[systematic review * examen critique systematique

[systematic review * examen critique systematique

ANTENATAL PSYCHOSOCIAL RISK FACTORS ASSOCIATEDWITH ADVERSE POSTPARTUM FAMILY OUTCOMES

Lynn M. Wilson, MD, CCFP; Anthony J. Reid, MD, MSc, CCFP; Deana K. Midmer, BScN, MEd, FACCE;Anne Biringer, MD, CCFP; June C. Carroll, MD, CCFP; Donna E. Stewart, MD, FRCPC

Objective: To determine the strength of the association between antenatal psychosocial risk factors andadverse postpartum outcomes in the family, such as assault of women by their partner, child abuse,postpartum depression, marital dysfunction and physical illness.

Data sources: MEDLINE, Cinahl, Famli, Psych Abstracts and the Oxford Database of Perinatal Trialswere searched for relevant articles published from Jan. 1, 1980, to Dec. 31, 1993, with the use ofMeSH terms "depression, involutional," "child abuse," "child neglect," "domestic violence," "family,""marital adjustment," "family health," "newborn health," "child health," "physical illness," "social sup-port," "psychosocial risk," "prediction," "risk factors," "obstetrics" and "prenatal care." Further articleswere identified from bibliographies.

Study selection: Of the 370 articles identified through the search, 1 18 were included for review. Studieswere included if they examined the association between psychosocial risk factors and the outcomes ofinterest. Articles were excluded if they were reviews of poor quality or they had one or more of the fol-lowing features: insufficient description of the sample, a high attrition rate, a lack of standardized out-come measures, outcomes other than the ones of interest or results that had already been reported in a

previous study.Data extraction: The strength of evidence of each study was evaluated. On the basis of the evidence, each

risk factor was assigned a rating of the strength of its association with each of the postpartum out-comes. The ratings were class A (good evidence of association), class B (fair evidence) and class C (noclear evidence). Of the 129 antenatal psychosocial risk factors studied, 15 were found to have a class Aassociation with at least one of the postpartum outcomes.

Data synthesis: Child abuse and abuse of the mother by her partner were most strongly correlated (classA evidence) with a history of lack of social support, recent life stressors, psychiatric disturbance in themother and an unwanted pregnancy. Child abuse was also strongly associated with a history of child-hood violence in the mother or her partner, previous child abuse by the mother's partner, a poor rela-tionship between the mother and her parents, low self-esteem in the mother and lack of attendance at

prenatal classes. Postpartum abuse of the mother was also associated with a history of abuse of themother, prenatal care not started until the third trimester and alcohol or drug abuse by the mother or

her partner (class A evidence). Child abuse had a fair (class B) association with poor marital adjustmentor satisfaction, current or past abuse of the mother and alcohol or drug abuse by the mother or herpartner. There was class B evidence supporting an association between abuse of the mother and poor

marital adjustment, traditional sex-role expectations, a history of childhood violence in the mother or

her partner and low self-esteem in the mother. Postpartum depression was most strongly associatedwith poor marital adjustment, recent life stressors, antepartum depression (class A evidence), but was

also associated with lack of social support, abuse of the mother and a history of psychiatric disorder inthe mother (class B evidence). Marital dysfunction was associated with poor marital adjustment beforethe birth and traditional sex-role expectations (class A evidence), and physical illness was correlatedwith recent life stressors (class B evidence).

Conclusions: Psychosocial risk factors during the antenatal period may herald postpartum morbidity. Re-search is required to determine whether detection of these risk factors may lead to interventions thatimprove postpartum family outcomes.

Drs. Wilson, Reid, Biringer and Carroll and Ms. Midmer are assistant professors in the Department of Family and Community Medicine, and Dr. Stewart is professor in the departments ofPsychiatry, Obstetrics and Gynecology, Surgery, Anaesthesia and Family and Community Medicine at the University of Toronto, Toronto, Ont. Dr.Stewart also holds the Lillian Love Chair of Women's

at Toronto Hospital, Toronto, Ont.

Reprint requests to:Dr. Lynn M. Wilson, St. Joseph's Health Center, 30 the Queensway, Toronto ON M6R 1B5

<- For prescribing information see page 947 CAN 1996; 154 (6) 785<-- For prescribing information see page 947 CAN MED ASSOC J * MAR. 15, 1996; 154 (6) 785

Objectif: Determiner le degre dassociation entre les facteurs de risque psychosociaux prenatals et les re-sultats nuisibles postpartum dans la famille, par exemple, agressions contre les femmes par leur parte-naire, mauvais traitement aux enfants, depression postpartum, problemes matrimoniaux et maladiephysique.

Sources de donn&ees: On a consulte MEDLINE, Cinahl, Famli, Psych Abstracts et la base de donneesd'Oxford sur les etudes perinatales pour y recenser les articles publies entre le 1er janv. 1980 et le 31dec. 1993 comportant les termes MeSH suivant: (<depression, involutional)>, ((child abuse>>, x(child ne-glect)>, ((domestic violence>>, <family>>, x(marital adjustment>>, <<family health>>, ((newborn health>),((child health>>, ((physical illness>>, x<social support>>, (<psychological risk>), ((prediction>), (<risk factors>>,((obstetrics>> et (<prenatal care)>. On a aussi releve dautres articles provenant de bibliographies.

