Article Text
Abstract
Introduction The measure of sexual orientation and gender identity (SOGI) data in electronic health records (EHR) has been critical for addressing health disparities and inequalities, especially for HIV care. Given that gender and sexual minorities (eg, transgender, men who have sex with men and intersex) are key groups in people living with HIV (PLWH), SOGI data can facilitate a more accurate understanding about the HIV outcomes (eg, viral suppression) among this key group and then lead to tailored therapeutic services. The two-step SOGI collection method as an emerging gender measurement can be used to measure SOGI status in medical settings. Using the statewide cohort of PLWH in South Carolina (SC), this project aims to: (1) integrate statewide PLWH cohort data with their birth certificate data to evaluate SOGI measurements from multiple EHR sources; and (2) examine differences in viral suppression based on SOGI measurements.
Methods and analysis Our EHR database includes several HIV data sources with patients’ gender information, such as SC Department of Health and Environmental Control Centre (DHEC), Health Sciences South Carolina (HSSC) and Prisma as well as birth certificate data to retrieve the sex at birth. The SC Enhanced HIV/AIDS Reporting System (e-HARS) from DHEC will provide longitudinal viral load information to define a variety of viral suppression status. Datasources like the SC office of Revenue and Fiscal Affairs (RFA) will extract longitudinal EHR clinical data of all PLWH in SC from multiple health systems; obtain data from other state agencies and link the patient-level data with county-level data from multiple publicly available data sources.
Ethics and dissemination The study was approved by the Institutional Review Board at the University of South Carolina (Pro00129906) as a Non-Human Subject study. The study’s findings will be published in peer-reviewed journals and disseminated at national and international conferences and through social media.
- Electronic Health Records
- HIV & AIDS
- Sexual and Gender Minorities
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
The data integration from multiple statewide data sources in South Carolina to form cohort analyses allows us to generate a robust evaluation of the impact of sexual orientation and gender identity (SOGI) on viral suppression with a long follow-up.
The two-step method of data collection will be utilised as a reference approach and framework in this programme to conduct the three-way comparison analysis across three data sources (ie, birth certificate gender, recorded SOGI in electronic health record (EHR) and reported SOGI in Enhanced HIV/AIDS Reporting System) and better understand patients’ special needs in HIV care context, and subsequently lead to improved HIV treatment outcomes.
We are expecting to have missing data in both the birth certificate and other EHR data; thus, caution may be needed when interpreting the results.
Introduction
Gender, a social and cultural variable, encompasses several domains, each of which influences health: gender identity and expression, gender roles and norms, gender relations, structural sexism, power, and equality and equity. For example, gender socialisation and norms of masculinity influence boys’ and men’s health-seeking behaviours; and structural gender inequalities limit girls’ and women’s access to health services and contribute to health inequities. Other social variables—including race, ethnicity, socioeconomic status and State and Federal policies—may additionally interact with gender to influence health, highlighting the importance of an intersectional approach to health research.1
Accurate sexual orientation and gender identity (SOGI) measurements hold significant implications in data collection and is important for bringing visibility to SGM communities.2 Measuring SOGI of patients in clinical settings and entering the data into electronic health record (EHR) has been recommended by the Institute of Medicine,3 the Joint Commission4 and other medical and policy experts as key steps to measuring and addressing multiple health disparities among sexual and gender orientation minorities (SGM; eg, transgender, bisexual, gay men and men who have sex with men (MSM), lesbians, intersex, two-spirit, non-binary and gender-fluid).5 Routine SOGI data, particularly for SGM populations, are critical for population health management and the improvement of clinical decision support and culturally affirming, patient-centred care.6–10 Knowing the accurate SOGI data can facilitate healthcare providers to develop tailored prevention and improved rapport, and receive more related trainings in effective communication, SOGI core concepts and SGM healthcare best practices.7 8 Consequently, SOGI data are required both to better care for the individual patient by capturing a fuller picture of the patient’s life and experiences than is available from clinical data, and to support population-based analyses of clinical data that can inform clinical care for patients,11 leading to reduced stigma and discrimination, improved healthcare access and improved health outcomes.12 13 It should be noted that among these SGM populations, transgender individuals in clinical settings need more attention due to inherent flaws in their gender identity data measurement.14–16
