KARINA M.

SHREFFLER Oklahoma State University

ARTHUR L. GREIL Alfred University*

JULIA MCQUILLAN University of Nebraska – Lincoln**

Pregnancy Loss and Distress Among

U.S. Women

Although pregnancy loss—especially miscar- Approximately 14% of all clinically recognized

riage—is a relatively common experience among pregnancies in the United States result in
reproductive-aged women, much of our under- miscarriage, defined as a loss during the
standing about the experience has come from first 20 weeks of pregnancy, another 0.5%
small clinic-based or other nonrepresentative result in stillbirth, a loss after the 20th
samples. We compared fertility-specific distress week (Saraiya, Berg, Shulman, Green, &
among a national sample of 1,284 women who Atrash, 1999). Pregnancy loss is often a
have ever experienced a stillbirth or miscar- devastating experience for parents. Research
riage. We found that commitment/attachment to suggests that women experience a variety
pregnancy that ended in loss as well as current of psychological distress outcomes following
childbearing contexts and attitudes were asso- miscarriage, including grief, anxiety, depression,
ciated with distress following pregnancy loss. and guilt (Lok, Yip, Lee, Sahota, & Chung, 2010;
Practitioners working with women or couples Thapar & Thapar, 1992), and that—although
who have experienced pregnancy loss should effects usually diminish within 6 months (Brier,
be aware of the importance of characteristics 2008)—these outcomes are often sustained
associated with higher distress, such as whether over years (Janssen, Cuisinier, & Hoogduin,
the pregnancy had been planned, recency of the 1996; Stinson, Lasker, Lohmann, & Toedter,
loss, no subsequent live births, having a medical 1992). Less is known about stillbirths than
explanation for the loss, a history of infertility, early pregnancy loss; however, research on
current childbearing desires, importance of stillbirth indicates that bereaved mothers, as
motherhood, and locus of control over fertility. a group, manifest significantly higher rates of
psychological distress than mothers of living
infants for at least 30 months after their loss
(Boyle, Vance, Najman, & Thearle, 1996). There
is some evidence that the adverse effects of
Oklahoma State University, 700 N. Greenwood Ave., Tulsa, stillbirth on mental health persist throughout the
OK 74106 ([email protected]). life course (Bernazzani & Bifulco, 2003).
*Alfred University, 1 Saxon Drive, Alfred, NY 14802. Although pregnancy loss, especially mis-
**University of Nebraska-Lincoln, 706 Oldfather Hall, carriage, is a relatively common experience
Lincoln, NE 68588. among reproductive-aged women, much of our
Key Words: attachment, commitment, distress, fertility, understanding about the experience of preg-
miscarriage, pregnancy loss, stillbirth. nancy loss has come from small clinic-based or
342 Family Relations 60 (July 2011): 342 – 355
DOI:10.1111/j.1741-3729.2011.00647.x
Pregnancy Loss and Distress 343

other nonrepresentative samples. Furthermore, to conceive, fertility history of the couple,

the various factors (individual, familial, eco- previous fertility help-seeking behaviors, age
nomic, medical, and cultural) that affect the of the mother, number of previous pregnancy
psychological response to pregnancy loss are not losses, number of living children, relationship
well understood (Bennett, Litz, Lee, & Maguen, quality between the parents of the child, and
2005), possibly limiting the effectiveness that outside influence and expectations about having
family professionals could have with regard a child. The factors in this list have not
to reducing psychological distress. We there- been evaluated through comprehensive studies.
fore advance understanding of the responses Research focusing on the importance of these
to pregnancy loss by examining factors asso- factors individually, however, has determined
ciated with distress using a population-based a number of pregnancy loss and fertility
sample of women who had stillbirths or mis- history characteristics that are associated with
carriages or both. Making use of the concepts distress. Our goal is to advance understanding of
of commitment and attachment, we endeav- differences in reactions to pregnancy loss among
ored to enhance understanding of the effect a probability-based sample of American women.
of pregnancy loss on fertility-specific distress Variations in distress following pregnancy
(FSD) using data from the National Survey of loss are often attributed to variations in levels of
Fertility Barriers (NSFB), a probability-based commitment (Lydon, Dunkel-Schetter, Cohan,
study of 4,796 American women of reproduc- & Pierce, 1996) and attachment (Robinson,
tive age. We assessed differences in distress Baker, & Nackerud, 1999) to pregnancy. The
by pregnancy commitment and attachment (e.g., concepts of commitment and attachment have
gestation length and whether the pregnancy had both been employed to describe the process
been planned); experiences since the loss (e.g., of identification with the motherhood role
recency of loss, having a medical explanation during pregnancy (Kemp & Page, 1987; Lydon
for the loss, giving birth after the loss, and et al.; Muller, 1992). Studies on commitment
having experienced multiple losses); current fer- use the concept to refer to an investment
tility context (e.g., infertility history, currently in a particular line of action (Becker, 1960;
wanting a baby, importance of motherhood, and Hirschi, 1969) and to a tendency to feel
pregnancy locus of control); and background psychologically attached to a role, a relationship,
characteristics (e.g., education level, age, rela- or an organization (Kanter, 1972; Lydon et al.;
tionship status, and race/ethnicity). Stryker, 1968). Pioneering work on commitment
in the 1960s and 1970s established support for
the fundamental tenet of commitment theory:
INVESTMENT IN PREGNANCY AND THE Greater investment in a line of action or a
RELEVANCE OF PREGNANCY LOSS relationship is associated with greater likelihood
There is a good deal of variability in how of continuing in that line of action or working
women respond to pregnancy loss. Evidence to maintain that relationship (Becker; Hirschi;
from research on pregnancy loss indicates that Kanter). A corollary, therefore, is that greater
anywhere from 20 to 55% of women report investment in a line of action or relationship will
elevated levels of depressive symptoms in the be associated with greater distress if that line of
months immediately following pregnancy loss action or relationship is lost.
(Janssen et al., 1996; Lok & Neugebauer, 2007). Attachment has been defined as a ‘‘relatively
A recent study (Lok et al., 2010) revealed that enduring emotional tie to a specific other
over a quarter of women scored high enough person’’ (Maccoby, 1980, p. 53). The concept
on the Beck Depression Inventory immediately of attachment calls to mind Bowlby’s (1969)
after miscarriage to be rated as probably having attachment theory, but those who study the role
a depressive disorder. Women with higher initial of attachment in reactions to pregnancy loss
depressive scores continued to report depressive are not necessarily making use of Bowlby’s
symptoms over the course of a year. attachment theory. Bowlby created attachment
Bennett et al. (2005) provided a long list of theory as an alternative to psychoanalytic
prebirth events or attitudes that are likely to theories of object relations and intended it as
be associated with distress following pregnancy an explanation of the development of separation
loss, including investment in and meaning of anxiety in children (Bretherton, 1985). Those
the pregnancy, time and energy spent trying who employ the concept in research on perinatal
344 Family Relations

