Development of an electronic medical record-based algorithm to identify patients with unknown HIV status

Automated HIV Case Identification from the MIMIC-IV Database

Automatic HIV phenotyping is needed for HIV research based on electronic health records (EHRs). MIMIC-IV, an extension of MIMIC-III, contains more than 520,000 hospital admissions and has become a valuable EHR database for secondary medical research. However, there was no prior phenotyping algorithm to extract HIV cases from MIMIC-IV, which requires a comprehensive knowledge of the database. Moreover, previous HIV phenotyping algorithms did not consider the new HIV-1/HIV-2 antibody differentiation immunoassay tests that MIMIC-IV contains. Our work provided insight into the structure and data elements in MIMIC-IV and proposed a new HIV phenotyping algorithm to fill in these gaps. The results included MIMIC-IV's data tables and elements used, 1,781 and 1,843 HIV cases from MIMIC-IV's versions 0.4 and 2.1, respectively, and summary statistics of these two HIV case cohorts. They could be used for the development of statistical and machine learning models in future studies about the disease.

Association of HIV and viral suppression status with hospital acute kidney injury in the era of antiretroviral therapy

In the modern era, it is unknown if people that are virally suppressed with HIV (PWH) are at increased risk for acute kidney injury (AKI) compared to people without HIV and no studies have compared the risk of AKI by viral suppression status. Here, we determined the associations of HIV status and AKI among PWH with and without viral suppression compared to people without HIV. An observational cohort study of PWH and people without HIV hospitalized in a large New York City health system between 2010-2019 was conducted. Multivariable Cox proportional hazards models were used to determine associations between HIV status and risk of AKI, severe AKI and development of chronic kidney disease (CKD). Among 173,884 hospitalized patients, 4,718 had HIV; 2,532 (53.7%) were virally suppressed and 2,186 (46.3%) were not suppressed. Compared to people without HIV, PWH with and without viral suppression were at increased risk of AKI (adjusted hazard ratio 1.27, 95% confidence interval 1.15, 1.40 and 1.73, 1.58, 1.90, respectively) and AKI requiring kidney replacement therapy (1.89, 1.27, 2.84 and 1.87, 1.23, 2.84, respectively). Incremental, graded associations were observed between HIV status and Stage 2 or 3 AKI, and among AKI survivors, and incident CKD. The elevated risk of AKI across ages of PWH was similar in magnitude to older people without HIV. Thus, regardless of virologic control, HIV is an independent risk factor for AKI among hospitalized patients. Future studies should determine the mechanisms by which HIV increases susceptibility to AKI and identify strategies to prevent AKI in PWH.

Computational phenotyping with the All of Us Research Program: identifying underrepresented people with HIV or at risk of HIV

Objective

This study aims to identify the people living with HIV (PWH) and pre-exposure prophylaxis (PrEP) users in the All of Us (AoU) database by integrating information from both electronic health record (EHR)- and self-reported survey data.

Methods

We identified PWH and PrEP users if they met the inclusion criterion by conditions, lab measurements, or medications related to HIV in EHR data or confirmed questions in the Survey data.

Results

We evaluated the latest data release through July 1, 2022 in AoU. Through computational phenotyping, we identified 4575 confirmed and 3092 probable adult PWH and 564 PrEP users. PWH was most identified by a combination of medications and conditions (3324, 43.4%) and drug exposure alone (2191, 28.6%), then less commonly by survey data alone (608, 7.9%) and lab alone (81, 1.1%).

Discussion and conclusion

Our methods serve as an overall framework for other researchers using AoU data for conducting HIV-related research.

Brief Report: Role of Gender-Affirming Hormonal Care in HIV Care Continuum Outcomes When Comparing Transgender Women With Cisgender Sexual Minority Men

BACKGROUND

Transgender women (transwomen) are frequently conflated with cisgender sexual minority men (cis-SMM) in HIV research. We examined the impact of socioeconomic and health conditions, and gender-affirming hormones in comparing HIV-related outcomes between cis-SMM and transwomen.

SETTING

Large tertiary care health system in the Bronx, NY.

METHODS

Retrospective cohort study of people with HIV receiving care in 2008-2017. We compared retention in care, antiretroviral therapy (ART) prescription, and viral suppression between cis-SMM and transwomen, using modified Poisson regression, adjusting for demographic and clinical factors. Transwomen were further stratified by receipt of estrogen prescription.

RESULTS

We included 166 transwomen (1.4%), 1936 cis-SMM (17%), 4715 other cisgender men (41%), and 4745 cisgender women (41%). Transwomen were more likely to have public insurance (78% vs 65%) and mental health (49% vs 39%) or substance use (43% vs 33%) diagnoses than cis-SMM. Compared with cis-SMM, transwomen prescribed estrogen (67% of transwomen) were more likely to be retained [adjusted risk ratio (aRR) 1.15, 95% confidence interval (CI) 1.08 to 1.23), prescribed ART (aRR 1.06, CI 1.01 to 1.11), and virally suppressed (aRR 1.08, CI 1.01 to 1.16). Transwomen not prescribed estrogen were less likely to be retained (aRR 0.92, CI 0.83 to 1.02), prescribed ART (aRR 0.90, CI 0.82 to 0.98), or virally suppressed (aRR 0.85, CI 0.76 to 0.95).

CONCLUSIONS

In the context of HIV, socioeconomic factors, comorbidities, and gender-affirming care distinguish transwomen from cis-SMM. Compared with cis-SMM, transwomen who were prescribed estrogen had better HIV care continuum outcomes; transwomen not prescribed estrogen had worse outcomes. These differences should be accounted for in HIV-related research.

Comparison of algorithms for identifying people with HIV from electronic medical records in a large, multi-site database

Objective

As electronic medical record (EMR) data are increasingly used in HIV clinical and epidemiologic research, accurately identifying people with HIV (PWH) from EMR data is paramount. We sought to evaluate EMR data types and compare EMR algorithms for identifying PWH in a multicenter EMR database.

Materials and Methods

We collected EMR data from 7 healthcare systems in the Chicago Area Patient-Centered Outcomes Research Network (CAPriCORN) including diagnosis codes, anti-retroviral therapy (ART), and laboratory test results.

