Neighborhood evictions, household children, and preterm birth among Black birthing people Authors: Lea Ghastine, Shawnita Sealy-Jefferson, Elizabeth S McClure, Cathrine Hoyo, Chantel L Martin Published: 2025 ABSTRACT AbstractBackgroundNeighborhood evictions are a key component of community-level housing instability and disproportionately impact communities of color. Increased neighborhood evictions may be associated with increased preterm birth (PTB). Previous research suggests familial distress as a potential driver of increased PTB among Black birthing parents exposed to high neighborhood eviction rates. In this study, we explore whether the association of neighborhood evictions with PTB differs depending on the presence or absence of children in the household during pregnancy.MethodsWe used information from 400 Black participants in the Newborn Epigenetics Study in Durham, NC. Our primary exposure, neighborhood eviction rate, was defined as the participant’s block group eviction rate during the year of last menstrual period. Preterm birth, defined as gestational age < 37 completed weeks at delivery was based on information abstracted from medical records at delivery. Using logistic regression, we estimated the change in odds of preterm birth associated with a 1 SD increase in neighborhood eviction rate. Models were adjusted for educational attainment, relationship status, participant age at delivery, and proportion of block group households below the federal poverty line. To determine whether associations between neighborhood evictions and preterm birth differ depending on household composition, we stratified the study sample by presence of children in the household during pregnancy (0, ≥ 1 children).ResultsAmong the total sample, we observed an increase in odds of preterm birth for every 1-SD increase in neighborhood eviction rate (OR: 1.21, 95% CI: 0.83, 1.73). Association between neighborhood eviction rate and preterm birth differed by household composition. Increased neighborhood eviction rates were associated with increased odds of PTB among participants with children in the household (OR: 1.89, 95% CI: 1.09, 3.28), but not among participants without children in the household (OR: 0.85, 95% CI: 0.49, 1.39).ConclusionsWe found that higher neighborhood evictions increase odds of preterm birth among Black birthing parents living with children. Our findings support the prioritization of Black families and communities for interventions to prevent evictions and promote housing stability.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12889-025-25577-w.Keywords: Evictions, Neighborhood evictions, Neighborhood environment, Household composition, Pregnancy, Preterm birth Introduction Evictions, defined as involuntary displacement of occupants from a rented residence, affect millions of Americans each year due to profit prioritization in the housing market [1], rising housing costs, paltry federal assistance, and increasing income inequality [2]. Eviction perpetuates poverty and is associated with increased risk of mental illness, poor birth outcomes, chronic disease, and poor self-rated health [3–8]. At the community-level, evictions are a key component of neighborhood housing instability, an important neighborhood characteristic with implications for neighborhood disadvantage, community stability, social cohesion, and health [9, 10]. Ecologic studies have demonstrated associations of area-level evictions with poor birth outcomes, infant mortality, and sexually transmitted infections, even after controlling for socioeconomic indicators [4, 11–13]. Due to structural racism, Black neighborhoods are disproportionately burdened by evictions and residential instability in the United States (US) [14–16]. Structural racism, a fundamental cause of racial and ethnic health inequities, refers to the ways in which societal structures perpetuate disadvantage in socially constructed racial groups through reinforcing inequitable systems (in housing, education, employment, finance, criminal justice, etc.) that in turn contribute to disparities in health, among many other inequities [17]. Structural racism harms health in racially marginalized groups through various mechanisms, including through concentration of disadvantage in majority non-white neighborhoods [18] and biological harm caused by chronic stress [19–21]. Neighborhood housing instability inflicted via evictions and foreclosures is a neighborhood-level manifestation of structural racism. In the U.S., evictions tend to be concentrated in majority non-white neighborhoods [14–16], and the burden of evictions is especially high among low-income, single, Black mothers with young children [22]. Three recent studies demonstrated that individual exposure to increased county or block group evictions is associated with increased preterm birth (PTB), defined as birth before 37 completed gestational weeks [23–25]. The positive association of area-level evictions with PTB appears to be stronger among Black birthing people, suggesting that Black birthing people may experience greater harm compared to birthing people of other races exposed to a similar eviction burden [23, 25]. Increased evictions may impact PTB risk by reducing community investment in the area, thereby contributing to reduced access to resources in the neighborhood, increased chronic stress, and decreased healthy behaviors among residents. In addition, residents of neighborhoods where