loss are not trying to account for the development simply a matter of the length of gestation.
of specific attachments or of attachment styles. Although pregnancy is obviously a biological
Rather, they are asserting that women with state, it is also a socially constructed reality.
stronger attachments to their pregnancies should Women assign meanings to their pregnancies as
experience greater distress from pregnancy loss intended or unintended, welcome or unwelcome,
(e.g., Peppers & Knapp, 1980). life-changing or relatively routine, and so forth.
Thus, there is much overlap between the Looking at pregnancy loss from the perspectives
concepts of commitment and attachment as they of attachment and commitment suggests that
are employed in the pregnancy loss literature. women will become more attached to children
Researchers have used both terms to suggest who are more ‘‘real’’ to them (Klier, Geller,
that the stronger psychological investment in & Ritscher, 2002; Muller, 1992). If commitment
pregnancy should lead to higher distress from a and attachment depend on the meaning attributed
pregnancy loss. Furthermore, researchers have to a pregnancy, then a pregnancy loss occurring
used both terms to make similar assertions for women who had been trying to get pregnant
about the factors related to variations in the should be more distressing than a loss for women
experience of pregnancy loss (e.g., Lydon et al., who were not trying or expecting to get pregnant.
1996; Peppers & Knapp, 1980; Robinson et al.,
1999). In the discussion that follows, we employ
the specific term that researchers have used, VARIATIONS IN DISTRESS AND CURRENT
understanding that both terms are employed in FERTILITY CONTEXT
an effort to show how investment in a pregnancy There is much evidence to suggest that distress
is related to the experience of pregnancy loss. resulting from pregnancy loss decreases over
time. Thus, we expect to find that women whose
losses occurred in the more recent past will report
VARIATIONS IN DISTRESS BY PREGNANCY
higher levels of distress (Janssen et al., 1996;
ATTACHMENT AND COMMITMENT Lok & Neugebauer, 2007; Lok et al., 2010).
The process of investment in a relationship A recent longitudinal study on miscarriage and
with an infant starts well before birth. Peppers psychological distress (Lok et al.) found that
and Knapp (1980, p. 59) described nine events initial elevated distress declined steadily over a
that contribute to mother-infant attachment: year, so that at the end of the year, differences
‘‘(a) planning the pregnancy; (b) confirming in distress between the women who miscarried
the pregnancy; (c) accepting the pregnancy; and the women in the comparison group were
(d) feeling fetal movement; (e) accepting the no longer significant.
fetus as an individual; (f) giving birth; (g) seeing Pregnancy loss often occurs in the absence
the baby; (h) touching the baby; and (i) giving of an obvious explanation. Women who have
care to the baby.’’ They point out that five medical explanations for miscarriage have less
of these events occur prenatally. The concept difficulty coping than women who do not
of commitment as investment suggests that the have medical explanations (Simmons, Singh,
more emotional and physical effort one spends Maconochie, Doyle, & Green, 2006). Lack
attempting to achieve a goal, the more committed of information about the reason for the loss
to that goal one becomes (Kanter, 1972). Thus, contributes to increased anxiety in subsequent
we expected that the longer the duration of a pregnancies (Theut, Pederson, Zaslow, &
pregnancy, the greater the attachment to the Rabinovich, 1988), although research has not
pregnancy and, hence, the greater the distress demonstrated that knowing the specific reason
experienced after a pregnancy loss (Robinson for a pregnancy loss reduces distress levels.
et al., 1999). Indeed, Lydon et al. (1996) We expected that knowing the reason for a
reported that expectant mothers’ commitment pregnancy loss would be associated with lower
to a pregnancy progresses as the pregnancy distress.
proceeds. Pregnancy histories are also relevant consid-
Variations in distress following a pregnancy erations for distress following pregnancy loss.
loss are not fully explained by gestation (length) Women without children in the household have
at time of loss, however. Although commitment higher personal significance associated with the
to a pregnancy may be greater later in the pregnancy (Swanson, 2000) and increased dis-
pregnancy, it is unlikely that commitment is tress when they lose a child compared with
Pregnancy Loss and Distress 345