Selection d'etudes: Des 370 articles trouves 'a la recherche, on en a choisi 1 18 pour recension. Une etudeetait retenue si elle examinait lassociation entre les facteurs de risque psychosociaux et les resultatsrecherches. Les articles etaient rejetes si lexamen etait de mauvaise qualite ou s'il comportait une ouplusieurs des caracteristiques suivantes description insuffisante de lI'chantillon, taux dattrition eleve,absence de mesures normalisees des resultats, resultats hors du domaine d'interet ou ayant dej'a etcites dans une etude anterieure.

Extraction de donnees : On a evalue la solidite des preuves apportees par chaque etude. On a ensuite at-tribue 'a chaque facteur de risque une note selon son degre d'association avec chacun des resultats post-partum. II y a eu trois notes: categorie A (donnees probantes de bonne qualite), categorie B (donneesprobantes de qualite moyenne). et categorie C (aucune donnee probante claire). Des 129 facteurs derisque psychosociaux prenatals etudies, on a juge que 15 presentaient une association A avec au moinsun des r6sultats postpartum.

Synthese des donnees: Les mauvais traitements faits aux enfants et les agressions contre la mere par sonpartenaire presentaient une tres forte correlation (donnees probantes de categorie A) avec un soutiensocial inexistant, de recents evenements stressants dans la vie, un desequilibre psychiatrique chez lamere et une grossesse non desiree. On a aussi releve une forte association entre les mauvais traitementsfaits aux enfants et des ant&cedents de violence chez la mere ou son partenaire dans leur enfance re-spective, des mauvais traitement faits anterieurement 'a des enfants par le partenaire de la mere, une re-lation peu harmonieuse entre la mere et ses parents, une pietre estime de soi de la mere et une non-participation 'a des cours prenatals. On a aussi associe les agressions contre la mere postpartum a desantec6dents du meme genre chez celle-ci, 'a labsence de soins prenatals avant le troisieme trimestre eta un probleme dalcool ou de drogue chez la mere ou son partenaire (donnees probantes de categorieA). On a repere une association de qualite moyenne (categorie B) entre les mauvais traitements faitsaux enfants et un niveau peu eleve de satisfaction ou d'adaptation 'a la vie matrimoniale, des agressionscourantes ou passees contre la mZere et des problemes dalcool ou de drogue chez la mere ou son parte-naire. Des donnees probantes de categorie B ont permis d'associer les agressions contre la mere 'a unepietre adaptation 'a la vie matrimoniale, a des attentes classiques au sujet du r6le des conjoints, a desepisodes de violence subis par la mere ou son partenaire pendant leur enfance respective et a une pietreestime de soi de la mere. On a constate une tres forte association entre la depression postpartum et unemauvaise adaptation 'a la vie matrimoniale, de recents evenements stressants dans la vie et une depres-sion antepartum (donnees probantes de categorie A), mais on la aussi associee 'a une absence de sou-tien social, 'a des agressions contre la mere et 'a des ant&cedents psychiatriques chez la mere (donneesprobantes de categorie B). On a associe les problemes matrimoniaux 'a une mauvaise adaptation 'a la viematrimoniale avant la naissance et aux attentes classiques au sujet des r6les de Ihomme et de la femme(donnees probantes de categorie A), et on a correle une maladie physique 'a de recents evenementsstressants dans la vie (donnees probantes de categorie B).

Conclusions: Les facteurs de risque psychosociaux au cours de la periode prenatale peuvent entrainer unemorbidite postpartum. Des recherches s'imposent pour determiner si le depistage de ces facteurs derisque peut deboucher sur des interventions qui amelioreront la situation familiale apres une naissance.

MA odern prenatal care has focused almost exclusively adian and US national guidelines for prenatal care recog-on the detection of medical and obstetric prob- nize this area.3 The guidelines state that the "prenatal

lems. This emphasis on clinical surveillance of the physi- period provides an opportunity to look beyond preg-cal health of the mother and the well-being of the fetus nancy and delivery and to marshall the resources essen-has been accompanied by a lack of attention to the so- tial for further healthy development of mother, infantcial and psychological needs and resources of the and family.", Adverse postpartum psychosocial out-mother and her partner.' However, there is increasing comes, such as postpartum depression,2 child abuse andrecognition of the importance of psychosocial issues neglect, and family violence,' are highlighted in theseduring the antenatal period. For example, recent Can- guidelines. In a thorough review of the association be-

786 CAN MED ASSOC J * 15 MARS 1996; 154 (6)

tween antenatal psychosocial factors and biomedicaloutcomes such as low birth weight, Culpepper and Jack4recommended that comprehensive assessment of antena-tal psychosocial risk be integrated into medical and ob-stetric care. They categorized psychosocial risks as de-mographic or social characteristics, psychological factorsand adverse health habits. They also reviewed preven-tion and management strategies for women at risk ofhaving psychosocial problems. However, neither the na-tional guidelines for prenatal care nor the review exam-ined the relation between antenatal psychosocial riskfactors and adverse postpartum family outcomes. Theydo not, therefore, provide clinicians with guidance in an-tenatal identification of individual parents or populationsat risk of having these problems.