evictions are commonplace may face the threat of housing instability themselves, which has been linked to various adverse health outcomes [26–28]. PTB is the leading cause of infant mortality in Black infants [29]. In the US in 2021, 14.8% of births to Black birthing parents were preterm compared to around 10% of births to white and Hispanic birthing parents [30]. Unfortunately, the burden of PTB among non-Hispanic Black individuals has not declined over the past several years [31–33]. Differences in socioeconomic status (SES), health behaviors, and prenatal care do not fully account for disparate burden of adverse birth outcomes among Black birthing parents [34–37]. Furthermore, adverse birth outcomes are not elevated to the same extent in African-born Black pregnant individuals when compared to American-born Black pregnant individuals [38, 39], ruling out etiology related to genetic ancestry and implicating exposure to structural racism across the life course [40, 41]. Neighborhood evictions may contribute to racial inequity in PTB given the disparately increased burdens of both evictions and PTB in Black communities and recent findings linking area-level eviction rates to increased PTB among Black birthing parents [23–25]. However, very few studies investigate the impact of evictions at the neighborhood level among Black birthing parents. As evictions are a modifiable neighborhood characteristic, it is important to further explore this pathway to prioritize avenues for interventions to prevent evictions and PTB among Black individuals. Our study builds upon this work by assessing whether the association of neighborhood evictions with PTB differs according to the presence of children in the household during pregnancy among Black birthing parents. Our study cohort resides in Durham, North Carolina, a city with entrenched history of racial residential segregation experiencing a decades long eviction crisis that disproportionately affects Black residents [42, 43]. Decades of residents’ experiences, as well as research, have demonstrated the cascading harm caused by eviction with respect to family financial, social, emotional, and physical wellbeing, including PTB. The purpose of this study is to investigate whether the relationship between neighborhood evictions and PTB is modified by the presence of children in the household during pregnancy among 400 singleton births to Black pregnant people in the Newborn Epigenetics Study (NEST) from 2009 to 2011. Methods Study population NEST is an ongoing prospective birth cohort study with 2,681 pregnant people recruited in two waves between 2005 and 2011. Recruitment is described in detail elsewhere [44, 45]. Participants were recruited from prenatal clinics serving Duke University Hospital and Durham Regional Hospital Obstetrics facilities in Durham, North Carolina. Eligible participants were: (1) pregnant (2), at least 18 years of age (3), English speaking, and (4) intending to deliver at Duke University Hospital or Durham Regional Hospital. Individuals with HIV or intending to give up custody of their offspring were excluded. A baseline survey administered to participants during the second wave of enrollment collected data including demographics, medical history, health behaviors, diet, perceived stress, depressive symptoms, environmental exposures, and home address during pregnancy. Data on household occupants were collected from participants recruited during the second wave, but not from those recruitedthe first wave of enrollment, so first wave participants were excluded (n = 1,134). Thus, the current study includes participants recruited during the second wave of enrollment (2009–2011) (n = 1,547). Participants were excluded if they lived outside of Durham County (n = 309), did not identify as Black or African American (n = 858), had multiple gestations (n = 72), or were missing a geocoded address (n = 132). There were 456 Black or African American participants with singleton pregnancies who lived in Durham County, North Carolina. To achieve a complete cases analysis, participants were excluded if they were missing covariates including relationship status (n = 10), educational attainment (n = 8), participant age at delivery (n = 36), household children during pregnancy (n = 3), or gestational age at delivery (n = 38). The resulting complete cases sample included in this analysis was 400 participants with singleton pregnancies who identified as Black or African American and resided in Durham County, North Carolina. Please see Fig. 1 for our cohort flow diagram. Ethical approval was obtained from the Institutional Review Board at the University of North Carolina – Chapel Hill. Cohort Flow Diagram Preterm birth assessment Gestational age at birth was calculated based on participant’s self-reported date of last menstrual period and date of delivery, which was abstracted from medical records at delivery. The primary outcome, PTB, is defined as gestational age < 37 completed weeks at delivery. Neighborhood evictions assessment Neighborhood eviction rate, the primary exposure, is defined as the participant’s block group eviction rate during the year of last menstrual period (LMP). To estimate the burden of evictions in each block group, we rely on summary ejectments, the first of two court processes required to formally evict a tenant in Durham, North Carolina. Annual summary ejectments and number of renter-occupied households standardized to 2010 