Research Article, J Womens Health Vol: 14 Issue: 3

Relations between Social and Institutional Support and Women’s Adjustment Following Pregnancy Loss

Alex M. Gillham^1*, Jessica P Abdalla², Nicki L. Aubuchon-Endsley^1,2

¹Department of Psychology, The University of Tulsa, Tulsa, OK, United States
²Department of Psychology, Idaho State University, Pocatello, United States

*Corresponding Author:

Alex M. Gillham

Department of Psychology,

The University of Tulsa,

Tulsa, OK,

United States,
Email: amg5038@utulsa.edu

Received date: 06 August, 2024, Manuscript No. JWHIC-24-144750;
Editor assigned date: 09 August, 2024, PreQC No. JWHIC-24-144750 (PQ);
Reviewed date: 23 August, 2024, QC No. JWHIC-24-144750;
Revised date: 10 June, 2025, Manuscript No. JWHIC-24-144750 (R);
Published date: 17 June, 2025, DOI: 10.4172/2325-9795.1000547.

Citation: Gillham AM, Abdalla JP, Aubuchon-Endsley NL (2025) Relations between Social and Institutional Support and Women’s Adjustment Following Pregnancy Loss. J Womens Health 14:3.

Abstract

Keywords: Mental health, Pregnancy loss, Support

Introduction

Approximately 10-20% of pregnancies result in miscarriage and 1% result in stillbirth [1,2]. In the United States (U.S.), 48-51% of these women experience psychiatric difficulties following their loss, and 22-44% exhibit clinically significant anxiety and depression [3]. Women who have experienced miscarriage or stillbirth are also at risk for experiencing traumatic stress and grief [4,5].

These symptoms of grief, depression, anxiety, and trauma [6] may be persistent and distressing, warranting additional research to understand associated risk and resiliency factors [7]. Specifically, women’s grief may go unrecognized by healthcare providers and/or loved ones, resulting in a lack of support [8]. A lack of recognition of a woman’s grief surrounding pregnancy loss can increase negative emotional consequences, while social and healthcare support aids with psychological adjustment [9]. Even if a physician and patient do not see the pregnancy loss as a significant source of distress, the loss could add to a woman’s traumatic burden and increase their vulnerability for mental health issues during subsequent pregnancies [10].

Conceptually, Bronfenbrenner’s ecological systems theory [11] and the stress buffering model of social support suggested by Cohen and Wills [12] highlight the importance of considering both a person’s context (e.g., timing of pregnancy loss) and microsystem factors (e.g., women’s experiences with family/friends/partners and healthcare institutions) to understand important life events (e.g., pregnancy loss). While prior literature supports the relationship between pregnancy loss and psychological and social outcomes, few recent studies have explored complex, multivariate relations among Gestational Age (GA) at the time of loss, different types of support provided, and various psychological outcomes [13].

Though there are mixed findings regarding timing of loss and psychosocial outcomes, women suffering early, or late term pregnancy loss are at a significant risk for negative psychological and social consequences. A better understanding of these relations would result in clarifying which types of support and other psychosocial interventions are most beneficial based on women’s psychological experiences corresponding to time of loss. In addition, there is a paucity of research examining these relations via quantitative, psychometrically sound measures validated or used with women experiencing pregnancy loss [14] and/or measures accessible and feasible (e.g., cost and administration time) for researchers and clinicians.

Proposed models and hypotheses

This study aims to examine relations between pregnancy loss timing and several indices of psychological maladjustment while considering important moderators (i.e., social and institutional support) using quantitative measures validated or commonly used with this population. To test hypotheses, eight moderation models were proposed. We first hypothesized that increased GA during loss is related to greater psychological symptoms (a=depression, b=anxiety, c=post-traumatic stress symptoms, and d=grief) at 2 years following the loss (Hypothesis 1 a-d). Most studies reviewed timeframes between 1-2 years following pregnancy loss, which informed the present study’s 2-year cut-off [6,15]. Swanson, et al. highlighted mixed research results and an unclear association between GA at the time of loss and psychological outcomes suggesting that more research is needed in this area. Additionally, we aimed to address gaps in the literature by assessing social and institutional support and clarifying the prevalence rates and persistence of psychological outcomes [3-6].

We predicted a significant interaction between GA and perceived social support, such that women with greater GA and less perceived social support would report more adverse psychological symptoms (Hypothesis 2a-d). For Hypotheses 1 and 2, an a priori power analysis using GPower (version 3.1), including a medium effect size (f2=0.15), two-tailed p-value of .05, and power of .80, suggested a sample of 98 participants for sufficient power [16].