Although many pregnant women and their familiesexperience significant difficulties during the postpartumperiod, the systematic detection of psychosocial riskfactors and incorporation of psychosocial informationinto routine prenatal care has not yet become universal.As a step toward this goal, we conducted a comprehen-sive literature review to determine the strength of theassociation of antenatal psychosocial risk factors withthe following adverse postpartum family outcomes:child abuse or neglect, assault of women by their part-ner, postpartum depression, marital dysfunction andphysical illness.No national data on reports of child maltreatment are

available in Canada.' The reported rate of suspectedchild abuse or neglect in the United States in 1987 was34.0 cases per 1000 children. It is difficult to estimatethe number of unreported and undocumented cases ofabuse. Existing estimates of the prevalence of child abusediffer, in part because of a lack of agreement on the defi-nition of maltreatment and on methods for measuring it.

In a recent Canadian study involving women seen in avariety of clinical settings, the prevalence of assault ofwomen by their partner during pregnancy was estimatedat 6.6% on the basis of an anonymous self-report ques-tionnaire.6 Less than 3% of the physically abused preg-nant women were identified by their prenatal careproviders, although injuries were often visible. In a fol-low-up to this study, one of us (D.E.S.) found that themean number of incidents of physical abuse per abusedwoman during the 3 months after delivery was signifi-cantly higher than the mean number of incidents du-ring each trimester or during the 3 months beforeconception.'

Studies in which clinical depression was measuredthrough semistructured clinical interviews and by stan-dardized diagnostic criteria have shown that approxi-mately 10% of women experience major depressionpostpartum.'0 Postpartum depression may increase therisk of later depression in the mother, which is associ-

ated with an increased risk of behaviour problems in thechild." As well, a prospective study involving a sampleselected during the antenatal period showed that chil-dren whose mothers experienced depression during thefirst year of the childs life obtained lower scores on testsof cognitive functioning than other children.2

Several prospective cohort studies involving randomor volunteer samples have shown that the transition toparenthood is often accompanied by a decline, whichvaries in duration, in the quality and functioning of themarital relationship.'3-'5 Women may experience morenegative changes than their partners. 15,16

Psychosocial stressors are associated with increases inindicators of early childhood morbidity, such as thenumber of medical consultations`7,8 and attendance at ahospital for treatment of lower respiratory tract illnesses,gastroenteritis, accidents, burns, scalds and accidentalpoisonings. 11

In preparing this article, five of us were each assignedto review the literature on one of the following postpar-tum outcomes: child abuse, assault of women by theirpartner, postpartum depression, marital dysfunction andphysical illness. These outcomes were chosen for studyon the basis of a preliminary review of the literature,which showed that they were significant concerns dur-ing the postpartum period. We searched MEDLINE,Cinahl, Famli, Psych Abstracts and the Oxford Databaseof Perinatal Trials for relevant articles published between

Pmsy it, ;a,jus.tm,n* -o- ;a-i-. tPoor marital adjustment or satisfaction-Lack of social suppoMtTraditional, rigid sex-role expectationsRecent stressful life events

DOmesic violence -:

Mother or her partner experiencdd orwitnssec violnce in.childhoodCurrent or past abuse of the woman byg tew partner :

Partner suspected of child abuse in past or of harsh discipline ofchildren

MatealPbor.relationship with parentsLow self-esteemPast or presentpsychiatric disorderUnwanted.pregnancy (after 20 weeks).Antepartum depresionPrenatal care not started until third trimesterRefused or quit prenatal classesSubsbnce abuseBy the mother or her partner

CAN MED ASSOC J * MAR. 15, 1996; 154 (6) 787

Jan. 1, 1980, and Dec. 31, 1993. The following MeSHheadings were used to search these databases: "depres-sion, involutional," "child abuse," "child neglect," "domes-tic violence," "family," "marital adjustment," "familyhealth," "newborn health,""child health," "physical ill-ness," "social support," "psychosocial risk," "prediction,""risk factors," "obstetrics" and "prenatal care." Further ar-ticles were identified through follow-up of the bibliogra-phies of these studies.

Articles were included for review if they examinedthe association between antenatal psychosocial risk fac-tors and the postpartum outcomes of interest. Many arti-cles were excluded because they only described the out-come, how to recognize or diagnose it or theoriesconcerning its cause. In addition, articles were excludedif they were traditional reviews (overviews), rather thansystematic reviews, or if they had one or more of the fol-

lowing features: insufficient description of the sample, ahigh rate of attrition of subjects from the study, a lack ofstandardized outcome measures, outcomes other thanthe ones of interest (e.g., postpartum "blues" as well asdepression) or results that had already been reported in aprevious study. In total 370 articles (88 on child abuse,97 on assault of women by their partner, 73 on postpar-tum depression, 70 on marital dysfunction and 42 onphysical illness) were identified through the search. Af-ter reviewing the articles with the use of the exclusioncriteria, we retained 118 articles (22 on child abuse, 34on assault of women by their partner, 27 on postpartumdepression, 24 on marital dysfunction and 11 on physi-cal illness).We developed an analytic framework for the evalua-

tion of the articles by modifying the approach of theCanadian Task Force on the Periodic Health Examina-

