Screening for Symptoms of Postpartum Traumatic Stress in Mothers with Preterm Infants

Issues in Mental Health Nursing, 35:198–206, 2014
Copyright © 2014 Informa Healthcare USA, Inc.
ISSN: 0161-2840 print / 1096-4673 online
DOI: 10.3109/01612840.2013.853332
Screening for Symptoms of Postpartum Traumatic Stress
in a Sample of Mothers with Preterm Infants
Richard J. Shaw, MD
Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Palo Alto,
California, USA
Emily A. Lilo, MPH
Department of Pediatrics, University of New Mexico, Albuquerque, New Mexico, USA
Amy Storfer-Isser, PhD
Statistical Research Consultants, LLC, Perrysburg, Ohio, USA
M. Bethany Ball, BS, Melinda S. Proud, RCP, RRT, Nancy S. Vierhaus, RN, MSN,
RNC, Audrey Huntsberry, RN, MSN, CNS, and Kelley Mitchell, BA
Division of Neonatology and Developmental Medicine, Stanford University School of Medicine,
Palo Alto, California, USA
Marian M. Adams, MD
Department of Pediatrics, Stanford University School of Medicine, Palo Alto, California, USA
Sarah M. Horwitz, PhD
Department of Child and Adolescent Psychiatry, New York University Medical School, New York,
New York, USA
There are no established screening criteria to help identify
mothers of premature infants who are at risk for symptoms of
emotional distress. The current study, using data obtained from
recruitment and screening in preparation for a randomized con-
trolled trial, aimed to identify potential risk factors associated with
symptoms of depression, anxiety and posttraumatic stress in a sam-
ple of mothers with premature infants hospitalized in a neonatal
intensive care unit. One hundred, thirty-five mothers of preterm in-
fants born at 26–34 weeks of gestation completed three self-report
measures: the Stanford Acute Stress Reaction Questionnaire, the
Beck Depression Inventory (2nd ed.), and the Beck Anxiety In-
ventory to determine their eligibility for inclusion in a treatment
intervention study based on clinical cut-off scores for each mea-
sure. Maternal sociodemographic measures, including race, eth-
nicity, age, maternal pregnancy history, and measures of infant
medical severity were not helpful in differentiating mothers who
screened positive on one or more of the measures from those who
screened negative. Programs to screen parents of premature in-
fants for the presence of symptoms of posttraumatic stress, anxi-
ety, and depression will need to adopt universal screening rather
Address correspondence to Richard J. Shaw, Department of Psy-
chiatry and Behavioral Sciences, Stanford University School of
Medicine 401 Quarry Road, Palo Alto, CA 94305-5719 USA. E-mail:
[e-mail]
than profiling of potential high risk parents based on their sociode-
mographic characteristics or measures of their infant’s medical
severity.
It is now well established that mothers of premature infants
are at increased risk of symptoms of psychological distress, in-
cluding posttraumatic stress disorder (PTSD), postpartum anxi-
ety, and postpartum depression (Bener, Gerber, & Sheikh, 2012;
Kersting et al., 2004; Miles, 1989; Peebles-Kleiger, 2000; Ross
& McLean, 2006). Data from community samples, for ex-
ample, suggest that prevalence rates of women who screen
positive for the diagnostic criteria for PTSD after childbirth
range from 1.7–9% (Beck et al., 2011) while mothers of pre-
mature infants hospitalized in the neonatal intensive care unit
(NICU) are at even higher risk, with prevalence rates as high as
23–35% in the immediate postpartum period (Feeley et al., 2011;
Lefkowitz et al., 2010; Vanderbilt, Bushley, Young, & Frank,
2009). However, despite the existence of several well-validated
measures useful for screening for PTSD (Brewin, 2005) as well
as other common postpartum disorders (Cox, Holden, & Sagov-
sky, 1987), guidelines for screening in this high risk population
of NICU parents are absent.
198

SYMPTOMS OF POSTPARTUM TRAUMATIC STRESS 199
To date, interest in screening parents of physically ill and
hospitalized children has been focused primarily in the area of
pediatric oncology. Kazak et al. (2011), for example, have de-
veloped the Psychiatric Assessment Tool, a screening measure
that assesses family psychosocial risk for clinically significant
distress during treatment in 18 separate domains that include
parent and family functioning, psychosocial resources, social
support, and illness factors. The instrument is based on the
premise that the majority of families dealing with the diagno-
sis of pediatric cancer are competent and adaptively organized
while a smaller proportion of families have risk factors that place
them at greater risk for poor outcomes. Scores on this measure
are also postulated to help in decisions about the allocation of
psychosocial treatment interventions. Other instruments, also
from the pediatric oncology literature and recently reviewed by
Kazak et al. (2012), include the Distress Thermometer (CAU
National Comprehensive Cancer Network, 2003), a brief uni-
dimensional rating of distress that has been used to screen par-
ents to determine eligibility for a behavioral intervention during
a child’s cancer treatment (Warner et al., 2011). In addition, the
SCREEM (Social, Cultural, Religious, Economic, Educational,
Medical Resources) Family Resources Survey (Panganiban &
Medina, 2011) has been used to help assess psychosocial needs
in families of children with cancer.
