Prevention of Traumatic Stress in Mothers ofPreterms: 6-Month Outcomes

WHAT’S KNOWN ON THIS SUBJECT: Interventions based onprinciples of trauma-focused cognitive behavior therapy havebeen shown to reduce symptoms of trauma and depression inmothers of premature infants. It is not known whether thesebenefits are sustained at long-term follow-up.

WHAT THIS STUDY ADDS: A brief, cost-effective 6-sessionmanualized intervention for parents of infants in the NICU waseffective in reducing symptoms of parental trauma, anxiety, anddepression at 6-month follow-up. There were no added benefitsfrom a 9-session version of the treatment.

abstractOBJECTIVE: Symptoms of posttraumatic stress disorder are a well-recognized phenomenon in mothers of preterm infants, with impli-cations for maternal health and infant outcomes. This randomizedcontrolled trial evaluated 6-month outcomes from a skills-basedintervention developed to reduce symptoms of posttraumatic stressdisorder, anxiety, and depression.

METHODS: One hundred five mothers of preterm infants were ran-domly assigned to (1) a 6- or 9-session intervention based on principlesof trauma-focused cognitive behavior therapy with infant redefinitionor (2) a 1-session active comparison intervention based on educationabout the NICU and parenting of the premature infant. Outcomemeasures included the Davidson Trauma Scale, the Beck DepressionInventory II, and the Beck Anxiety Inventory. Participants wereassessed at baseline, 4 to 5 weeks after birth, and 6 months afterthe birth of the infant.

RESULTS: At the 6-month assessment, the differences between the in-tervention and comparison condition were all significant and sizableand became more pronounced when compared with the 4- to 5-weekoutcomes: Davidson Trauma Scale (Cohen’s d =20.74, P, .001), BeckAnxiety Inventory (Cohen’s d = 20.627, P = .001), Beck DepressionInventory II (Cohen’s d = 20.638, P = .002). However, there were no dif-ferences in the effect sizes between the 6- and 9-session interventions.

CONCLUSIONS: A brief 6-session intervention based on principles oftrauma-focused cognitive behavior therapy was effective at reducingsymptoms of trauma, anxiety, and depression in mothers of preterminfants. Mothers showed increased benefits at the 6-month follow-up,suggesting that they continue to make use of techniques acquiredduring the intervention phase. Pediatrics 2014;134:e481–e488

AUTHORS: Richard J. Shaw, MD,a Nick St John, PhD,b EmilyLilo, MPH,c Booil Jo, PhD,a William Benitz, MD,b David K.Stevenson, MD,b and Sarah M. Horwitz, PhDd

aDepartment of Psychiatry and Behavioral Sciences, bDivision ofNeonatology, Stanford University School of Medicine, Palo Alto,California; cDepartment of Pediatrics, University of New Mexico,Albuquerque, New Mexico; and dDepartment of Child andAdolescent Psychiatry, New York University Medical School, NewYork, New York

KEY WORDSneonatal intensive care, premature infants, posttraumatic stressdisorder, intervention, PTSD, preterm infants, neonatal ICU,intervention

ABBREVIATIONSASD—acute stress disorderBAI—Beck Anxiety InventoryBDI-II—the Beck Depression Inventory IICBT—cognitive behavior therapyDSM-IV-TR—Diagnostic and Statistical Manual of MentalDisorders, Fourth Edition, Text RevisionDTS—Davidson Trauma ScaleMINI—Mini-International Neuropsychiatric InterviewPSS:NICU—Parental Stressor Scale: NICUPTSD—posttraumatic stress disorderSASRQ—Stanford Acute Stress Reaction Questionnaire

Dr Shaw conceptualized and designed the study and drafted theinitial manuscript; Dr St John contributed to the manualdevelopment, supervised the infant sessions, and reviewed andrevised the manuscript; Ms Lilo contributed to the manualdevelopment, coordinated and supervised the data collection,and reviewed and revised the manuscript; Dr Jo participated inthe grant application, carried out and drafted the data analysis,and reviewed and revised the manuscript; Dr Benitz facilitatedthe implementation of the study and reviewed and revised themanuscript; Dr Stevenson participated in the planning of study,facilitated the implementation of the study, and reviewed andrevised the manuscript; Dr Horwitz conceptualized and designedthe study and reviewed and revised the manuscript; and allauthors approved the final manuscript as submitted.