census boundaries for all 152 Durham County block groups were obtained from Durham Neighborhood Compass [43] and National Historical GIS [46], respectively. Given that summary ejectments represent one step in the formal eviction process that may or may not precede an eviction judgement, they likely overestimate the number of formal evictions and underestimate the number of total evictions (formal and informal combined). Research suggests formal evictions may comprise less than a quarter of total evictions [10]. Here, we assume summary ejectments positively correlate with total evictions in a block group. Annual block group eviction rates were calculated using a method described by The Eviction Lab (total number of summary ejectments within the block group divided by the number of renter-occupied households within the block group). A multiplier of 100 was used to obtain the number of summary ejectments per 100 renter-occupied households in the block group (hereafter “neighborhood eviction rate”) [47, 48]. Compared to eviction counts, eviction rates better capture the impact on overall neighborhood housing stability by accounting for the total number of renter-occupied households in the block group; for example, 50 evictions in a block group with 200 renter-occupied households represents a drastically different neighborhood environment compared to 50 evictions in a block group with 1,000 renter-occupied households. Covariates We stratified the study sample by presence of children in the household during pregnancy (0, ≥ 1 children). Number of children in the household excludes the current gestation. In a self-administered baseline survey, participants were asked “Did you live with children when you got pregnant?” (yes/no). In the baseline survey, participants were asked to specify their race from pre-determined categories (Black/African American, non-Hispanic White, Hispanic White, Asian/Pacific Islander, American Indian/Native American, Biracial or Multiracial, or Other). Separate questions asked participants whether they were of Hispanic ethnicity (yes/no) or born in the U.S. (yes/no). Our sample includes all participants identifying as Black or African American, including Hispanic, non-Hispanic, U.S.-born, and non-U.S.-born participants. However, it does not include participants who identified as Biracial or Multiracial, as no further information was collected regarding the racial identity of these individuals. Individual-level covariates for this analysis were collected through self-completed or interviewer-administered surveys at baseline and include participant age at delivery, educational attainment (less than high school, high school graduate/GED, some college, or college graduate), and relationship status (never married, married, widowed, living with partner, divorced/separated, or other). Block group poverty (proportion of household incomes below the federal poverty line), a neighborhood-level covariate, was obtained from U.S. Census data. Participant age and block group poverty were modeled continuously. Due to sparse data, we dichotomized educational attainment (no college education vs. any college education) and relationship status (married or cohabiting vs. neither married nor cohabiting). We conceptualized baseline health status and pregnancy health-status as being on the causal pathway between our exposure and outcome, so we did not adjust for these variables. Statistical analysis To examine how the effect of neighborhood evictions may differ depending on the household composition, we stratified the study sample by presence of children in the household during pregnancy (0, ≥ 1 children). We evaluated the validity of this stratification through a test of interaction between neighborhood eviction Z-score and presence of household children in unadjusted and adjusted models. Participant sociodemographic characteristics were described by presence of children in the household using means and standard deviations for continuous variables and frequencies for categorical variables. To estimate the association between higher neighborhood eviction rate and PTB, we rescaled continuous neighborhood evictions by the standard deviation (SD = 18.6 evictions per 100 renter-occupied households). We used logistic regression models to compute crude and adjusted odds ratios (ORs) and associated 95% confidence intervals (CIs) to estimate the change in odds of PTB associated with a 1 SD increase in neighborhood eviction rate. We adjusted for covariates comprising a sufficient adjustment set according to our directed acyclic graph of the relationship between neighborhood evictions and PTB (Supplemental Fig. 1). Adjusted models included participant dichotomous educational attainment and relationship status variables, participant age at delivery (continuous), and proportion of block group households below the federal poverty line (continuous). In addition, we examined the effects of participant clustering within block groups by calculating the intraclass correlation coefficient (ICC) from a null linear mixed model and used generalized estimating equations (GEEs) to ensure the robustness of our logistic regression results to heteroskedasticity and non-independence of observations within the same block group. The ICC showed the effects of clustering to be negligible, and GEEs yielded similar ORs and standard errors to logistic regression models. Therefore, we report the results from logistic regression