Results

In total, 13 935 patients had a positive laboratory test for HIV; 33 412 patients had a diagnosis code for HIV; and 17 725 patients were on ART. Only 8576 patients had evidence of HIV-positive status for all 3 data types (laboratory results, diagnosis code, and ART). A previously validated combination algorithm identified 22 411 patients as PWH.

Conclusion

EMR algorithms that combine laboratory results, administrative data, and ART can be applied to multicenter EMR data to identify PWH.

Phenotypic Presentations of Heart Failure Among Patients With Chronic Inflammatory Diseases

Objective

Characterize incident heart failure (HF) phenotypes among patients with various chronic inflammatory diseases (CIDs).

Background

Several CIDs are associated with increased HF risk, but differences in HF phenotypes across CIDs are incompletely understood. No prior studies to our knowledge have manually adjudicated HF phenotypes across a CID spectrum.

Methods

We screened for patients with—and controls without—CIDs who had possible HF, then hand-adjudicated HF endpoints. Possible HF resulted from a single HF administrative code; HF was deemed definite/probable vs. absent using standardized, validated criteria. We queried adjudicated HF patients' charts to define specific HF phenotypes, then compared clinical, demographic, and HF phenotypic characteristics for HF patients with specific CIDs vs. non-CID controls using Fisher's exact test.

Results

Out of 415 possible HF patients, 192 had definite/probable HF. Significant differences in HF phenotypes existed across CIDs. Isolated right-sided HF was present in 27.8% of patients with SSc and adjudicated HF, which is more than twice as common as it was in any other CID. Left ventricular systolic dysfunction was most common in patients with HIV and lupus (SLE); mean LVEF was 45.0% ± 18.6% for HIV and 41.3% ± 17.1% for SLE, but was 57.7% ± 10.7% for SSc. Those with HIV and multiple CIDs were most likely to have coronary artery disease.

Conclusions

Different CIDs present with different phenotypes of physician-adjudicated HF, potentially reflecting different underlying inflammatory pathophysiologies. Larger studies are needed to confirm these findings, as are mechanistic studies focused on understanding specific immunoregulatory contributors to HF.

A Phenotyping Algorithm to Identify People With HIV in Electronic Health Record Data (HIV-Phen): Development and Evaluation Study

Background

Identification of people with HIV from electronic health record (EHR) data is an essential first step in the study of important HIV outcomes, such as risk assessment. This task has been historically performed via manual chart review, but the increased availability of large clinical data sets has led to the emergence of phenotyping algorithms to automate this process. Existing algorithms for identifying people with HIV rely on a combination of International Classification of Disease codes and laboratory tests or closely mimic clinical testing guidelines for HIV diagnosis. However, we found that existing algorithms in the literature missed a significant proportion of people with HIV in our data.

Objective

The aim of this study is to develop and evaluate HIV-Phen, an updated criteria-based HIV phenotyping algorithm.

Methods

We developed an algorithm using HIV-specific laboratory tests and medications and compared it with previously published algorithms in national and local data sets to identify cohorts of people with HIV. Cohort demographics were compared with those reported in the national and local surveillance data. Chart reviews were performed on a subsample of patients from the local database to calculate the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the algorithm.

Results

Our new algorithm identified substantially more people with HIV in both national (up to an 85.75% increase) and local (up to an 83.20% increase) EHR databases than the previously published algorithms. The demographic characteristics of people with HIV identified using our algorithm were similar to those reported in national and local HIV surveillance data. Our algorithm demonstrated improved sensitivity over existing algorithms (98% vs 56%-92%) while maintaining a similar overall accuracy (96% vs 80%-96%).

Conclusions

We developed and evaluated an updated criteria-based phenotyping algorithm for identifying people with HIV in EHR data that demonstrates improved sensitivity over existing algorithms.

Comparative Risk of Incident Coronary Heart Disease Across Chronic Inflammatory Diseases

Background: Chronic inflammatory diseases (CIDs) are considered risk enhancing factors for coronary heart disease (CHD). However, sparse data exist regarding relative CHD risks across CIDs. Objective: Determine relative differences in CHD risk across multiple CIDs: psoriasis, rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), human immunodeficiency virus (HIV), systemic sclerosis (SSc), and inflammatory bowel disease (IBD). Methods: The cohort included patients with CIDs and controls without CID in an urban medical system from 2000 to 2019. Patients with CIDs were frequency-matched with non-CID controls on demographics, hypertension, and diabetes. CHD was defined as myocardial infarction (MI), ischemic heart disease, and/or coronary revascularization based on validated administrative codes. Multivariable-adjusted Cox models were used to determine the risk of incident CHD and MI for each CID relative to non-CID controls. In secondary analyses, we compared CHD risk by disease severity within each CID. Results: Of 17,049 patients included for analysis, 619 had incident CHD (202 MI) over an average of 4.4 years of follow-up. The multivariable-adjusted risk of CHD was significantly higher for SLE [hazard ratio (HR) 1.9, 95% confidence interval (CI) 1.2, 3.2] and SSc (HR 2.1, 95% CI 1.2, 3.9). Patients with SLE also had a significantly higher risk of MI (HR 3.6, 95% CI 1.9, 6.8). When CIDs were categorized by markers of disease severity (C-reactive protein for all CIDs except HIV, for which CD4 T cell count was used), greater disease severity was associated with higher CHD risk across CIDs. Conclusions: Patients with SLE and SSc have a higher risk of CHD. CHD risk with HIV, RA, psoriasis, and IBD may only be elevated in those with greater disease severity. Clinicians should personalize CHD risk and treatment based on type and severity of CID.

Optimizing Identification of People Living with HIV from Electronic Medical Records: Computable Phenotype Development and Validation

BACKGROUND

Electronic health record (EHR)-based computable phenotype algorithms allow researchers to efficiently identify a large virtual cohort of Human Immunodeficiency Virus (HIV) patients. Built upon existing algorithms, we refined, improved, and validated an HIV phenotype algorithm using data from the OneFlorida Data Trust, a repository of linked claims data and EHRs from its clinical partners, which provide care to over 15 million patients across all 67 counties in Florida.