This trial has been registered at www.clinicaltrials.gov(identifier NCT01307293).

www.pediatrics.org/cgi/doi/10.1542/peds.2014-0529

doi:10.1542/peds.2014-0529

Accepted for publication Apr 29, 2014

Address correspondence to Richard J. Shaw, MD, Department ofPsychiatry and Behavioral Sciences, Stanford University School ofMedicine, 401 Quarry Rd, Palo Alto, CA 94305-5719. E-mail:rjshaw@stanford.edu

(Continued on last page)

PEDIATRICS Volume 134, Number 2, August 2014 e481

ARTICLE

by guest on June 2, 2018www.aappublications.org/newsDownloaded from

Although survival rates for prematureinfants (,32 weeks’ gestation) haveimproved dramatically in the past 10years,1 parents, particularly mothers,continue to have very high rates ofpsychological distress beyond thoseobserved in parents of full-terminfants.2 In addition to high rates ofdepression,3–5 researchers have iden-tified acute stress disorder (ASD) andposttraumatic stress disorder (PTSD)as features of parental psychologicalreactions.6–12 To date, interventionsdeveloped to reduce psychologicaldistress in parents of preterm infantshave been primarily supportive or ed-ucational in nature.13–15 More recently,a small number of studies have evalu-ated intervention programs focused onreducing parental stress16 and traumasymptoms.17

Recently, we reported findings froma randomized controlled trial of a6-session, skills-based intervention de-velopedtoreducesymptomsofparentaldepression, anxiety, and trauma inparents of preterm infants.18 Mothersin the intervention group reported agreater reduction in both trauma symp-toms and depression but not anxietycompared with the comparison groupimmediately postintervention. The in-tervention, which incorporated com-ponents of trauma-focused cognitivebehavior therapy (CBT), was found tobe feasible and easily delivered in theNICU environment and received highratings of maternal satisfaction.19 How-ever, whether these symptomatic ben-efits continue after the infants returnhome is unknown. In this report, wedescribe the 6-month follow-up out-come data of our study participants andhypothesize that mothers in the in-tervention group will continue to showstatistically significant reductions insymptoms of trauma, anxiety, and de-pression compared with mothers in thecomparison group. In addition, we hy-pothesize that a subset of mothers who

receive 3 additional sessions specifi-cally developed to target trauma andparenting issues will report additionalbenefits compared with the group re-ceiving the 6-session intervention.

METHODS

Participants

Participants were mothers of pre-mature infants hospitalized in 1 of the 4participatingNICUsaffiliatedwithLucilePackard Children’s Hospital in northernCalifornia who had developed symp-toms of trauma, anxiety, or depressionrelated to their traumatic experienceof their infant’s preterm birth and NICUhospitalization. The demographic andclinical information collected at base-line has been previously described(Table 1). Inclusion criteria were as fol-lows: (1) English- or Spanish-speakingmothers .18 years of infants aged 25to 34 weeks weighing.600 g and bornat or transferred to 1 of 4 four par-ticipating NICUs within the first weekof delivery; (2) mothers who scoredabove the clinical cutoff on 1 of 3screening instruments administeredat baseline, the Beck Anxiety Inventory(BAI),20 the Beck Depression InventoryII (BDI-II),21 and the Stanford AcuteStress Reaction Questionnaire (SASRQ).22

Exclusion criteria were (1) mothers ofchildren with developmental abnor-malities or awaiting cardiac surgeryand those assessed as being unlikelyto survive and (2) mothers with psy-chotic symptoms or suicidal/infanticidalideation.

Procedure

The Stanford University InstitutionalReview Board approved the protocol.After receiving written informed con-sent, participants completed an intakeassessment that included screening forclinically significant symptoms of ASD(SASRQ score $3 for the requirednumber of questions in $2 of the

symptom categories), depression (BDI-II score $20), and anxiety (BAI score$16). Participants meeting the cutoffon$1 of these measures were invitedto participate in the intervention. Arandom-number generator was usedto assign eligible participants, un-blinded, to either (1) 6 or 9 sessions ofthe manualized treatment interventionor (2) an active comparison group. Theintervention lasted 3 to 4 weeks withone or two 45- to 55-minute sessionsadministered weekly. Participants inboth conditions were assessed atbaseline (1–2 weeks after the birth ofthe infant), 1 week after the completionof the first 6 sessions of the in-tervention or 4 5 weeks postbirth forthe comparison group, and 6 monthsafter the birth of the infant for the 6-session, 9-session, and comparisongroups. Recruitment and delivery ofthe intervention took place betweenJuly 1, 2011, and December 31, 2012(Fig 1). Data collection was completedon May 30, 2013.

Measures

Traumatic Events Questionnaire

The Traumatic Events Questionnaire,23

an 11-item questionnaire administeredat baseline, assesses specific traumaexperiences capable of eliciting post-traumatic stress symptoms. The 2-week test-retest reliability for numberof events is 0.91 and for specific eventsis 0.72 to 1.0.

Davidson Trauma Scale

The Davidson Trauma Scale (DTS),4 a 17-item scale administered at baselineand follow-up, assesses Diagnostic andStatistical Manual of Mental Disorders,Fourth Edition, Text Revision (DSM-IV-TR), symptoms of PTSD. The scale hassolid test-retest reliability (R = 0.86) andgood internal consistency (Cronbach’sa = 0.99). In our own sample, the in-ternal consistency estimate of the DTSis 0.93. At a score of 4, the DTS achieves

e482 SHAW et al by guest on June 2, 2018www.aappublications.org/newsDownloaded from

an 83% diagnostic accuracy with theStructured Clinical Interview for DSM-IV Axis I Disorders.