.Outcome; class of association"Assault of

mother by her: Postpartum Marital PhysicatAntenatallactor;: -- Child abuse partner deipression dysfunctions gilnsPoor ^;ntarftaadjustfientor satisfAction B223.24 A'0', A'5"s No tiies:

lavailable,Lack of sociat support A B-A7B273'4 No studies C

available

Traditional,' rigid sex*role expectations No studies B2324."AMZ * No studies AM553w,.5 :No studiavailable available available -

A22,37^4',F"-7 A.t49 A24e0161' No studies B.7available

Mother or her partner experienced A22v7342? B"""".'6 No studies No studies . Nq iSror witnessed violence in childhood.. available available ailable

Current or past emotional, physical, B37'm A6.62,65 B7 No studies .INostu,di.sexual or verbal-abuse of the mother available availableby her partner

Partner suspected of child abuse in A22,37238,55 57 No studies No studies No studies No studiespast or of harsh discipline of children available available available avalAble

Poor relatiohshliof motier with her A'"' No studies C'0"""' No studi No;tudiesparents available;- availabe aeiila ;

Low selfpsteerp inthe-mother A"""5" B; No studies Nostud'es No sidl;s. % - .. I . .. . ' -! d ;, ' -;ava|lable ava>ilabe ; a;;lhilable , t

Past or 4resent psychiatric disorder A2240A2A3.s, A.M.5,2MJo8."-'"':' Notit4 A ItostB în the m9ther ; avaRable> available

Prency unwanted bthe mother A22t ,. A" %)."$G6 t)5 i$C B ' NW " V*0 JUi -':(after.20 weeks) 4 v` 'a v . .., . r . a

Antepartum depression No studies N iAstu1s? r"A0 'PIPhA 1f ilestu 'iacd

5 - 5J <aveilabl&; 't;4*ila4l -i *yf8114b1;,3, ir9;3i-ts9V$!$- r&c,3- 2Prenatal care not started'until the thiid. No studies: A:'7'5 C No studies Nost4etrimester .avalable .. ., .......,. avaiFl -i-7 e

,Mother refused,c 6quitp a p es., 1 ¶A .stuim \'!A b$ N*flI i

Aloolodu abse .by.the ote r '".g,J '14W 3)cA tA~; M

or her partner . n41 2 a4ailbe - avatla .,'*Class A=gd evidence of a iation c 'CB = fa idenc andcla c =-,n cii- tiidnc


Aw 11.

(l -I--) P croetropoliktan Denve'r

M m. er et

evens1volece-i1n. chil d-

U*Io Tl

viio lence. chi ,k;ihoodW'of t her, chil.d

StraussL:W oOcotrl, 01114 ..br~(I1-2it - tjw~~~~~~tfntb wttee* efrpr,ed the aatle

Miner. et. aP o-onrl I0sM c Ss a ---ii$*' Al$ oprsc lda.e

uban ~ T9t ttufet o.*iiaw

o n & . peer tetnSchfiWtO r~trto? itt nf 1. blt

weLl-baby clinicsv ~'~h~.p~,lI,bt *ara1~tem

n Oai#ontro 410(01-2) '-4b Eb4A 'irdlf t Is -

i&09ft,&I!+

and bad a-% urcPt~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~......1~ ~ ~ ~~ ~~~"t F*~ TW..$ W0 -t

easchildhood, psychiatrc -:r~~~~~~~~~~~~~~~~.tG o o3Q4t;'