METHODS

Our computable phenotype examined information from multiple EHR domains, including clinical encounters with diagnoses, prescription medications, and laboratory tests. To identify an HIV case, the algorithm requires the patient to have at least one diagnostic code for HIV and meet one of the following criteria: have 1+ positive HIV laboratory, have been prescribed with HIV medications, or have 3+ visits with HIV diagnostic codes. The computable phenotype was validated against a subset of clinical notes.

RESULTS

Among the 15+ million patients from OneFlorida, we identified 61,313 patients with confirmed HIV diagnosis. Among them, 8.05% met all four inclusion criteria, 69.7% met the 3+ HIV encounters criteria in addition to having HIV diagnostic code, and 8.1% met all criteria except for having positive laboratories. Our algorithm achieved higher sensitivity (98.9%) and comparable specificity (97.6%) relative to existing algorithms (77-83% sensitivity, 86-100% specificity). The mean age of the sample was 42.7 years, 58% male, and about half were Black African American. Patients' average follow-up period (the time between the first and last encounter in the EHRs) was approximately 4.6 years. The median number of all encounters and HIV-related encounters were 79 and 21, respectively.

CONCLUSION

By leveraging EHR data from multiple clinical partners and domains, with a considerably diverse population, our algorithm allows more flexible criteria for identifying patients with incomplete laboratory test results and medication prescribing history compared with prior studies.

Brief Report: Acute Kidney Injury in People Living With HIV Hospitalized With Coronavirus Disease 2019: Clinical Characteristics and Outcomes

BACKGROUND

Data on clinical characteristics and outcomes of people living with HIV (PLWH) hospitalized with coronavirus disease 2019 (COVID-19) who develop acute kidney injury (AKI) are limited.

SETTING

Large tertiary health care system in the Bronx, NY.

METHODS

We performed a retrospective cohort study of 83 PLWH and 4151 patients without HIV hospitalized with COVID-19 from March 10, 2020, to May 11, 2020. We compared the clinical characteristics and outcomes associated with AKI by HIV serostatus and evaluated HIV-related factors for AKI among PLWH. AKI was defined and staged using Kidney Disease Improving Global Outcomes criteria.

RESULTS

The incidence of AKI in hospitalized patients with COVID-19 did not differ significantly by HIV serostatus (54.2% in PLWH vs 49.5% in patients without HIV, P = 0.6). Despite a higher incidence of stage 3 AKI (28.9% vs 17.1% P = 0.05) in PLWH compared with those without HIV, there was no significant difference in the need for renal replacement therapy (22.2% vs 13.4% P = 0.12), renal recovery (76.9% vs 82.5% P = 0.61), or dependence on renal replacement therapy (7.7% vs 3.8% P = 0.27). CD4 T-cell count, HIV-1 RNA viral suppression, and antiretroviral therapy use were not associated with AKI. AKI was associated with increased need for invasive ventilation and in-hospital death, but HIV was not an independent risk factor of in-hospital death after AKI [adjusted hazard ratio 1.01 (95% CI: 0.59 to 1.72), P = 0.98].

CONCLUSIONS

HIV-related factors were not associated with increased risk of AKI in PLWH hospitalized with COVID-19. PLWH hospitalized with COVID-19 had more stage 3 AKI, but outcomes after AKI were similar to those without HIV.

The associations of CD4 count, CD4/CD8 ratio, and HIV viral load with survival from non-small cell lung cancer in persons living with HIV

HIV status may influence survival from non-small cell lung cancer (NSCLC). Among NSCLC patients in the Bronx, NY, we assessed (1) associations of CD4 count, CD4/CD8 ratio and HIV viral load (VL) with survival and (2) prognostic factors among persons living with HIV (PLWH). We compared survival from NSCLC diagnosis (2004-2017) between HIV-negative persons (HIV-, n=2,881) and PLWH (n=88) accounting for clinical and sociodemographic factors. HIV-survival was also compared with PLWH, dichotomized by CD4 (<200 vs. ≥200cells/µL), CD4/CD8 (median, <0.43 vs. ≥0.43) and VL (<75 vs. ≥75copies/mL) at NSCLC diagnosis. Among PLWH, we assessed the relationships of CD4, CD4/CD8, and VL with survival, adjusting for age, sex, and cancer stage. PLWH with CD4< 200cells/µL had lower survival than HIV- [hazard ratio, 95% confidence interval [HR(95%CI)]=1.86(0.98-3.55)]. Survival was similar between PLWH with CD4≥ 200cells/µL and HIV- [HR(95%CI) = 0.90(0.61-1.33)]. Results were similar when categorizing PLWH by CD4/CD8 [vs. HIV-: low CD4/CD8: HR(95%CI) = 1.74(1.07-3.89); high CD4/CD8: HR(95%CI) = 0.63(0.37-1.07)] and VL [vs. HIV-: <75copies/mL: HR(95%CI) = 0.74(0.46-1.21), ≥75copies/mL: HR(95%CI) = 1.41(0.88-2.27)]. Among PLWH, CD4< 200cells/µL was associated with worse survival [vs. CD4≥ 200cells/µL: HR(95%CI) = 2.37(1.14-4.92)]. CD4, CD4/CD8, and VL may be prognostic markers for PLWH with NSCLC, suggesting immune status may be important in NSCLC survival among PLWH.

Clinical Outcomes and Inflammatory Markers by HIV Serostatus and Viral Suppression in a Large Cohort of Patients Hospitalized With COVID-19

BACKGROUND

Limited data exist about clinical outcomes and levels of inflammatory and immune markers among people hospitalized with COVID-19 by HIV serostatus and by HIV viral suppression.

SETTING

Large tertiary care health system in the Bronx, NY, USA.

METHODS

We conducted a retrospective cohort study of 4613 SARS-CoV-2 PCR-positive patients admitted between March 10, 2020, and May 11, 2020. We examined in-hospital intubation, acute kidney injury (AKI), hospitalization length, and in-hospital mortality by HIV serostatus, and by HIV-viral suppression and CD4 counts among people living with HIV (PLWH) using adjusted competing risks regression. We also compared immune and inflammatory marker levels by HIV serostatus and viral suppression.