SASRQ

The SASRQ,22 a 30-item self-reportquestionnaire used to assess DSM-IV-TR symptoms of ASD, was administeredat baseline. Psychometric properties ofthe SASRQ show good internal consis-tency, test-retest reliability, and pre-dictive validity. In our sample, theinternal consistency estimate of theSASRQ was 0.91.

Parental Stressor Scale: NICU

The Parental Stressor Scale: NICU (PSS:NICU),25 a 34-item scale administered

at baseline, evaluates parental per-ceptions of the stressors due to theinfant’s NICU stay in 3 areas: physicalenvironment of the NICU, infant’sbehavior/appearance, and alterationsin parental role. The PSS:NICU is inter-nally consistent (Cronbach’s a .0.70for all scales). In our sample, the inter-nal consistency estimate of the PSS:NICUwas 0.94.

BDI-II

The BDI-II,21 a 21-item questionnaireadministered at baseline and follow-up, assesses depressive symptomswith a reliability of 0.92. In our sample,the internal consistency estimate ofBDI-II was 0.88.

BAI

The BAI,20 a 21-item self-report mea-sure administered at baseline andfollow-up, assesses symptoms of anxi-ety. The scale has good internal con-sistency and a 1-week test-retestreliability of 0.75. In our sample, theinternal consistency estimate of the BAIwas 0.90.

Mini-International NeuropsychiatricInterview

The Mini-International Neuropsychiat-ric Interview (MINI),26 a structured in-terview diagnosis administered atbaseline and follow-up, was used toestablish the DSM-IV-TR diagnoses ofmajor depressive episode, any anxietydisorder, and PTSD. A multicenter studythat compared the diagnoses by gen-eral practitioners obtained by usingthe MINI with the diagnoses obtainedby psychiatrists using nonstructuredinterviews found a k coefficient be-tween 0.41 and 0.68, sensitivity be-tween 0.41 and 0.86, and specificitybetween 0.84 and 0.97.

Illness Health Severity Index

Aprobability of death index (range: 0–1)was calculated by using a multivari-able risk adjustment model, designedto capture important factors related topatient risk based on the Vermont Ox-ford Network model27,28 by using mod-ifications tailored to the CaliforniaPerinatal Quality Care Collaborativedata. The model includes terms forgestational age, gestational age squared,race, gender, location of birth, mul-tiple birth, prenatal care, 5-minuteApgar score, small size for gestationalage (lowest 10th percentile), majorbirth defect, and California Children’sServices NICU level. The CaliforniaChildren’s Services NICU level is de-termined in the California PerinatalQuality Care Collaborative databaseby using a regional NICU comparisonchart.

TABLE 1 Demographic and Baseline Clinical Characteristics

Characteristics Intervention (n = 62) Comparison (n = 43) Group Difference

DTS, mean (SD) 49.40 (25.49) 42.35 (27.05) t(103) = 1.360, P = .177BDI-II, mean (SD) 20.60 (9.48) 17.49 (10.68) t(103) = 1.568, P = .120BAI, mean (SD) 21.97 (11.83) 20.30 (12.49) t(103) = 0.693, P = .490Infant gestational age,

mean (SD), wk30.90 (3.00) 31.56 (2.60) t(103) = 1.161, P = .248

Infant severity score, mean (SD) 0.047 (0.05) 0.034 (0.03) t(103) = 1.438, P = .154Mother’s age, mean (SD), y 33.76 (6.25) 30.70 (5.50) t(103) = 2.590, P = .011Mother’s education less than

college degree, n (%)19 (30.6) 19 (44.2) x2(1) = 2.016, P = .156

Mother’s race: white (vsnonwhite), n (%)

42 (67.7) 22 (51.2) x2(1) = 2.932, P = .087

Mother’s ethnicity: Hispanic(vs non-Hispanic), n (%)

16 (25.8) 14 (32.6) x2(1) = 0.567, P = .451

Mother is US-born, n (%) 37 (59.7) 23 (53.5) x2(1) = 0.397, P = .529Married/partnered

(vs single/divorced), n (%)58 (93.5) 43 (100) x2(1) = 2.884, P = .089

Household income, n (%)a

,$50 000 15 (25.0) 9 (25.7) x2(1) = 0.006, P = .938$50 000–$99 000 10 (16.7) 5 (14.3) x2(1) = 0.094, P = .759$$100 000 35 (58.3) 21 (60.0) x2(1) = 0.025, P = .873

Interview language inEnglish (vs Spanish), n (%)

55 (88.7) 37 (86.0) x2(1) = 0.166, P = .684

Traumatic EventsQuestionnaire yes, n (%)

