https://orcid.org/0000-0002-3226-4516
Abstract
The lack of sex-specific variables, such as reproductive system history (RSH), in cardiovascular research studies is a missed opportunity to address the cardiovascular disease (CVD) burden, especially among women who face sex-specific risks of developing CVD. Collecting RSH data from women enrolled in research studies is an important step towards improving women’s cardiovascular health. In this paper, we describe two approaches to collecting RSH in CVD research: extracting RSH from the medical record and participant self-report of RSH. We provide specific examples from our own research and address common data management and statistical analysis problems when dealing with RSH data in research.
Become familiar with the breadth of reproductive system history (RSH) factors that should be considered in cardiovascular research.
Learn methods to collect RSH from the electronic medical record.
Apply best practices in creating an RSH participant self-report questionnaire.
Understand ways to address missing data when collecting RSH in research.
The problem
Historically, cardiovascular disease (CVD) management has been informed by research biased towards the male sex. In an effort to better understand sex differences, and to expand to populations beyond historically male-dominated research studies, multiple funding agencies across North America and Europe have issued notices either recommending or requiring the consideration of sex as a biological variable (SABV) in grant applications.1 Furthermore, many journals in cardiovascular science require the consideration of SABV in submitted manuscripts,2 while other journals dedicate special issues focused on women’s health or sex/gender differences.3–5 As such, there is a noticeable uptick in published research in this space, including in the European Journal of Cardiovascular Nursing (EJCN).6,7 Advancing women’s cardiovascular health is more than simply including women in studies, however; it requires incorporating sex-specific factors relevant to CVDs, as CVD remains the leading cause of death for women worldwide.8 Nearly half of all adult women in the USA have some form of CVD ranging from hypertension to advanced heart failure,9 and the incidence of CVD is rising in younger adults, with the greatest increases in women aged 35–54 years.10 This increase cannot be explained solely by traditional risk factors,11 but may also include risks specific to female biologic sex, such as reproductive system history (RSH).
Defined here as events across the female reproductive system lifespan (i.e. gynaecology and obstetrics history), RSH has been linked with cardiovascular function and outcomes among women.12–16 Recent reviews have clarified which RSH factors impact cardiovascular health.17,18 For example, there are sex-specific risk factors (i.e. endometriosis, pre-eclampsia, menopause) for developing any CVD including hypertension,14,19–21 coronary heart disease,17,19 cerebrovascular disease,17,19 and heart failure.21–23 Yet, collecting RSH is often an overlooked step in advancing the science around women’s cardiovascular health, and there are no standardized approaches currently. Reasons for not including RSH in cardiovascular research are multifactorial and may include poor investigator awareness of the impact of RSH on cardiovascular health,23 lack of gender diversity among primary investigators,24 and concerns regarding significant missing data.25 The purpose of this article is to describe two approaches to consider when collecting RSH in CVD research (Central illustration). For each approach, we (i) explain the RSH data collection method, (ii) provide an example using our research, and (iii) discuss benefits and barriers associated with it. Finally, we briefly provide guidance on addressing common statistical concerns when analysing lifespan RSH.
Steps to include reproductive system history in cardiovascular nursing research.
Potential solution #1: extract reproductive system history from the electronic medical record
Extracting data from the electronic medical record (EMR) is one way to ascertain RSH. With larger sample sizes (e.g. ≥ 500), it is possible to use machine learning to extract RSH from the EMR, as described elsewhere.26–28 In smaller samples, however, machine learning is not a reliable option for collecting RSH due to risk for overfitting the data,27 and so manual data extraction is necessary. Extracting data from the EMR requires addressing structured (i.e. medications, vital signs, or diagnostic codes) and unstructured (i.e. free-form, such as clinician notes) data.29
Developing a systematic approach to manual retrieval of RSH data from EMR is important to ensure rigour. First, search available literature to identify reproductive system factors that may contribute to the outcome of interest. For example, menopause is significantly associated with both incident frailty30 and heart failure22—and frailty is highly prevalent among women with heart failure;31 thus, when researching frailty among women with heart failure, it is important to consider the menopause stage as a variable. Next, creating a data-abstraction protocol for manual data abstraction maximizes reproducibility. The protocol should list RSH search terms that might be included in EMR (i.e. ‘post-menopause’, ‘pre-menopause’). Finally, researchers need to specify where the data are located in the EMR (i.e. the problem list, the clinician notes) and record on a data abstraction form for future clarification and verification.