women who already have a child (Janssen, suggests they should be relevant, as both
Cuisinier, & de Brauw, 1997; Schwerdtfeger pregnancy loss and infertility are barriers to
& Shreffler, 2009; Thapar & Thapar, 1992). childbearing for women who want to have
Motherhood does not alleviate the distress from children. In a study on infertile couples, Abbey,
a loss, however, even after the subsequent birth Andrews, and Halman (1992) reported that
of a healthy baby. Approximately one third of greater importance of children is associated with
mothers who experienced pregnancy loss prior greater distress. Miles, Keitel, Jackson, Harris,
to a live birth continue to report symptoms that and Licciardi (2009) found that infertile women
place them at a high risk for depression (Arm- who report greater pressure to become mothers
strong, 2007). Still, we expected that distress score higher on a general measure of distress.
would be lower for women who have had a live These findings suggest that women who view
birth following their (most recent) loss. motherhood as more important should report
Approximately 1 – 2% of reproductive-aged higher distress following a pregnancy loss.
women experience recurrent pregnancy loss Current pregnancy intentions should also be
(Kutteh, 2005), which typically refers to three relevant for distress following a loss. In a study
or more consecutive pregnancy losses. Women comparing infertile women with and without
who have experienced prior losses attach pregnancy intent, Greil, McQuillan, Johnson,
more significance to a miscarriage (Swanson, Blevins-Slauson, and Shreffler (2010) found that
2000), and research indicates that recurrent infertility is only distressing for women who
pregnancy loss is associated with significant report that they currently want to have a baby.
psychological distress (Adeyemi, 2008; Magee, Thus, we expected that pregnancy loss would be
2003). Therefore, we expected that women who more distressing for women who currently want
have experienced more than one loss would to have a baby.
report higher FSD. Another consideration is how confident
If investment of time and energy increases women are that they can and will get pregnant
commitment and attachment to pregnancy, then and have a baby when the time is right—that
women who have difficulty conceiving or car- is, a childbearing locus of control. Infertility
rying a pregnancy to term should experience research has shown that having lower locus of
pregnancy loss as more distressing than women control is associated with greater likelihood of
who had no challenges conceiving or carrying a seeking medical help to become pregnant (Greil
pregnancy. Women experiencing high-risk preg- & McQuillan, 2004), suggesting that women
nancies need to invest more time and energy who feel more confident in their ability to get
in their pregnancy and report higher levels pregnant or carry a pregnancy to term should
of attachment to their pregnancies (Mercer & feel less distress.
Ferketich, 1990; Stainton, McNeil, & Harvey, The vast majority of studies that have looked
1992). Previous studies have shown that invol- at the influence of demographic characteristics
untary childlessness is associated with higher (e.g., age, socioeconomic status, race/ethnicity,
distress than voluntary childlessness or infertility and union status) on the relationship between
(Janssen et al., 1997; McQuillan, Greil, White, pregnancy loss and psychological distress have
& Jacob, 2003; Schwerdtfeger & Shreffler, 2009; not reported significant findings (Brier, 2008;
Toedter, Lasker, & Alhadeff, 1988), although the Klier et al., 2002; Lok & Neugebauer, 2007).
association between difficulty conceiving and The major exception to this generalization has
psychological distress among women who have to do with parity. As noted above, pregnancy
had a pregnancy loss has not been explored. loss appears to be significantly more distressing
We expected to find that infertility would be among women with no children. This is
related to higher distress among women who consistent with what one would expect on the
have experienced a pregnancy loss. basis of findings concerning the relationship
In addition, women who are more eager for between the meaning of a pregnancy and
motherhood should be more committed to their the distress resulting from pregnancy loss
pregnancy and report more distress than women (Klier et al., 2002; Muller, 1992). It must
who are less eager for motherhood. Importance be remembered, however, that most studies
of motherhood and current fertility intentions have used relatively small numbers of women
have not been assessed in the pregnancy loss who self-select into studies. An appropriate
literature, but evidence from infertility research evaluation of demographic characteristics and
346 Family Relations

distress following pregnancy loss requires a the pregnancy had been planned were used
representative sample. Despite the limits of to create four groups. We compared women
existing research, some studies have found an who have experienced miscarriages that were
association between demographic variables and unplanned pregnancies with women who expe-
distress. Toedter et al. (1988) reported that lower rienced miscarriages that were planned, women
socioeconomic status is associated with higher who experienced stillbirths that were unplanned,
levels of distress. We expected older women and women who experienced stillbirths of
to have more concern about pregnancy loss planned pregnancies to assess the hypothesis
because of fertility age limits, but we are aware that losses from longer pregnancies (stillbirths)
of only one study that found this association would be associated with higher distress than
(Janssen et al., 1997). Because demographic losses from shorter pregnancies (miscarriages)
factors may influence the incidence of pregnancy and that losses of pregnancies that had been
loss experiences as well as fertility contexts, it is planned are more distressing than losses of preg-
important to control for these variables to avoid nancies that were not planned. Third, we used
spurious findings. a fertility-specific measure of distress to capture
variations in the experience of pregnancy loss.
Fourth, we included numerous measures rele-
STATEMENT OF THE PROBLEM
vant to the meaning of the pregnancy and loss
Most studies using measures specifically related to women: time since the most recent loss, birth
to distress associated with pregnancy loss, such since most recent loss, knowing the reason for a
as the Perinatal Grief Scale (Toedter et al., loss, multiple losses, self-identification of a fer-
1988) and the Perinatal Bereavement Grief Scale tility problem, currently wanting a baby, impor-
(Ritsher, 2002) have found that those who had tance of motherhood, and pregnancy locus of
shorter pregnancies have lower grief scores than control. Fifth, we controlled for many variables
those who had longer pregnancies (Franche, that might explain why some women experience
2001; Janssen et al., 1997; Lasker & Toedter, more distress from pregnancy loss than other
2000). Studies that have employed more general women, including education, race/ethnicity, age,
measures of psychological functioning—such and union status.
as depression—as the dependent variable, These concepts and associations are depicted
however, have failed to find a conclusive in the conceptual map provided in Figure 1. This
link between length of gestation and distress figure shows that demographic variables lay a
levels (Klier, Geller, & Neugebauer, 2000; foundation for experiencing pregnancy loss and
Neugebauer et al., 1992), although women that current pregnancy context should mediate at
with late losses exhibit more symptoms of least some of the association between pregnancy
depression than women with earlier losses commitment and attachment and FSD. Finally,
(Neugebauer et al.). This suggests that measures guided by insights from the commitment and
specifically related to fertility may be more attachment perspectives, we used multiple
sensitive to variation in distress following regression to assess the associations between
pregnancy loss than more general measures. FSD by commitment to the lost pregnancy,
Infertility researchers have made a similar controlling for measures of the experiences since
argument for studying psychological distress the pregnancy loss, current fertility context, and
related to infertility (Greil, Slauson-Blevins, demographic variables.
& McQuillan, 2010; Schmidt, 2009). We,
therefore, utilized a measure of FSD to provide
a more sensitive measure of emotional reactivity METHOD
to pregnancy loss than a general measure of
Sample
psychological distress.
We provide several contributions to research The sample for this study comes from the NSFB,
on the associations between pregnancy loss and a random-digit-dial nationally representative
distress. First, we used a nationally represen- data set of 4,796 women of childbearing
tative sample of women of childbearing age age (25 – 45) and a subset of their partners
who have experienced pregnancy loss. Sec- which includes oversamples of women with
ond, we examined the importance of pregnancy fertility problems and census tracts with minority
commitment; length of gestation and whether (African American and Hispanic) populations
Pregnancy Loss and Distress 347

FIGURE 1. CONCEPTUAL MODEL OF COMMITMENT/ATTACHMENT TO THE PREGNANCY ENDING IN LOSS AND

FERTILITY-SPECIFIC DISTRESS.