HIV continues to be a significant public health issue, where populations with SOGI are disproportionately affected by the disease. The Institute of Medicine’s report on the health of SGM individuals highlighted the health disparities among them, such as higher prevalence of sexually transmitted infections and HIV.17 SGM individuals are more likely to be infected with HIV than the general population.18–24 For example, a meta-analysis demonstrated that transgender women had a staggering 66 times higher odds of being infected with HIV compared with HIV rates for people 15 years and older in the general population, while for transgender men, this was 6.8 times higher.25 In addition, these vulnerable populations (eg, transgender individuals and MSM) are reported to have more difficulties achieving better HIV treatment cascade (eg, viral suppression)26 and usually have worse HIV treatment outcomes than general populations.27 Studies have found that transgender individuals are less likely to be linked to care, retained in care and achieve higher levels of viral suppression.28–31 Thus, SOGI data measurement for PLWH are essential for healthcare providers (eg, case managers, physicians and nurses) to assess health risks and disparities of this key populations, facilitate HIV treatment cascade outcomes (eg, viral suppression) and build a truly patient-centred healthcare system.11
The two-step method of data collection is recommended as a framework for conducting accurate SOGI measurement.32 This method as a suggested SOGI measurement for assessing sex, gender and sexual orientation has been widely used in clinical and healthcare settings, particularly for SGM patients.1 33–37 In the first step, the natal sex of the individual is recorded; in the second step, the individual is provided with a list of terms which to identify gender.38 This two-step model has been extensively evaluated and found to be superior to single questions (eg, ‘What is your gender identity? Female, male, transgender, other’) because it allows researchers to classify participants as gender diverse by identifying an incongruence between sex and gender identity, even if individuals do not identify with specific terms such as ‘transgender’.38–40 It has been proposed as being useful for monitoring the health of transgender and other gender minority populations with enhanced accuracy and inclusiveness.41 42 In addition, the two-step SOGI measurement has been endorsed by the Gender Identity in US Surveillance Group by the Williams Institute at the University of California, Los Angeles, the Centre for Excellence in Transgender Health at the University of California, San Francisco and the World Professional Association of Transgender Health Electronic Medical Record Working Group.6 43–47
The EHR data (eg, administrative and billing data and electronic medical records) have demonstrated the potential to facilitate a large-scale cohort study over a longer timespan with rich healthcare information.48 With the rise of electronic medical and administrative records, clinical cohort studies exist that capitalise using these data for healthcare outcomes improvement.49 EHR remains an important source of structured clinical data to explore PLWH’s healthcare and treatment outcomes and its linkage with clients’ individual characteristics, including SOGI.6 50 Routine and standardised documentation of patients’ SOGI data in the EHR can empower healthcare organisations to deliver more patient-centred care and identify, monitor and address health disparities among patients,33 especially PLWH.
Since 1 January 2018, all EHR systems certified under the federal Meaningful Use Stage 3 Incentive Programme are required to have the capacity to capture SOGI data.11 Despite the guidelines and recommendations of measuring SOGI have been acknowledged for several years, the majority of existing EHR systems have not complied with the guidelines and not expanded data fields that involve all recommended aspects of SOGI structured data.51 Relevant research is needed to enhance the effectiveness of the two-step method of SOGI data collection using EHR data.
Accurate SOGI measurement in EHR is playing an important role in improving the understanding of HIV treatment outcomes among SGM.52 Gender is recorded in multiple places in the EHR, such as in the birth certificate, all payer claim system and electronic medical record, and integrating multiple EHR might provide a unique opportunity to define the SOGI measure via two step method. These SOGI measurements can assist in examining the potential differences in HIV treatment outcomes (eg, viral suppression) among this key group, and ultimately guide the health and treatment outcomes of PLWH. Therefore, there is a critical need to evaluate the two-step SOGI measurement of PLWH in EHR system and then investigate its impact on HIV health outcomes.
