This Therapeutic Area Data Standards User Guide for HIV (TAUG-HIV) was developed under the Coalition for Accelerating Standards and Therapies (CFAST) initiative.

The purpose of the TAUG-HIV is to describe how to use Clinical Data Interchange Standards Consortium (CDISC) standards to represent data pertaining to HIV or data commonly collected in HIV trials. These include studies of both HIV treatment and HIV prevention, as well as studies that may include HIV-infected individuals.

1. First, read the Study Data Tabulation Model (SDTM) Version 1.7, SDTM Implementation Guide (SDTMIG) Version 3.3, Clinical Data Acquisition Standards Harmonization (CDASH) Model Version 1.0, CDASH Implementation Guide (CDASHIG) Version 2.0, Analysis Data Model (ADaM) Version 2.1, and ADaM Implementation Guide (ADaMIG) Version 1.1. These standards are available at http://www.cdisc.org/.
2. As needed, visit Section 1.3, CDASH Metadata and Annotated CRFs, of this guide for an explanation of how to read the annotated CRF.
3. Read the SDTMIG for Associated Persons (SDTMIG-AP), SDTMIG for Pharmacogenomics/Genetics (SDTMIG-PGx), and SDTMIG for Medical Devices (SDTMIG-MD) to gain some familiarity with representing non-study subjects, genetics, and medical devices in the SDTM. These standards are available from: http://www.cdisc.org/.
4. Read the Introduction to Therapeutic Area Standards (http://wiki.cdisc.org/x/SSy8AQ) and/or take CDISC's free training module, TA001 - Overview of Therapeutic Area User Guides (https://www.cdisc.org/education), to understand what to expect from a therapeutic area user guide.
5. Read this guide all the way through (without skipping any sections) at least once.
6. Finally, revisit any sections of particular interest.

Draft standards of interest to this document are listed at: Draft Standards of Interest to TAUG-HIV ( https://wiki.cdisc.org/x/OJ2FAg ).

For information and specifications for the content of data sets that should be submitted as part of the sponsor or applicant's application for drugs intended to treat HIV, refer to the FDA's HIV Technical Specifications Guidance: https://www.fda.gov/downloads/ForIndustry/DataStandards/StudyDataStandards/UCM603323.pdf.

All general caveats for the standards given in the Introduction to Therapeutic Area Standards (http://wiki.cdisc.org/x/SSy8AQ) apply to this document.

This document is divided into the following sections:

• Section 1, Introduction, provides an overall introduction to the purpose and goals of the HIV project.
• Section 2, Overview of HIV, provides an overview of what HIV is and whom it affects.
• Section 3, Subject and Disease Characteristics, covers data that are usually collected once at the beginning of a study.
• Section 4, Baseline Medical History and Physical Exams of Special Interest, contains examples and discussion on baseline menstrual history and pelvic exam.
• Section 5, Mother-Infant Pairs in HIV Studies, covers the management of mother-infant relationships, with respect to the representation of data collected on either the mother or infant or both.
• Section 6, Interventions for the Prevention and Treatment of HIV, contains information about and examples of HIV interventions.
• Section 7, Disease Assessments, provides information on data that are typically collected multiple times during a study.
• Section 8, Analysis Data, includes key data analysis concepts for an HIV study.
• Appendices provide additional background material and describe other supplemental material relevant to HIV.

CDASH examples include both metadata tables and sample case report forms (CRFs). Each table of CDASH metadata corresponds to an example annotated CRF (aCRF), built directly from the metadata. The annotations show the variables associated with each field in the context of data collection (CDASH) and submission (SDTM). The color of the variable name denotes the applicable context. SDTMIG variable names are shown in RED. If the same variable name is used for both the SDTMIG variable and the CDASHIG variable, only the SDTMIG variable is shown. If the CDASHIG variable differs from the one defined in the SDTMIG, the CDASHIG variable is also shown in GRAY. Data collected but not submitted in SDTM-based datasets are denoted as. Not submitted

The following diagram illustrates how to interpret the annotations.