models below. All analyses were performed in R version 4.1.0. Results Characteristics of study participants are summarized in Table 1. Of 400 participants in the study population, 208 (52.0%) reported having children living in their households during pregnancy. Overall, 9.5% of births were preterm. The proportion of preterm births was lower among those with children in the household (7.7%) compared to those without children in the household (11.5%). People with children in the household had an older mean age at delivery (27.3 years) compared to those without children in the household (24.6 years). Participants were mostly non-Hispanic (95.8%) and U.S. born (96.4%). Among people with children in the household, 41.8% attended college and 42.8% reported being married or living with their partner. Among people without children in the household, 52.1% attended college and 35.4% were married or living with their partner. Mean block group percent of households with income below the federal poverty line was 24% among people with children in the household and 19.8% among people without children in the household. Mean neighborhood eviction rate was higher for participants with children in the household (42 evictions 100 per renter-occupied households) compared to those without children in the household (38.8 evictions per 100 renter-occupied households). Table 1Sociodemographic characteristics of black NEST participants residing in Durham, North Carolina 2009–2011No household children(N = 192)Household children(N = 208)Overall(N = 400) Preterm birth No170 (88.5%)192 (92.3%)362 (90.5%) Yes22 (11.5%)16 (7.7%)38 (9.5%) Participant age at delivery (years) Mean (SD)24.6 (5.59)27.3 (5.18)26.0 (5.54) Hispanic ethnicity No186 (96.9%)197 (94.7%)383 (95.8%) Yes0 (0%)3 (1.4%)3 (0.8%) Missing6 (3.1%)8 (3.8%)14 (3.5%) Born in the U.S. No7 (3.6%)9 (4.3%)16 (4.0%) Yes185 (96.4%)199 (95.7%)384 (96.0%) Attended college Yes100 (52.1%)87 (41.8%)187 (46.8%) No92 (47.9%)121 (58.2%)213 (53.3%) Married or living with partner Yes68 (35.4%)89 (42.8%)157 (39.3%) No124 (64.6%)119 (57.2%)243 (60.8%) Block group percent of households below federal poverty level Mean (SD)19.8 (17.3)24.0 (19.1)22.0 (18.4) Neighborhood eviction rate Mean (SD)38.8 (18.0)42.0 (19.1)40.4 (18.6) Interquartile Range25.3–50.527.0–50.926.2–50.5 Sociodemographic characteristics of black NEST participants residing in Durham, North Carolina 2009–2011 Our logistic regression results demonstrate that in the overall study group, a 1-SD increase in neighborhood eviction rate was associated with 1.21 times the odds of PTB (95% CI: 0.83, 1.73) adjusted for participant age at delivery, relationship status, educational attainment, and neighborhood poverty (Fig. 2, Supplemental Table 1). In participants living with children during pregnancy, a 1-SD increase in neighborhood eviction rate was associated with nearly twice the odds of PTB (OR = 1.89; 95% CI: 1.09, 3.28) adjusted for covariates. This relationship was not observed among participants who were not living with children during pregnancy (OR = 0.85; 95% CI: 0.49, 1.39). Our test of the interaction term between neighborhood eviction rate and household children yielded a p-value of 0.053 in the unadjusted model and a p-value of 0.042 in the adjusted model. Our sensitivity analyses using GEE to calculate ORs yielded similar results (see Supplemental Table 1). In addition, our ICC calculated from a null linear mixed model was 0.097, showing that the effects of clustering were small and most of the total variance is accounted for by individual-level differences. Adjusted OR and 95% CI for one SD increase in neighborhood eviction rate by presence of household children in the home. Models are adjusted for educational attainment, relationship status, participant age at delivery, and proportion of block group households below the federal poverty line. Interaction term between neighborhood eviction rate and household children p-value: 0.04 Discussion Our study finds that higher neighborhood eviction rates are associated with higher odds of PTB among Black NEST participants living with children during pregnancy in Durham, NC from 2009 to 2011. This relationship was not observed among those who were not living with children during pregnancy, suggesting that Black families with children may be particularly susceptible to the harmful impact of neighborhood evictions on PTB. This complements previous work by Sealy-Jefferson et al. which found that higher block group evictions were associated with increased PTB among Black women who were married or living with their partner [24]. One possible explanation for these findings is that community socioeconomic disadvantage harms familial relationships [49, 50], potentially leading to adverse health outcomes through mechanisms such as chronic stress and maladaptive family health behaviors [51–53]. In addition, Black birthing parents with children who are living in areas with high eviction rates may experience increased threat of eviction themselves, as Black mothers with children are especially likely to face eviction [54]. The reproductive justice framework is useful to contextualize the implications of our findings. The reproductive justice framework coined in 1994 by the Women of African Descent for Reproductive Justice, is defined as: “ [1] the right not to have a child; [2] the right to have a child; and [3] the right to parent children in safe and healthy environments” [55]. The