RESULTS

Most patients were either non-Hispanic Black (36%) or Hispanic (37%); 100/4613 (2.2%) were PLWH, among whom 15 had detectable HIV viral load. PLWH compared to patients without HIV had increased intubation rates (adjusted hazard ratio 1.73 [95% CI: 1.12 to 2.67], P = 0.01). Both groups had similar rates of AKI, length of hospitalization, and death. No (0%) virally unsuppressed PLWH were intubated or died, versus 21/81 (26%, P = 0.04) and 22/81 (27%, P = 0.02) of virally suppressed PLWH, respectively. Among PLWH, higher CD4 T-cell counts were associated with increased intubation rates. C-reactive protein, IL-6, neutrophil counts, and ferritin levels were similar between virally suppressed PLWH and patients without HIV, but significantly lower for unsuppressed PLWH (all P < 0.05).

CONCLUSIONS

PLWH had increased risk of intubation but similarly frequent rates of AKI and in-hospital death as those without HIV. Findings of no intubations or deaths among PLWH with unsuppressed HIV viral load warrant further investigation.

Electronic Medical Record-Based Case Phenotyping for the Charlson Conditions: Scoping Review

BACKGROUND

Electronic medical records (EMRs) contain large amounts of rich clinical information. Developing EMR-based case definitions, also known as EMR phenotyping, is an active area of research that has implications for epidemiology, clinical care, and health services research.

OBJECTIVE

This review aims to describe and assess the present landscape of EMR-based case phenotyping for the Charlson conditions.

METHODS

A scoping review of EMR-based algorithms for defining the Charlson comorbidity index conditions was completed. This study covered articles published between January 2000 and April 2020, both inclusive. Embase (Excerpta Medica database) and MEDLINE (Medical Literature Analysis and Retrieval System Online) were searched using keywords developed in the following 3 domains: terms related to EMR, terms related to case finding, and disease-specific terms. The manuscript follows the Preferred Reporting Items for Systematic reviews and Meta-analyses extension for Scoping Reviews (PRISMA) guidelines.

RESULTS

A total of 274 articles representing 299 algorithms were assessed and summarized. Most studies were undertaken in the United States (181/299, 60.5%), followed by the United Kingdom (42/299, 14.0%) and Canada (15/299, 5.0%). These algorithms were mostly developed either in primary care (103/299, 34.4%) or inpatient (168/299, 56.2%) settings. Diabetes, congestive heart failure, myocardial infarction, and rheumatology had the highest number of developed algorithms. Data-driven and clinical rule-based approaches have been identified. EMR-based phenotype and algorithm development reflect the data access allowed by respective health systems, and algorithms vary in their performance.

CONCLUSIONS

Recognizing similarities and differences in health systems, data collection strategies, extraction, data release protocols, and existing clinical pathways is critical to algorithm development strategies. Several strategies to assist with phenotype-based case definitions have been proposed.

Algorithm to identify transgender and gender nonbinary individuals among people living with HIV performs differently by age and ethnicity

PURPOSE

HIV research among transgender and gender nonbinary (TGNB) people is limited by lack of gender identity data collection. We designed an EHR-based algorithm to identify TGNB people among people living with HIV (PLWH) when gender identity was not systematically collected.

METHODS

We applied EHR-based search criteria to all PLWH receiving care at a large urban health system between 1997 and 2017, then confirmed gender identity by chart review. We compared patient characteristics by gender identity and screening criteria, then calculated positive predictive values for each criterion.

RESULTS

Among 18,086 PLWH, 213 (1.2%) met criteria as potential TGNB patients and 178/213 were confirmed. Positive predictive values were highest for free-text keywords (91.7%) and diagnosis codes (77.4%). Confirmed TGNB patients were younger (median 32.5 vs. 42.5 years, P < .001) and less likely to be Hispanic (37.1% vs. 62.9%, P = .03) than unconfirmed patients. Among confirmed patients, 15% met criteria only for prospective gender identity data collection and were significantly older.

CONCLUSION

EHR-based criteria can identify TGNB PLWH, but success may differ by ethnicity and age. Retrospective versus intentional, prospective gender identity data collection may capture different patients. To reduce misclassification in epidemiologic studies, gender identity data collection should address these potential differences and be systematic and prospective.

Differential Associations of Chronic Inflammatory Diseases With Incident Heart Failure

OBJECTIVES

The purpose of this study was to compare the risks of incident heart failure (HF) among a variety of chronic inflammatory diseases (CIDs) and to determine whether risks varied by severity of inflammation within each CID.

BACKGROUND

Individuals with CIDs are at elevated risk for cardiovascular diseases, but data are limited regarding risk for HF.

METHODS

An electronic health records database from a large urban medical system was examined, comparing individuals with CIDs with frequency-matched controls without CIDs, all of whom were receiving regular outpatient care. Rates of incident HF were determined by using the Kaplan-Meier method and subsequently used multivariate-adjusted proportional hazards models to compare HF risks for each CID. Exploratory analyses determined HF risks by proxy measurement of CID severity.

RESULTS

Of 37,636 patients (n = 18,278 patients with CIDs; and n = 19,358 controls without CIDs) there were 960 incident HF cases over a median of 3.6 years. Risks for incident HF were significantly or borderline significantly elevated for patients with systemic sclerosis (hazard ratio [HR]: 7.26; 95% confidence interval [CI]: 5.72 to 9.21; p < 0.01), systemic lupus erythematosus (HR: 3.15; 95% CI: 2.41 to 4.11; p < 0.01), rheumatoid arthritis (HR: 1.39; 95% CI: 1.13 to 1.71; p < 0.01), and human immunodeficiency virus (HR: 1.28; 95% CI: 0.99 to 1.66; p = 0.06). There was no association between psoriasis or inflammatory bowel disease and incident HF, although patients with those CIDs with higher levels of C-reactive protein had higher risks for HF than controls.

CONCLUSIONS

Systemic sclerosis and systemic lupus erythematosus were associated with the highest risks of HF, followed by rheumatoid arthritis and HIV. Measurements of inflammation were associated with HF risk across different CIDs.