When viewing sample aCRFs, bear in mind that:

• Sample CRFs are provided to illustrate data collection instruments. These examples are not meant to imply that any particular layout is preferable over another.
• Example CRFs include annotations that refer to SDTM and CDASH data elements. Refer to these standards for more information about the use of these elements (https://www.cdisc.org/standards/foundational).
• Most example CRFs do not include header information such as subject identifier and visit, where applicable. Information in these headers may be populated in a variety of ways depending on the data management system.
• Sponsors are responsible for understanding and implementing CDISC Controlled Terminology (https://www.cancer.gov/research/resources/terminology/cdisc) where applicable.
• CDASH variable names for denormalized variables are examples. Sponsors may use other conventions for creating denormalized CDASH names.
• SDTM variable names separated by "/" indicate that either SDTM variable may be used when creating the SDTM datasets.

When viewing the CRF metadata, bear in mind that:

• Some CDASH metadata attributes are not included.
• Some non-CDASH metadata attributes necessary for rendering the CRF are included (eg, Hidden, Pre-specified Values, etc).

• Non-standard variables: This document has adopted the practices outlined in the proposed SDTMIG Section 8.4.4, Alternative Representation of Non-Standard Variables (available in draft form at: http://wiki.cdisc.org/x/Ui68AQ). Accordingly, SDTM-based examples containing sample data requiring the use of a variable outside the standard set of variables included in the SDTM v1.7 are represented not with Supplemental Qualifier records, but with non-standard variables (NSVs) appended to the parent domain, followed by sample variable-level metadata for the NSVs.

  In order to avoid confusion between standard variables and NSVs, NSVs have been rendered visually distinct, as shown below, with white text on black in the header row, and separated from the standard variables by a small space. Metadata for the NSVs, from the define.xml file that would accompany the submission, are tabulated below the example. Only those attributes or elements that assist the example are included. (For more information on variable-level metadata in general, see the Define-XML Specifications found at: https://www.cdisc.org/standards/data-exchange/define-xml ). A list of all NSVs used in this document, and the variable-level metadata that might become normative for the NSVs should they be promoted to standard variables, is given in Appendix C, Non-Standard Variables.

  

• Representation of prespecified findings: This guide shows an example in which prespecified findings from a pelvic exam are represented (see Section 4.2, Pelvic Examination). Within the CDISC community there is an ongoing discussion regarding the best way to represent prespecified findings. Historically, prespecified findings have been coordinated in the --TESTCD/--TEST variables (i.e., the prespecified target of interest is named in an indicator TEST, and the result is simply Y/N to indicate whether or not it was found). However, this can lead to a proliferation of controlled terms for these variables. Other options, including a new variable to represent the prespecified finding of interest (--RESTRG), are currently being explored. Previous therapeutic area user guides (TAUGs) have represented prespecified findings in the NSV --RESTRG. However, there has been a lack of consensus regarding this approach. Due to the lack of consensus, TAUGs will continue to coordinate prespecified findings in the --TESTCD/--TEST variables. This decision may change in the future.

• Representation of subject characteristics collected more than once during a study: The concepts of gender identity and gestational age may be collected more than once during a study. Although these concepts are generally thought to belong in the Subject Characteristics (SC) domain, the SDTMIG v3.3 indicates that subject characteristics collected more than once should be represented in the Subject Status (SS) domain. However, representing gender identity in both SC and SS, depending on how often it is collected, conflicts with the principle that data should be consistently represented in only a single place. This issue was evaluated by the CDISC Global Governance Group, which decided that the assumptions of the SC domain should be modified to allow the representation of subject characteristics regardless of the number of times the characteristic is collected. This decision will be published in a future version of the SDTMIG. In this TAUG-HIV, Sections 3.2 (Representation of Gender Identity over Time) and 5.4 (Gestational Age) reflect this modeling decision.