framework emphasizes access to community-based resources – including a living wage, a healthy environment, health care, housing, and education – as a human right [55]. From a reproductive justice perspective, the inequitable burden of PTB, infant mortality, and other adverse birth outcomes among Black birthing people constitutes reproductive oppression – the regulation or exploitation of individuals’ procreative capacity [56]. Reproductive justice intersects with other social justice issues such as poverty, health inequity, prisoners’ rights, environmental justice, and housing, the latter of which is the focus of this paper. Neighborhood environments that increase adverse birth outcomes and impede a parent’s ability to provide safe and healthy environments for their children (including stable housing, safe recreation, nutritious food, strong education, and ample employment opportunities) constitute reproductive injustice. Under the reproductive justice framework, our results support that increased neighborhood evictions, by harming reproductive health outcomes, are a form of reproductive oppression. This research builds upon a growing body of work investigating the impact of neighborhood or county evictions on individual birth outcomes by race and ethnicity, which has largely shown that increased area-level evictions are associated with increased adverse birth outcomes [23, 25]. Our results are also consistent with previous findings showing associations of neighborhood housing instability with adverse health outcomes in non-white communities [57], further emphasizing the role of neighborhoods in perpetuating racial and ethnic health inequities. Residing in neighborhoods with high eviction rates may contribute to adverse birth outcomes through the embodiment of stress, the process by which the material and social environment promote harmful physiological adaptation, increasing risk for adverse health outcomes [19, 58, 59]. People living in neighborhoods with high eviction burdens may face increased stress for several reasons including the detrimental impact on the neighborhood environment and the threat of being evicted themselves. Such neighborhood environments may negatively impact family cohesion. It is possible that the negative impact on family cohesion may be more pronounced on households with children due to additional stressors, adding an additional layer of stress that may contribute to the detrimental effect of neighborhood evictions on PTB specific to individuals living in households with children. However, current research evaluating the impact of socioeconomic circumstances on family relationships largely focuses on romantic relationships or the effect of parental stress on children [50, 53]. For effective interventions, further research is needed to understand the pathways connecting socioeconomic disadvantage, household stress, and adverse birth outcomes. Neighborhood context is a powerful force in promoting reproductive justice and shaping overall wellbeing, as demonstrated by the spatial clustering of myriad outcomes such as infant mortality, low birth weight, accidental injury, homicide, and incarceration rates that defies explanation by individual behaviors or socioeconomic disadvantage [9, 60–62]. Evictions contribute to neighborhood housing instability, an important component of the neighborhood context. For individual households, housing stability provides a foundation for psychological stability and allows residents to invest in their homes, relationships, and communities [63]. Housing instability may harm community health by disrupting neighborhood collective efficacy, defined by Sampson et al. as “the capacity of residents to achieve social control over the environment and to engage in collective action for the common good” [64]. Lower neighborhood collective efficacy is associated with greater burdens of obesity [65], depression [66], poor self-rated health [67], poor diet [68], asthma [69], and low birth weight [70]. Compared to voluntary moves, the threat of evictions causes exceptional household stress and material hardship [3], resulting in additional barriers to community integration and causing more damage to neighborhood collective efficacy. Evictions are a particularly salient indicator of neighborhood context in Durham, NC. For the past 20 years, Durham County has seen 10,000 summary ejectments each year. Before onset of the COVID-19 pandemic, Durham was seeing a similar pattern, with over 2,000 summary ejectments filed by March 2020 [71]. The CDC eviction moratorium and a subsequent North Carolina executive order were temporary mitigations to displacement in Durham, but evictions have been steadily rising again since these moratoria lapsed in July 2021 [72]. Durham’s eviction and displacement crisis can largely be traced to tenants’ inability to make rent due to economic conditions. Durham’s unemployment rates have rapidly increased during the pandemic, and service work wages have declined over the past two decades [73]. Durham evictions largely occur in rental units owned by local landlords who operate older apartment complexes marketed toward working class People of Color and large scale, global real estate investment firms [71]. Local community organizations are working to reform local eviction and tenant protection policies as well as raise funds to offset the costs of evictions to tenants [74]. This study adds to the mounting evidence for