In certain EHR data sets, such as Enhanced HIV/AIDS Reporting System (e-HARS) system from the Department of Health and Environmental Control (DHEC), All-Payer Claims Databases, Health Sciences South Carolina (HSSC) and Prisma, the responses of the gender variable only include male/female/unknown. Based on ICD-10 code in EHR, we can measure gender identity disorder (GID), but this measure might pose challenges to accurately measure patients’ SOGI. These challenges suggest that a better understanding of SOGI measure using multiple EHR sources is critically needed. To test gender terminology (eg, woman, man and non-binary) for measuring current gender identity as part of the two-step method of data collection (sex assigned at birth and current gender identity), we will fill this knowledge gap by leveraging the South Carolina statewide HIV databases and birth certificate data and adding new measures of SOGI based on gender/gender disorder information in EHR. Specifically, we aim to perform data acquisition and linkage process between different data sources, leading to a three-way study to compare and describe gender and its potential change from the original birth certificate, gender in all payer claims system and gender in electronic medical records, and gender in e-HARS. In addition, we will also evaluate the gender disorder based on ICD-10 code. Differences in PLWH’s viral suppression on SOGI measurements will then be explored. With the data integration, the current exploratory study has the following specific aims:
Aim 1: Expand the current data set to involve the original birth certificate data to assess and conduct a three-way comparison analysis (ie, birth certificate gender, recorded SOGI in EHR and reported SOGI in e-HARS and gender disorder based on ICD-10 codes) on consistency of SOGI measurements.
Aim 2: Examine whether different SOGI measurement methods affect PLWH’s viral suppression via exploring the association between SOGI and viral load.
The proposed project is significant since it can provide valuable information for developing more tailored guidelines for the two-step gender data measurement in EHR and it also adds value to the parent grant via expanding the existing database, adding new measure of SOGI and investigating the impact of SOGI on viral suppression. The results will provide valuable information for tailored guidelines for measuring PLWH’s SOGI, contribute to HIV care management and ultimately contribute to the improvement of HIV treatment outcomes in SC and beyond.
Methods and analysis
Overview of the study design
Our EHR database includes several HIV data sources with HIV patients’ gender information, such as SC DHEC, HSSC and Prisma, which are detailed in figure 1. The HIV diagnosis was recorded in DHEC-CASES, which is cohort data with 12 171 people diagnosed with HIV between 2005 and 2020 with a residence of SC. Our first goal is to characterise the utility of different EHR data types to identify SOGI data and SGM population. Low cost and efficient means of identifying transgender patients in a large EHR have shown that SOGI information can be identified through keywords (demographic data), diagnosis codes or their combination.53 Keyword-based identification alone would have overestimated the number of individuals with SOGI identity; while diagnosis codes might underestimate such rate.53 For example, in Robin et al’s study, they used digitised free-text clinical notes: ‘transgender’, ‘transsexual’, ‘transsexual’, ‘gender dysphoria’, etc, to identify transgender individuals. Using the keywords only, it only identified 45% of true transgender individuals.53 We are going to use SOGI-specific diagnosis codes (eg, GID) and SOGI-demographic data in multiple EHR sources, then investigate its impact on viral suppression (figure 1). We hypothesise that cisgender sexual minority man, people with GID assigned female at birth of any sexual orientation and people with GID assigned male at birth of any sexual orientation might have worse HIV outcomes, such as suboptimal viral suppression.
Data sources
This project will integrate data from both individual-level and county-level. The individual-level data are integrated from several state agencies/systems, like:1 SC e-HARS54,55,2 SC Department of Alcohol and Other Drug Abuse Services (DAODAS);3 SC RFA integrated data warehouse;55,4 Health Sciences South Carolina (HSSC);5 SC Department of Mental Health (SCDMH) and6 Prisma Health System. The county-level data sources include American Community Survey (ACS),56 Area Health Resources File (AHRF), Behavioural Risk Factor Surveillance System (BRFSS) and county health rankings and roadmaps programme. All participating SC agencies will submit their EHR data of the PLWH cohort to SC RFA. The SC RFA will link patient records from all sources and generate a linked data set, which will include longitudinal observations of hospital visits, medication, claims data, mental health visits and other relevant data for the study cohort.57 In compliance with HIPAA regulations, SC RFA deidentified system-generated number ensures confidentiality but allows the study to conduct data mining at both the individual and aggregated data levels. There is a total of 12 171 people diagnosed with HIV during this time period. Table 1 lists the basic demographic of this population. Black, male and urban, MSM have a higher rate of HIV diagnosis.