31 (50.0) 14 (32.6) x2(1) = 3.154, P = .076

SASRQ yes, n (%) 56 (90.3) 40 (93.0) x2(1) = 0.236, P = .627PSS:NICU, mean (SD)Global index 3.02 (0.72) 2.65 (0.86) t(103) = 2.369, P = .020Sights and sounds of the NICU 2.88 (0.84) 2.66 (1.12) t(103) = 1.150, P = .253Infant behavior and appearance 3.12 (0.96) 2.58 (1.00) t(103) = 2.794, P = .006Alteration in parental role 3.86 (0.82) 3.48 (1.11) t(103) = 2.012, P = .047

MINI, n (%)Major depressive episode,

current51 (82.3) 36 (83.7) x2(1) = 0.038, P = .845

Dysthmia 2 (5.4) 1 (3.4) x2(1) = 0.144, P = .705Anxiety disorder, current 7 (11.3) 2 (4.7) x2(1) = 1.428, P = .232

N = 105.a Ten individuals (8 control, 2 intervention) with missing income information were not included.

ARTICLE

PEDIATRICS Volume 134, Number 2, August 2014 e483 by guest on June 2, 2018www.aappublications.org/newsDownloaded from

Intervention

The development and content of the6-session treatment manual, which in-cludes components of trauma-focusedCBT and infant redefinition, have beenpreviously described.19 An additional 3sessions were developed for a subsetof the intervention group, with contentdirected at identifying triggers associ-ated with the development of parentaltrauma symptoms as well as educationabout parenting patterns associatedwith the aspects of the vulnerablechild syndrome.29 Participants in theinformation/usual-care comparisongroupreceived one 45-minute information

session on the policy, procedures, andenvironment of the NICU, with educa-tion about parenting the prematureinfant. Mothers were referred to theexisting parent mentor program forsupport and coping strategies to helpensure that the contact was similar tothe intervention group and would ap-proximate an attention-matched com-parison condition. Mothers also receivedusual NICU care including contactswithsocial workers, chaplaincy, and devel-opmental psychologists.

Analyses

For sample descriptive statistics,meansand SDs were used to summarize

continuous variables, whereas countsand proportions were used to summa-rize categorical measures. Two-samplet tests (for continuous variables) andx2 tests (for categorical variables)were used to compare baseline mea-sures across the intervention and com-parison groups (2-sided, a = 0.05).Standard linear mixed-effects model-ing30,31 was used to model longitudinaltrajectories of main outcomes fromthe baseline to postintervention to6-month assessments. Specifically, weused a random intercept model as-suming a quadratic trend over time.In line with the intention-to-treat prin-ciple, we included all randomly as-signed individuals in the analyses forwhom data were available from atleast 1 of the 3 assessments. Datapoints that were not available weretreated as missing at random, con-ditional on observed information usinga maximum likelihood estimation.32

All 105 individuals randomly assignedto either the intervention (n = 62) orto the comparison condition (n = 43)were included in the mixed-effectsmodeling because all of them hadthe baseline data available. Withinthe group of 62 participants in the in-tervention group, 28 received 9 ses-sions and 34 received 6 sessions. Theretention rate was high in all groups.By the postintervention assessment,there were 5 dropouts (8%) who wereassigned to the intervention and 2dropouts (5%) who were assigned tothe comparison group. By the 6-monthassessment, there were 5 dropouts(8%) who were assigned to the inter-vention and 5 dropouts (12%) who wereassigned to the comparison group. Weincorporated the MacArthur frame-work33,34 in this longitudinal model-ing framework for our exploratorymoderator/mediatoranalysis. TheMplus35

program version 7.11 was used to con-duct themaximum likelihood estimationfor all of the longitudinal mixed-effectsanalyses.

FIGURE 1CONSORT Diagram.

e484 SHAW et al by guest on June 2, 2018www.aappublications.org/newsDownloaded from

RESULTS

Baseline Characteristics

Baseline demographic characteristicsand clinical features of mothers andinfants, which have been previouslydescribed, are presented in Table 1.18

There were no statistical differencesbetween the intervention and compar-ison groups onmost baseline variablesincluding maternal ratings of trauma(SASRQ and DTS), anxiety (BAI), anddepression (BDI-II); previous traumahistory (Traumatic Events Question-naire); and current psychiatric diag-noses of major depression, dysthymia,or anxiety assessed using the MINI Di-agnostic Interview.

Primary Outcomes

We examined whether women in theintervention group were different fromwomen in the information/usual ser-vices group in terms of trauma symp-toms (DTS), depression (BDI), andanxiety (BAI) at the 6-month assess-ment. Figure 2 shows estimated longi-tudinal trajectories of primary (DTS)and secondary (BDI, BAI) outcomesbased on mixed-effects modeling. All 3outcomes declined more over timeunder the intervention condition com-pared with the comparison condition,resulting in more noticeable differ-ences between the 2 groups by 6months.