Demographic Characteristics:
Education
Age
In a union
Race/Ethnicity

Commitment/Attachment to the Fertility

Pregnancy Ending in Loss Specific
Planned/Unplanned Distress
Stillbirth/Miscarriage

Current Fertility Context

Experiences Since Loss
Years since loss
Birth since loss
Medical Explanation for loss
Multiple losses
Pregnancy-Related Attitudes
Perceives a Fertility Problem
Wants a Baby
Importance of Motherhood
Pregnancy Locus of Control

greater than 40%. This current sample of scale from 0 (no distress) to 1 (high distress),
women who have experienced a pregnancy loss then logged to reduce skew. Cronbach’s α for
(N = 1,284) included women who experienced the FSD scale is .80 for women who have
miscarriage(s) only (n = 1,152) and women experienced a pregnancy loss, and the logged
who have had at least one stillbirth (n = 132). scale ranges from 0 to .69.
Data were weighted so that the sample is
representative of the population. Commitment/attachment to the pregnancy loss.
For the descriptive analyses, type of pregnancy
loss is measured by an indicator variable for
Measures ever stillbirth, with miscarriages only as the
Fertility-specific distress. Although there are a reference category. Respondents were classi-
number ways to measure distress resulting from fied in the miscarriage group if they had ever
pregnancy loss, the NSFB includes a broad had at least one miscarriage but no stillbirths.
measure of FSD that could be applied to a The respondents in the stillbirth group had
wide range of fertility barriers. Respondents who had at least one stillbirth, but many also had
reported any type of fertility problem (such as experienced miscarriages as well. Women self-
pregnancy loss or infertility) were asked a series identified their type of loss, as our data do
of questions regarding whether they experienced not include the exact gestation at which the
certain reactions to their fertility problem(s), loss occurred. For the regression analysis, we
including pregnancy loss. These include (a) ‘‘I operationalized commitment or attachment as
felt cheated by life,’’ (b) ‘‘I felt that I was being including both gestation length and whether the
punished,’’ (c) ‘‘I felt angry at God,’’ (d) ‘‘I felt lost pregnancy had been planned. Respondents
inadequate,’’ (e) ‘‘I felt seriously depressed,’’ were asked ‘‘When you got pregnant, were you
and (f) ‘‘I felt like a failure as a woman.’’ These trying to get pregnant, trying not to get preg-
items are dichotomous (1 = yes; 0 = no). The nant, or were you okay either way?’’ about
mean of available items were used to create a each pregnancy. Women who reported that they
348 Family Relations

were ‘‘trying to’’ get pregnant for the preg- constructed by combining responses to four
nancy that resulted in a loss were coded 1; questions. Four items are measured on Likert
other responses were coded 0. We coded respon- scales (strongly disagree to strongly agree): (a)
dents into four groups using these two concepts: ‘‘Having children is important to my feeling
miscarriage/unplanned refers to women whose complete as a woman,’’ (b) ‘‘I always thought
miscarriage had been an unplanned pregnancy; I would be a parent,’’ (c) ‘‘I think my life will
miscarriage/planned includes women who had be or is more fulfilling with children,’’ and (d)
planned their pregnancy that resulted in mis- ‘‘It is important for me to have children.’’ The
carriage; stillbirth/unplanned refers to women Cronbach’s α is .80 for the current sample, and
who had a stillbirth of an unplanned pregnancy; the mean of available items were used to create
and stillbirth/planned includes women who had a scale ranging from 1 to 4. Pregnancy locus of
planned the pregnancy that ended in stillbirth. control was measured by agreement to two items
Note that many of the respondents had more than (strongly disagree to strongly agree: (a) ‘‘I think
one loss; these groups were created based on the (or thought for those not currently intending to
gestation length and planning data for the most get pregnant) I would get pregnant when the time
recent pregnancy that had been lost if there were was right’’ and (b) ‘‘I think (or thought) if it’s
multiple losses. In effect, these variables are God’s will, I would get pregnant.’’ The mean of
interaction terms that combine type of loss and available items was used to create a scale ranging
attitude toward pregnancy at the time of the loss. from 1 to 4 with an α of .76 for this sample.
We used the four-variable approach because it
simplifies interpretation of the associations. Background variables. Education (in years) is a
continuous variable, ranging from 2 to 22 in our
sample. Age is a continuous variable and ranges
Current Fertility Context from 25 to 45 in our sample. In a union is a
Experiences since loss. Years since loss refers dichotomous variable, with 1 indicating that the
to the length of time (in years) since the (most respondent is currently married or cohabiting.
recent) loss. The variable was mean centered for Race/ethnicity is included as dummy variables
regression analyses. Respondents were coded as for Black, Hispanic, and ‘‘Other race,’’ with
having a birth since loss if the year of their White respondents as the reference category.
most recent live birth was more recent than
the year of their (last) pregnancy loss. Medical
explanation is a dichotomous variable indicating Analytic Strategy
that the respondent had medical evaluation Descriptive analyses estimated differences by
following the pregnancy loss that resulted type of pregnancy loss (miscarriage(s) only
in an explanation for the loss. Respondents and ever stillbirth). For continuous variables,
were coded as 1 if they received a medical means and standard deviations were provided;
explanation; respondents who did not have an t tests were conducted to determine the sig-
evaluation or received an evaluation but given nificance of differences between means. For
the diagnosis of ‘‘unexplained’’ were coded categorical variables, differences in propor-
as 0. Multiple loss is a dichotomous variable tion tests were used to provide indication
indicating whether the respondent experienced of differences between groups. Ordinary least
more than one loss. squares regression (OLS) models the associa-
tions between the independent variables, includ-
Pregnancy-related attitudes. Perceives a fertil- ing commitment/attachment (length of gestation
ity problem is a dichotomous variable indicating and planning), pregnancy loss experiences, cur-
that the respondent identifies herself as ‘‘some- rent fertility contexts, and FSD. Continuous
one who has or has had fertility problems’’ or variables in the OLS models were mean centered
thinks of herself as ‘‘someone who has, has for the analyses.
had, or might have trouble getting pregnant.’’
Wants to have a baby is a continuous variable
indicating the extent to which the respondent RESULTS
‘‘would like to have a(nother) baby,’’ and Table 1 shows the minimum, maximum, mean,
responses range from 1 (definitely no) to 4 and standard deviation for each variable by
(definitely yes). Importance of motherhood was pregnancy loss outcome (miscarriage(s) only or
Pregnancy Loss and Distress 349