Evaluation of Algorithms Used for PrEP Surveillance Using a Reference Population From New York City, July 2016-June 2018

OBJECTIVE

Daily tenofovir disoproxil fumarate/emtricitabine (TDF/FTC) use as HIV preexposure prophylaxis (PrEP) is monitored by identifying TDF/FTC prescriptions from pharmacy databases and applying diagnosis codes and antiretroviral data to algorithms that exclude TDF/FTC prescribed for HIV postexposure prophylaxis (PEP), HIV treatment, and hepatitis B virus (HBV) treatment. We evaluated the accuracy of 3 algorithms used by the Centers for Disease Control and Prevention (CDC), Gilead Sciences, and the New York State Department of Health (NYSDOH) using a reference population in Bronx, New York.

METHODS

We extracted diagnosis codes and data on all antiretroviral prescriptions other than TDF/FTC from an electronic health record database for persons aged ≥16 prescribed TDF/FTC during July 2016-June 2018 at Montefiore Medical Center. We reviewed medical records to classify the true indication of first TDF/FTC use as PrEP, PEP, HIV treatment, or HBV treatment. We applied each algorithm to the reference population and compared the results with the medical record review.

RESULTS

Of 2862 patients included in the analysis, 694 used PrEP, 748 used PEP, 1407 received HIV treatment, and 13 received HBV treatment. The algorithms had high specificity (range: 98.4%-99.0%), but the sensitivity of the CDC algorithm using a PEP definition of TDF/FTC prescriptions ≤30 days was lower (80.3%) than the sensitivity of the algorithms developed by Gilead Sciences (94.7%) or NYSDOH (96.1%). Defining PEP as TDF/FTC prescriptions ≤28 days improved CDC algorithm performance (sensitivity, 95.8%; specificity, 98.8%).

CONCLUSIONS

Adopting the definition of PEP as ≤28 days of TDF/FTC in the CDC algorithm should improve the accuracy of national PrEP surveillance.

Adjudicated Heart Failure in HIV-Infected and Uninfected Men and Women

Background HIV is associated with elevated risk of heart failure ( HF ). Despite poor agreement between automated, administrative code-based HF definitions and physician-adjudicated HF , no studies have evaluated incident adjudicated HF for people living with HIV ( PLWH ). Methods and Results We analyzed PLWH and uninfected controls receiving care in an urban medical system from January 1, 2000, to July 12, 2016. Physicians reviewed data from medical records to adjudicate HF diagnoses. We used multivariable-adjusted Cox models to analyze incident HF for PLWH versus controls and HIV -related factors associated with incident HF . We also analyzed the performance of automated versus physician-adjudicated HF definitions. Incident adjudicated HF occurred in 97 of 4640 PLWH (2.1%; mean: 6.8 years to HF ) and 55 of 4250 controls (1.3%; mean: 6.7 years to HF ; multivariable-adjusted hazard ratio: 2.10; 95% confidence interval, 1.38-3.21). Among PLWH , higher HIV viral load ( hazard ratio per log₁₀ higher time-updated viral load: 1.22; 95% confidence interval, 1.11-1.33) was associated with greater HF risk and higher CD 4+ T cell count was associated with lower HF risk ( hazard ratio per 100 cells/mm³ higher time-updated CD 4 count: 0.80; 95% confidence interval, 0.69-0.92). In exploratory analyses, the most accurate automated HF definitions had sensitivities of 67% to 75% and positive predictive values of 54% to 60%. Conclusions In a cohort with rigorous HF adjudication, PLWH had greater risks of HF than uninfected people after adjustment for demographics and cardiovascular risk factors. Higher HIV viral load and lower CD 4+ T cell count were associated with higher HF risk among PLWH . Automated methods of HF ascertainment exhibited poor accuracy for PLWH and uninfected people.

Development and validation of an electronic medical record (EMR)-based computed phenotype of HIV-1 infection

Background

Electronic medical record (EMR) computed algorithms allow investigators to screen thousands of patient records to identify specific disease cases. No computed algorithms have been developed to detect all cases of human immunodeficiency virus (HIV) infection using administrative, laboratory, and clinical documentation data outside of the Veterans Health Administration. We developed novel EMR-based algorithms for HIV detection and validated them in a cohort of subjects in the Duke University Health System (DUHS).

Methods

We created 2 novel algorithms to identify HIV-infected subjects. Algorithm 1 used laboratory studies and medications to identify HIV-infected subjects, whereas Algorithm 2 used International Classification of Diseases, Ninth Revision (ICD-9) codes, medications, and laboratory testing. We applied the algorithms to a well-characterized cohort of patients and validated both against the gold standard of physician chart review. We determined sensitivity, specificity, and prevalence of HIV between 2007 and 2011 in patients seen at DUHS.

Results

A total of 172 271 patients were detected with complete data; 1063 patients met algorithm criteria for HIV infection. In all, 970 individuals were identified by both algorithms, 78 by Algorithm 1 alone, and 15 by Algorithm 2 alone. The sensitivity and specificity of each algorithm were 78% and 99%, respectively, for Algorithm 1 and 77% and 100% for Algorithm 2. The estimated prevalence of HIV infection at DUHS between 2007 and 2011 was 0.6%.

Conclusions

EMR-based phenotypes of HIV infection are capable of detecting cases of HIV-infected adults with good sensitivity and specificity. These algorithms have the potential to be adapted to other EMR systems, allowing for the creation of cohorts of patients across EMR systems.

Differences in statin utilization and lipid lowering by race, ethnicity, and HIV status in a real-world cohort of persons with human immunodeficiency virus and uninfected persons

BACKGROUND

Risks for cardiovascular diseases, including myocardial infarction and stroke, are elevated in people with HIV infection (PWH). However, no trials of statin utilization with clinical cardiovascular disease (CVD) end points have been completed in PWH, and there are sparse real-world data regarding statin use and lipid-lowering effectiveness. We therefore used a unique cohort of PWH and uninfected controls to evaluate (1) differences in statin types used for PWH versus uninfected persons; (2) lipid lowering achieved by statin use for PWH versus uninfected persons; and (3) racial and ethnic disparities in appropriate statin use among PWH and uninfected persons.

