital infections.
Our findings and methods could be replicable in other hospitals that are similar to ours: communities that have many low-income and underinsured patients that use an ED as their source of primary care. Beyond the Homestead community, other hospitals in the Baptist Health South Florida system, which preside in more urban, insured, and mixed income areas, are closely monitoring the results of this work. One additional hospital is currently implementing the ED-based HIV/HCV screening program, with the goal of having routine screening available at all Baptist Health South Florida hospitals in the future. Although the ED may be the ideal location for the identification of patients with syphilis from low-income and/or underinsured backgrounds, we must start thinking about how other settings can reach a larger patient population. For example, because Baptist Health South Florida is a large health system with more than 50 outpatient centers, a variation of this screening algorithm could be incorporated into the urgent care offices. Patients who screen positive for STIs, such as gonorrhea or chlamydia, could be automatically flagged for follow-up screening for HIV and/or syphilis and vice versa.
Increasing access to screening and care through EDs in comparable communities should be a significant public health objective because EDs play a critical role in identifying undiagnosed patients in high-prevalence areas throughout the United States. The EDs access difficult to reach populations and can seamlessly integrate screening programs into their workflow using electronic health record system algorithms. Emergency departments can scale-up screening with high volumes, and the approach is more efficient in comparison to the more costly, targeted point-of-care screening programs.12 Although the utility of wider (including universal screening) syphilis screening in the ED is ideal, the funding for universal testing is a barrier for implementation. With our algorithm-based model, a lower volume of tests is generated, which is more cost-efficient given the financial constraints. Further research is required to examine the extent of the algorithm-based model's limitations and restrictions.
In addition, implementing universal screening in an ED setting requires major systems change. Embracing technology by using the EMR and obtaining support from department leadership can help to seamlessly integrate screening into the existing clinical workflow. A facilitator to our success was the buy-in from key stakeholders: the ED and laboratory staff, information technology department, leadership, and the hospital's Medical Executive Committee. Stakeholders were part of the design and implementation of the algorithm, which increased their ownership and engagement with the program. Because the screening orders are automated (i.e., autopopulated at triage) and generated without a physician “sign-off” or “authentication,” the majority of the work is done by the EMR system. This little to nonexistent impact to clinical workflows is central to stakeholder buy-in. Future replications of the algorithm must consider a) stakeholders' clinical expertise in the development of the algorithm and b) stakeholders' buy-in by highlighting the ease of implementation in workflow. Funding to cover the cost of reagents and testing is not currently available to support the ideal, universal syphilis screening or a universal STI panel for ED patients. With fiscal resources constrained, our syphilis smart screening algorithm enhanced our existing screening program to identify those at higher risk of syphilis infections.
Per the CDC, the risks of syphilis include all pregnant women, men who have sex with men, those who are living with HIV, or those who have partners who have tested positive for syphilis.13 Our algorithm triggers a test for all pregnant women as mentioned by the CDC, STI infections (date of visit or historical), and a broad category of medical visits. A key limitation to the algorithm is missing risk factors that could be included in the calculation such as the ones recommend by the CDC. For example, social determinants of health (e.g., income and education), sexual identity or orientation (e.g., men who have sex with men), substance abuse, and/or mental health disorders have been highly correlated with STIs.14 Interestingly, men who have sex with men are one of the high-risk groups for syphilis,15 yet our findings were mostly in women. If the algorithm was triggered by these risk factors, in addition to the clinical event (e.g., reason for medical visit or pregnancy), we may be able to identify more patients. In contrast, our HIV/HCV model of opt-out screening will trigger a test for all patients regardless of the risk factors. Therefore, we are able to identify more patients for HIV and/or HCV in comparison to our syphilis screening program. It is shown in our data that our positive findings mostly involved women due in part to the algorithm automatically selecting all pregnant women. Our population consists of diverse patient demographics including uninsured women who use our hospital as their primary care physician/obstetrician-gynecologist, typically not receiving any prenatal care including STI screening. Furthermore, the smart algorithm could be improved upon through statistical techniques to better understand how risk factors relate and/or predict risk for syphilis. We understand that out smart algorithm will miss many cases because syphilis infections are often painless and asymptomatic. However, potential machine learning could mine and analyze the ED data. As another example, latent class analysis could classify the patient population into subgroups who are most likely to be affected by STIs. Lastly, the relinkage alert system could be expanded. For patients who screened positive for HIV as a result of routine HIV/HCV testing and who were not flagged for syphilis testing via the smart algorithm, follow-up syphilis testing could be incorporated as part of our LTC program to rule out coinfection.