We also hypothesized that women who experienced a miscarriage (≤ 20 weeks GA) would endorse less healthcare and social support than women who experienced a stillbirth (>20 weeks GA; Hypothesis 3). This may be because many women who experience pregnancy loss prior to 20 weeks’ gestation may not require associated medical care or may not reveal their pregnancy to friends and family prior to the end of the first trimester. Another a priori power analysis using GPower (version 3.1) was conducted for Hypothesis 3 with results suggesting a size of 128 participants was needed to conduct an Independent Samples t-tests with a medium effect size of d=.05 and equal group sizes to achieve a power of .80. [17]. Beyond group differences, we hypothesized that increased GA during loss is positively related to social and healthcare support (Hypothesis 4). A GPower a priori power analysis suggested that a total sample size of 67 was needed to conduct Hypothesis 4’s correlation analysis and achieve a power of .80 and a .3 medium effect size.

Materials and Methods

Participants and recruitment

The study was approved by Idaho State University’s Human Subjects Committee. Participants (76 women who experienced a miscarriage and 73 who experienced a stillbirth) were recruited from Amazon’s Mechanical Turk (MTurk) to access a sociodemographically diverse sample [17,18]. We included women who had experienced pregnancy loss within the last 2 years and received related healthcare. The study excluded women with planned abortions [19]. because literature suggests that they may experience differences in healthcare and social and institutional support/stigma and therefore may exhibit different psychological adjustment.

Participants completed informed consent and responded to openended questions and attention checks throughout the survey to verify participant and response validity. Survey piloting suggested an average completion time of 60 minutes, which was doubled (120 minutes) to account for variability in the study sample. Data from participants who completed the survey in <10 minutes were excluded as this likely reflects random responding [20]. Respondents were permitted to complete the survey once, and universally neutral responses (e.g., variance and standard deviation of data=0) were excluded. Women were offered $1 upon completion of the survey. To facilitate data collection for the stillbirth group, compensation was raised to $2.

Participants were debriefed after each survey via a document including an explanation of the study’s aims and resources for those interested in psychosocial support.

Quantitative analyses

Analyses were conducted using Hayes’ PROCESS macro in SPSS. Statistically significant moderation coefficients were probed via simple slopes analyses using the pick-a-point technique. Simple slopes of the relations between psychological symptoms and social support were examined separately for women who experienced a miscarriage or a stillbirth or at the 16, 50, and 84th percentiles of social support. The proposed relations in Hypothesis 3 were tested via independent samples t-tests. Proposed relations in Hypothesis 4 were analyzed by Spearman's rank order correlations. All data were checked for regression assumptions and assumptions were met after using grand mean centering.

Measures

Variables included sociodemographics, GA during pregnancy loss, perceived healthcare and social support, post-traumatic stress, anxiety, depression, and grief symptoms. Potential covariates included prior psychiatric diagnoses, time since pregnancy loss, multiple pregnancy losses, other children in the home, maternal age at loss, rurality, income, and health insurance. If a woman experienced more than 1 pregnancy loss within the last 2 years, she was asked to answer questions about the loss she felt was most salient within that period.

Sociodemographic questionnaire: The sociodemographic questionnaire included items about age, ethnicity, race, education, annual income, religious/spiritual identity, type of insurance, and state and zip code of residence to determine rural/urban status during miscarriage. Race and ethnicity items were adapted from the U.S. Census Bureau. The insurance coverage question was adapted from the Pregnancy Risk Assessment Monitoring System (PRAMS) questionnaire. Participants’ average age was 31 years old (SD=7 years; see Table 1 for sociodemographic and reproductive history descriptives). Fifty-six percent of women lived in a Health Professional Shortage Area (HPSA), which is an area of shortage in medical or mental health access. Women reported an average total annual household income of $61,461 during their pregnancy loss. Most participants identified as White (87%), with partial college education (42%), and Christian (56%). Women reported an average of 1.57 lifetime pregnancy losses with average GA at 17.86 weeks. Forty-two percent of women reported between 1 to 5 prior and/or current psychiatric diagnoses. Half of women felt they were to blame for the pregnancy loss.

Table 1: Sociodemographic data and reproductive history.