METHODS

We analyzed a cohort of 5,039 PWH and 10,011 uninfected demographically matched controls who received care at a large urban medical center between January 1, 2000, and May 17, 2017. Medication administration records, prescription data, and validated natural language processing algorithms were used to determine statin utilization. Statins were categorized by generic active ingredient name and intensity (high, moderate, or low). Lipid values collected in routine clinical care were available for analysis. The first set of analyses was restricted to PWH and uninfected matched controls taking statins and compared (1) differences in statin type and (2) difference in cholesterol levels after versus before statin initiation by HIV status. For the second set of analyses, we first used prevalent CVD risk factors to determine participants with statin indications and then determined how many of these participants were taking statins. We then compared statin utilization among persons with indications for statins by race/ethnic group for PWH and uninfected matched controls using multivariable-adjusted logistic regression.

RESULTS

Among people prescribed statins, PWH were more likely than controls to have ever taken pravastatin (34.8% vs 12.3%, P < .001) or atorvastatin (72.2% vs 65.6%, P = .002) and less likely to have ever taken simvastatin (14.2% vs 39.5%, P < .001). Among PWH with indications for statin utilization, 55.7% of whites, 39.4% of blacks, and 45.8% of Hispanics were prescribed statins (P < .001). These differences in statin prescription by race/ethnicity remained significant after adjustment for demographics (including insurance status), cardiovascular risk factors, antiretroviral therapy use, HIV viremia, and CD4 count. These racial/ethnic disparities in statin utilization were less pronounced among uninfected persons.

CONCLUSIONS

Among PWH with statin indication(s), blacks and Hispanics were less likely than whites to have been prescribed a statin. These racial/ethnic disparities were less pronounced among uninfected persons. There were significant differences in type of statin used for PWH compared to uninfected matched controls. Future efforts addressing disparities in CVD prevention among PWH are warranted.

Ventricular Ectopy and Arrhythmia Characteristics for Persons Living with HIV and Uninfected Controls

BACKGROUND

Persons with HIV have elevated risk for cardiovascular disease, but little is known about the risk of ventricular ectopy and ventricular tachycardia (VE/VT) for HIV-infected (HIV+) persons.

METHODS

We evaluated the presence and anatomic origin of VE/VT for HIV+ persons and controls by screening a cohort using International Classification of Diseases codes and adjudicating positive screens by chart review. We sought to evaluate (1) presence of VE/VT and (2) likely anatomic origin of the VE/VT based on electrocardiogram.

RESULTS

There was no significant difference in the prevalence of VE/VT for HIV+ or uninfected persons. Among HIV+ persons, worse HIV control was associated with significantly greater odds of VE/VT. Exploratory analyses suggested that HIV+ persons may have a greater likelihood of VE/VT originating from the left ventricle.

CONCLUSION

Although worse HIV control was associated with higher odds of VE/VT among persons with HIV, odds of VE/VT were not higher for persons with HIV than uninfected persons.

Development and Validation of an Electronic Medical Record-Based Algorithm to Identify Patient Milestones in the Hepatitis C Virus Care Cascade

Disease-specific care cascades are important public health and organizational tools to characterize gaps in care and target resources, but they are labor-intensive to maintain. Using data available from the electronic medical record, we developed an algorithm with high accuracy for correctly representing an individual’s status in the hepatitis C virus care cascade.

Atrial arrhythmia prevalence and characteristics for human immunodeficiency virus-infected persons and matched uninfected controls

BACKGROUND

Human Immunodeficiency Virus-Infected (HIV+) persons have elevated risks for various manifestations of cardiovascular disease (CVD). No studies to our knowledge have compared atrial fibrillation (AF) and atrial flutter (AFL) prevalence and associated characteristics for HIV+ persons and matched uninfected controls.

METHODS AND FINDINGS

Persons with diagnoses of HIV receiving care at a large urban academic medical center were frequency-matched 1:2 on age, sex, race, zip code, and clinic location with uninfected persons. Possible AF/AFL was screened for using administrative codes and diagnoses of AF/AFL were subsequently adjudicated using electrocardiography and physician notes; adjudication was performed given the inconsistent validity of administrative code-derived AF diagnoses found in previous studies. There were 101 confirmed AF/AFL cases (2.00%) among 5,052 HIV+ patients and 159 confirmed AF/AFL cases (1.57%) among 10,121 uninfected controls [Odds Ratio (OR) 1.27, 95% Confidence Interval (CI) 0.99-1.64; p = 0.056]. The association between HIV serostatus and AF/AFL was attenuated after adjustment for demographics and CVD risk factors. Among HIV+ persons, nadir CD4+ T cell count <200 cells/mm3 was associated with approximately twofold elevated odds of AF/AFL even after adjustment for demographics and CVD risk factors (Multivariable-adjusted OR 1.98, 95% CI 1.21-3.25). There was no significant association between log10 of peak HIV viral load and AF/AFL (Multivariable-adjusted OR 1.03, 95% CI 0.86-1.24). Older age, diabetes, hypertension, and chronic obstructive pulmonary disease were associated with similarly elevated odds of AF/AFL for HIV+ persons and uninfected controls.

CONCLUSION

HIV-related immunosuppression (nadir CD4 T cell count <200 cells/mm3) and traditional CVD risk factors are associated with significantly elevated odds of AF/AFL among HIV+ persons. Although atrial fibrillation and flutter was more common among HIV+ versus uninfected persons in this cohort, this difference was attenuated by adjustment for demographics and CVD risk factors.