Example of manually abstracting reproductive system history from electronic medical record
We used a method of manual data abstraction32 to characterize menopause status (i.e. pre- or post-menopause) in women enrolled in the Gender Associated Phenotypes in Physical Frailty in Heart Failure (GAP-FRAIL-HF) study.33 First, we specified the relevant terms using the nomenclature defined by the Stages of Reproductive Aging Workshop + 10—a multinational group of scientists who developed the gold-standard criteria for ovarian ageing criteria through menopause.13 Next, we examined each EMR problem list for RSH terms related to menopause. Then, we used the same terms and EMR search function to identify any relevant RSH documentation in narrative notes. For records where the search term resulted in hits, we read the narrative notes to determine if the participant had the defined RSH. Finally, we documented the RSH in our data abstraction form.
Results
In our sample of 66 women (Table 1) with heart failure, we found documentation of menopause status [n = 59 (89%)], menopause details [n = 44 (67%)], and medical interventions related to menopause (hysterectomy and cancer treatment). Seven (10.6%) did not have any documentation of the reproductive life stage. We addressed missing data by including only observations that included menopause status in our statistical analyses.
Reproductive system data abstracted from electronic medical record in GAP-FRAIL-HF
| n (%) | |
|---|---|
| Reproductive system life stage (n = 66) | |
| Missing | 7 (10.6) |
| Pre-menopause | 12 (17.9) |
| Menopause transition | 3 (4.5) |
| Post-menopause | 44 (67.7) |
| Menopause details (n = 44) | |
| Age of final menstrual period | 24 (54.5) |
| Natural menopause | 14 (31.9) |
| Surgical menopause | 10 (22.7) |
| Reproductive system medical interventions related to menopause | |
| Hysterectomy | 28 (41.8) |
| Age of hysterectomy | 20 (71.4) |
| Ovaries retained with hysterectomy | 8 (28.6) |
| Treatment for reproductive system cancer | 7 (10.4) |
| n (%) | |
|---|---|
| Reproductive system life stage (n = 66) | |
| Missing | 7 (10.6) |
| Pre-menopause | 12 (17.9) |
| Menopause transition | 3 (4.5) |
| Post-menopause | 44 (67.7) |
| Menopause details (n = 44) | |
| Age of final menstrual period | 24 (54.5) |
| Natural menopause | 14 (31.9) |
| Surgical menopause | 10 (22.7) |
| Reproductive system medical interventions related to menopause | |
| Hysterectomy | 28 (41.8) |
| Age of hysterectomy | 20 (71.4) |
| Ovaries retained with hysterectomy | 8 (28.6) |
| Treatment for reproductive system cancer | 7 (10.4) |
Reproductive system data abstracted from electronic medical record in GAP-FRAIL-HF
| n (%) | |
|---|---|
| Reproductive system life stage (n = 66) | |
| Missing | 7 (10.6) |
| Pre-menopause | 12 (17.9) |
| Menopause transition | 3 (4.5) |
| Post-menopause | 44 (67.7) |
| Menopause details (n = 44) | |
| Age of final menstrual period | 24 (54.5) |
| Natural menopause | 14 (31.9) |
| Surgical menopause | 10 (22.7) |
| Reproductive system medical interventions related to menopause | |
| Hysterectomy | 28 (41.8) |
| Age of hysterectomy | 20 (71.4) |
| Ovaries retained with hysterectomy | 8 (28.6) |
| Treatment for reproductive system cancer | 7 (10.4) |
| n (%) | |
|---|---|
| Reproductive system life stage (n = 66) | |
| Missing | 7 (10.6) |
| Pre-menopause | 12 (17.9) |
| Menopause transition | 3 (4.5) |
| Post-menopause | 44 (67.7) |
| Menopause details (n = 44) | |
| Age of final menstrual period | 24 (54.5) |
| Natural menopause | 14 (31.9) |
| Surgical menopause | 10 (22.7) |
| Reproductive system medical interventions related to menopause | |
| Hysterectomy | 28 (41.8) |
| Age of hysterectomy | 20 (71.4) |
| Ovaries retained with hysterectomy | 8 (28.6) |
| Treatment for reproductive system cancer | 7 (10.4) |
Benefits and barriers to this approach