ever stillbirth). t tests for the difference in means had a miscarriage than a stillbirth. Some of
for the continuous variables and difference in these findings (e.g., trying to conceive, having
proportions tests for the categorical variables multiple losses, and knowing the cause of
showed that women in these two groups differ the problem) suggest that women who have
on many characteristics. The mean logged FSD had stillbirths should have higher distress than
was higher for women who have had stillbirths women who have had miscarriages. Other
than for women who have had miscarriages, but findings (e.g., having had a birth since the loss
the range of values and standard deviation were and lower importance of motherhood scores)
very similar. Women who have had stillbirths suggest that women who have had stillbirths
were also more likely to have planned their should have lower distress. There are several
pregnancy. In addition, more have had a birth characteristics that do not differ by pregnancy
since the loss, more know the problem that outcome status. Although the maximum value
caused the pregnancy outcome, and more have for time since the focal pregnancy was higher for
had multiple pregnancy losses. Women who women who had a miscarriage than women who
have had miscarriages, however, were more had a stillbirth, the mean time since the focal
likely to report a fertility problem than women pregnancy did not differ by pregnancy outcome.
who have had stillbirths. Women who have had Additionally, mean strength of desire for a child,
miscarriages also had slightly but significantly mean pregnancy locus of control, mean age, and
higher importance of motherhood scores and mean education were not statistically different
were more likely to be in a relationship. Among between the two groups.
Hispanic women, a higher proportion reported a The descriptive statistics revealed that there
stillbirth than a miscarriage. Among women in is considerable variation among the women who
the ‘‘other’’ race category, a higher proportion have had pregnancy losses. For example, the

Table 1. Fertility-Specific Distress, Pregnancy Loss Experiences, Current Fertility Contexts and Attitudes, and Demographic
Characteristics by Type of Pregnancy Loss (N = 1,284)

Miscarriage(s) Only (n = 1,152) Ever Stillbirth (n = 132)

Variables Minimum Maximum Mean SD Minimum Maximum Mean SD

Fertility-specific distress, logged 0.00 0.69 0.17 0.23 0.00 0.69 0.22 0.26 ∗

Pregnancy loss experiences

Years since loss 0.01 29.08 9.99 6.74 0.05 25.12 11.13 6.55
Planned pregnancy 0.00 1.00 0.38 0.49 0.00 1.00 0.48 0.50 ∗∗

Had a birth since the loss 0.00 1.00 0.27 0.44 0.00 1.00 0.33 0.47 ∗∗

Know the problem 0.00 1.00 0.06 0.24 0.00 1.00 0.12 0.33 ∗∗∗
∗∗∗
Multiple losses 0.00 1.00 0.30 0.46 0.00 1.00 0.44 0.50
Current fertility contexts and attitudes
Perceives a fertility problem 0.00 1.00 0.35 0.48 0.00 1.00 0.27 0.45 ∗∗∗

Wants a baby 1.00 4.00 2.14 1.20 1.00 4.00 2.07 1.21
Importance of motherhood 1.00 4.00 3.37 0.58 1.67 4.00 3.35 0.64 ∗

Pregnancy locus of control 1.00 4.00 3.10 0.63 1.00 4.00 3.13 0.49
Demographic characteristics
Education 2.00 22.00 13.46 2.78 2.00 22.00 12.55 2.57
Age 25.00 45.00 36.02 5.80 25.00 45.00 36.53 5.85
∗∗
In a union 0.00 1.00 0.79 0.41 0.00 1.00 0.72 0.45
White 0.00 1.00 0.64 0.48 0.00 1.00 0.60 0.49
Black 0.00 1.00 0.16 0.36 0.00 1.00 0.15 0.36
Hispanic 0.00 1.00 0.16 0.37 0.00 1.00 0.23 0.42 ∗∗

Other race 0.00 1.00 0.05 0.22 0.00 1.00 0.02 0.14 ∗∗

Note: t tests were conducted for continuous variables and difference in proportion tests for categorical variables. National
Survey of Fertility Barriers, women ages 25 – 45.
∗
p < .05. ∗∗ p < .01. ∗∗∗ p < .001.
350 Family Relations

standard deviation for FSD was larger than the characteristics were associated with FSD.
mean. Fewer than half of the women were We assessed whether the characteristics that
trying to conceive the pregnancy that they differ among the four groups of women—un-
lost. About a third have had a child since planned/miscarriage, planned/miscarriage, un-
the loss, although very few knew the cause planned/stillbirth, and planned/stillbirth—
of the pregnancy loss. Although all the women helped explain the differences in FSD or whether
have had a problem with a pregnancy (i.e., at there was a unique effect related to pregnancy
least one pregnancy loss), only about a third commitment after other factors were controlled.
saw themselves as having a fertility problem.
Overall, the women had high average scores on
the ‘‘importance of motherhood’’ scale (over 3 Pregnancy Loss and Fertility-Specific Distress
on a scale from 1 to 4), but scores that covered The results for the multiple regression analyses
the full range from very low to very high. of FSD are displayed in Table 2. Model 1
These descriptive statistics, from a population- examined demographic characteristics and FSD.
based sample, showed that pregnancy loss occurs Because the dependent variable is logged, the
across a broad spectrum of women. This suggests coefficients can be interpreted as percentages. Of
that the subjective experience of pregnancy loss the demographic characteristics, only education
is likely to vary among women as well. was associated with FSD. Each year increase in
In the multivariate analysis, we examined education was associated with 1% lower FSD.
which characteristics of pregnancy commit- Model 2 examined how commitment/attach-
ment/attachment, experiences since loss, atti- ment to the lost pregnancy was associated with
tudes, childbearing desires, and background FSD. Whether the pregnancy was planned was

Table 2. Ordinary Least Squares Regression Analysis of Variables Predicting Fertility-Specific Distress (N = 1,284)

Model 1 Model 2 Model 3

Variables b SE Beta b SE Beta b SE Beta
∗∗∗ ∗∗∗ ∗∗
Constant .17 .02 .13 .02 .06 .02
Demographic characteristics
Education −.01 .00 −.07 ∗ −.01 .00 −.08 ∗∗ −.01 .00 −.12 ∗∗∗