The HH and DOH-Miami-Dade response to Miami's syphilis problem is innovative and replicable. We have shared our experience with pioneering this program for replication at other hospitals' EDs. The program embraces technology and enhances the routine screening model to treat syphilis and coinfections. The smart screening algorithm is facilitating a precision diagnosis, which is enabling LTC and interrupting disease transmission.
1. Centers for Disease Control and Prevention. New CDC report: STDs continue to rise in the U.S. Available at: https://www.cdc.gov/nchhstp/newsroom/2019/2018-STD-surveillance-report-press-release.html
. Accessed January 6, 2020.
3. Florida Department of Health, Bureau of Communicable Diseases. Congenital syphilis cases. Available at: http://www.flhealthcharts.com/charts/OtherIndicators/NonVitalSTDDataViewer.aspx?cid=0083
. Accessed January 6, 2020.
4. Centers for Disease Control and Prevention. Congenital syphilis—CDC fact sheet. Available at: https://www.cdc.gov/std/syphilis/cong-syph-feb-2017.pdf
. Accessed January 6, 2020.
5. Centers for Disease Control and Prevention. 2018 STD surveillance report. Available at: https://www.cdc.gov/nchhstp/newsroom/2019/2018-STD-surveillance-report.html
. Accessed January 6, 2020.
6. Jarzebowski W, Caumes E, Dupin N, et al. Effect of early syphilis infection on plasma viral load and CD4 cell count in human immunodeficiency virus–infected men: Results from the FHDH-ANRS CO4 cohort. Arch Intern Med 2012; 172:1237–1243.
7. Fleming DT, Wasserheit JN. From epidemiological synergy to public health policy and practice: The contribution of other sexually transmitted diseases to sexual transmission of HIV infection. Sex Transm Infect 1999; 75:3–17.
8. Buchacz K, Patel P, Taylor M, et al. Syphilis increases HIV viral load and decreases CD4 cell counts in HIV-infected patients with new syphilis infections. AIDS 2004; 18:2075–2079.
9. Baptist Health South Florida. Community health needs assessment spring 2019. Available at: https://baptisthealth.net/en/about-baptist-health/documents/hh-report.pdf
Accessed January 6, 2020.
10. US Preventive Services Task Force. Screening for syphilis infection in pregnant women: US Preventive Services Task Force reaffirmation recommendation statement. JAMA 2018; 320:911–917.
11. March of Dimes Foundation. Quick facts: Prenatal care. Available at: https://www.marchofdimes.org/Peristats/ViewTopic.aspx?reg=12086&top=5&lev=0&slev=6
. Accessed January 6, 2020.
12. Mugavero M, Amico K, Horn T, et al. The state of engagement in HIV care in the United States: From cascade to continuum to control. Clin Infect Dis 2013; 57:1164–1171.
13. Centers for Disease Control and Prevention. Syphilis—CDC fact sheet. Available at: https://www.cdc.gov/std/syphilis/stdfact-syphilis.htm
. Accessed January 6, 2020.
14. Centers for Disease Control and Prevention. National Center for HIV/AIDS, Viral Hepatitis. Available at: https://www.cdc.gov/nchhstp/health-disparities-risk.htm
. Accessed January 6, 2020.
15. Centers for Disease Control and Prevention. Sexually transmitted diseases (STDs). Available at: https://www.cdc.gov/std/syphilis/stdfact-msm-syphilis.htm
. Accessed January 6, 2020.