In the proposed study, we will update the database by including additional original birth certificate data of PLWH. The SC Revenue and Fiscal Affairs Office (RFA) will serve as the honest broker for the linkage of all identifiable data and remove all the identifiable information from the linked data before releasing it to the research team. In the USA, State laws require birth certificates to be completed for all births, and Federal law mandates national collection and publication of births and other vital statistics data.58 The National Vital Statistics System, the Federal compilation of this data, is the result of the cooperation between the National Centre for Health Statistics (NCHS) and the States to provide access to statistical information from birth certificates.58 59 SC DHEC holds a vital records and statistical integrated system which includes the sex at birth. Thus, for PLWH who was born in SC, we could obtain their sex at birth through the data linkage.
Aim 1: Two-step method and ICD-10 diagnostic code based approach
Gender information: Multiple EHR data provide the longitudinal collection of the gender information at birth, at HIV diagnosis and at each hospital encounter.
Sex assigned at the birth: The sex assigned at the birth will be obtained from the patients’ birth certificate. Gender at HIV diagnosis. The gender indicated in the e-HARS denotes the gender at the HIV diagnosis. Current gender. Gender information is input into the EHR during or after a clinical encounter by physicians, nurse practitioners, physician assistants, nurses and other clinical providers. Patients who have a clinical encounter with gender-affirming surgical providers may enter their own gender as well through a survey in the patient portal (MyChart) at any time once they have patient portal access. Usually, the most recent input overwrites previous input and no record of changes is retained by the system if data are overwritten.
Two step method. The two-step method of measuring SOGI of PLWH includes but not limited to transgender, gay, bisexual and other MSM, intersex, non-binary and gender-fluid. This data will be retrieved from EHR and e-HARS data sets, and the additional original birth certificate database that will be updated based on our ongoing projects. Considering both the SOGI information, we will categorise the patients Cisgender heterosexual woman, Cisgender heterosexual man, Cisgender sexual minority woman, Cisgender sexual minority man, People with GID assigned female at birth of any sexual orientation and people with GID assigned male at birth of any sexual orientation. ICD-10 diagnostic codes. ICD-10 diagnostic codes specific to transgender patients include GID in adolescence and adulthood (F64.0), GID unspecified (F64.9), GID of childhood (F64.2), other GID (F64.8) and personal history of sex reassignment (PHSR; Z87.890). F64.0 and F64.1 are grouped as a single output in our query because following ICD-10 code implementation, a provider diagnosis of ‘gender dysphoria’ is mapped to either code in our system for billing purposes. ICD-10 codes that were active, resolved or deleted were queried. ICD-10 codes generated from admission diagnosis, billing, clinical encounters and clinician problem lists were all queried. The psychiatric diagnosis of ‘gender dysphoria’ maps to F64.0. The previously utilised psychiatric diagnosis of ‘GID’ was mapped to either F64.0, F64.2 or F64.8 before the Diagnostic and Statistical Manual of Mental Disorders, Version 5 publication and adoption.60 To improve the accuracy of identifying transgender people, we will involve (1) Endoctine Disorder Not Otherwise Specified (Endo NOS) codes and a transgender-related procedure code or (2) Receipt of sex hormones not associated with the sex recorded in the patient’s chart (‘sex-discordant hormone therapy’) and an Endo NOS code or transgender-related procedure code.61
Statistical analysis: comparison of two-step methods and gender disorder (aim 1)
We will conduct the data cleaning and management for the integrated data set and then conduct the correlation analysis. The distributions of demographic variables for the HIV cohort will be summarised (mean, SD, counts), and compared using the t-test, analysis of variance (ANOVA) test or χ2 test as appropriate. The differences of the distributions of demographic variables between samples identified by demographic data alone and sample identified by ICD diagnosis code alone will be compared using the t-test, ANOVA test or χ2 test as appropriate. If test assumptions are not satisfied, non-parametric tests (Wilcoxon rank test and Kruskal-Wallis Test) will be applied. The box plot and heat map will be used to depict the difference of continuous measures over time and a bar graph will be applied to the categorical measures. The number of individuals identified by demographic data alone or EHR diagnosis code alone will be calculated and compared. The number of individuals identified by both demographic data and EHR diagnosis code will be calculated as well. In the samples identified by ICD-10 diagnostic codes, the most common ICD-10 diagnosis code will be recorded (eg, GID in adolescence and adulthood, 996). In the two samples identified, the two-step gender identity questions will be applied separately. In the two-step identity questions, sex at birth and current gender identity will be reviewed at the same time to determine the GID status (eg, male SAB and female GI, or female SAB and male GI) if their GI (either male or female) information is discordant with their SAB. The sexual orientation information will be accounted for to stratify the SOGI status of each individual (eg, people with GID assigned male at birth of any sexual orientation). The Venn diagram will provide visualisation of demographic data alone or EHR diagnosis code alone.