GA of pregnancy loss and reproductive history: GA at pregnancy loss was determined by participants’ retrospective reporting of which week of gestation their pregnancy ended. Reproductive history questions assessed the number of pregnancy losses; parity; gravidity; days since pregnancy loss; number of children living with woman during the pregnancy loss; use of reproductive assistance technology; whether the lost pregnancy was planned; premorbid mental health diagnoses; current pregnancy status; whether the woman had given birth to a child since the most recent loss; type of healthcare organization used; use of online support groups following loss; whether the woman knew what caused/contributed to her pregnancy loss; and whether the woman blamed herself for the loss adapted from PRAMS phase 8 standard questions.

Grief: The Perinatal Bereavement Grief Scale (PBGS) includes 15 items assessing symptoms of yearning and grief related to pregnancy loss and a rating scale examining the past week. The rating scale is coded from 1 to 4 points with 1 meaning rarely/none of the time, 2 some of the time, 3 a moderate amount of time, and 4 most or all of the time.23 The total score ranges from 15 to 60 points, with higher scores indicating a greater degree of yearning and grief. Internal consistency reliability is high within 6 months following pregnancy loss (a=.89). Test-retest reliability is .69 between 2-6 weeks after the loss, .67 between 6-26 weeks, and .48 between 2-26 weeks. In the present sample, the PBGS demonstrated good internal consistency reliability (Cronbach’s a=.88).

Depressive symptoms: To assess depressive symptoms, participants were administered the Center for Epidemiological Studies Depression Scale-Revised (CESD-R)24, a 20-item questionnaire with a Likert scale widely used in perinatal research. Among perinatal samples, pooled sensitivity and specificity estimates for a diagnosis of Major Depressive Disorder with the CESD and CESD-R are between 84-90% and 78-80%, respectively. The Likert scale is 5 points (i.e., rarely or none of the time, some or a little of the time, occasionally or a moderate amount of time, most or all of the time, and nearly every day). It assesses the constructs of depressive symptomology including: sadness, dysphoria, anhedonia, appetite, sleep, thinking/concentration, guilt/worthlessness, tired/fatigue, movement/agitation, and suicidal ideation. Scores are summed across the 20 items with higher scores indicating more depressive symptomatology present. In the present sample, the CESD-R demonstrated excellent internal consistency reliability (Cronbach’s a=.94).

Anxiety symptoms: The Perinatal Anxiety Screening Scale (PASS) is a 31-item self-report questionnaire used to screen for a range of anxiety symptoms. It includes four factors: Acute anxiety and adjustment, general worry and specific fears, perfectionism/control/ trauma, and social anxiety. It has been validated in perinatal samples but not widely used among pregnancy loss samples. Because of this item 2 (fear that harm will come to the baby) was modified to ensure the appropriateness of the full measure for this sample given that it uses language that assumes the successful live birth of the baby. Item 2 with modifications read, “fear of harm to fertility/ability to have a baby,” to generalize the fear to a woman’s ability to conceive and maintain a successful pregnancy. The current study used the total global anxiety score as a continuous measure of post-pregnancy loss anxiety. In the present sample, the PASS demonstrated excellent internal consistency reliability (Cronbach’s a=.96).

Post-traumatic stress symptoms: To assess post-traumatic stress symptoms, the PTSD Checklist for DSM-5 Standard Form (PCL-5) was administered. The PCL-5 is a 20-item self-report checklist corresponding to DSM-5 PTSD symptoms. Items are rated on a scale of 0 to 4 for each symptom, with 0 indicating “not at all” experiencing the symptom, 1 indicating a “little bit,” 2 indicating “moderately,” 3 indicating “quite a bit,” and 4 indicating “extremely." The present project utilized the form without criterion a items or the extended life events checklist, as inclusion criteria required the presence of a traumatic event. In the present sample, the PCL-5 demonstrated excellent internal consistency reliability (Cronbach’s a=.96).

Perceived healthcare support: The short-form Patient Satisfaction Questionnaire (PSQ-18) is an 18-item self-report measure that broadly assesses satisfaction with healthcare services. Items are rated on a fivepoint scale (1=strongly agree, 2=agree, 3=uncertain, 4=disagree, 5=strongly disagree). The technical quality scale was used for primary analyses in this study and the remaining scales were used to describe the sample. No single subscale of the PSQ-18 captures all areas of satisfaction with healthcare, but the Technical Quality scale appeared to touch on more of these themes than other scales (e.g., continuity of care, provider knowledge, thoroughness, and general provider ability to provide care). In the present sample, the PSQ-18 subscales demonstrated adequate to very good internal consistency reliability (Cronbach’s a=.60-.83).