Clinical characteristics of HIV-infected patients with adjudicated heart failure

Aims HIV-infected persons may have elevated risks for heart failure, but factors associated with heart failure in the modern era of HIV therapy are insufficiently understood. Despite substantial disagreement between physician-adjudicated heart failure and heart failure diagnosis codes, few studies of HIV cohorts have evaluated adjudicated heart failure. We evaluated associations of HIV viremia, immunosuppression, and cardiovascular risk factors with physician-adjudicated heart failure. Methods and results We analyzed clinical characteristics associated with heart failure in a cohort of 5041 HIV-infected patients receiving care at an urban hospital system between 2000 and 2016. We also evaluated characteristics of HIV-infected patients who screened negative for heart failure, screened positive for possible heart failure but did not have heart failure after adjudication, and had adjudicated heart failure. HIV-infected patients with heart failure ( N = 216) were older and more likely to be black, hypertensive, and have diabetes than HIV-infected patients without heart failure; heart failure with reduced ejection fraction was more common than heart failure with preserved ejection fraction. In our primary analyses restricted to HIV-infected patients whose heart failure diagnoses did not precede their HIV diagnoses ( N = 149), peak HIV viral load ≥100,000 copies/mL (odds ratio (OR) 2.12, 1.28-3.52) and nadir CD4 T-cell count <200 cells/mm³ (OR 2.35, 1.04-5.31) were associated with significantly elevated odds of heart failure. Overall, 30.6% of patients with any diagnosis code of heart failure had adjudicated heart failure. Conclusion Higher peak HIV viremia and lower CD4 cell nadir are associated with significantly elevated odds of heart failure for HIV-infected persons. Physician adjudication of heart failure may be helpful in HIV cohorts.

Expanded HIV Testing Strategy Leveraging the Electronic Medical Record Uncovers Undiagnosed Infection Among Hospitalized Patients

BACKGROUND

Routine HIV testing of hospitalized patients is recommended, but few strategies to expand testing in the hospital setting have been described. We assessed the impact of an electronic medical record (EMR) prompt on HIV testing for hospitalized patients.

METHODS

We performed a pre-post study at 3 hospitals in the Bronx, NY. We compared the proportion of admissions of patients 21-64 years old with an HIV test performed, characteristics of patients tested, and rate of new HIV diagnoses made by screening while an EMR prompt recommending HIV testing was inactive vs. active. The prompt appeared for patients with no previous HIV test or a high-risk diagnosis after their last HIV test.

RESULTS

Among 36,610 admissions while the prompt was inactive, 9.5% had an HIV test performed. Among 18,943 admissions while the prompt was active, 21.8% had an HIV test performed. Admission while the prompt was active was associated with increased HIV testing among total admissions [adjusted odds ratio (aOR) 2.78, 95% confidence interval (CI): 2.62 to 2.96], those without a previous HIV test (aOR 4.03, 95% CI: 3.70 to 4.40), and those with a previous negative test (aOR 1.52, 95% CI: 1.37 to 1.68) (P < 0.0001 for all). Although the prompt was active, testing increased across all patient characteristics. New HIV diagnoses made by screening increased from 8.2/100,000 admissions to 37.0/100,000 admissions while the prompt was inactive and active, respectively (OR 4.51 95% CI: 1.17 to 17.45, P = 0.03).

CONCLUSIONS

An EMR prompt for hospitalized patients was associated with a large increase in HIV testing, a diversification of patients tested, and an increase in diagnoses made by screening.

Case-Control Study of Herpes Simplex Eye Disease: Bronx Epidemiology of Human Immunodeficiency Virus Eye Studies

PURPOSE

To determine whether human immunodeficiency virus and acquired immune deficiency syndrome (HIV/AIDS), diabetes mellitus, and atopic disease are associated with herpes simplex virus (HSV) eye disease and to examine the characteristics of patients with HIV infection and HSV eye disease.

DESIGN

Retrospective case-control study.

METHODS

A hospital-based control group was matched to outpatient visits from June 1, 2010, through May 31, 2014, at Montefiore Medical Center (Bronx, NY). Inclusion criteria were a diagnosis of HSV eye disease during the study period and residency in the Bronx. Associations evaluated included age, sex, HIV/AIDS, diabetes mellitus, and atopic disease.

RESULTS

HSV eye disease was confirmed in 70 patients, who were compared with 280 controls. Patients with ocular HSV had a greater prevalence of HIV/AIDS compared with controls (8.6% and 2.9%, respectively). Using multivariate analysis to control for age and sex, atopic dermatitis [odds ratio (OR) 3.08, 95% confidence interval (CI), 0.84-11.20] and diabetes with chronic complications (OR 2.25, 95% CI, 0.91-5.61) approached significance, whereas HIV/AIDS (OR 3.37, 95% CI, 1.09-10.40), an age less than 45 years (OR 2.89, 95% CI, 1.54-5.41), and male sex (OR 1.85, 95% CI, 1.07-3.18) were significant. In patients with HIV infection and HSV eye disease, 3 of 6 (50%) had confirmed AIDS at the time of ocular HSV diagnosis.

CONCLUSIONS

Patients with HIV infection have more than a 3-fold increased risk of developing ocular HSV.

Increased Antiretroviral Therapy Use and Virologic Suppression in the Bronx in the Context of Multiple HIV Prevention Strategies

Multiple population-based HIV prevention strategies from national, state, local, and institutional levels have been implemented in the Bronx, which has one of the highest HIV prevalences in the U.S. We examined changes in antiretroviral therapy (ART) use and associated outcomes between 2007 and 2014 among patients seen at one of >20 outpatient clinics affiliated with the largest Bronx HIV care provider. Among eligible HIV-infected patients age ≥13 years, we examined annual trends in ART use, mean HIV RNA level, and virologic suppression (<200 copies/ml) overall and among prespecified subgroups. In a subset with suppressed HIV RNA at the end of each year, we determined the percentage whose levels remained suppressed within the next year. Regression models assessed disparities in outcomes. Among 7,196 patients (median age 50, 47% Hispanic, 45% black), we identified consistent increases over time in the percent prescribed ART (78% in 2007 to 93% in 2014) and with virologic suppression (58% to 80%), as mean HIV RNA decreased (351 to 73 copies/ml) (all p < .001). Sustained virologic suppression improved markedly beginning in 2011, coinciding with local test-and-treat initiatives and adoption of expanded treatment guidelines. While disparities among population groups were most pronounced for sustained virologic suppression, those aged 13-24 fared relatively poorly for all outcomes examined (e.g., rate ratio 0.57 for virologic suppression, 95% confidence interval 0.52-0.62, vs. age 65+). Population-wide HIV prevention strategies coincided with improvements in virologic suppression among most population groups. However, more attention is needed to address continued disparities in the HIV care continuum among young people.