Table 2 summarizes the estimated in-tervention effects (group differences)on the basis of the longitudinal analy-ses shown in Fig 2. The postinter-vention 4- to 5-week outcomes, whichhave been previously reported, showeda significant effect of the interventionon symptoms of both trauma and de-pression and depression but not foranxiety.18 At the 6-month assessment,the differences between the inter-vention and control condition becamemore pronounced with respect totrauma (DTS: Cohen’s d = 20.74, P ,.001), depression (DTS: Cohen’s d=20.55,

P = .002), and anxiety (DTS: Cohen’s d =20.63, P = .001). None of the 3 primaryoutcomes showed significant differ-ences between the 6- and 9-sessioninterventions at either the 1-month(DTS: Cohen’s d = 0.152, P = .483; BDI-II:Cohen’s d = 0.354, P = 2.136; BAI:Cohen’s d = 0.117, P = .664) or 6-month(DTS: Cohen’s d = 20.184, P = .395;BDI-II: Cohen’s d = 20.129, P = 2.615;BAI: Cohen’s d = 20.179, P = .477)follow-up.

Moderator Analysis

Eleven baseline variables (infant med-ical severity score, length of stay,Traumatic Events Questionnaire, NICUoverall stress, major depressive epi-sode current, education, maternal age,income, white/nonwhite, Hispanic, andUS born) were examined as potentialmoderatorsof the interventioneffect onthe DTS by using analytical criteriaconsistent with the MacArthur ap-proach.33,34 According to our mixed-effects modeling, where the keyparameter of interest is the effect ofinteraction between the interventionstatus and a potential moderator onDTS at the 6-month assessment, none ofthe variables were identified as mod-erators. However, mother’s education,race (white/nonwhite), and PSS:NICUscore were found to be nonspecificpredictors of the outcome. Less edu-cated mothers (P = .033), white moth-ers (P = .025), and mothers with higherbaseline NICU stress (P , .001) hadlower DTS scores at the 6-month as-sessment irrespective of whether theywere in the intervention or comparisongroup.

Mediator Analysis

We considered the possible role ofmothers’ use of psychotherapy afterthe intervention (but before the6-month outcome assessment) as apotential mediator of the interventioneffect. Approximately 25% of mothers

in each group self-reported the addi-tional use of mental health services.However, the intervention had littleimpact on the mothers’ use of psycho-therapy (association between therapyuse frequency and the interventionstatus: r = 0.066, P = .526); and as aresult, mothers’ use of psychotherapydid not meet the eligibility criteria forbeing a moderator.34 This finding sug-gests that the observed decrease insymptoms is due to the direct effects ofthe intervention rather than any addi-tional mental health services.

DISCUSSION

This is the first study to our knowledgeto evaluate the reasonably long-termeffects of an intervention based onprinciples of trauma-focused CBT ina sample of parents of prematureinfants. The intervention was found tohave a significant and strong effect onmaternal traumasymptoms (d=20.74,P , .001) at the 6-month follow-up as-sessment in addition to sizable benefitswith respect to symptoms of bothanxiety and depression. Although Jotzoand Poets17 were able to show an effecton trauma pathology, their outcomeswere short term; and whereas Kaaresenet al16 had longer-term outcomes, theirintervention assessed parental stressnot PTSD symptoms. Our data suggestthat our manualized treatment inter-vention, which incorporates principlesof psychoeducation, cognitive restruc-turing, and trauma exposure, is effec-tive in reducing symptoms that occurin response to medical trauma, spe-cifically for mothers who have experi-enced the trauma of preterm birth andNICU hospitalization.

Although our study included partic-ipants who did not meet full criteria forPTSDat the timeof entry, our results arecomparable to those of studies inindividuals exposed to military andnonmedical civilian trauma. A recentCochrane review36 of early psychological

ARTICLE

PEDIATRICS Volume 134, Number 2, August 2014 e485 by guest on June 2, 2018www.aappublications.org/newsDownloaded from

interventions to treat acute traumaticstress symptoms found that trauma-focused CBT was superior to waitinglist/usual care or supportive counsel-ing in individuals who were symptom-atic at the time of entry into the study interms of a reduction in symptoms ofPTSD, anxiety, and depression. How-ever, 1 significant difference in the datafrom our study is that the magnitude ofthe effect size for the 6-month out-comes for PTSD and anxiety actuallyincreased when compared with the4- to 5-week outcomes. By contrast,Roberts et al36 reported no strong ef-fect of trauma-focused CBT in thestudies that assessed participants ateither 3 to 5 months or 12 to 18 months

after the intervention. Similarly, Kornøret al,37 in a meta-analysis of earlytrauma-focused CBT to prevent chronicPTSD and related symptoms, found only5 RCTs with outcome data .3 monthsposttreatment and that the efficacy oftrauma-focused CBT compared withsupportive counseling was significantonly for only 1 research group at 3to 6 months and with no differencesfound at 9 months or beyond. Our find-ing of increased benefits at the 6-monthfollow-up suggests that mothers, as theyreported in exit interviews, continued topractice and make use of techniquesacquired during the intervention phaseto address their trauma and anxietysymptoms.