red0~-uictive abilityw

se

~~~In ild in

theslniy PItA.5'i4iMor.Ir

unwanted

64ip

Shr7 wijhery

'Sh:ort 2yer. follow-~up; res.ults


*5m.text for. definition.of classe 6f eldebnce. . .*.tFwctcrsfumTa'ble:1 veg- iven & fesert * .. *

.1"111'..-.., -t. %r-M Ir 1, :,. .,'tc%, o-.. -. g. -- 1 `. ,, .. '. .1' ... 1.._ ., :- ..: .- ... f: 1. .,, p-.. :.. .:-- SiC. -;. ;,;:. ;,, n -1- :.,-.;.t-.. ..52: .,i.-

. .- - 1. .' f m, . '. .- k le 1 I. -. . -

... :I 'I'!' ",. '. .:... .-- :..!:... 11 :- ... :.,-

mi l' .... 1, ., .. -..i

4 4 1:ik, -i.i me..'mi"99 -11.. i.. -"-.7

-,

Lk5.1

tion.'9 The strength of evidence presented in each study * Class II-i: evidence from prospective cohort stud-was classified as follows. ies.* Class I: evidence from randomized controlled trials or * Class II-2: evidence from case-control studies or sys-

well-designed natural'history studies.20 tematic literature reviews.2'

of- evidence); Sample f.ac -. wi tstors- :LimitSlot>f};;i;|--Stevvsrt at ar Prospective 548 mothers seen at antenatal A * frRecailbias114)- cohort clinics in various locations in ihitfldtsdder,

Ontario i*i thbuse;NOt a :pplicableL :arital: 1:t, sex , eSO liin. ;dTt*s

(1444" s.ht;i4 vbtbf 82 ; tS ;vt ence S; gychologi ilittitwC t9tgwit'>^'< rsl tI i - childod towoef-;

4S~~I¶)~W~ ru~~Jkngpj~lg esteemw Cs tr 10 subjects,in Orange and Los. Msta*umnt, sex, Somenonvalidate~a t'bW

fr;ot"tV '' f Angelescounties; n Catifornia.; lAoeo inme* t-|'v i~ a. cases were couples in counselling childhood, alcohol or gop differed fami

tor.rwif assult, contrls were drug abuse incoenonviolent couples

#S Descriptive 49 women seen in:an.emergency .oia supp..tsex roles, Selection bias,noCOIl¾)'t;al'iA(*$ ;~,'; rU'-sdepartment in Sweden vs;iol Wechdhood, group

tiiSit'f#e0'cerwol 50 subjects; aseswerefrom a '. a upport,sex rolesi. Snats*r'i*ai1 $X~.e; .S4 J(W ,t._ r spouse-abuse shelter,controls bias Wi 'A

were s.een.t ..a health clinic ip oly vd 1 w* 4; 'Tampa, Fla: with a low socioe omic

status

,,i.¢~-'p~pp~tive 1243 mainrily r black or ocialqport, life Resls no ge& 1 >(11Z*&¢> qyR:star Hispanic, iner-city mothers seen-:e,nt-. sychiatic

at prenatal clinics associated with, disorder, unwantedthe Boston City Hospital preg ancy

Campbell;- Case-control 488 wornen in a metropolitan areaS'ctalpt, Results-notguetar'(W2t) :- of the midwestern United States, psycbidtr isorder, late hig(4 ).!X-DdC2 f ti-.S::iseen-2 to5 day post um rea acohol ore;.. t. e

-----------S---ampl M I

.antort 7a Case-control. 2033 women interviewed by rin Seifrpott.1J12) telephone in a national probability < .hI*4&*oNlor

sarnmpe of randomly selected. - dru abuseuhoseholds:in theUnited Sttes ., : -

Buti#.> $ >i$t teseAo4irol 797 subjectsee n a(tPlanned f "eit ric Dm h=

l'arnthdd linis i Teas dlbdrcases ad cenrl ocomprd

: >::9-: ;;; '-' ;; c m:i~t .&,k ia A

J W-r ae'CftO 14t) suDjcs-cases-C were Trorn L1.T eveuc, jJtalrIdI. oinauq;wni'(ltt 1j r'1{ ' .} . Onai sh3tr it abse dile boifttelo Fi! for^ ngtv IIt l;--*F8---S;n - wometcntroliie- recruited events

fthough .a newspopeadverisement'., naaw&aAas, 'l-'tHr--Inr .. '' > .e'0in . . . | l.,.s


! l ' ' $ j w _ __, ,; ' j g 0 | 8 Z_el S _ _

* Class II-3: evidence from studies comparing differentsites and populations or from descriptive studies.

* Class III: evidence from traditional reviews(overviews) or from opinions of experts in' the field.

4A"MJ!4b t'r ~~~~1 tt,

(11:2)T

Each article was reviewed and classified by one of us.To ensure that the classification procedure was reliable, theother authors independently read and classified three tofive of the articles in each of the outcome areas. For each

*

.

ospitat,131 wna* t"

gruency~~wbttt8;i~~A- cCaw'uk9flr$ ~16'1

&.pturi--`

Stewart Case-conrol 22puestp.pit

Bland. etE ')C(cN$' fo

r -~~~~~~6117r, (11~~ ~ t t!clE sro Id 2et. p 00~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~k

*~~~~~~~~~~~~IA 1 .1. ''- 2ii3~Slkf $I)<!

n, fl.' *bni $tJ aictjjrAlcoho o'r abuse..

zy~~~~~ooo~ ;t4 sinf:ntII4~Berthb*PSV ~~~~~~~~~O'~~~be#tol SQLi' tic i~~~~~~~~~~~~~~.I4 " *t~~~~~~~ -' ~~~~btpo4~~~~~'hhizabW4l<C'~~~~~~~J

Miller~~~~~~~~~~~~~~~~~~ieta"Ce'ttt 5 uj0tb'Wtt;wA rht&kte $n1fi~t;cssad(It.2t casewere ~~~~~aWht '~~"~~ cp*tulsdiftued~R

*8 .. * *f4la dtfX. ~ ~ ~ ~ ~ *.i

----- '* c * . - v -


IA.

*Ii

1------ - --- _. i ` k, 1 '. ---P ., - 1g o . r. '. oi i.

., - .. 1 1. ..M.

cl J 1 -- J.

article, we prepared a summary that included the site, thepopulation, the design, the classification of strength of evi-dence and limitations. Antenatal psychosocial risk factorsfound to be associated with the relevant postpartum out-come were then listed and given the classification of therelevant article.

After all of the articles concerning a particular postpar-tum outcome had been reviewed, the article summarieswere circulated to the other authors. Each antenatal psy-chosocial risk factor was assessed by the whole group andassigned a rating according to the class of evidence sup-porting its association with each of the five different post-partum outcomes. Consensus was easily achieved in morethan 90% of these cases. More difficult cases were re-solved after discussion. Nodence.