The impact of nurse-driven targeted HIV screening in 8 emergency departments: study protocol for the DICI-VIH cluster-randomized two-period crossover trial

BACKGROUND

In 2010, to reduce late HIV diagnosis, the French national health agency endorsed non-targeted HIV screening in health care settings. Despite these recommendations, non-targeted screening has not been implemented and only physician-directed diagnostic testing is currently performed. A survey conducted in 2010 in 29 French Emergency Departments (EDs) showed that non-targeted nurse-driven screening was feasible though only a few new HIV diagnoses were identified, predominantly among high-risk groups. A strategy targeting high-risk groups combined with current practice could be shown to be feasible, more efficient and cost-effective than current practice alone.

METHODS/DESIGN

DICI-VIH (acronym for nurse-driven targeted HIV screening) is a multicentre, cluster-randomized, two-period crossover trial. The primary objective is to compare the effectiveness of 2 strategies for diagnosing HIV among adult patients visiting EDs: nurse-driven targeted HIV screening combined with current practice (physician-directed diagnostic testing) versus current practice alone. Main secondary objectives are to compare access to specialist consultation and how early HIV diagnosis occurs in the course of the disease between the 2 groups, and to evaluate the implementation, acceptability and cost-effectiveness of nurse-driven targeted screening. The 2 strategies take place during 2 randomly assigned periods in 8 EDs of metropolitan Paris, where 42 % of France's new HIV patients are diagnosed every year. All patients aged 18 to 64, not presenting secondary to HIV exposure are included. During the intervention period, patients are invited to fill a 7-item questionnaire (country of birth, sexual partners and injection drug use) in order to select individuals who are offered a rapid test. If the rapid test is reactive, a follow-up visit with an infectious disease specialist is scheduled within 72 h. Assuming an 80 % statistical power and a 5 % type 1 error, with 1.04 and 3.38 new diagnoses per 10,000 patients in the control and targeted groups respectively, a sample size of 140,000 patients was estimated corresponding to 8,750 patients per ED and per period. Inclusions started in June 2014. Results are expected by mid-2016.

DISCUSSION

The DICI-VIH study is the first large randomized controlled trial designed to assess nurse-driven targeted HIV screening. This study can provide valuable information on HIV screening in health care settings.

TRIAL REGISTRATION

ClinicalTrials.gov: NCT02127424 (29 April 2014).

Increased HIV testing among hospitalized patients who declined testing in the emergency department

Health-care systems have serial encounters with many of the same patients across care settings; however, few studies have examined the role of reoffering HIV testing after a patient declines. We assessed whether an intervention to increase HIV testing among hospitalized patients was associated with increased testing among those who declined a test while in the Emergency Department (ED). We studied 8-week periods pre- and post-implementation of an electronic medical record (EMR)-based intervention to increase HIV testing among hospitalized patients. We included all patients 21-64 years old who had no prior HIV test, declined HIV testing in the ED, and were subsequently hospitalized. We used logistic regression to test for an association between time of hospital admission (pre- vs. post-intervention) and whether an HIV test was performed prior to discharge. Pre- and post-implementation, 220 and 218 patients who declined HIV testing in the ED were hospitalized, respectively. There were no significant demographic or clinical differences among patients pre- and post-implementation. Pre- and post-implementation, the median proportion of patients tested weekly was 6.7% (IQR 6.5%, 10.0%) and 41.4% (IQR 33.3%, 41.9%), respectively (aOR 6.2: 95%CI: 3.6, 10.6). HIV testing increased among hospitalized patients who declined a test in the ED after implementation of an EMR-based intervention. Almost half of the patients who declined testing in the ED ultimately underwent testing after it was reoffered during hospitalization, suggesting that the decision to undergo HIV testing is a dynamic process. Leveraging EMR resources may be an effective tool for expanding HIV testing, and testing should be reoffered to patients who previously declined.

How next-generation sequencing and multiscale data analysis will transform infectious disease management

We should integrate next-generation sequencing data from pathogen specimens with phenotypes from electronic medical records to create quantitative, predictive models of infectious disease. Precision infection control and antimicrobial interventions can address urgent global problems, including healthcare-associated infections and multidrug resistance.

Recent reviews have examined the extent to which routine next-generation sequencing (NGS) on clinical specimens will improve the capabilities of clinical microbiology laboratories in the short term, but do not explore integrating NGS with clinical data from electronic medical records (EMRs), immune profiling data, and other rich datasets to create multiscale predictive models. This review introduces a range of “omics” and patient data sources relevant to managing infections and proposes 3 potentially disruptive applications for these data in the clinical workflow. The combined threats of healthcare-associated infections and multidrug-resistant organisms may be addressed by multiscale analysis of NGS and EMR data that is ideally updated and refined over time within each healthcare organization. Such data and analysis should form the cornerstone of future learning health systems for infectious disease.

Unknown HIV Status in the Emergency Department: Implications for Expanded Testing Strategies

OBJECTIVES

The authors sought to determine the prevalence of unknown HIV status among emergency department (ED) patients, how it has changed over time, and whether it differs according to patient characteristics.

METHODS

The authors used electronic medical record data to identify whether HIV status was known or unknown among patients aged ≥13 seen in the ED of a large, urban medical center between 2006 and 2011. The authors used multivariate logistic regression to identify the characteristics associated with unknown HIV status.

RESULTS

The prevalence of unknown HIV status decreased each year, from 87.7% in 2006 to 74.9% in 2011 (P < .001). Characteristics associated with unknown HIV status included being nonblack, in the youngest and oldest age-groups, and nonpublically insured. Compared to men, women without prior pregnancy were equally likely to have unknown HIV status, but women with prior pregnancy were significantly less likely to have unknown HIV status.

CONCLUSION

The prevalence of unknown HIV status is decreasing, but in 2011 75% of ED patients aged ≥13 still had unknown status, and it was associated with specific patient characteristics. Understanding the trends in the prevalence of unknown HIV status and how it is associated with patient characteristics should inform the design and implementation of expanded HIV-testing strategies.