the urgency and importance of addressing eviction and displacement in Durham, especially among Black residents. It suggests that in addition to addressing the already unjust process of eviction-related displacement, reducing evictions in Durham could also have health benefits for residents of neighborhoods with high concentrations of evictions and address the disparately high burden of PTB in Black birthing parents. This study has several limitations. Our study relied on formal eviction records, which notoriously underestimate the true number of evictions due to the large number of informal evictions that take place outside the court system [10, 75]. For example, a tenant may move preemptively following an informal threat of eviction from a landlord to avoid a formal eviction tarnishing their record. Furthermore, we were unable to determine whether participants experienced evictions themselves, and increased neighborhood eviction rate is likely associated with individual probability of eviction. Unfortunately, we cannot differentiate between the impact of neighborhood evictions and individual evictions. In addition, we did not have data on length of time in current residence, so we were unable to determine the role of exposure duration. This measurement error is likely to be larger among those on the higher end of exposure due to housing instability. Moreover, it is possible for homeowners to be impacted by high neighborhood eviction rates differently than renters. Unfortunately, our study did not collect information on whether participants were homeowners or renters, and we were unable to assess potential differential effects. Additionally, selection bias may have been introduced by the exclusion of several populations. Studyparticipants were recruited from prenatal care settings, so individuals who did not seek prenatal care were excluded. Individuals who identified as biracial/multiracial were not included in this study. People with HIV and non-English speaking individuals were excluded from the parent study. The exclusion of these populations (particularly those who did not seek prenatal care and have limited English proficiency) that may experience increased preterm birth may potentially bias our results toward the null. There is also the potential of residual confounding by unmeasured variables. Because we defined PTB using gestational age based on LMP, it is possible that there was outcome misclassification. Generalizability of our results to the broader Durham population and other metropolitan areas also may be limited due to the overrepresentation of highly educated individuals in our study population. Lastly, our study used census data to measure neighborhood characteristics. Census block groups are an imperfect proxy for what an individual perceives as their “neighborhood”, and census data may not accurately reflect residents’ perceptions of neighborhood characteristics [76, 77]. Conclusion This study demonstrates that increased neighborhood evictions increase odds of PTB among Black birthing parents living with children. PTB is a source of lifelong health consequences and the primary cause of mortality among Black infants despite decades of research. Stable housing provides a foundation for economic, psychological, and social stability for individual households and communities; thus, it is essential for reproductive justice. The outsized burden of housing instability in Black neighborhoods contributes to the reproduction of poor health, poverty, and disadvantage across generations [54]. Our findings support the prioritization of Black families and communities for interventions to prevent evictions and promote housing stability. Supplementary Information Supplementary Material 1. Acknowledgements We thank the NEST participants for generously contributing their data and time for this study. Authors’ contributions LG conceptualized and designed the study, drafted the manuscript, conducted the analysis. CLM obtained the data, contributed to writing the manuscript, and supervised all aspects of the study. CH led the parent study. SSJ, ESM, CH contributed to conceptualizing the study, writing the manuscript, and interpreting results. All authors read and approved the final manuscript. Funding This research was supported by the following grant mechanisms: K99MD012808/R00MD012808 from the National Institute of Minority Health and Health Disparities (CLM) and P30ES010126 from the National Institute of Environmental Health Sciences (CLM). The NEST study was funded by the National Institute on Aging (R21AG041048) and the National Institute of Environmental Health Sciences (R01ES016772). Data availability To maintain participant confidentiality, the datasets analyzed for the current study are not publicly available due to the sensitive and identifiable nature of the data, but are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate This study was conducted in accordance with the Declaration of Helsinki and approved by Duke University, Durham Regional Hospital, and the University of North Carolina at Chapel Hill Institutional Review Boards and all participants provided written informed consent. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement To maintain participant confidentiality, the datasets analyzed for the current study are not publicly available due to the sensitive and identifiable nature of the data, but are available from the corresponding author on reasonable request.