Aim 2: association of SOGI measures with viral suppression
Definition of key variables
VL indicators. Following the ‘US Guidelines for the Use of Antiretroviral Agents in Adults and Adolescents with HIV’,62 the following definitions will be used to describe and examine the different levels of virologic response. (1) Viral suppression is defined as a confirmed VL level below the lower limit of the detection of available assays (eg, 200 copies/mL in our data set); (2) Sustained viral suppression is defined generally as a viral load persistently (eg, ≥40 months) below the level of detection depending on the assay used (eg, 200 copies/mL); (3) Virologic failure is defined as the inability to achieve or maintain suppression of viral replication to an HIV RNA level<200 copies/mL; (4) Viral rebound is defined as confirmed HIV RNA level≥200 copies/mL after initial viral suppression; (5) Viral blip is defined as after viral suppression, an isolated detectable HIV RNA level that is followed by a return to viral suppression. Isolated blips are defined as viral load transiently detected at low levels, typically HIV RNA<400 copies/mL and (6) Low-level viremia (LLV) is defined as confirmed detectable HIV RNA level<1000 copies/mL (eg, from 201 to 999 copies/mL). The primary variable will be measured in continuous variable, binary or categorial variables.
Covariates. Data suggested that sexual minority populations or transgender individuals are at increased risk for poor health outcomes due to syndemic factors such as discrimination, substance use, limited access to care and poor mental health.63 64 Since some of the information cannot be retrieved at the individual level, the area-level indicators will be employed as a proxy of these measures. County-level data include demographic characteristics such as age, race, income, education, obesity, smoking, unemployment, access to healthy foods and other socioeconomic/demographic data. Substance use, such as tobacco use, alcohol use and illicit drug use (eg, cannabis, hallucinogen, opioid, cocaine, amphetamine and antidepressant) will also be considered as covariates. To construct the longitudinal measure, use/abuse of any substance is defined as at least one diagnosis of a specific substance use during the follow-up period.
Statistical analysis (Aim 2)
We will employ generalised linear mixed regression (ie, logistic and multinomial) with different prespecified correlation matrix as appropriate such as autoregression covariance matrix and choose the best model based on quasi-likelihood under the independence model information criterion (QIC) to evaluate the differences of the probability of viral suppression on different SOGI measurements adjusting key demographic characteristics and other variables listed above. We will adopt the stepwise selection of all variables. The lasso regression will be used if the standard stepwise selection does not work due to the high dimension of risk factors.
Individual level model
We will focus on all individual level variables and the model will be built sequentially by (1) including SOGI only for the crude model; (2) add other demographics of PLWH individuals; (3) add the HIV risk factors other than the outcomes, such as CD4 counts, duration of HIV diagnosis, and HIV engagement factors (eg, retention in care) and (4) the interaction between the PLWH population and SOGI measurements.
Multilevel modelling
The integrated data set includes individual-level and county-level data. Multilevel modelling will be used to account for county-level, time and individual longitudinal patterns. Appropriate logistic/multinomial regression will be developed. Random effects will be considered for each model to consider the repeated measure at the individual level and the cluster effects from the same county. Multivariate normal distribution will be used to model the repeated measure and the multilevel models will be fitted through the maximum likelihood method.
Model selection. The best model will be selected based on Akaike information criterion or Bayesian information criterion criteria. Potential interactions with SOGI will be examined, and the log-likelihood ratio test will be used to determine statistical significance. If significant, stratified modelling will be used. Depending on the sample size of subset of interest in the integrated data, we could (1) conduct the stratify analysis for each PLWH population with different SOGI using similar generalised linear mixed regression models, and (2) add the interaction term between sexual orientation and gender minority disorder variables. For the continuous variables, if the linear relationship is not satisfied, other relationships will be considered such as quadratic association or P-spline. We will report OR and 95% CI for logistic regressions and the forest plots will be used to visualise the OR.