studies involved class I evi-

On the basis of this review, we assigned ratings toeach risk factor with the use of the following guidelines.* Class A: There is good evidence of an association be-

tween the antenatal psychosocial risk factor and theoutcome. This includes any risk factor for which atleast one class 11-1 and one class 11-2 study demon-strate the association.

* Class B: There is fair evidence of an association be-tween the antenatal psychosocial risk factor andthe outcome. This includes any risk factor forwhich only class 11-2 studies or only one class 11-1study and no class 11-2 studies demonstrate the as-sociation.

* Class C: There is no clear evidence of an associationbetween the antenatal psychosocial risk factor andthe outcome. This includes any risk factor for which

Stdy (clqes;*M^4t . Sampe:^ *Aseo;i*4-.ctors imitatons,lt; 'â

Kumard titie 119patients seen at an iMaiiaWitnt, Resultsnotgenertlizab le;w t(Ill)*>; '' i' -*tR{qt/S)P#ti;))*>l{ antenatal clinic in London, m rp; may gh t ou clnic

England, most.of whom rbetiovl ; bcause of its reptatii.-werer e-oruppr-c .;l

(YHara &tISQC "%stive 362 patients seen at a Marialdu fil Included minor arid aJ ht(Il-1) cbhbtt public obstetrics/gynecology events,-ichiatric validity and reliabilityp e

(ob/gyncliicand o disdq a mesures no stated"4- \ x.r ; si.<t;>fprivate practices at a depression

university teaching hospitalin Iowa

WI$ffs1¶ -r.&.l. Propective 'I4 mothers recrid fm Marital ta4jrnntlfe Lw-.(7j>p s.r'i-(114>.} ).e--,.;; X>t; ! 5 ciohort public health prenatal event, nonrendent no ch iiqe4

--'-,r, ,;, s', classes in Vancouver depression'OF:Har et aP' lul5d 1.70 patients.seen at a Marital adjustment, social. Incl'uded mjr admno SS(il-0s t %j;sf r§i).fcoof publcob n cinic and . supi fe validity ; tiI

private practices ps d der newkmare unciwWin lowa

PayIti btilX aSe-nleatrol 120 white women attending Marital adjustment, life Two interiewers; no(11i-2) ^ ;t + i 5 - r thospital postnatal cl.inics in events, psychiatric . determinationofie e

London, England disorder reliability; validity, relality-,... *.

-

oi.I;fRa T ;r N

;'Scale and tev fO e?t Life Evens ndt std --

Watson etaP- - Pq*tive 128 women seen t an . MAitdt1ver lie M!hcta wi(I-l) . . .r.,-i)cohort antepartum clinic: in south events, p hiatric psychololog as mn .

London, England, before- 24 disorder, antepartum-nweks' gptatio depression

O.H v tAiiqptRW

practices in' f

Soyceat'ai" ,; Pqspeotive 170 patients seen at.a MaritaVNs3fit Women cu if they ,vp(114-)1,tis.: 'sm'- ;&A.public..atenity hsplta1 in:. trinw,; Eibr?

tntY4~~dtwt-v . ~~ Sydniey, Australia teapvsw

Edinburgh, Scotland' random


there is only class 111 evidence or extremely conflict-ing evidence.

RESULTS

We studied a total of 129 antenatal psychosocial riskfactors (Appendix 1) and their relation with the postpar-tum outcomes of interest. Of these factors, 15 were

found to have a class A association with at least one ofthe postpartum outcomes (Table 1). These risk factorscan be grouped into the general categories of family fac-tors, domestic violence, maternal factors and substanceabuse. The various definitions of these risk factors andthe instruments used in the studies to determine theirpresence are listed in Appendix 2.

For the 15 strongest risk factors, the strength of evi-

dence concerning their association with each of the fivepostpartum outcomes of interest in this review is pre-

sented in Table 2.

CHILD ABUSE

Child abuse was most strongly correlated (class A evi-dence) with a history of lack of social support, recent lifestressors, a history of childhood violence in the mother or

her partner, partner suspected of child abuse or of harsh dis-cipline of children, a poor relationship between the motherand her parents during childhood, low self-esteem in themother, psychiatric disturbance in the mother, unwantedpregnancy and lack of attendance at prenatal classes (Tables2 and 3). There was fair (class B) evidence of an associationbetween child abuse and poor marital adjustment or satis-

,.

il

.d ol

...; t:`.be

et

arid

'.4

Stein'(11-1), -' ., .. coh ...

O'Hara at61 Prospective

(I~~~~)s

1-) rt E @..v-..Sf

m'fatme

.,, atter%d.ng

-Soci$,2;P

,m ~ .`_&W *f .,ice, 0-4 "..I-itii* ._.,A .,L

hein.

a: quLife -t

lwa i} tt:j0g>st9/ 6 W 4~~pr

et As"-X., ;.;-¢: pr5. p, Cticea *1e

Zee*iid nt re-q'JrdO'Neill etalt2 Case-cont pu1bic. : b

O'Har'it'f ';lRizXi PIk ve 99 womFien PObicpublic A01-1) . c. .. GOh.or....gti *o iian4 te , .-~ ! ,

. ~~prtce in Iowa *-}>


(1l-1)~7kSi '-, ,sr of , ridon s,.|!t..........,, _ ........,.- ..._. ,......., ... .. ,,.......-.Ânw.

T * -

' 11, ',- 14&i,., 1.4 IAC 01 i.-ii.,r ii4 .a'l .WW'. 1 A` 4, 7.4 J+. i.tm".. -I .o ..7.

{'sit;, _'4 -A' I

'7;9' ..W-ZY.j

--- 1. -i, l,: 't lt"I -13 -11 .. a -- i11 t.:.c 4W

faction, current or past abuse of the mother by her partnerand a substance-use disorder in the mother or her partner.

