Development of a Machine Learning Modelling Tool for Predicting HIV Incidence Using Public Health Data from a County in the Southern United States

BACKGROUND

Recent advancements in Machine Learning (ML) have significantly improved the accuracy of models predicting HIV incidence. These models typically utilize electronic medical records and patient registries. This study aims to broaden the application of these tools by utilizing de-identified public health datasets for notifiable sexually transmitted infections (STIs) from a southern U.S. County known for high HIV incidence rates. The goal is to assess the feasibility and accuracy of ML in predicting HIV incidence, which could potentially inform and enhance public health interventions.

METHODS

We analyzed two de-identified public health datasets, spanning January 2010 to December 2021, focusing on notifiable STIs. Our process involved data processing and feature extraction, including sociodemographic factors, STI cases, and social vulnerability index (SVI) metrics. Various ML algorithms were trained and evaluated for predicting HIV incidence, using metrics such as accuracy, precision, recall, and F1 score.

RESULTS

The study included 85,224 individuals, with 2,027 (2.37%) newly diagnosed with HIV during the study period. The ML models demonstrated high performance in predicting HIV incidence among males and females. Influential predictive features for males included age at STI diagnosis, previous STI information, provider type, and SVI. For females, they included age, ethnicity, previous STIs information, overall SVI, and race.

CONCLUSIONS

The high accuracy of our ML models in predicting HIV incidence highlights the potential of using public health datasets for public health interventions such as tailored HIV testing and prevention. While these findings are promising, further research is needed to translate these models into practical public health applications.

Infectious Diseases Among People Experiencing Homelessness: A Systematic Review of the Literature in the United States and Canada, 2003-2022

Homelessness increases the risk of acquiring an infectious disease. We conducted a systematic review of the literature to identify quantitative data related to infectious diseases and homelessness. We searched Google Scholar, PubMed, and SCOPUS for quantitative literature published from January 2003 through December 2022 in English from the United States and Canada. We excluded literature on vaccine-preventable diseases and HIV because these diseases were recently reviewed. Of the 250 articles that met inclusion criteria, more than half were on hepatitis C virus or Mycobacterium tuberculosis. Other articles were on COVID-19, respiratory syncytial virus, Staphylococcus aureus, group A Streptococcus, mpox (formerly monkeypox), 5 sexually transmitted infections, and gastrointestinal or vectorborne pathogens. Most studies showed higher prevalence, incidence, or measures of risk for infectious diseases among people experiencing homelessness as compared with people who are housed or the general population. Although having increased published data that quantify the infectious disease risks of homelessness is encouraging, many pathogens that are known to affect people globally who are not housed have not been evaluated in the United States or Canada. Future studies should focus on additional pathogens and factors leading to a disproportionately high incidence and prevalence of infectious diseases among people experiencing homelessness.

Barriers and Facilitators to HIV Service Access among Hispanic/Latino Gay, Bisexual, and Other Men Who Have Sex with Men in Metropolitan Atlanta-a Qualitative Analysis

Hispanic/Latino persons are disproportionately impacted by HIV in the US, and HIV diagnoses among Hispanic/Latino men in Georgia have increased over the past decade, particularly in metropolitan Atlanta. In 2022, the Georgia Department of Public Health detected five clusters of rapid HIV transmission centered among Hispanic/Latino gay, bisexual, and other men who have sex with men (HLMSM) in metropolitan Atlanta. We conducted in-depth interviews with 65 service providers and 29 HLMSM to identify barriers and facilitators to HIV service access for HLMSM. Interviews were audio recorded, transcribed, and translated, if needed. Initial data analyses were conducted rapidly in the field to inform public health actions. We then conducted additional analyses including line-by-line coding of the interview transcripts using a thematic analytic approach. We identified four main themes. First, inequity in language access was a predominant barrier. Second, multiple social and structural barriers existed. Third, HLMSM encountered intersectional stigma. Finally, the HLMSM community is characterized by its diversity, and there is not a one-size-fits-all approach to providing appropriate care to this population. The collection of qualitative data during an HIV cluster investigation allowed us to quickly identity barriers experienced by HLMSM when accessing HIV and other medical care, to optimize public health response and action. Well-designed program evaluation and implementation research may help elucidate specific strategies and tools to reduce health disparities, ensure equitable service access for HLMSM, and reduce HIV transmission in this population.

Mpox and HIV: a Narrative Review

PURPOSE OF REVIEW

We reviewed the available literature on mpox in People with HIV (PWH). We highlight special considerations of mpox infection related to epidemiology, clinical presentation, diagnostic and treatment considerations, prevention, and public health messaging in PWH.

RECENT FINDINGS

During the 2022 mpox outbreak, PWH were disproportionally impacted worldwide. Recent reports suggest that the disease presentation, management, and prognosis of these patients, especially those with advanced HIV disease, can widely differ from those without HIV-associated immunodeficiency. Mpox can often be mild and resolve on its own in PWH with controlled viremia and higher CD4 counts. However, it can be severe, with necrotic skin lesions and protracted healing; anogenital, rectal, and other mucosal lesions; and disseminated organ systems involvement. Higher rates of healthcare utilization are seen in PWH. Supportive, symptomatic care and single or combination mpox-directed antiviral drugs are commonly used in PWH with severe mpox disease. Data from randomized clinical control trials on the efficacy of therapeutic and preventive tools against mpox among PWH are needed to better guide clinical decisions.