• Published gestational age controlled terminology: CDISC Controlled Terminology includes both RPTESTCD/RPTEST for EGESTAGE/Estimated Gestational Age and SCTESTCD/SCTEST for GSTABRTH/Gestational Age at Birth. Because gestational age is considered to be a subject characteristic of the infant, rather than a reproductive system finding, the published controlled terminology will be reviewed and updated. Section 5.4, Gestational Age, provides additional details.

• Use of Device In Use (DU) to represent vaginal ring drug measurements:According to published controlled terminology the DU domain is "a domain for the findings for the values of measurements and settings that are intentionally set on a device when it is in use. These are characteristics that exist for the device, and have a specific setting for a use instance." However, there is ambiguity regarding whether or not the domain is intended to represent:

  • a domain for the findings for the values of (measurements and settings) that are intentionally set on a device when it is in use, or
  • a domain for the findings for the (values of measurements) and (settings that are intentionally set) on a device when it is in use.

  The Device Team has agreed that the second interpretation is what was originally intended when creating the domain. The definition and intended use will be clarified in the next version of the SDMTIG-MD. Because of this, the HIV Team felt it was appropriate to represent the vaginal ring drug measurements within DU (Section 6.2.1, Vaginal Ring).

• Use of --RSDISC (Reason for Treatment Discontinuation) in the Device Exposure (DX) domain:In the vaginal ring example (Section 6.2.1, Vaginal Ring), the DXRSDISC variable is used to represent both the reason why each individual ring application ended and the overall reason for discontinuation of treatment with a vaginal ring.
• Representation of the date of a particular medical history episode: The variable MHEVDTYP specifies the aspect of the medical condition or event by which MHSTDTC and/or the MHENDTC is defined. Current controlled terminology values are "DIAGNOSIS", "EPISODE", "EXACERBATION", and "SYMPTOM ONSET". When the date is for an episode or exacerbation, which episode or exacerbation (e.g., first, most recent, worst) must be specified. It has not yet been decided whether this information should be included in the value of MHEVDTYP (e.g., giving a value such as "MOST RECENT EPISODE"), or in a separate variable. In Section 4.1, Baseline Menstrual History, Example 1, a separate NSV (MHCRNORD) has been used to represent the fact that the date collected was for the most recent menstrual period.
• Calculation of a single analysis value that is based on the data from more than 1 subject: The ADaM Basic Data Structure (BDS) class currently is designed to support subject-level analysis (as well as site-level analysis). However, grouping a number of subjects as an analysis unit has not been addressed. Example 1 in Section 8.3, Pregnancy Outcomes Analysis, proposes the variable USUBJGR1 (USUBJGRy as a generic variable) to allow for the grouping of subjects for analysis purposes. This allows for the calculation of a single analysis value that is based on the data from more than 1 subject. Because the ADOUTCOM dataset is summarized by family, based on the values across multiple subjects, a structure based on family and parameter is needed—which does not fit within the BDS definition. Therefore, it is proposed as an ADAM OTHER class of dataset.

Human immunodeficiency virus (HIV) causes a viral infection which, if not adequately controlled by treatment, may lead to acquired immunodeficiency syndrome (AIDS). HIV can weaken a person's immune system by reducing CD4+ T cells. AIDS is defined by the World Health Organization (WHO) as the occurrence of any 1 of more than 20 AIDS-defining clinical conditions, including opportunistic infections and HIV-related cancers.^[1]The U.S. Centers for Disease Control and Prevention (CDC) definition is similar, but includes alternative criteria of a CD4+ cell count of fewer than 200 10^6/L or a CD4+ T cell percentage of total lymphocytes of less than 14%.^[2] The 2 major types of HIV are HIV-1 and HIV-2. HIV-1 is the most common and most pathogenic. HIV-2 is less pathogenic and has been largely confined to Africa,although some cases have occurred in other parts of the world.