Definitions of child abuse differed among many of thestudies. Some data were obtained from case records ratherthan from structured interviews. The subjects of manystudies were limited to populations that were demographi-cally distinct (in terms of location, race, social status ormarital status). Many of these studies used a case-controldesign, and some did not specifically examine the pres-ence of these risk factors during pregnancy. However,many of the findings from these retrospective studies wereconfirmed by a prospective, double-blind cohort study in-volving 1400 pregnant women, which employed a stan-dardized interview format.57 Many of the studies of childabuse examined the sensitivity of checklists of risk factors,rather than of individual factors, in predicting child abuse.

ASSAULT OF WOMEN BY THEIR PARTNER

There was good (class A) evidence of an association

between assault of women by their partner and lack ofsocial support, recent stressful life events, current or pastabuse of the mother, current or past psychiatric disorderin the mother, unwanted pregnancy, inadequate prenatalcare and alcohol or drug abuse by the mother or herpartner (Tables 2 and 4). Abuse was also associated, al-though more weakly, with poor marital adjustment orsatisfaction, traditional, rigid sex-role expectations, a his-tory of childhood violence in the mother or her partner,and low self-esteem in the mother.

There were many methodological issues in this areaof research. Different definitions of abuse were used todefine the samples of abused women. Women were re-cruited through different settings (e.g., women's shelters,emergency departments or physicians' offices) andmethods (e.g., telephone surveys and advertisements).The accuracy of self-reporting may have been affectedby the fact that having been abused is considered so-cially undesirable, by the respondents' perceptions andby their faulty recall. Studies often neglected to report

:S, l:~~~~~~~~~~~~~~~~~~~~Y7 9

of.v$ane); Design Sample.. vr'rx Limitations

21,- ........... .. i i ;- 0 , fl~~~~Z -tl..".

i g >,se-.contri Randomsam re of,511 M.iiUadjsmei I"e reteof*%tt)5tx:(i>J4f18 $r9Fi JInl 2){;~ti.- i. - !'icouples in Minnesota madeiI tio,rnai*."-mi

nonrespondents eprinemore distreshnm esodet

Law.s~ (11 1) Prospectivp 38 couples from Timberlawn Mari d tentS.ma e -Jr^ l-t-eor>t Psychiaic R . i ;

Foundation, Dallas M: 5rk*........s; nbt

* j ,~~~~~chlre lsuffY. * t dehalt" -conenng sttnfins~~~~~~~~~~~~~~~~~~~~~~.. #;.St\;t8q>;;' ~*

Tomti r" ~ cte 96. volunteer.couples,.94% Of Marital, adijustnmet sex. Results ntgnrI~~~1 - --!----.-- --

- -----.. _.....- r ;fa+;i .... .....__-i - "lM

(Il-I)!;.l::,la'' PwhomwereWhiteahd'middle- rofile"s:class, attendingchildbirh

* . . tJ*J->@w;-; ;'Y' classes (location not specified)*.r.(i( ... ......$i i;XE.;r...............;'2dr,;r.i.

;.thi' .3.1>,*.W..~~~~~

3~ ~~~~~~~~~.*toxF :,'),@l S, ,1 e J tes

61 couples recruited Tor the R*'~esi.tS ea;-PdnmjIvtaniatn*idFamily . 1;small sample:Development Pr*t,'!l of -.whom were white, middle-class, : . -,iX .v -. -weId Pt ...J o;...t!t..gZ) d *1 5'vt

*^1i ti5un ig 5l;(@5\ - i. . j; d5L4 9iF} 2i9f.1 ti;ti2 s dl[.T P.1J095di-<g

ea i A. .::I<S*j.s

~ '' t ; :'

zeiaiand:- f o -' :"k;':.


- 1|j 1w__---1-...... . r, i0w.;-,--w-=.I.- . I. .. * !5 1;.' ',., _ ..'.-, --t -- yb-

- -~~-~~~--- Nl --B-----^-- -r -.'-- ;;î-t- 1

.1 C ' ll.---... 7T Z. ., .! ---, -.. 7.

negative findings. Some of the studies included in thisreview involved samples of women who were not preg-nant. Because violence may occur throughout a woman'slife and women who report abuse during pregnancy areoften abused preconception and postpartum as well,6these studies were retained.

POSTPARTUM DEPRESSION

There was good evidence of an association betweenthe following antenatal psychosocial risk factors andpostpartum depression: poor marital adjustment or satis-faction, recent stressful life events and antepartum de-pression (Tables 2 and 5). There was fair evidence of anassociation between postpartum depression and lack ofsocial support, current or past abuse of the mother by herpartner and current or past psychiatric disturbance in themother.

There were more prospective cohort studies concern-ing postpartum depression than concerning the otherfour postpartum outcomes in this review. However, manyof the studies of postpartum depression were conductedat a single hospital or clinic, restricting the generalizabil-ity of the results. Studies used different criteria to diag-nose depression, although we included only studies inwhich standardized instruments were employed. For ex-ample, some studies relied primarily on symptom-severityindices, which may be confounded by the normal physi-ological changes during pregnancy and the puerperium.In some studies, the repeated antenatal assessments mayhave provided social support and helped prevent postpar-tum depression. Results may have varied because womenwere recruited at different intervals before childbirth andbecause depression was diagnosed at different points dur-ing the postpartum period. Different risk factors may berelevant during different periods; as a result, the compari-son of results of the studies may be influenced by thetiming of the postpartum assessment.