Age .00 .00 −.04 .00 .00 −.06 .001 .00 .02
In a union .01 .02 .02 .00 .02 .00 .001 .02 .00
Black (White) −.03 .02 −.05 −.02 .02 −.03 −.001 .02 .00
Hispanic .01 .02 .01 .01 .02 .02 .03 .02 .04
Other race .01 .03 .01 .02 .03 .02 .03 .03 .03
Commitment to lost pregnancy (unplanned/miscarriage)
Planned/miscarriage .10 .01 .21 ∗∗∗ .05 .01 .11 ∗∗∗

Unplanned/stillbirth .05 .03 .05 .05 .03 .04

∗∗∗ ∗∗
Planned/stillbirth .14 .03 .13 .10 .03 .09
Experiences since loss
Years since loss −.004 .00 −.11 ∗∗
∗
Birth since loss −.04 .02 −.08
Medical explanation .11 .03 .12 ∗∗∗

Multiple losses .02 .01 .04

Current pregnancy-related attitudes
Perceives a fertility problem .10 .01 .21 ∗∗∗

Wants a baby .02 .01 .11 ∗∗

Importance of motherhood .04 .01 .09 ∗∗

Pregnancy locus of control −.04 .01 −.11 ∗∗∗

Adjusted R 2 .01 .05 .15

Note: Reference categories are in parentheses.
∗
p < .05. ∗∗ p < .01. ∗∗∗ p < .001.
Pregnancy Loss and Distress 351

linked to significant differences in FSD; women (Model 3). For all the four groups, FSD was par-
who had a miscarriage of a planned pregnancy tially explained by experiences since the loss and
has almost 10% higher FSD than women who the current pregnancy-related context, indicated
had a miscarriage of an unplanned pregnancy by lower levels of FSD in Model 3 compared
(b = .10, p<.001). Experiencing a stillbirth of with Model 2. Perceiving a fertility problem,
a planned pregnancy was associated with 14% wanting a child, and importance of motherhood
higher FSD than women who had a miscarriage were all associated with higher FSD, whereas
of an unplanned pregnancy (b = .14, p < .001). higher pregnancy locus of control was associated
Women who experienced a stillbirth of an with lower FSD.
unplanned pregnancy reported 5% higher FSD
than women who experienced a miscarriage
of an unplanned pregnancy, but the difference CONCLUSIONS
was not significant. The commitment/attachment
variables explain 4% of the variance in FSD. We studied variations in the effects of preg-
Adding women’s current fertility contexts, nancy loss on distress among women who have
including experiences since loss and pregnancy- had a pregnancy loss from a random sample
related attitudes, explained an additional 10% of American women of reproductive age. The
of the variance in FSD (Model 3). Having large NSFB sample provided an unprecedented
losses that occurred further in the past and hav- opportunity to examine pregnancy attachment
ing a subsequent birth were associated with and commitment, operationalized by gestation
lower distress. Women who know the reason length (late term pregnancy loss vs. early term
for the pregnancy loss had higher distress than pregnancy loss) and whether the lost pregnancy
women in the comparison groups. Most of these had been planned as well as numerous other
measures remained associated with FSD in sub- characteristics related to the pregnancy loss,
sequent models; only multiple losses ceased to perceptions and attitudes regarding childbearing,
be significant. Figure 2 shows the predicted log current intentions, and demographic characteris-
FSD by attachment/commitment category for the tics. Our analysis of FSD provides evidence that
model controlling for demographic characteris- pregnancy-relevant commitment and attachment
tics (Model 2) and also adds experiences since measures are associated with greater FSD. In
the loss and current pregnancy-related indicators particular, whether the pregnancy that resulted

FIGURE 2. PREDICTED FERTILITY-SPECIFIC DISTRESS LEVEL BY PREGNANCY COMMITMENT/ATTACHMENT CATEGORY

CONTROLLING FOR DEMOGRAPHIC CHARACTERISTICS (MODEL 2) AND CURRENT FERTILITY CONTEXT (MODEL 3) FROM THE
REGRESSION MODELS.
352 Family Relations

in loss had been planned appears to be particu- distressed report less control or whether feeling
larly salient. Women who experience losses of less in control of pregnancies increases distress.
planned pregnancies report greater FSD than Second, we would like to know the trajectory
women who lost pregnancies that were not of distress before, during, and following the
planned. Women who know the cause of their pregnancy loss, but we only have measures
pregnancy loss, women with self-identified fer- at the time of the interview. For example,
tility problems, women who currently desire a our findings show that distress decreases with
baby, and women who place a higher value on time following a loss, but it is unclear if the
motherhood are more distressed than women decline is linear. Finally, although we believe
not in these categories. More time since the one of the great strengths of this study is
loss, having a birth since the loss, and reporting the use of a nationally representative sample
more locus of control regarding getting pregnant of reproductive-age women, we are limited to
are associated with less distress. These results the measures included in the NSFB. There are
suggest that the context of women’s pregnancy potentially relevant concepts that we are unable
and fertility experiences as a whole and the to include in this study. For example, we use
meanings they attribute to their pregnancies are women’s reports of whether their loss was a
crucial in shaping the psychological response to miscarriage or a stillbirth, but there may be
pregnancy loss. discrepancy in these reports; whereas some
We were surprised that knowing the reason for women may use the medical definition of a
the pregnancy loss was associated with higher stillbirth as a loss after 20 weeks, others may
distress. This finding is counter to previous use the term ‘‘miscarriage’’ to refer to any loss
literature, which argued that unexplained loss that occurs during the pregnancy. Furthermore,
should be more distressing (Simmons et al., we do not have measures for some concepts that
2006). Rather than empowering women, perhaps may be important to understand the reduction
finding out the reason for a pregnancy loss allows in distress following a pregnancy loss. Day and
women to place ‘‘blame’’ on themselves even Hooks (1987) found that family cohesion and
if the loss was out of their control. Infertility adaptability are crucial for a faster recovery
research shows similar findings; in infertile from pregnancy loss. Unfortunately, the NSFB
couples, women express more distress and does not include measures for affect, ability to
internalize the infertility diagnosis even when it be flexible, or the ability to reframe and redefine
is their husband with the infertility ‘‘problem’’ stressful events. The NSFB also does not include
(Greil, 1991). We were also surprised that detailed information about the actual experience
recurrent pregnancy loss was not associated of the loss. For example, we cannot determine
with significantly higher distress. We explored if women who had a stillbirth and were able to
this association and found that the association hold their baby report more or less distress than
was significant until ‘‘perceiving a fertility women who did not process their loss in this way.
problem’’ was added to the model. We therefore Despite these limitations, however, this study is
conclude that multiple losses contribute to the first to provide an in-depth investigation of
distress primarily if they contribute to women factors that increase or reduce distress following
self-identifying a fertility problem. a pregnancy loss.