Leveraging the EMR to streamline data collection in observational studies is widely accepted and can be implemented in both large and small sample sizes.34 Benefits of extracting RSH from both structured and unstructured data in the EMR include decreasing participant responder burden, and can improve researcher efficiency and reduce costs for researchers.34 For large sample sizes, machine learning can expedite EMR data collection by using computer algorithms.35 In studies with smaller sample sizes, manual data extraction using search terms from EMR is an effective way to collect RSH data. Limitations to this method include the time commitment required to manually extract the data and other common barriers to obtaining history from EMR such as age limitations and clinician bias. Reproductive system care that predates EMR (i.e. due to age) could contribute to incomplete history, as recording of the RSH is dependent on clinicians and health systems not only asking women about RSH, but also including responses in EMR. Moreover, when care is delivered across multiple health systems with different EMR, it is not always feasible for researchers to access all data due to heterogeneity across EMR platforms and firewalls between healthcare systems.34 Other potential disadvantages of collecting RSH from the EMR include quality concerns such as coding errors by clinicians entering the data, researcher knowledge of what is or is not included in the EMR, and missing data due to death or other unforeseen circumstances.
Potential solution #2: develop a self-report reproductive system history questionnaire
Including self-report RSH in a questionnaire36,37 is a participant-centred approach to ascertain nuanced information that may not be found in EMR. Beall and Leslie38 noted that women are the most reliable source for ascertaining RSH compared with medical record review, as EMR are often lacking complete reproductive system data.25 In 2016, Ouyang et al.25 published a thoughtful guide describing strategies to study female-specific cardiovascular health. Similar to Potential Solution #1, an essential first step to creating a self-reported RSH questionnaire is a careful review of the literature relating RSH with the outcome of interest. Next, writing RSH questions with a life-course approach allows participants to reflect on their history starting with the earliest event (i.e. menarche). For each question, define terms that may be confusing (i.e. menopause) and use sub-questions to elicit more information (i.e. age of last menstrual period). It may be helpful to include a final open-ended question that asks for any other RSH that the participant feels like sharing.
Example of a participant self-report reproductive system history questionnaire
Based on our findings shared in Potential Solution #1, we sought to collect self-reported RSH among women previously enrolled in the GAP-FRAIL-HF study. We conducted a follow-up pilot study to assess the feasibility of asking participants to recall and report their lifespan RSH. We created a participant survey based on evidence linking reproductive system factors with heart failure21–23,39 or frailty40,41 that asked women to recall their life-course RSH. Response options for each survey question were ‘yes’, ‘no’, or ‘do not know’ (Table 2). First, we contacted women enrolled in GAP-FRAIL-HF33 who indicated interest in future research and who were listed in our database as living to ask if they would participate in this follow-up study, which was approved by our Institutional Review Board. Surveys were distributed by email or post, and receipt of the completed survey indicated consent. Survey response time was kept open for a period of 4 weeks. Upon receipt of the completed survey, participants received $10.00 in renumeration. We reported the count and percentage of all responses on the RSH questionnaire.