There are several populations that have a disproportionate burden of HIV. These groups include "men who have sex with men, transgender women in many settings and heterosexual men and women who have sexual partners with undiagnosed or untreated HIV infection".^[3] Another population of interest is pregnant women with HIV. There are various approaches to preventing the mother-to-infant transmission of HIV. In HIV prevention studies involving pregnant women, infants will receive HIV testing and may receive antiretroviral therapy (ART) based on the mother's HIV status.

HIV treatment guidelines suggest a combination of antiretroviral (ARV) drugs to decrease the viral load below the levels of detection of sensitive HIV-RNA assays.^[3,4] Antibody therapy may also be used to achieve this reduction in viral load. Current guidelines suggest lifelong treatment. Prevention of HIV in vulnerable populations includes the use of ARVs or treatment of potential sources (e.g., infected sexual partners, infected pregnant mothers to prevent perinatal transmission) prior to potential exposure to prevent transmission of HIV. Studies to develop vaccines for HIV have so far been unsuccessful. Vaccine trials are ongoing and may one day result in this being a viable option for prevention of HIV infection.^[5,6]

Co-infections and complications of HIV-infected populations such as tuberculosis, hepatitis, and cardiovascular disease are also of interest. CDISC has published several therapeutic area user guides specifically for these indications. These guides can be found on the CDISC website (https://www.cdisc.org/standards/therapeutic-areas ).

Information about subject and disease characteristics generally includes events and activities that have affected the subject prior to the study. In the TAUG-HIV, this includes:

• Diagnosis of HIV
• Representation of Gender Identity over Time
• Risk Factors
• Mode of Transmission

Diagnosing a subject with HIV requires an assay with a high degree of sensitivity and specificity. This is most easily achieved via a diagnostic algorithm that includes multiple redundant or complementary assays, as opposed to a single test.

The following concept map depicts a widely used algorithm for determining HIV status. Note that this algorithm does not apply to newborn infants of HIV-infected mothers.

The following example illustrates a subject's progression through the diagnostic algorithm depicted in the concept map (second-stage confirmatory test results not shown).

Transgender people are at high risk for HIV. Accurate capture and representation of gender identity in HIV prevention and treatment studies is important. Gender identity is the innermost concept of self as male, female, a blend of both, or neither. Gender is a spectrum, and not limited to just 2 possibilities. One's gender identity can be the same as the sex assigned at birth (cisgender) or different (transgender). A person may have a non-binary gender identity, meaning they do not identify strictly as male or female: They could identify as both, as neither, or as another gender entirely.^[7,8,9]

Gender identity has traditionally been included in the Subject Characteristics (SC) domain. However, gender identity can be a fluid characteristic, especially in populations studied for HIV; thus it may be collected more than once during a study. A recent decision at CDISC allows subject characteristics collected as repeated measure to be included in SC. Further detail on this decision can be found in Section 1.4, Known Issues.

Because the terminology used to describe gender identity is evolving, CDISC has decided not to provide controlled terminology for potential values at this time. However, in order to populate the topic variable for SC consistently, a new SCTESTCD/SCTEST of "GENIDENT"/"Gender Identity" has been submitted to the CDISC Controlled Terminology Team.

In addition, the Tuberculosis Therapeutic Area User Guide v2.0 introduced the concept of "Sex Reported at Birth," which is also represented in the SC domain. The variable SEX in the Demographics (DM) domain is not well defined (sex of the subject) but is generally accepted to mean the self-reported sex of the subject. In support of this assumption, the CDASH Standard states that SEX is "the self-reported sex of the individual and/or is the clinician's assignment based on a physical examination. This is not to be confused with a genotypic determination of a subjects' chromosomally determined gender, but a less scientifically controlled method of visual determination that HL7 has defined as 'administrative sex.'" The concept of "Sex Reported at Birth" is intended to be used in cases where sex at birth differs from the subject's self-reported sex. However, sex and gender identity are complex issues that are only partially covered by the above concepts; further standards development work is required to adequately represent these concepts.