797. Investigation of HIV Clusters Among Hispanic/Latino Gay or Bisexual Men in Metro Atlanta, Georgia

Background

In Metro Atlanta, Georgia, annual HIV diagnoses among Hispanic/Latino (H/L) adolescents and adults have increased since 2014 (Figure 1), and four HIV molecular clusters, consisting primarily of Hispanic/Latino gay or bisexual men and other men who have sex with men (HLGBM), were identified in 2021. In March 2022, the Georgia Department of Public Health, the Centers for Disease Control and Prevention, and four health districts of Metro Atlanta (Fulton, Gwinnett, DeKalb, and Cobb) launched an investigation to characterize the clusters, assess barriers to accessing HIV care and prevention services, and inform improvements to service delivery.

Methods

We described the four clusters using HIV surveillance data. We conducted semi-structured qualitative interviews with 29 HLGBM and 28 providers through purposive sampling. Iterative analyses were conducted daily, comparing findings across interviews to identify commonly mentioned barriers to accessing HIV care and prevention services.

Results

The four clusters varied in size (5–42 members) and proportion of H/L members (41–100%); one cluster included members reporting injection drug use (14%) (Table 1). Viral suppression among cluster members was high across clusters (87–100%). Overarching reported barriers to accessing medical services included lack of culturally and linguistically concordant services, fear of deportation, transportation, and financial barriers (Table 2). LGBTQ and HIV stigma, low STD/HIV awareness, low access to primary care and HIV screening in primary and urgent care settings, and limited community outreach and marketing were common barriers to accessing HIV prevention services. For pre-exposure prophylaxis (PrEP) specifically, fear of being perceived as promiscuous, limited PrEP knowledge, and concerns with side effects were barriers. Navigating the healthcare system was noted as the primary challenge in HIV care.

Conclusion

The investigation of four clusters affecting HLGBM identified opportunities to improve access to HIV prevention and PrEP services for this population. State and local partners are planning cluster response activities, including ways to provide low-barrier culturally and linguistically concordant services (Table 3).

Disclosures

All Authors: No reported disclosures.

617. Long Acting Lipoglycopeptide Use in Veterans for Serious Gram-Positive Infections in the COVID Era

Background

Dalbavancin and Oritavancin are semisynthetic lipoglycopeptides (LGP) that are FDA-approved for treatment of skin and soft tissue infections, but emerging data supports LGP use for other serious gram positive (GP) infections. We describe our experience with LGP during the COVID-19 pandemic.

Methods

We initiated a quality improvement project to assess the use of LGP for label and off-label indications at the Atlanta Veterans Affairs Health Care System. We define serious GP infections as infective endocarditis, osteomyelitis, joint infections, or bacteremia. Patients with serious GP infections that receivedLGP were selected at the treating physician's discretion. We reviewed medical records of all patients receiving at least one dose of long-acting LGP from March 1, 2020 - May 31, 2021. We described patient demographics, clinical information,and outcomes (90-day readmission).

Results

Nineteen patients with GP infections received LGP (table). Overall, the most common infection was cellulitis 7 (35%); 14 patients received LGPs for serious GP infections. All patients received at least one other non-LGP antibiotic for at least 2 days, majority vancomycin (60%) and cefazolin (30%). Overall, the median hospital stay among patients who received LGP was 8.5 days (range: 2-45 days), for those with serious GP infections the median hospital stay was 15 days (range: 4-45). 90% of patientswho received LGP were discharged home. Number of LGP doses ranged from 1 to 6 doses total, based on type of infection. Sixteen veterans (80%) followed up in outpatient clinicfollowing discharge within 2 weeks, two patients were discharged to home hospice due to complications of underlying malignancies and two patients were lost to follow up. Noadverse drug events were reported, and none with serious GP infections required rehospitalization at 90 days.

Conclusion

Our experience suggests that long-acting LGP may be valuable tools to treat serious gram-positive infections by optimizing theduration of hospitalization and preventing unnecessary admissions to acute care and nursing facilities for daily antibiotic infusions. These aspects of LGP use are especially important during the COVID-19 pandemic where nosocomial transmission has been documented.

Disclosures

All Authors: No reported disclosures

Soft Tissue, Bone, and Joint Infections in People Who Inject Drugs

Infections are a common complication among people who inject drugs (PWID). Skin and soft tissue infections (SSTI) as well as bone and joint infections comprise a significant source of morbidity and mortality among this population. The appropriate recognition and management of these infections are critical for providers, as is familiarity with harm-reduction strategies. This review provides an overview of the presentation and management of SSTI and bone and joint infections among PWID, as well as key prevention measures that providers can take.