Reproductive system history questionnaire used in GAP-FRAIL-HF reproductive
| 1 | At what age did you have your first menstrual period? |
| 1a | If unknown, please approximate your age were when you had your first menstrual period: □ Less than 10 years □ 10–12 years □ 13–15 years □ Greater than 15 years |
| 2 | Are you post-menopausal? |
| 2a | If yes, at what age did you have your final menstrual period? |
| 2b | If unknown, please approximate your age when you had your final menstrual period: □ <40 □ 40–45 □ 46–55 □ 56–61 □ >61 |
| 3 | How many total pregnancies have you had? |
| 4 | Have you experience pre-term delivery of a pregnancy? |
| 5 | Were you ever diagnosed with high blood pressure during pregnancy? |
| 6 | Were you ever diagnosed with pre-eclampsia during pregnancy? |
| 7 | Were you ever diagnosed with diabetes during pregnancy (Gestational diabetes)? |
| 8 | Have you ever breastfed an infant for any length of time? |
| 9. | Have you had a hysterectomy? |
| 10 | If yes, did you keep one or both of your ovaries when you had your hysterectomy? |
| 11 | Have you ever been diagnosed with endometriosis? |
| 12 | Have you ever been diagnosed with polycystic ovarian syndrome? |
| 13 | Have you received radiation or chemotherapy for breast, ovarian, uterine, or cervical cancer |
| 1 | At what age did you have your first menstrual period? |
| 1a | If unknown, please approximate your age were when you had your first menstrual period: □ Less than 10 years □ 10–12 years □ 13–15 years □ Greater than 15 years |
| 2 | Are you post-menopausal? |
| 2a | If yes, at what age did you have your final menstrual period? |
| 2b | If unknown, please approximate your age when you had your final menstrual period: □ <40 □ 40–45 □ 46–55 □ 56–61 □ >61 |
| 3 | How many total pregnancies have you had? |
| 4 | Have you experience pre-term delivery of a pregnancy? |
| 5 | Were you ever diagnosed with high blood pressure during pregnancy? |
| 6 | Were you ever diagnosed with pre-eclampsia during pregnancy? |
| 7 | Were you ever diagnosed with diabetes during pregnancy (Gestational diabetes)? |
| 8 | Have you ever breastfed an infant for any length of time? |
| 9. | Have you had a hysterectomy? |
| 10 | If yes, did you keep one or both of your ovaries when you had your hysterectomy? |
| 11 | Have you ever been diagnosed with endometriosis? |
| 12 | Have you ever been diagnosed with polycystic ovarian syndrome? |
| 13 | Have you received radiation or chemotherapy for breast, ovarian, uterine, or cervical cancer |
Based on this pilot study, we have revised the participant self-report to include more detailed definitions to facilitate participant understanding (e.g. ‘‘menopause transition’ means changes in periods, but have not gone 12 months in a row without a period’) and expanded the questions to include other relevant RSH factors such as hormone replacement therapy.
Reproductive system history questionnaire used in GAP-FRAIL-HF reproductive
| 1 | At what age did you have your first menstrual period? |
| 1a | If unknown, please approximate your age were when you had your first menstrual period: □ Less than 10 years □ 10–12 years □ 13–15 years □ Greater than 15 years |
| 2 | Are you post-menopausal? |
| 2a | If yes, at what age did you have your final menstrual period? |
| 2b | If unknown, please approximate your age when you had your final menstrual period: □ <40 □ 40–45 □ 46–55 □ 56–61 □ >61 |
| 3 | How many total pregnancies have you had? |
| 4 | Have you experience pre-term delivery of a pregnancy? |
| 5 | Were you ever diagnosed with high blood pressure during pregnancy? |
| 6 | Were you ever diagnosed with pre-eclampsia during pregnancy? |
| 7 | Were you ever diagnosed with diabetes during pregnancy (Gestational diabetes)? |
| 8 | Have you ever breastfed an infant for any length of time? |
| 9. | Have you had a hysterectomy? |
| 10 | If yes, did you keep one or both of your ovaries when you had your hysterectomy? |