Risk factors for contracting HIV include high-risk sexual behavior, intravenous drug use, prior sexually transmitted infections (STIs), and, for males, being uncircumcised.

Example 1 demonstrates a hypothetical CRF designed to capture the subject's history of HIV risk factors. The sponsor of this study chose to collect sexual behavior, sexually transmitted infections (STIs), intravenous drug use, and circumcision to assess the risk of HIV infection.

SDTM examples follow the aCRF.

In HIV studies, it may be important to collect information about how the subject contracted the HIV virus (i.e., mode of disease transmission). At the broadest level, the mode of disease transmission can be categorized as either vertical transmission or horizontal transmission. Vertical transmission means that the virus is passed from mother to infant during pregnancy, birth, or breastfeeding. Horizontal transmission means that the virus is passed between individuals through activities such as sexual encounters and the sharing of IV needles.

As described below, HIV infections may be represented in either the Clinical Events (CE) domain or the Medical History (MH) domain, depending on the focus of the study. In either case, the mode of HIV transmission can be represented in the non-standard variable (NSV) --MODTR (Mode of Disease Transmission). If a study allows a subject to report more than 1 mode of transmission, --MODTR can be populated with "MULTIPLE". Additional NSVs can be added using the naming convention of --MODTR1 to --MODTRnto represent each mode of transmission reported. The level of granularity collected around the mode of transmission may vary by study. For example, a study may collect a category of mode of transmission such as "HORIZONTAL TRANSMISSION" or "VERTICAL TRANSMISSION". However, another study may collect a more granular mode of transmission that falls into 1 of these categories (e.g., "BLOOD TRANSFUSION", a type of horizontal transmission). Submission values can be found in the SDTM codelist, Mode of Disease Transmission (--MODTRN), and include:

This section contains examples and discussion on baseline menstrual history and pelvic examinations.

HIV may cause more severe premenstrual symptoms for some women.^[10] It is also possible that some women and girls living with or at risk for HIV, including those participating in clinical research for HIV/AIDS, do not menstruate. Reasons may include premenarche, menopause (natural or induced), long-acting contraceptives, and conditions that cause amenorrhea. Therefore, HIV/AIDS studies may need to collect more comprehensive data about the subject's gynecological assessments in order to place menstrual information within appropriate context.

Example 1 illustrates collecting information on premenstrual symptoms such as fatigue, cramps, headache, dysphoric disorder, pain, menstrual spotting, and diarrhea. Example 2 includes the number of days between menses and the range of usual bleeding days for the subject.

During a pelvic examination, an investigator may check for prespecified abnormalities. When this occurs, the prespecified finding of interest is represented in RPTESTCD/RPTEST. The value in the variable RPORRES indicates whether the prespecified finding of interest was observed. Example 1 shows the representation of data from a prespecified pelvic exam for a single subject.

Given the high burden of HIV infection among women of childbearing potential, particularly in low- and middle-income settings, the study of HIV treatment and prevention in HIV-infected pregnant women and their infants is of high priority. When a subject is pregnant or becomes pregnant, it is important to track pregnancy outcomes and it may be necessary to collect both prenatal and postnatal data on the infant pertaining to overall health and HIV status.

In studies that collect perinatal data, the term used to describe the child changes over time, from fetus to newborn (or neonate) to infant. Within particular examples where the child is at a particular stage, the appropriate term will be used, according to definitions in Appendix B, Glossary and Abbreviations. Statements in this document that apply to multiple stages use the term infant.

In SDTM-based datasets, data about persons who are not study subjects (associated persons) are represented differently from data about study subjects. This means that representation of data about mothers and infants depends on whether they are study subjects or associated persons—which, in turn, depends on the study design and protocol. This section covers the representation of data collected about mothers and infants and the representation of the relationship between mother and infant in studies where one or the other or both are study subjects.