Although there is agreement on the importance of recogniz-
ing parental distress and trauma in the NICU, and a small num-
ber of interventions have been developed to target and prevent
trauma symptoms, what is missing from the literature is a rigor-
ous, systematic effort to compare mothers who screen positive
for symptoms of depression, anxiety, and traumatic stress with
those who do not. Although research on posttraumatic stress
symptoms after childbirth has established specific risk factors,
which include prior history of psychological problems (specifi-
cally of anxiety), prior trauma exposure, obstetrical difficulties,
and lack of social support (Olde, van der Hart, Kleber, & van
Son, 2006), it is not known whether factors related to the infant’s
medical history, including gestational age or pregnancy-related
complications place mothers at higher risk of psychological dis-
tress. In this brief report, we describe a sample of mothers of
premature infants who were screened for possible inclusion in a
randomized controlled trial designed to reduce symptoms of de-
pression, anxiety, and posttraumatic stress (Shaw, Brecht et al.,
2013; Shaw, St. John et al., 2013). Based on prior work that has
not shown a strong relationship between a mother’s psycholog-
ical distress and her infant’s medical history, we hypothesized
that maternal, rather than infant, variables would be associated
with maternal symptoms. Our specific aims were to determine
whether there are easily identifiable maternal sociodemographic
characteristics, aspects of pregnancy history, or factors related to
the infant’s medical history in postpartum mothers who screen
positive for symptoms of psychological distress so that guide-
lines for appropriate screening procedures in the NICU parent
population might be developed. Secondary aims were to exam-
ine sociodemographic and medical characteristics that differen-
tiate mothers with multiple categories of psychological distress
(i.e., depression, anxiety, and posttraumatic stress) and, given
the high prevalence of posttraumatic symptoms in NICU par-
ents (Shaw et al., 2006), whether it is possible to differentiate
mothers specifically based on these symptoms.
METHODS
Participants
135 English- and Spanish-speaking mothers of infants born
between 26 and 34 weeks old, weighing over 1000 grams, born
at the Lucile Packard Children’s and El Camino Hospitals in
northern California or transferred within 72 hours, without ma-
jor health complications such as congenital abnormalities, and
judged by physicians as likely to survive, were eligible for par-
ticipation in a treatment intervention study which has been pre-
viously described (Shaw, Brecht et al., 2013; Shaw, St. John
et al., 2013). During the recruitment process, 187 mothers were
approached, 30 of whom declined to meet to discuss the study.
Of the 157 mothers who were assessed for eligibility, an addi-
tional 14 declined to participate and 8 were considered ineligible
(two spoke neither English nor Spanish, two had infants who
were transferred to other hospitals, and four lived greater than
one hour from the hospital). Psychiatric risk factors, includ-
ing suicidal or homicidal ideation or the presence of psychotic
symptoms, were exclusion factors for the study; no mothers
endorsed any of these symptoms.
Study Design
The study was approved by the Stanford University Institu-
tional Review Board. Recruitment and delivery of the interven-
tion took place between July, 2011, and December, 2012. After
obtaining informed consent, mothers completed three screen-
ing measures within one week of the birth of their infants.
Mothers who screened positive on one of the three measures
were invited to participate in a separate treatment intervention
study (Shaw, Brecht et al., 2013; Shaw, St. John et al., 2013).
These well-validated self-report measures included the Beck
Depression Inventory, second edition (BDI-II; Beck, Brown, &
Steer, 1996), the Beck Anxiety Inventory (BAI; Beck, Epstein,
Brown, & Steer, 1988; Beck & Steer, 1993), and the Stanford
Acute Stress Reaction Questionnaire (SASRQ; Carde ̃na, Koop-
man, Classen Waelde, & Spiegel, 2000; Koopman, Classen, &
Spiegel, 1994).
Measures
Sociodemographic/Self-Reported Health Variable
Mothers’ date of birth, years of education, race/ethnicity,
family composition, family income, and employment were gath-
ered using questions employed in the NIMH-funded Connecti-
cut Early Development Project (Horwitz et al., 2007). Women
were asked to rate their overall physical health on a 5-point
Likert scale from “poor” to “excellent.” In addition, history of

200 R. J. SHAW ET AL.
the mother’s pregnancy and obstetrical complications were col-
lected from the medical record along with data on their infant’s
birth and medical history. These data were used to compute the
Illness Health Severity Index.
Illness Health Severity Index (IHSI)
A probability of death index (range 0–1) was calculated us-
ing a multivariable risk adjustment model, designed to capture
important factors related to patient risk based upon the Vermont
Oxford Network (VON) model (Horbar, 1999; Zupancic et al.,
2007) using modifications tailored to the California Perinatal
Quality Care Collaborative (CPQCC) data. The model includes
terms for gestational age, gestational age squared, race, sex,
location of birth, multiple birth, prenatal care, 5-minute Apgar
score, small size for gestational age (lowest 10th percentile),
major birth defect, and California Children’s Services (CCS)
NICU level. The CCS NICU level is determined in the CPQCC
database using a Regional NICU comparison chart.