Findings from the study revealed noaddedbenefit formotherswhoreceived9 rather than 6 sessions. The effect oflength of treatment has also beenevaluated in meta-analyses of trauma-focused CBT. Roberts et al36 foundthat the effect size was reduced instudies that offered only 4 to 5 sessionsof treatment. Although conventionaltrauma-focused CBT typically offers 12to 16 sessions, the optimal length of theintervention has not been fully estab-lished. Data from our study suggestthat it is possible to deliver the corecomponents of the intervention in just6 sessions, with no added benefitsfor additional sessions. By contrast,mothers in general reported greatersatisfaction with the shorter treatmentlength in part due to the demands as-sociated with having their infant in theNICU.

The moderator analysis revealed thatno sociodemographic variables influ-enced the intervention response, sug-gesting that the intervention is usefulfor mothers of different ethnic back-grounds and socioeconomic status.Maternal ratings of baseline NICUstress, race, and mother’s educationwere found to be nonspecific pre-dictors of the outcome. Women withhigher ratings of parental stress aremore likely to show greater declinein DTS scores regardless of their

FIGURE 2Estimated trajectories of primary and secondary outcomes based on longitudinal mixed-effects analysis (n = 43 for control and n = 62 for intervention). base,baseline; cont, control; int, intervention.

TABLE 2 Estimated Intervention Effects (Group Difference) at Postintervention (4–5 Weeks) and at6-Month Assessment

Postintervention (1 Month) Follow-up (6 Months)

DTSIntervention effect 27.347 215.996P .041 ,.00195% CI 214.394 to 20.301 223.128 to 28.863Effect size 20.333 20.741

BDI-IIIntervention effect 24.144 25.119P .002 .00295% CI 26.826 to 21.463 28.381 to 21.857Effect size 20.546 20.638

BAIIntervention effect 21.697 25.308P .353 .00195% CI 25.275 to 1.881 28.358 to 22.258Effect size 20.193 20.627

Values are based on mixed-effects longitudinal analysis. Effect size represents estimated group difference/estimated SD ofeach outcome at postintervention and at 6-month assessment. CI, confidence interval.

e486 SHAW et al by guest on June 2, 2018www.aappublications.org/newsDownloaded from

intervention status, suggesting per-haps that more highly stressed womenwere motivated to practice, and po-tentially benefit from, treatment or, forthe comparisons, seek care for theirsymptoms. In addition, white mothersalso showed greater decline in DTSscores, a finding we cannot fully ex-plain, although it is possible that otherfactors (eg, differences in social sup-port) may be implicated.38 Finally, lesseducated mothers showed a greaterdecline in DTS scores irrespective ofwhether they were in the interventionor comparison group. It is possible thatfor less educated women, these tech-niques were novel and that, as a result,they practiced and made use of them

more frequently. Several limitationsshould be noted. First, our sample sizewas relatively small, and althoughrepresentative of the population inour hospital catchment area, African-American mothers were underrepre-sented. However, although our samplewas skewed toward mothers of highsocioeconomic status, the moderatoranalysis did not show any effect ofmaternal age, education, income, orethnicity in terms of response to theintervention. Second, results of thestudy are applicable only to the pop-ulation of NICU mothers who screenedpositive for inclusion in the study anddo not address potential needs ofmothers who were not symptomatic

early in the course of their infants’hospitalization.

CONCLUSIONS

The results of the study reveal that (1)a brief intervention based on principlesof trauma-focused CBT was effective atreducing symptoms of trauma, anxiety,and depression in a highly stressedpopulation of mothers of preterm in-fants; (2) mothers showed increasedbenefits at the 6-month follow-up, sug-gesting that they continued tomake useof techniques acquired during the in-tervention phase; and (3) therewere nobenefits from additional sessions be-yond the original 6-session intervention.

REFERENCES

1. Berger TM, Steurer MA, Woerner A, Meyer-Schiffer P, Adams M; Swiss Neonatal Net-work. Trends and centre-to-centre variabilityin survival rates of very preterm infants(,32 weeks) over a 10-year-period in Swit-zerland. Arch Dis Child Fetal Neonatal Ed.2012;97(5):F323–F328

2. Brandon DH, Tully KP, Silva SG, et al. Emo-tional responses of mothers of late-pretermand term infants. J Obstet Gynecol NeonatalNurs. 2011;40(6):719–731

3. Davis L, Edwards H, Mohay H, Wollin J. Theimpact of very premature birth on thepsychological health of mothers. Early HumDev. 2003;73(1–2):61–70

4. Howland LC, Pickler RH, McCain NL, GlaserD, Lewis M. Exploring biobehavioral out-comes in mothers of preterm infants. MCNAm J Matern Child Nurs. 2011;36(2):91–97

5. Singer LT, Fulton S, Davillier M, Koshy D,Salvator A, Baley JE. Effects of infant riskstatus and maternal psychological distresson maternal-infant interactions during thefirst year of life. J Dev Behav Pediatr. 2003;24(4):233–241