Potential challenges and considerations
Incompleteness of birth certificate. First, we may miss certain populations with PLWH who were not born at SC since their birth records will not be in the vital records. We will record this proportion. Although the proportion is uncertain, the risk will be minimum since we will focus on two approaches of SOGI measure where the gender disorder based on ICD 10 will not rely on the completeness of the birth certificate. We also expect this proportion to be small.
Missing data. With the linkage of longitudinal real-world data from different sources, the issue of missing data will be inevitable. If data are missing for other individual characteristics, depending on the missing rate, we will include individuals when few variables are missing and impute the missing variable with mean or median, assuming the missing value at random. If the main variable of interest is missing, multiple imputation procedures will be implemented. Based on consultation with the research team, the regression model will be built to impute the missing variables. Additionally, we will conduct sensitivity analyses comparing results with and without imputation.
Dissemination. To materialise the anticipated methodological and clinical benefits of the proposed research and to maximise their impact on HIV clinical care, we will disseminate our study through presentations at academic conferences and publish peer-reviewed articles with titles such as ‘Using data for linkage: an application of data analytics for risk prediction of viral load suppression among SOGI’. We will capitalise on professional networks that can increase the reach and accessibility of findings such as open access publication, webinars, files and videos available on websites and publicly available channels to increase visibility and impact of the scientific publications and presentations.
Alternative strategies. In the event that we cannot retrieve enough sample size or some additional information (eg, birth certificate) are not available in the statewide merged data sources, we will use the nationwide data from the All of Us research programme, which is a NIH sponsored programme aiming to recruit underrepresented populations (eg, sexual minorities) in biomedical research. Since 2022, we have already initiated several HIV related projects and accumulated extensive research experience in this platform.
Patient and public involvement
None.
Ethics and dissemination
The study team applied for data from different sources and submitted individual Institutional Review Board (IRB) application to the University of South Carolina (USC) IRB. The study was approved by the IRB at the University of South Carolina (Pro00129906) as a Non-Human Subject study. The SC Office of Revenue and Fiscal Affairs (SC RFA) is authorised by SC state law to collect and merge data from these different sources and to ensure the privacy of all people living with HIV. Researchers will get only the encrypted deidentified data set to prevent any breach of privacy in the data transfer, management and analysis processes. Neither Principal Investigators nor other investigators will have access to identifiable data from any state agencies. Due to the use of secondary data, the USC IRB has granted the planned study a ‘non-human subject research designation’. These SOGI measurements can assist in examining the potential differences in HIV treatment outcomes (eg, viral suppression) among this key group, and ultimately guide the health and treatment outcomes of PLWH. The proposed supplement will add value to the parent grant via expanding the database, adding the new measure of SOGI and allowing an examination of its impact on viral suppression. The results will inform tailored guidelines for measuring PLWH’s SOGI, and ultimately contribute to better HIV treatment outcomes. To protect the identity of PLWH, researchers will only receive linked deidentified data from the RFA. The SC RFA will coordinate the efforts among relevant state agencies (eg, SC DHEC, Health Sciences South Carolina (HSSC)) to link the data and to provide us with only the deidentifiable data for analysis. Extensive data agreements ensuring data security and patient confidentiality for the deidentified linked data have been established and are stringently adhered to. No investigators will have access to identifiable data from any state agencies.
We will publish the findings in peer-reviewed scientific journals and present the study findings at national and international professional conferences and through appropriate social media outlets. We will capitalise on social media and professional networks that can increase the reach and accessibility of findings, such as open access publication, webinars, files and videos available on websites and publicly available channels (eg, YouTube), to increase visibility and impact of the scientific publications and presentations. The dissemination efforts of this project will extend beyond the scientific arena and also target our stakeholders in healthcare system and policy makers in the USA at local (SC DHEC, Prisma Health) and national levels (CDC) through various policy forums, policy papers and special presentations.
Ethics statements
Patient consent for publication
References
Footnotes
Contributors XY, BO and JZ contributed to the conception and design of the study. XY led the writing of this protocol manuscript. SW and XL contributed significantly to the editing of this manuscript. All authors reviewed and provided comments to improve the manuscript. All authors contributed to the editing and final approval of the protocol.
Funding This work was supported by the US Department of Health and Human Services, National Institutes of Health, National Institute of Allergy And Infectious Diseases (grant number R01AI164947-03S1). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.