The roles that mothers and infants play in a study determine how pregnancy and perinatal data are represented in the SDTM. This TAUG-HIV includes examples that illustrate data in one of 3 categories of study: those that enroll women of childbearing potential, studies of mothers and infants, and studies of infants.

Studies that Enroll Women of Childbearing Potential

If a woman of childbearing potential becomes pregnant during a study, the protocol may call for collection of data about the pregnancy and minimal data about the infant(s). In such a study, the infant is unlikely to be treated as a study subject, but rather as an associated person or a non-study subject. Any data about the infant would be represented in an associated persons (AP) dataset, as described in the SDTMIG-AP.

An associated person is given an associated person identifier (APID), rather than a unique subject identifier (USUBJID). AP datasets include the APID, the USUBJID of the study subject with whom they are associated (RSUBJID), and a variable (SREL) that describes the relationship of the associated person to the study subject. SREL would be a value such as "CHILD, BIOLOGICAL" for infants of mothers in a study in this category.

Studies of Mothers and Infants

Some studies enroll pregnant women and treat a mother and her infant(s) as study subjects. In this case, each person has a unique subject identifier (USUBJID). Relationships between study subjects can be represented in the Related Subjects dataset (RELSUB).

Studies of Infants

There are occasional studies of neonates which collect data about mothers, but in which the mothers are associated persons, rather than study subjects. Maternal data would be represented in AP domains.

The mother-infant relationship is represented in the SDTM in a RELSUB dataset if both mother and the infant are study subjects, and in an Associated Persons Related Subjects (APRELSUB) dataset if only 1 is a study subject. The APRELSUB dataset is not necessary unless there are associated persons who have either relationships with multiple study subjects or multiple relationships with the same study subject. However, the APRELSUB dataset can be helpful even when not required, as a single dataset that shows all relationships of associated persons to study subjects.

Because multiple gestations during a single pregnancy increase the complexity of data modeling, many of the examples shown in this section are based on a mother who is pregnant with or has given birth to twins. In the examples for scenarios in which the mother is a study subject, the infant identifiers are shown as the mother's study identifier with the letter "A" or "B" appended. In the scenario where the mother is an associated person, the mother's identifier is based on the identifier of an infant.

When the mother is the study subject and the infant is an associated person, the relationship of the infant to the mother is represented in each record in an AP dataset, since AP datasets include the variables RSUBJID (Related Subject Identifier) and SREL (Subject Relationship). The relationship can also be represented in the APRELSUB dataset. Relationships are described as the relationship of the associated person to the study subject, so only 1 record is required for each relationship.

When the infant and the mother are both study subjects, they are each assigned a unique subject identifier (USUBJID) and the relationship is defined in the RELSUB dataset. The RELSUB dataset always shows each relationship between 2 study subjects using 2 records, since relationships (represented in SREL) are generally not symmetrical.

When the infant is the study subject and the mother is an associated person, the relationship can be represented in the APRELSUB dataset. Relationships are described as the relationship of the associated person to the study subject, so only 1 record is required for each relationship.

Pregnancies are represented as events. A woman could become pregnant more than once during a long study. In this TAUG-HIV, --LNKID is used as a pregnancy identifier, to link data within and across domains for the same pregnancy.

The amount of data collected about pregnancy outcome is likely to depend on the time at which the pregnancy ends. In this guide, 20 weeks has been used as the dividing line between early and late pregnancy outcomes. The 20-week division is derived from the definitions of spontaneous abortion (also known as miscarriage) as something that happens before 20 weeks, and late fetal death as something that happens after 20 weeks. ( Complications of pregnancy are covered in Section 5.8, Routine Data for Mother-Infant Pairs.)