Limitations Implications for Practice

The current study has several limitations. Some Our findings increase knowledge about the psy-
of these limitations are the result of the cross- chological impact of pregnancy loss by pro-
sectional data. First, we would have stronger viding evidence that distress differs for women
causal certainty if we had more data points in depending on both their commitment/attachment
the process of dealing with pregnancy loss. We to their pregnancies and their current fertility
have causal ordering in that the loss happened contexts. This study has practical implications
before our measures of distress, but we do not for women who have experienced pregnancy
know about other sequences, such as the ordering loss and for family professionals who work
of distress and pregnancy locus of control. We with them. It may be difficult for practition-
do not know whether women who are more ers to effectively reduce negative psychological
Pregnancy Loss and Distress 353

consequences of pregnancy loss without under- article was presented at the 2009 annual meeting of the
standing the factors that shape the experience American Sociological Association in San Francisco, CA.
and meaning of pregnancy loss for women. Our
findings highlight the importance of women’s REFERENCES
fertility histories and their childbearing desires.
Women who planned for the pregnancy that Abbey, A., Andrews, F. M., & Halman, L. J. (1992).
resulted in a loss are more distressed, regard- Psychosocial, treatment, and demographic predic-
tors of the stress associated with infertility. Fertility
less of whether the loss was a miscarriage
and Sterility, 57, 122 – 128.
or stillbirth. As expected, women who expe- Adeyemi, A. (2008). Depressive symptoms in a
rienced a stillbirth of a planned pregnancy are sample of women following perinatal loss.
the most distressed, but women who experienced Journal of the National Medical Association, 100,
a miscarriage of a planned pregnancy are also 1463 – 1468.
significantly distressed. In addition, women who Armstrong, D. S. (2007). Perinatal loss and parental
know the reason for the loss, experienced more distress after the birth of a healthy infant. Advances
than one loss, have experienced infertility, want in Neonatal Care: Official Journal of the National
to have a baby, and value motherhood as more Association of Neonatal Nurses, 7, 200 – 206.
important are more distressed following a preg- Becker, H. S. (1960). Notes on the concept of
commitment. American Journal of Sociology, 66,
nancy loss. Time since the loss, having a birth
32 – 40.
after the loss, and viewing oneself as having Bennett, S. M., Litz, B. T., Lee, B. S., & Maguen, S.
more control over pregnancies are associated (2005). The scope and impact of perinatal loss:
with less distress. Practitioners working with Current status and future directions. Professional
women and couples who have experienced preg- Psychology: Research and Practice, 36, 180 – 187.
nancy loss could provide more targeted support Bernazzani, O., & Bifulco, A. (2003). Motherhood as
and effective treatment if they assess FSD, com- a vulnerability factor in major depression: The role
mitment/attachment to a pregnancy that ended in of negative pregnancy experiences. Social Science
loss, and current fertility context and pregnancy- and Medicine, 56, 1249 – 1260.
related attitudes for a contextual understanding Bowlby, J. (1969). Attachment and loss: Attachment.
of the meaning of the loss. New York: Basic Books.
Boyle, F. M., Vance, J. C., Najman, J. M., & Thearle,
Much of the information currently available to M. J. (1996). The mental health impact of stillbirth,
practitioners regarding the psychological conse- neonatal death or SIDS: Prevalence and patterns
quences of pregnancy loss posits that attachment of distress among mothers. Social Science and
increases further along as a pregnancy pro- Medicine, 43, 1273 – 1282.
gresses. Our findings suggest, however, that the Bretherton, I. (1985). Attachment theory: Retrospect
meaning that women place on their pregnancy, and prospect. Monographs of the Society for
along with their fertility histories, attitudes, and Research in Child Development, 50(1/2), 3 – 35.
intentions are as important as gestation length Brier, N. (2008). Grief following miscarriage:
at the time of the loss. This study suggests A comprehensive review of the literature. Journal
support for a social constructionist approach to of Women’s Health, 17, 451 – 463.
Day, R. D., & Hooks, D. (1987). Miscarriage:
understand how pregnancy loss affects women’s
A special type of family crisis. Family Relations,
mental health. A more nuanced understanding of 36, 305 – 310.
the consequences of pregnancy loss for women Franche, R. (2001). Psychological and obstetric
includes the type of loss and whether the loss predictors of grief during pregnancy or after
occurred for a planned pregnancy as well as other perinatal death. Obstetrics and Gynecology, 97,
pregnancy loss and fertility-related characteris- 597 – 602.
tics such as subsequent childbearing, infertility, Greil, A. L. (1991). Not yet pregnant: Infertile couples
childbearing desires or intentions, and impor- in contemporary America. New Brunswick, NJ:
tance of motherhood. Rutgers University Press.
Greil, A. L., & McQuillan, J. (2004). Help-seeking
patterns among infertile women. Journal of Repro-
NOTE ductive and Infant Psychology, 22, 305 – 319.
This research was supported in part by Grant R01-HD044144
Greil, A. L., McQuillan, J., Johnson, K., Blevins-
‘‘Infertility: Pathways and Psychosocial Outcomes’’ funded Slauson, K., & Shreffler, K. M. (2010). The hidden
by NICHD (Lynn White and David R. Johnson, Co-PIs). infertile: Infertile women without pregnancy intent
The authors wish to thank Carolyn Henry for her helpful in the United States. Fertility and Sterility, 93,
comments on an earlier draft. A previous version of this 2080 – 2083.
354 Family Relations