Perceived social support: The Multidimensional Scale of Perceived Social Support (MSPSS) is a 12-item questionnaire that assesses perceived social support from friends, family, and significant others. Items are ratedon a seven-point scale (1=very strongly disagree, 2=strongly disagree, 3=mildly disagree, 4=neutral, 5=mildly agree, 6=strongly agree, and 7=very strongly agree). Mean scores may be calculated for each source of support (i.e., family, friends, or significant other) or an average total scale score may be used for analyses. The present study used the total scale score to capture broad perceived social support. In the present sample, the MSPSS total score demonstrated excellent internal consistency reliability (Cronbach’s a=.91).

Results

Hypothesis 1a-d: GA, social support, and psychological symptoms

The main and interactive effects of GA and perceived social support and the covariate of maternal age explained a statistically significant amount of variance in each outcome across the four regression analyses (Table 2). There were statistically significant positive relationships between GA and depression symptoms, grief, and post-traumatic stress, but not anxiety symptoms. Additionally, there were statistically significant negative relationships between perceived social support and depression symptoms, grief, post-traumatic stress, and anxiety symptoms in the model. There was a statistically significant negative relationship between the interaction term and depression symptoms, grief, posttraumatic stress, and anxiety in the regression. For participants having experienced miscarriage, the negative slope of the relation between perceived social support and depressive symptoms (b=-5.71, t=-4.54, SE=1.26, p<.0001), grief (b=-3.01, t=-3.74, SE=.81, p=.003), post-traumatic stress (b=-6.28, t=-3.79, SE=1.66, p=.0002), and anxiety (b=-6.99, t=-3.86, SE=1.81, p<.0002) significantly differed from 0, but these relationships were not statistically significant for those who experienced a stillbirth.

Table 2: Summary of regression analyses for hypotheses 1a-d.

Finally, maternal age was significantly negatively related to depression symptoms, post-traumatic stress, and anxiety symptoms, but not grief while considering all other variables.

Hypothesis 2a-d: GA, healthcare support, and psychological symptoms

The main effects of GA and perceived healthcare support, the interaction between these variables, and the covariate of maternal age explained a statistically significant amount of the variance in each outcome across the four regression analyses (Table 3). There was a statistically significant positive relationship between GA and grief, but not depressive, post-traumatic stress, or anxiety symptoms. There was a statistically significant negative relationship between perceived healthcare support and depressive symptoms in the model, however no significant relationships were found for grief, post-traumatic stress, or anxiety symptoms. There was not a statistically significant relationship between the interaction term and depression, grief, post-traumatic stress, or anxiety symptoms in the models. Finally, maternal age was significantly negatively related to depression, post-traumatic stress, and anxiety symptoms, though not grief while considering all other model variables.

Table 3: Summary of regression analyses for hypotheses 2a-d.

Hypothesis 3 and 4: GA and social/healthcare support

The Independent samples t-test comparing perceived healthcare support between groups was not statistically significant. In addition, mean differences between perceived social support between groups was not statistically significant. The Spearman’s rank-order correlations between GA and perceived social support and GA and perceived healthcare were not statistically significant.

Discussion

The purpose of this study was to examine relations among GA, perceptions, impact, and mental health outcomes of healthcare and social support following pregnancy loss. Several statistically and practically significant relations were found with important implications for research and clinical practice.

Hypothesis 1

Greater GA was related to more symptoms of depression, grief, and post-traumatic stress consistent with findings in literature. Although research to date suggests a relationship between GA and anxiety, the present study did not find this significant main effect. A modified version of the PASS was utilized, which has not been done in existing literature, and could explain the null relationship as there could be variables in the model that explain a larger variance in anxiety symptoms. Nonetheless, the modified measure demonstrated excellent excellent reliability, and future studies should continue to examine it as a measure of pregnancy-specific anxiety.

In moderation models, more social support during pregnancy loss was related to fewer symptoms of depression, grief, post-traumatic stress, and anxiety. Notably, some of these main effects are also supported by literature examining depression and grief, while the relations between anxiety, post-traumatic stress, and social support is less investigated and understood, necessitating more research in these areas.

Finally, greater maternal age during loss was related to better psychological adjustment outcomes, except when grief was the outcome variable. It is possible that older mothers were more likely to have had successful pregnancies and children in the home, which could have served as a buffer against symptoms of grief. Women of more advanced age may also have greater access to more quality healthcare and other resources than younger mothers.