| 11 | Have you ever been diagnosed with endometriosis? |
| 12 | Have you ever been diagnosed with polycystic ovarian syndrome? |
| 13 | Have you received radiation or chemotherapy for breast, ovarian, uterine, or cervical cancer |
| 1 | At what age did you have your first menstrual period? |
| 1a | If unknown, please approximate your age were when you had your first menstrual period: □ Less than 10 years □ 10–12 years □ 13–15 years □ Greater than 15 years |
| 2 | Are you post-menopausal? |
| 2a | If yes, at what age did you have your final menstrual period? |
| 2b | If unknown, please approximate your age when you had your final menstrual period: □ <40 □ 40–45 □ 46–55 □ 56–61 □ >61 |
| 3 | How many total pregnancies have you had? |
| 4 | Have you experience pre-term delivery of a pregnancy? |
| 5 | Were you ever diagnosed with high blood pressure during pregnancy? |
| 6 | Were you ever diagnosed with pre-eclampsia during pregnancy? |
| 7 | Were you ever diagnosed with diabetes during pregnancy (Gestational diabetes)? |
| 8 | Have you ever breastfed an infant for any length of time? |
| 9. | Have you had a hysterectomy? |
| 10 | If yes, did you keep one or both of your ovaries when you had your hysterectomy? |
| 11 | Have you ever been diagnosed with endometriosis? |
| 12 | Have you ever been diagnosed with polycystic ovarian syndrome? |
| 13 | Have you received radiation or chemotherapy for breast, ovarian, uterine, or cervical cancer |
Based on this pilot study, we have revised the participant self-report to include more detailed definitions to facilitate participant understanding (e.g. ‘‘menopause transition’ means changes in periods, but have not gone 12 months in a row without a period’) and expanded the questions to include other relevant RSH factors such as hormone replacement therapy.
Results
Nineteen out of 40 (47.5%) eligible participants returned the survey. Participants ranged in age from 33 to 91 years old and all had non-ischaemic heart failure aetiology. These respondent characteristics did not differ significantly from non-respondent characteristics. All participants reported at least one reproductive system events with known associations with heart failure or frailty (Table 3). One participant did not remember if they had a hysterectomy and a different participant could not recall if they had been diagnosed with adverse pregnancy events.
Reproductive system history ascertained from participant self-report
| n (%) | |
|---|---|
| Reproductive system life stage (n = 19) | |
| Remembers age of first menstrual period | 15 (79.0) |
| Estimates age of first menstrual period | 4 (21.1) |
| Menarche <10 years | 1 (5.3) |
| Post-menopausal | 13 (68.4) |
| Menopause ≤40 years | 4 (30.1) |
| Remembers age of final menstrual period | 10 (76.9) |
| Estimates age of final menstrual period | 2 (15.4) |
| Pregnancy-related history (n = 19) | |
| Nulliparous | 4 (21.1) |
| Three or more pregnancies | 11 (73.3) |
| Pre-term birth (<37 weeks) | 4 (36.4) |
| Unknown pre-eclampsia, hypertension, gestational diabetes | 1 (9.1) |
| Pre-eclampsia | 1 (9.1) |
| Gestational diabetes | 1 (9.1) |
| Breast/chest fed ever | 10 (90.9) |
| Reproductive system medical interventions and diagnoses (n = 18) | |
| Unknown hysterectomy | 1 (5.3) |
| Hysterectomy | 5 (27.8) |
| Ovaries retained with hysterectomy | 4 (80.0) |
| Endometriosis | 4 (21.1) |
| Polycystic ovarian syndrome (PCOS) | 3 (15.8) |
| Radiation or chemotherapy for reproductive system cancer | 1 (5.3) |
| n (%) | |
|---|---|
| Reproductive system life stage (n = 19) | |
| Remembers age of first menstrual period | 15 (79.0) |
| Estimates age of first menstrual period | 4 (21.1) |
| Menarche <10 years | 1 (5.3) |
| Post-menopausal | 13 (68.4) |
| Menopause ≤40 years | 4 (30.1) |
| Remembers age of final menstrual period | 10 (76.9) |
| Estimates age of final menstrual period | 2 (15.4) |
| Pregnancy-related history (n = 19) | |