6. Kersting A, Dorsch M, Wesselmann U, et al.Maternal posttraumatic stress responseafter the birth of a very low-birth-weight in-fant. J Psychosom Res. 2004;57(5):473–476

7. Holditch-Davis D, Bartlett TR, Blickman AL,Miles MS. Posttraumatic stress symptomsin mothers of premature infants. J ObstetGynecol Neonatal Nurs. 2003;32(2):161–171

8. Peebles-Kleiger MJ. Pediatric and neonatalintensive care hospitalization as traumatic

stressor: implications for intervention. BullMenninger Clin. 2000;64(2):257–280

9. Vanderbilt D, Bushley T, Young R, Frank DA.Acute posttraumatic stress symptomsamong urban mothers with newborns inthe neonatal intensive care unit: a pre-liminary study. J Dev Behav Pediatr. 2009;30(1):50–56

10. Pierrehumbert B, Nicole A, Muller-Nix C,Forcada-Guex M, Ansermet F. Parental post-traumatic reactions after premature birth:implications for sleeping and eating prob-lems in the infant. Arch Dis Child FetalNeonatal Ed. 2003;88(5):F400–F404

11. Shaw RJ, Deblois T, Ikuta L, Ginzburg K,Fleisher B, Koopman C. Acute stress dis-order among parents of infants in theneonatal intensive care nursery. Psycho-somatics. 2006;47(3):206–212

12. Shaw RJ, Bernard RS, Deblois T, Ikuta LM,Ginzburg K, Koopman C. The relationshipbetween acute stress disorder and post-traumatic stress disorder in the neonatalintensive care unit. Psychosomatics. 2009;50(2):131–137

13. Preyde M, Ardal F. Effectiveness of a parent“buddy” program for mothers of very pre-term infants in a neonatal intensive careunit. CMAJ. 2003;168(8):969–973

14. Roman LA, Lindsay JK, Boger RP, et al.Parent-to-parent support initiated in theneonatal intensive care unit. Res NursHealth. 1995;18(5):385–394

15. Anderson GC, Chiu SH, Dombrowski MA,Swinth JY, Albert JM, Wada N. Mother-

newborn contact in a randomized trial ofkangaroo (skin-to-skin) care. J ObstetGynecol Neonatal Nurs. 2003;32(5):604–611

16. Kaaresen PI, Rønning JA, Tunby J, NordhovSM, Ulvund SE, Dahl LB. A randomizedcontrolled trial of an early interventionprogram in low birth weight children:outcome at 2 years. Early Hum Dev. 2008;84(3):201–209

17. Jotzo M, Poets CF. Helping parents copewith the trauma of premature birth: anevaluation of a trauma-preventive psycho-logical intervention. Pediatrics. 2005;115(4):915–919

18. Shaw RJ, St John N, Lilo EA, et al. Pre-vention of traumatic stress in mothers withpreterm infants: a randomized controlledtrial. Pediatrics. 2013;132(4). Available at:www.pediatrics.org/cgi/content/full/132/4/e886

19. Shaw RJ, Sweester CJ, St John N, et al.Prevention of postpartum traumatic stressin mothers with preterm infants: manualdevelopment and evaluation. Issues MentHealth Nurs. 2013;34(8):578–586

20. Beck AT, Epstein N, Brown G, Steer RA. Aninventory for measuring clinical anxiety:psychometric properties. J Consult ClinPsychol. 1988;56(6):893–897

21. Beck AT, Steer RA, Brown GK. Beck De-pression Inventory-II (BDI-II). San Antonio,TX; Psychological Corporation 1996

22. Cardeña E, Koopman C, Classen C, Waelde LC,Spiegel D. Psychometric properties of theStanford Acute Stress Reaction Questionnaire

ARTICLE

PEDIATRICS Volume 134, Number 2, August 2014 e487 by guest on June 2, 2018www.aappublications.org/newsDownloaded from

(SASRQ): a valid and reliable measure ofacute stress. J Trauma Stress. 2000;13(4):719–734

23. Vrana S, Lauterbach D. Prevalence of trau-matic events and post-traumatic psycho-logical symptoms in a nonclinical sampleof college students. J Trauma Stress. 1994;7(2):289–302

24. Davidson JR, Book SW, Colket JT, et al. As-sessment of a new self-rating scale forpost-traumatic stress disorder. PsycholMed. 1997;27(1):153–160

25. Miles MS, Funk SG, Carlson J. ParentalStressor Scale: neonatal intensive careunit. Nurs Res. 1993;42(3):148–152

26. Sheehan DV, Lecrubier Y, Sheehan KH, et al.The Mini-International Neuropsychiatric In-terview (M.I.N.I.): the development and vali-dation of a structured diagnostic psychiatricinterview for DSM-IV and ICD-10. J Clin Psy-chiatry. 1998;59(20 suppl 20):22–33; quiz 34–57