In this TAUG, pregnancy-related data are modeled as data about the mother, which means that pregnancies with multiple infants complicate the modeling when the outcome and the way in which the fetuses separate from the mother can differ. It is assumed that in an early pregnancy termination (before 20 weeks), the pregnancy end is the same for all fetuses, so separate records for multiple fetuses are not needed. A non-standard qualifier, Fetus/Infant Identifier (–FTINID), has been used to distinguish between multiple maternal records for pregnancies that end after 20 weeks. Although the SDTM includes the variable FETUSID, the variable is not intended for use in human clinical trials, and was intended only to distinguish between records for a particular test, not to be consistent across tests. The --FTINID in this guide is intended to identify fetuses consistently across records at a particular time (e.g., at birth), but not necessarily across times (e.g., across ultrasounds at various times during pregnancy).

The outcome of a pregnancy is modeled as an attribute of the mother's pregnancy as "Early pregnancy termination", "Late fetal death", "Stillbirth", or "Live birth", although there is an outcome for each fetus. This combines survival of the fetus with the stage at which a non-surviving fetus died.

The way in which a fetus is separated from the mother may be natural, as in a miscarriage or an unassisted vaginal delivery, or by means of an intervention such as an induced abortion or a Cesarean delivery. In the examples below, all ways in which a pregnancy ends are modeled as events, with the rationale that all (with the possible exception of live birth) are things which might be reported as adverse events. Events which can also be considered interventions present a choice for the sponsor, similar to the choice of whether to create a Procedures (PR) domain record for procedures whose main purpose is to gather data reported in findings domains. In the examples below, surgical abortion, surgical termination of an ectopic pregnancy, Cesarean delivery, forceps-assisted vaginal delivery, and vacuum-assisted vaginal delivery are modeled as procedures; medication-induced abortion and termination of an ectopic pregnancy are modeled as a concomitant medication administrations. Some of the data collected in the examples below, such as the urgency of a cesarean delivery or the medication used to induce pregnancy termination, are interventions data, and are best represented in an interventions record or a supplemental qualifier to an interventions record. If such data are not collected, the sponsor can decide whether to create interventions records for interventions that separate a fetus from its mother.

This modeling is illustrated in the following concept map, which includes SDTM domains and selected variables.

The following examples show pregnancy-related data as it would appear in the categories of studies described in Section 5.2, Identifiers and Relationships for Mother-Infant Pairs.

In a study in which mothers and infants were study subjects, pregnancy-related data would be represented as in the case when only the mother is a study subject. Note that since the information is represented in records for the mother, this information is associated with infants only indirectly. When a pregnancy includes only 1 fetus, the association between maternal and infant data would be adequately represented by the mother-infant relationships in RELSUB. For multiple pregnancies, the relationship between a fetal identifier in a maternal record and a USUBJID in an infant record (CEFTINID in the example above) cannot currently be represented in the SDTM. This relationship would be explained in a comment to USUBJID and/or SUBJID in the Define-XML document and/or in the Clinical Study Data Reviewers Guide (cSDRG).

In the studies on which the TAUG-HIV was based, only live-born infants were considered study subjects and prenatal data were treated as data about the mother. Other studies might take a different approach and treat fetuses as study subjects.

Estimations of gestational age may be based on examinations performed on a pregnant woman or on a newborn infant. Note that gestational age can be estimated at any time during pregnancy, although after birth "gestational age" is commonly used to refer to the gestational age at birth. Example 1 shows the representation of estimated gestational age, and Example 2 the representation of various kinds of data from which gestational age may be estimated.

If data to support an estimate of gestational age are collected, the domain in which those data are represented depends on the nature of the supporting data. If study data for estimated gestational age did have supporting data, the relationship between the estimated gestational age record in SC and the supporting data in another domain would be represented in RELREC.

The Ballard score is an assessment of physical characteristics of a newborn from which gestational age may be estimated. This is a clinical classification that would be represented in the Disease Response and Clinical Classification (RS) domain. Development of a supplement for this scale has been requested. See Section 7.6, Questionnaires, Ratings, and Scales, for further information about QRS supplements.