| Nulliparous | 4 (21.1) |
| Three or more pregnancies | 11 (73.3) |
| Pre-term birth (<37 weeks) | 4 (36.4) |
| Unknown pre-eclampsia, hypertension, gestational diabetes | 1 (9.1) |
| Pre-eclampsia | 1 (9.1) |
| Gestational diabetes | 1 (9.1) |
| Breast/chest fed ever | 10 (90.9) |
| Reproductive system medical interventions and diagnoses (n = 18) | |
| Unknown hysterectomy | 1 (5.3) |
| Hysterectomy | 5 (27.8) |
| Ovaries retained with hysterectomy | 4 (80.0) |
| Endometriosis | 4 (21.1) |
| Polycystic ovarian syndrome (PCOS) | 3 (15.8) |
| Radiation or chemotherapy for reproductive system cancer | 1 (5.3) |
Reproductive system history ascertained from participant self-report
| n (%) | |
|---|---|
| Reproductive system life stage (n = 19) | |
| Remembers age of first menstrual period | 15 (79.0) |
| Estimates age of first menstrual period | 4 (21.1) |
| Menarche <10 years | 1 (5.3) |
| Post-menopausal | 13 (68.4) |
| Menopause ≤40 years | 4 (30.1) |
| Remembers age of final menstrual period | 10 (76.9) |
| Estimates age of final menstrual period | 2 (15.4) |
| Pregnancy-related history (n = 19) | |
| Nulliparous | 4 (21.1) |
| Three or more pregnancies | 11 (73.3) |
| Pre-term birth (<37 weeks) | 4 (36.4) |
| Unknown pre-eclampsia, hypertension, gestational diabetes | 1 (9.1) |
| Pre-eclampsia | 1 (9.1) |
| Gestational diabetes | 1 (9.1) |
| Breast/chest fed ever | 10 (90.9) |
| Reproductive system medical interventions and diagnoses (n = 18) | |
| Unknown hysterectomy | 1 (5.3) |
| Hysterectomy | 5 (27.8) |
| Ovaries retained with hysterectomy | 4 (80.0) |
| Endometriosis | 4 (21.1) |
| Polycystic ovarian syndrome (PCOS) | 3 (15.8) |
| Radiation or chemotherapy for reproductive system cancer | 1 (5.3) |
| n (%) | |
|---|---|
| Reproductive system life stage (n = 19) | |
| Remembers age of first menstrual period | 15 (79.0) |
| Estimates age of first menstrual period | 4 (21.1) |
| Menarche <10 years | 1 (5.3) |
| Post-menopausal | 13 (68.4) |
| Menopause ≤40 years | 4 (30.1) |
| Remembers age of final menstrual period | 10 (76.9) |
| Estimates age of final menstrual period | 2 (15.4) |
| Pregnancy-related history (n = 19) | |
| Nulliparous | 4 (21.1) |
| Three or more pregnancies | 11 (73.3) |
| Pre-term birth (<37 weeks) | 4 (36.4) |
| Unknown pre-eclampsia, hypertension, gestational diabetes | 1 (9.1) |
| Pre-eclampsia | 1 (9.1) |
| Gestational diabetes | 1 (9.1) |
| Breast/chest fed ever | 10 (90.9) |
| Reproductive system medical interventions and diagnoses (n = 18) | |
| Unknown hysterectomy | 1 (5.3) |
| Hysterectomy | 5 (27.8) |
| Ovaries retained with hysterectomy | 4 (80.0) |
| Endometriosis | 4 (21.1) |
| Polycystic ovarian syndrome (PCOS) | 3 (15.8) |
| Radiation or chemotherapy for reproductive system cancer | 1 (5.3) |
Benefits and barriers to this approach
Participant self-report of RSH may provide additional RSH details missing from the EMR. Studies have shown that self-report is mostly reliable for pregnancy-related events42 or hysterectomy,43 but validation of self-reported RSH is missing from the literature. There may be limitations with recall and understanding of RSH, potentially limited by the effectiveness of provider communication and generational cohort effects.42 Recall bias may skew study results when participant reporting of gynaecologic or obstetric history is limited by memory, time since event, or history not being adequately communicated to the patient. Recall bias can be partially mitigated by generating carefully worded RSH questions that clearly define RSH events and comparing EMR and self-report RSH. Non-response bias can threaten the validity of the results and may be mitigated by using evidence-informed strategies to achieve a higher response rate44 Bias from participant underreporting—either intentionally or unintentionally—can be mitigated by ensuring questions are specific and include clear language when developing individual RSH questions.