27. Horbar JD. The Vermont Oxford Network:evidence-based quality improvement for neo-natology. Pediatrics. 1999;103(1 suppl E):350–359

28. Zupancic JA, Richardson DK, Horbar JD,Carpenter JH, Lee SK, Escobar GJ; VermontOxford Network SNAP Pilot Project Partic-ipants. revalidation of the score for neonatalacute physiology in the Vermont OxfordNetwork. Pediatrics. 2007;119(1). Available at:www.pediatrics.org/cgi/content/full/119/1/e156

29. Thomasgard M, Metz WP. The vulnerablechild syndrome revisited. J Dev BehavPediatr. 1995;16(1):47–53

30. Raudenbush S, Bryk A. Hierarchical LinearModels: Applications and Data AnalysisMethods. Thousand Oaks, CA: Sage; 2002

31. Singer JD, Willett JB. Applied LongitudinalData Analysis: Modeling Change and EventOccurrence. New York, NY: Oxford Univer-sity Press; 2003

32. Little RJA, Rubin DB. Statistical Analysiswith Missing Data. New York, NY: J Wiley &Sons; 2002

33. Kraemer HC, Wilson GT, Fairburn CG, AgrasWS. Mediators and moderators of treat-ment effects in randomized clinical trials.Arch Gen Psychiatry. 2002;59(10):877–883

34. Kraemer HC, Kiernan M, Essex M, Kupfer DJ.How and why criteria defining moderatorsand mediators differ between the Baron &Kenny and MacArthur approaches. HealthPsychol. 2008;27(2 suppl):S101–S108

35. Muthén LK, Muthén BO. MPlus user’s guide.Los Angeles, CA: Muthén & Muthén; 1998–2013

36. Roberts NP, Kitchiner NJ, Kenardy J, BissonJI. Early psychological interventions totreat acute traumatic stress symptoms.Cochrane Database Syst Rev. 2010;(3):CD007944

37. Kornør H, Winje D, Ekeberg Ø, et al. Earlytrauma-focused cognitive-behaviouraltherapy to prevent chronic post-traumaticstress disorder and related symptoms:a systematic review and meta-analysis.BMC Psychiatry. 2008;8:81

38. Kaniasty K, Norris FH. Help-seeking comfortand receiving social support: the role ofethnicity and context of need. Am J Com-munity Psychol. 2000;28(4):545–581

(Continued from first page)

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

Copyright © 2014 by the American Academy of Pediatrics

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

FUNDING: Supported by National Institutes of Mental Health grant RO1-MH086579A to Drs Shaw and Horwitz and by the National Center for Research Resources andthe National Center for Advancing Translational Sciences, National Institutes of Health, through grant UL1 RR025744. Funded by the National Institutes of Health(NIH).

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

e488 SHAW et al by guest on June 2, 2018www.aappublications.org/newsDownloaded from

DOI: 10.1542/peds.2014-0529 originally published online July 21, 2014; 2014;134;e481Pediatrics 

Stevenson and Sarah M. HorwitzRichard J. Shaw, Nick St John, Emily Lilo, Booil Jo, William Benitz, David K.

Prevention of Traumatic Stress in Mothers of Preterms: 6-Month Outcomes

ServicesUpdated Information &

http://pediatrics.aappublications.org/content/134/2/e481including high resolution figures, can be found at:

Referenceshttp://pediatrics.aappublications.org/content/134/2/e481#BIBLThis article cites 31 articles, 6 of which you can access for free at:

Subspecialty Collections

y_subhttp://www.aappublications.org/cgi/collection/psychiatry_psychologPsychiatry/Psychologyhttp://www.aappublications.org/cgi/collection/neonatology_subNeonatologysubhttp://www.aappublications.org/cgi/collection/fetus:newborn_infant_Fetus/Newborn Infantfollowing collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

http://www.aappublications.org/site/misc/Permissions.xhtmlin its entirety can be found online at: Information about reproducing this article in parts (figures, tables) or

Reprintshttp://www.aappublications.org/site/misc/reprints.xhtmlInformation about ordering reprints can be found online:

by guest on June 2, 2018www.aappublications.org/newsDownloaded from

DOI: 10.1542/peds.2014-0529 originally published online July 21, 2014; 2014;134;e481Pediatrics 

Stevenson and Sarah M. HorwitzRichard J. Shaw, Nick St John, Emily Lilo, Booil Jo, William Benitz, David K.

Prevention of Traumatic Stress in Mothers of Preterms: 6-Month Outcomes

http://pediatrics.aappublications.org/content/134/2/e481located on the World Wide Web at:

The online version of this article, along with updated information and services, is

ISSN: 1073-0397. 60007. Copyright © 2014 by the American Academy of Pediatrics. All rights reserved. Print the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,has been published continuously since 1948. Pediatrics is owned, published, and trademarked by Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it

by guest on June 2, 2018www.aappublications.org/newsDownloaded from