Stanford Acute Stress Reactions Questionnaire (SASRQ)
The Stanford Acute Stress Reactions Questionnaire
(SASRQ) is a 30-item self-report questionnaire that was used
to assess the DSM-IV-TR symptoms of acute stress disorder
(ASD) related to the traumatic experience of the mother’s pre-
mature birth and her infant’s NICU hospitalization. The SASRQ
was chosen since screening occurred within the first week of the
infant’s birth and ASD is a more appropriate diagnostic category
given that symptoms need to be present for a longer period to
meet criteria for PTSD. Each item is rated on an ordinal scale of
1 to 5 to indicate the frequency of experiencing each symptom.
The SASRQ has been shown to have a high internal consistency
for ASD in a range of survivor samples (α = 0.80–0.95), good
test-retest reliability as assessed during a 3- to 4-week retest pe-
riod (α = 0.69), and sound convergent and discriminant validity
(Carde ̃na et al., 2000). In addition, symptoms measured using
the SASRQ are strongly related to later PTSD symptomatology
(Carde ̃na et al., 2000). In the current study, individuals who re-
ported a score of ≥ 3 (“sometimes experienced”) on at least two
of the four ASD symptom categories were considered to have
screened positive for ASD based on a scoring algorithm devel-
oped for the SASRQ (Koopman, Gore-Felton, Classen, Kim, &
Spiegel, 2001).
Beck Depression Inventory-Second Edition (BDI-II)
The second edition of the Beck Depression Inventory (BDI-
II) is a 21-item self-report questionnaire used to assess maternal
symptoms of depression in the two-week period prior to en-
rolment in the study. The reliability of the BDI-II is excellent
(coefficient alpha = .92; Beck et al., 1996), and it both corre-
lates well with other screening measures and is highly sensitive
when compared with a diagnostic assessment, particularly in the
late postpartum period. Because women in this study were to
be followed past the postpartum period, the BDI-II was selected
to screen for depressive symptoms (Ji et al., 2011). A BDI-II
score of ≥ 20 (indicating at least moderate depression based on
criteria in the scoring manual) was used as the cut-off value to
assign mothers who screened positive (Beck et al., 1996).
The Beck Anxiety Inventory (BAI)
The Beck Anxiety Inventory (BAI) is a 21-item self-report
measure used to assess maternal symptoms of anxiety. The scale
has good internal consistency and one week test-retest reliability
of .75 (Beck et al., 1988). A BAI score of ≥ 16 (indicating at
least moderate anxiety based on criteria in the scoring manual)
was used as the cut-off value to assign mothers who screened
positive (Beck & Steer, 1993).
Statistical Analysis
Study data were collected and managed using REDCap
(hosted at the Stanford Center for Clinical Informatics). RED-
Cap (Research Electronic Data Capture: Harris et al., 2009) is a
secure, web-based application designed to support data capture
for research studies. Categorical responses were summarized
using counts and frequencies, normally distributed variables
were described using means and standard deviations, and the
median and interquartile range were reported for non-normally
distributed measures. Mothers who screened positive for depres-
sion, anxiety, or ASD were coded as having a positive screen-
ing and were eligible for inclusion in the treatment interven-
tion study. Bivariate statistics, including the two-sample t-test,
Wilcoxon rank-sum test, the Chi-square test, and Fisher’s exact
test, were used to compare maternal sociodemographic and med-
ical characteristics for mothers who screened positive to those
who screened negative (primary analyses). Effect sizes were
calculated to facilitate the interpretation of the results; Cohen’s
D, ρ, and Cramer’s V are reported for normally distributed, non-
normally distributed, and categorical measures, respectively. To
further understand variations in maternal characteristics by the
number of posttraumatic stress symptoms, a three-level vari-
able was created to indicate whether mothers screened positive
on zero, one, or two–three instruments. ANOVA, the Kruskal-
Wallis test, and the Mantel-Haenszel Chi-Square test were used
to compare these three groups. Secondary analyses also included
comparing mothers who screened positive on the SASRQ versus
those who did not among the subset of mothers who screened
positive on at least one instrument. SAS version 9.2 (SAS Insti-
tute Inc., Cary, NC) was used for data analysis.
RESULTS
One hundred and five (77.8%) mothers out of the total sample
of 135 mothers screened positive on at least one of the three
measures (SASRQ, BDI-II, or BAI). Nearly three-fourths (n =
96; 71.1%) of the sample met screening criteria for ASD. Almost
half (n = 64; 47.4%) of the mothers screened positive for anxiety
symptoms and over one-third (n = 48; 35.6%) screened positive
for depressive symptoms.
Baseline demographic characteristics of the mothers in the
two groups are shown in Table 1. With the exception of one
single variable (born in the US), no maternal sociodemographic