Analysing and reporting reproductive system history in research
Analysing outcomes based on reproductive age [i.e. pre- or post-menopause either by self-report, average age of menopause (i.e. 51 years), or medical records], rather than chronologic age, may provide a more nuanced understanding of the impact of RSH on the outcome of interest. However, analysing data that includes RSH beyond reproductive age and across the lifespan can present unique challenges of missing data from cohort effects or truncation. Cohort effects (i.e. participants living through different time periods of availability of medical information) can result in missing data because participants either were never informed of or do not remember details of their RSH. To address missing data because a participant did not report all or part of their RSH, researchers should first contact the participant to ensure the participant understood the question. If this does not resolve the problem, analysts should decide to (i) exclude the observation from the analysis, (ii) impute the data from other subjects with similar life-course RSH, or (iii) use an estimation approach that does not require complete observations on each case (e.g. full information maximum likelihood). Imputation, however, is only appropriate when estimating age of menarche or menopause, but not to estimate the presence of other RSH events, such as endometriosis or pregnancy outcome.
Truncation [i.e. data on a given event of RSH cannot be analysed from all participants because not all have experienced it (e.g. pregnancy or menopause)] is likely in samples that include participants across the lifespan. For example, it is common for women in our studies range in age from 18 to 100 years,33 meaning that women are at different stages of reproductive senescence. Moreover, even in participants at a similar reproductive age, there may be vast differences in RSH (e.g. some participants have a diagnosis of endometriosis, but others do not). Listwise deletion might be the best approach when examining outcomes associated with specific RSH factors. For example, when examining the relationship between hypertensive disorders of pregnancy and HF, analysts should exclude participants who have not been pregnant. Importantly, when reporting outcomes that include RSH, the authors should specify mechanisms of missingness and how missing data were handled. Moreover, for specific RSH, reporting details such as average age of event or diagnosis (i.e. menopause) provides additional perspective important to cardiovascular outcomes.
Comparing medical record and self-report reproductive system history for congruency
Comparing medical records of RSH with participant self-report is a potential solution to obtaining the most detailed RSH and addressing reliability and bias in RSH data collection. Some studies report the reliability of recall compared with medical records. In one study validating a maternal recall of pregnancy complications questionnaire, Carter et al.45 found maternal recall of infant birthweight and gestational diabetes was highly accurate, but recall of hypertensive disorders of pregnancy was less accurate. These important factors individually been shown to be signals for subsequent development of Type 2 diabetes, hypertension, heart failure, and stroke.16 In a systematic review of maternal recall of hypertensive disorders of pregnancy, Stuart et al.42 found that among 10 studies that compared maternal recall with medical records (structured and unstructured data), recall periods (e.g. the length of time from pregnancy) did not predict recall quality; however, sensitivity was relatively low, indicating that maternal recall of hypertensive disorders of pregnancy may not always be reliable. In a study comparing participant report of endometriosis diagnosis to documentation of endometriosis in medical record, Shafrir et al.46 found high congruency between medical record and self-report. For researchers considering assessing congruency between self-report and EMR, consideration of participant age, life stage, and healthcare utilization should occur. Methods for resolving discrepancies between EMR and participant self-report should be determined prior to data analysis and reported with data analysis.
Conclusion
In cardiovascular research, studying female-specific outcomes or sex differences should include RSH. Here, we provide details on two approaches to collecting RSH and included examples from our recent research in heart failure. We also provided guidance on data management for statistical analysis. Regardless of the method implemented, collecting RSH is an essential variable when studying female-specific or sex/gender difference outcomes in CVD. Future research in this area includes clinician understanding of RSH impact on cardiovascular health, and interventions to improve RSH data collection in the cardiovascular clinical and research settings.
Funding
This work was funded by the Office of Research on Women’s Health and the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health (8K12AR084221). The work reported in this paper was also supported, in part, by the National Center for Advancing Translational Sciences of the National Institutes of Health (UL1TR002369).
Data availability
The data underlying this article will be shared on reasonable request to the corresponding author.
References
Author notes
Conflict of interest: None declared.
Comments