Hypothesis 2

Greater GA at loss was associated with more grief symptoms, which is supported by research finding relationships between greater GA and more symptoms of depression, post-traumatic stress, and anxiety. This may suggest that GA affects women’s experiences of grief more so than other psychological outcomes. Furthermore, maternal age explained a significant portion of the variance in moderation models with post-traumatic stress and anxiety as outcomes.

Existing literature highlights that grief is more common than other psychological symptoms following pregnancy loss. It is possible that regardless of the type of support provided, many women experience some degree of grief, but this may not occur with the development of other psychological symptoms.

Healthcare support as a buffering variable to the relationship between GA during pregnancy loss and psychological outcomes did not reveal a statistically significant interaction.

The technical quality subscale of the PSQ-1830 may not capture enough variability in women’s perceptions of healthcare. Alternatively, a dosage effect may explain this result. Maternal age was related to depression, anxiety, and post-traumatic stress symptoms even while controlling for other variables in the regression equations. Maternal age may play a significant role in our understanding of psychological adjustment following pregnancy loss and should be considered in future research. Alternatively, these findings may suggest that it is important to consider all the variables in the regression analyses comprehensively when studying psychological adjustment following pregnancy loss.

Regarding the main effect of healthcare support, a statistically significant negative relationship with depression symptoms was found in the moderation model, a finding that is consistent with some literature. However, no significant relations were observed in models with grief, post-traumatic stress, or anxiety as outcome variables. Research regarding the impact of healthcare support on psychological adjustment following pregnancy loss is somewhat mixed.

This suggests that future research should examine the nature and quality of support provided in healthcare settings as they relate to psychological outcomes. Other variables examined in the regression models may also contribute to psychological adjustment besides perceived healthcare support. It appears that aside from grief reactions, maternal age serves as an important variable for consideration when examining psychological outcomes following pregnancy loss.

Hypotheses 3 and 4

Neither Hypothesis 3 nor 4 resulted in statistically significant findings. Limited variability in perceived social and healthcare support, or an overall lack of relationship between GA and support following pregnancy loss are possible explanations for these null findings.

Alternative variables, such as differences in women’s perceived attachment to the lost child, coping styles following the loss, and whether social and healthcare systems’ support match women’s coping needs, may also impact perceptions.

Conclusion

Based on the study’s findings, patients and healthcare providers would benefit from screening for perceptions of support and psychosocial outcomes following pregnancy loss to make recommendations about appropriate interventions, and it appears feasible to do so using current study measures. Healthcare providers should also be aware of women’s mental health histories, as it may relate to psychological adjustment to pregnancy loss. Screening broadly for psychological maladjustment may be most appropriate. For women who miscarry, monitoring anxiety following pregnancy loss appears relevant despite the dominance of depression in literature.

Women endorsed symptoms of grief following loss regardless of maternal age, suggesting clinicians should utilize measures that are able to differentiate between normative grief reactions and clinical distress.

Miscarriage was related to negative psychosocial outcomes highlighting that researchers should investigate socially-based interventions with those experiencing early pregnancy loss. Community-based interventions are recommended to target the patient and broader social context levels. Future studies and practitioners should consider screening for self-blame and social stigma on psychological adjustment, in addition to maternal age, to examine impact.

Strengths and Limitations

The present study design included several strengths that may inform future research on pregnancy loss. The study was novel in its use of MTurk for data collection. Most participants lived in a HPSA and represented 36 unique U.S. states, demonstrating geographic diversity. Future projects should consider offering both online and hard copies of the survey to increase accessibility. In considering women’s contexts across numerous societal levels, healthcare institutions and providers can play an important role in these outcomes and should be studied further to inform assessment and treatment of pregnancy loss.

Study limitations should also be considered when interpreting current findings. First, there was a shift in study procedures, including a title change and increasing compensation to recruit an adequate number of participants experiencing a stillbirth. The title was changed to specifically target women who had experienced stillbirth and the compensation was increased to $2 to further incentivize participation. Future research should consider using more specific study titles and higher compensation when utilizing MTurk with similar samples. Regarding survey content, more specific and comprehensive measures could be used with healthcare support, social support, pregnancy loss, and reproductive history. A more expansive definition of social support may also help to understand which forms of support and members of the social network are most important to identify intervention targets. Understanding unique impacts of pregnancy loss on other family members and dynamics may serve to clarify the importance of these roles.

Additionally, future projects should examine women’s categorization of their pregnancy loss and compare these conventions to the field.