Greil, A. L., Slauson-Blevins, K., & McQuillan, J. McQuillan, J., Greil, A. L., White, L., & Jacob, M. C.
(2010). The experience of infertility: A review of (2003). Frustrated fertility: Infertility and psy-
recent literature. Sociology of Health and Illness, chological distress among women. Journal of
32, 140 – 162. Marriage & Family, 65, 1007 – 1018.
Hirschi, T. (1969). Causes of delinquency. Berkeley: Mercer, R. T., & Ferketich, S. L. (1990). Predictors
University of California Press. of parental attachment during early parenthood.
Janssen, H. J., Cuisinier, M. C. & de Grauw, K. P. Journal of Advanced Nursing, 15, 268 – 280.
(1997). A prospective study of risk factors Miles, L. M., Keitel, M., Jackson, M., Harris, A., &
predicting grief intensity following pregnancy loss. Licciardi, F. (2009). Predictors of distress in
Archives of General Psychiatry, 54, 56 – 61. women being treated for infertility. Journal
Janssen, H. J., Cuisinier, M. C., & Hoogduin, K. A. L. of Reproductive and Infant Psychology, 27,
(1996). A critical review of the concept of 238 – 257.
pathological grief following pregnancy loss. Muller, M. E. (1992). A critical review of prenatal
attachment research. Scholarly Inquiry for Nursing
Journal of Death and Dying, 33, 21 – 42.
Practice: An International Journal, 6, 5 – 21.
Kanter, R. M. (1972). Commitment and community:
Neugebauer, R., Kline, J., O’Connor, P., Johnson, J.,
Communes and utopias in social perspective.
Skodol, A., Wicks, J., & Susser, M. (1992).
Cambridge, MA: Harvard University Press. Depressive symptoms in women in the six months
Kemp, V. H., & Page, C. K. (1987). Maternal prenatal after miscarriage. American Journal of Obstetrics
attachment in normal and high-risk pregnancies. and Gynecology, 166, 104 – 109.
Journal of Obstetrics, Gynecologic, and Neonatal Peppers, L. G., & Knapp, R. J. (1980). Motherhood
Nursing, 16, 195 – 206. and mourning: Perinatal death. New York:
Klier, C. M., Geller, P. A., & Neugebauer, R. (2000). Praeger.
Minor depressive disorder in the context of Ritsher, J. B. (2002). Perinatal bereavement grief
miscarriage. Journal of Affective Disorders, 59, scale: Distinguishing grief from depression fol-
13 – 21. lowing miscarriage. Assessment, 9, 31 – 40.
Klier, C. M., Geller, P. A., & Ritsher, J. B. (2002). Robinson, M., Baker, L., & Nackerud, L. (1999). The
Affective disorders in the aftermath of miscarriage: relationship of attachment theory and perinatal
A comprehensive review. Archives of Women’s loss. Death Studies, 23, 257 – 270.
Mental Health, 5, 129 – 149. Saraiya, M., Berg, C. J., Shulman, H., Green, C. A.,
Kutteh, W. H. (2005). Recurrent pregnancy loss. In & Atrash, H. K. (1999). Estimates of the annual
B. R. Carr, R. E. Blackwell, & R. Azziz (Eds.), number of clinically recognized pregnancies in the
Essential reproductive medicine (pp. 585 – 592). United States, 1981 – 1991. American Journal of
New York: McGraw-Hill. Epidemiology, 149, 1025 – 1029.
Lasker, J. N., & Toedter, L. J. (2000). Predicting Schmidt, L. (2009). Social and psychological con-
outcomes after pregnancy loss: Results from sequences of infertility and assisted reproduc-
studies using the Perinatal Grief Scale. Illness, tion—What are the research priorities? Human
Crisis and Loss, 8, 350 – 372. Fertility, 12, 14 – 20.
Lok, I. H., & Neugebauer, R. (2007). Psychological Schwerdtfeger, K. L., & Shreffler, K. M. (2009).
morbidity following miscarriage. Best Practice & Trauma of pregnancy loss and infertility for
Research, Clinical Obstetrics and Gynaecology, mothers and involuntarily childless women in the
21, 229 – 247. contemporary United States. Journal of Loss and
Trauma, 14, 211 – 227.
Lok, I. H., Yip, A. S., Lee, D. T., Sahota, D., &
Simmons, R. K., Singh, G., Maconochie, N.,
Chung, T. K. (2010). A 1-year longitudinal study
Doyle, P., & Green, J. (2006). Experience of mis-
of psychological morbidity after miscarriage.
carriage in the UK: Qualitative findings from the
Fertility and Sterility, 93, 1966 – 1975. National Women’s Health Study. Social Science
Lydon, J., Dunkel-Schetter, C., Cohan, C. L., & & Medicine, 63, 1934 – 1946.
Pierce, T. (1996). Pregnancy decision making as Stainton, M. C., McNeil, D., & Harvey, S. (1992).
a significant life event: A commitment approach. Maternal tasks of uncertain motherhood. Maternal-
Journal of Personality and Social Psychology, 71, Child Nursing Journal, 20, 113 – 123.
141 – 151. Stinson, K. M., Lasker, J. N., Lohmann, J., & Toedter,
Maccoby, E. E. (1980). Social development: Growth L. J. (1992). Parents’ grief following pregnancy
and the parent-child relationship. New York: loss: A comparison of mothers and fathers. Family
Harcourt Brace Jovanovich. Relations, 41, 218 – 223.
Magee, P. L. (2003). Psychological distress in recur- Stryker, S. (1968). Identity salience and role per-
rent miscarriage: The role of prospective thinking formance: The relevance of symbolic interaction
and role and goal investment. Journal of Repro- theory for family research. Journal of Marriage
ductive and Infant Psychology, 21, 35 – 47. and the Family, 4, 558 – 564.
Pregnancy Loss and Distress 355

Swanson, K. M. (2000). Predicting depressive symp- Theut, S. K., Pederson, F. A., Zaslow, M. J., & Rabi-
toms after miscarriage: A path analysis based on novich, B. A. (1988). Pregnancy subsequent to
the Lazarus paradigm. Journal of Women’s Health perinatal loss: Parental anxiety and depression.
& Gender-Based Medicine, 9, 191 – 206. Journal of the American Academy of Child and
Thapar, A. K., & Thapar, A. (1992). Psychological Adolescent Psychiatry, 27, 289 – 292.
sequelae of miscarriage: A controlled study using Toedter, L. J., Lasker, J. N., & Alhadeff, J. M. (1988).
the general health questionnaire and the hospital The Perinatal Grief Scale: Development and initial
anxiety and depression scale. British Journal of validation. American Journal of Orthopsychiatry,
General Practice, 42, 94 – 96. 58, 435 – 449.
Copyright of Family Relations is the property of Wiley-Blackwell and its content may not be copied or emailed
to multiple sites or posted to a listserv without the copyright holder's express written permission. However,
users may print, download, or email articles for individual use.