Gestational age may be estimated from the date of the last menstrual period, which is the start date of the event "Menstrual period". In a study that enrolled pregnant women, the most recent menstrual period would be represented in MH, using a form similar to that shown in Section 4.1, Baseline Menstrual History. In a study where the last menstrual period occurred during the study, it would be represented in the Clinical Events (CE) domain.

Fundal height, which is the distance from the top of the pubic bone to the top of the uterus assessed by palpation, can be used to estimate gestational age. Because this test is a measurement of the uterus, it is represented in the Reproductive System Findings (RP) domain.

Gestational age can also be estimated from ultrasound measurements of a fetus in utero. Body measurements such as head circumference and bone lengths are represented in the Vital Signs (VS) domain.

In studies in which the infant is considered an associated person, demographics data will be limited. Many of the date variables in the Demographics (DM) domain are dates that refer to study participation, which do not apply to an associated person and would not be included in an APDM dataset.

Demographics data include multiple dates related to study participation. In studies that recruit pregnant women and in which mothers and infants are study subjects, the meaning of these dates for an infant subject should be defined in the protocol:

• The date of informed consent (i.e., informed assent for the infant), RFICDTC, will be before birth.
• The dates of start and end of study treatment, RFXSTDTC and RFXENDTC, would depend on whether infants receive study treatment directly and whether indirect exposure in utero or via breast milk is included.
• The study reference start date, RFSTDTC, should have the same definition for all study subjects, and have a meaning that will make --DY variables that reference this date useful. This choice may be difficult in studies that include both mothers and infants as study subjects.

Study participation dates in infant demographic data for studies in which only the infant is a study subject may be somewhat easier to define.

The examples in this section assume that the infant is a study subject. Data collected for an infant who was an associated person would be represented in the corresponding AP domain and the dataset would include APID, RSUBJID, and SREL, rather than USUBID.

Apgar scores and Ballard Maturational Assessments may be performed shortly after birth. Examples of data from these scales are not included here, as examples would be developed by the Questionnaires, Ratings, and Scales (QRS) Team. See Section 7.6, Questionnaires, Ratings, and Scales for additional information about these and other scales relevant to HIV studies.

Congenital anomalies would be recorded as adverse events for the infant.

Infant weight may be compared to a growth chart to determine whether an infant is small for gestational age (generally defined as having a weight less than the 10th percentile for gestational age).

Fontanelle closure may also be assessed at birth.

In studies where the infant is a study subject, data may be collected about infant feeding. Data about breastfeeding may be used in combination with maternal medication data to assess infant exposure to medications in breast milk.

The SDTMIG provides advice on the collection of data such as vital signs, adverse events, and medications, and the SDTMIG-AP provides information on how to represent these kinds of data for associated persons. The following are a few points to consider in a study that involves mother-infant pairs.

Vital Signs

There is no specific test for "birth weight". Birth weight would be represented in a Vital Signs (VS) or APVS domain in a record with VSDTC equal to the birth date.

Medical Conditions

For study subjects, medical conditions are represented in Medical History (MH), Adverse Events (AE), or Clinical Events (CE), depending on when they occurred (MH for events that started before the study) and on whether the event is reportable as an adverse event in the particular study (the distinction between AE and CE). For associated persons, the distinctions are less clear. APMH tends to be used for medical conditions before the study subject started the study. For mothers who are associated persons for an infant study subject, however, the domain in which to represent events that occurred after the pregnant woman was recruited might be represented in CE or AE, even though the infant may not be considered to start the study until after birth.

Medications

Maternal medications during pregnancy and lactation are likely to be of interest when the infant is a study subject. Which medications were taken during pregnancy and lactation may be ascertained by comparing intervention record dates with the date of birth and with dates of breastfeeding. When the mother is an associated person rather than a study subject, limited concomitant medication data might be collected, and the data might be collected in separate modules for medications taken during pregnancy and those taken during lactation. If so, then the collection module of origin could be represented in a value of the CMCAT or CMSCAT variable.