References

1. Athey J, Spielvogel AM (2000) Risk factors and interventions for psychological sequelae in women after miscarriage. Prim Care Update Ob Gyns 7: 64-69.

   [Crossref] [Google Scholar] [PubMed]

2. Heazell AE, Siassakos D, Blencowe H, Burden C, Bhutta ZA, et al. (2016) Stillbirths: Economic and psychosocial consequences. Lancet 387: 604-616.

   [Crossref] [Google Scholar] [PubMed]

3. Quenby S, Gallos ID, Dhillon-Smith RK, Podesek M, Stephenson MD, et al. (2021) Miscarriage matters: the epidemiological, physical, psychological, and economic costs of early pregnancy loss. Lancet 397: 1658-1667.

   [Crossref] [Google Scholar] [PubMed]

4. Bhat A, Byatt N (2016) Infertility and perinatal loss: When the bough breaks. Curr Psychiatry Rep 18: 1-11.

   [Crossref] [Google Scholar] [PubMed]

5. Janssen HJ, Cuisinier MC, Hoogduin KA (1996) A critical review of the concept of pathological grief following pregnancy loss. Omega-J Death Dying 33: 21-42.

   [Crossref] [Google Scholar]

6. Watson MA, Jewell VD, Smith SL (2018) Journey interrupted: A phenomenological exploration of miscarriage. Open J Occup Ther 6.

   [Crossref] [Google Scholar]

7. Lee C, Slade P (1996) Miscarriage as a traumatic event: a review of the literature and new implications for intervention. J Psychosom Res 40: 235-244.

   [Crossref] [Google Scholar] [PubMed]

8. Giannandrea SA, Cerulli C, Anson E, Chaudron LH (2013) Increased risk for postpartum psychiatric disorders among women with past pregnancy loss. J Womens Health (Larchmt) 22: 760-768.

   [Crossref] [Google Scholar] [PubMed]

9. Bronfenbrenner U (1994) Ecological models of human development. Int Encycl Educ 3: 37-43.

   [Google Scholar]

10. Cohen S, Wills TA (1985) Stress, social support, and the buffering hypothesis. Psychol Bull 98: 310.

    [Crossref] [Google Scholar] [PubMed]

11. Swanson KM, Connor S, Jolley SN, Pettinato M, Wang TJ (2007) Contexts and evolution of women's responses to miscarriage during the first year after loss. Res Nurs Health 30: 2-16.

    [Crossref] [Google Scholar] [PubMed]

12. Campbell-Jackson L, Horsch A (2014) The psychological impact of stillbirth on women: A systematic review. Illness, Crisis Loss 22: 237-256.

    [Crossref] [Google Scholar]

13. Cullen S, Coughlan B, Casey B, Power S, Brosnan M, et al. (2017) Exploring parents' experiences of care in an Irish hospital following second-trimester miscarriage. Br J Midwifery 25: 110-115.

    [Crossref] [Google Scholar]

14. Faul F, Erdfelder E, Lang AG, Buchner A (2007) G* Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods 39: 175-191.

    [Crossref] [Google Scholar] [PubMed]

15. Buhrmester MD, Talaifar S, Gosling SD (2018) An evaluation of Amazon’s Mechanical Turk, its rapid rise, and its effective use. Perspect Psychol Sci 13: 149-154.

    [Crossref] [Google Scholar] [PubMed]

16. Mason W, Suri S (2012) Conducting behavioral research on Amazon’s Mechanical Turk. Behav Res Methods 44: 1-23.

    [Crossref] [Google Scholar] [PubMed]

17. Nikcevic AV, Tunkel SA, Nicolaides KH (1998) Psychological outcomes following missed abortions and provision of follow�up care. Ultrasound Obstet Gynecol 11: 123-128.

    [Crossref] [Google Scholar] [PubMed]

18. Bardos J, Hercz D, Friedenthal J, Missmer SA, Williams Z (2015) A national survey on public perceptions of miscarriage. Obstet Gynecol 125: 1313-1320.

    [Crossref] [Google Scholar] [PubMed]

19. Preacher KJ, Curran PJ, Bauer DJ (2006) Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. J Educ Behav Stat 31: 437-448.

    [Crossref] [Google Scholar]

20. Sharp LK, Lipsky MS (2002) Screening for depression across the lifespan: a review of measures for use in primary care settings. Am Fam Physician 66: 1001-1009.

    [Google Scholar] [PubMed]