786. Improving Access by Prescreening Self-Referred Patients to an Infectious Disease Clinic

Background

Studies show that 35% of patients obtaining care in a specialty clinic are self-referred, of these, 60% do not have a PCP. Frequent changing of physicians without provider referral contributes to high costs and healthcare inefficiencies. Many subspecialty clinics do not require provider referral for access, so-called: ‘self-referral’; becoming an issue when patients discern a need for a subspecialty that does not align with their disease process. The Infectious Disease (ID) Department of the Cleveland Clinic Main Campus began prescreening all self-referred patients in July 2016. (Figure 1) We reviewed this process after sustained implementation.

Methods

Retrospective cohort study examining “self-referred” patients (lacking an internal referral order) to the ID Clinic from December 2016–October 2018. Grouped patients by chief complaint using system-based categories, subsequently reviewed records received for prior workup and identified if they had been seen by a specialist or received antibiotics in the past 6 months prior to visit. Those seen in our department were evaluated for: (1) additional workup ordered (2) if antibiotics were prescribed, and (3) referral to another specialty.

Results

1449 patients self-referred. By prescreening them, waiting time for appointment fell from >40 to 13 days, no-shows steadily drop from 11% to 8%. We audit 105 (40%) of patients seen, most common reasons for referral: Skin 19 (18%), Bone/Joint 18 (17%), Concern for Lyme Disease 16 (15%). 73% had been previously seen by a subspecialty; majority by ID. 75% received antibiotics in the past 6 months. After the visit, 53% had new work up ordered, 21% were prescribed antibiotics and 26% patients were referred to another specialty.

Conclusion

Prescreening self-referred patients contributed to improved access to subspecialty care and decreased”no-show’ appointments. The majority of self-referred patients had previous evaluation, frequently by the specialty being requested.

Disclosures

All authors: No reported disclosures.

2371. A Multicenter Cohort Study of the Natural History of Clostridioides difficile Colonization and Infection

Background

Asymptomatic carriage of toxigenic Clostridioides difficile strains is common in healthcare settings. However, the natural history of C. difficile colonization and infection is not well understood, particularly for patients with new acquisition of carriage.

Methods

In 3 tertiary care hospitals and affiliated long-term care facilities (LTCFs), we conducted a 6-month cohort study to identify patients with new acquisition of rectal carriage of toxigenic C. difficile and determined the duration and burden of carriage. Asymptomatic carriage was defined as transient if only 1 culture was positive with negative cultures before and after or persistent if 2 or more cultures were positive; clearance was defined as 2 consecutive negative rectal cultures.

Results

Of 4180 patients with negative initial cultures, 144 (3%) acquired asymptomatic carriage of toxigenic C. difficile, and 19 (13%) of these carriers subsequently were diagnosed with CDI. Of 50 asymptomatic carriers analyzed for duration of carriage, 33 (66%) had transient carriage of toxigenic C. difficile and 17 (34%) had persistent carriage. For persistent carriers, the estimated median time to clearance of colonization was 76 days (range, 41 to 95 days from acquisition). Ten of 17 (59%) persistent carriers had a high burden of carriage (defined as > 25 colonies recovered from 1 or more swabs) vs. only 1 of 33 (3%) transient carriers (P < 0.001).

Conclusion

Acquisition of asymptomatic carriage of toxigenic C. difficile carriage was common among patients in healthcare facilities, but most carriers had transient low-level carriage. Additional studies are needed to determine whether a higher burden of carriage predicts subsequent risk of transmission.

Disclosures

All authors: No reported disclosures.

126. Cascade of Care for Opioid Use Disorder in Patients with Infective Endocarditis

Background

The term “Cascade of Care” has been used to analyze care delivered by a health system for conditions such as HIV, hepatitis C, tuberculosis, and diabetes. It outlines sequential steps required to reach a specific outcome (i.e., viral suppression in the case of HIV). This allows to estimate the proportion of patients achieving each step and to identify gaps in care. Medication-assisted treatment (MAT) is integral in the treatment of patients with infective endocarditis (IE) and opioid use disorder (OUD). We propose a Cascade of Care aiming to identify fundamental milestones in the management of these patients.

Methods

A retrospective cohort study examined patients with IE in the setting of OUD hospitalized between July 1, 2007 and January 1, 2015 to the Cleveland Clinic. We identified 4 key steps along the treatment cascade of these patients and estimated the proportion of patients: (1) evaluated by an addiction treatment service, (2) prescribed MAT while in-patient, (3) prescribed MAT at discharge, and (4) continued MAT at least 90 days after discharge.

Results

Of 273 patients with IE in the setting of OUD, 134 (49%) were evaluated by an addiction treatment service; 45 (17%) were prescribed MAT while in-patient; only 22 (8%) were prescribed MAT at discharge. At 90 days following discharge, there was evidence of continuing MAT for all 22 patients (8%).

Conclusion

Describing the process of addiction treatment for patients with IE and OUD in the format of a cascade of care provides a powerful quantitative method to identify gaps in care and can be used as a resource to implement interventions to address losses. We found only 8% of these patients continued MAT in the community after discharge. This study provides an estimate of how compromised the potential benefits from medical and surgical treatment for IE are by the lack of an effective approach to OUD after hospital discharge.

Disclosures

All authors: No reported disclosures.