MARITAL DYSFUNCTION

Postpartum marital dysfunction was associated withpoor antenatal marital adjustment and traditional, rigidsex-role expectations (class A evidence, Tables 2 and 6).Many of the samples consisted of white, middle-class,married volunteers, often recruited from prenatal classesor antenatal clinics. Therefore, the generalizability ofthese results to other groups is limited.

PHYSICAL ILLNESS

There were very few studies examining the relationbetween antenatal psychosocial stressors and physicalillness in family members during the postpartum period.

There was fair (class B) evidence of an association be-tween recent stressful life events and illness in childrenfrom a cohort study of 1082 children 1 to 4 years ofage'7 (Tables 2 and 6).

DISCUSSION

This comprehensive, critical appraisal of the literatureyields 15 antenatal psychosocial risk factors that are as-sociated with an increased likelihood of one or more ad-verse postpartum family outcomes. Unlike most studiesof antenatal psychological and social factors, which tendto focus on biomedical outcomes, the outcomes of inter-est were primarily psychosocial. This review is a step to-ward broadening routine prenatal care to include sys-tematic, evidence-based detection of psychosocial riskfactors and to incorporate psychosocial information.

The limitations of this review should be noted. Thereare no agreed-upon "gold standards" for the definitionsof the risk factors or the outcomes. Many of the studieswere limited in regard to the site (often a tertiary carecentre), the population examined (e.g., inner-city, black,poor children in studies of child abuse or middle-class,white, married couples in studies of marital dysfunction)or the number of subjects. Many of the studies of childabuse examined whether the outcome could be pre-dicted with the use of existing scales or tools involvingmultiple factors. It was therefore difficult to determinethe influence of one risk factor in isolation. The studiesof abuse of women by her partner mainly involvedwomen who were not pregnant.

Furthermore, the risk factors identified are those forwhich there were sufficient data. There has been a ten-dency to study maternal factors and to place much lessemphasis on paternal, family and community influences.Because of this tendency, there is the potential for"mother-blaming" and underestimation of the role of thefather. As shown in Table 2, many areas have not beenstudied.

Priorities for future research include determining theprevalence and distribution of adverse postpartum out-comes in general populations, case finding for these ad-verse outcomes in high-risk populations, determiningfactors associated with positive postpartum outcomes(which may serve as buffers when psychosocial risk fac-tors are also present), and developing and evaluatingpreventive interventions.

Association does not necessarily imply causality. Aswe have mentioned, some of the risk factors have notbeen studied independently, and many of them are inter-related.

This review summarizes the associations between an-tenatal psychosocial risk factors and postpartum familyoutcomes. It will be important to focus future efforts on


systematic identification of mothers and families at riskas well as on interventions offered to them. As a.step.to-ward this goal, we are developing a form to enablehealth care providers to screen routinely for psycho-social risk factors during the prenatal period. Ideally, thisenquiry will become a standard part of preconceptioncare. We hope that such a form will help educate healthcare providers, serve as a reminder to ask patients aboutrelevant psychosocial problems, facilitate the collectionand recording of information, help target allocation ofresources and funds to high-risk populations and ulti-mately lead to interventions that improve postpartumfamily outcomes.

1. Elbourne D, Oakley A, Chalmers 1: Social and psychologicalsupport during pregnancy. In Chalmers 1, Enkin M, KeirseMJN (eds): Effective Care in Pregnancy and Childbirth, OxfordUniversity Press, New York, 1989: 221-236

2. Family-Centred Maternity and Newborn Care, Department of Na-tional Health and Welfare, Ottawa, 1987

3. Rosen MG: Caring for Our Future: the Content of Prenatal Care: aReport of the Public Health Service Expert Panel on the Content of Pre-natal Care, Public Health Service, Department of Health andHuman Services, Washington, 1989

4. Culpepper L, Jack B: Psychosocial issues in pregnancy. PrimCare 1993; 20: 599-619

5. MacMillan HL, MacMillan JH, Offord DR and the Can-adian Task Force on the Periodic Health Examination: Peri-odic health examination, 1993 update: 1. Primary preven-tion of child maltreatment. Can Med Assoc J 1993; 148:151-163

6. Stewart DE, Cecutti A: Physical abuse in pregnancy. CanMedAssocJ 1993; 149: 1257-1263

7. Stewart DE: Incidence of postpartum abuse in women with ahistory of abuse during pregnancy. Can Med Assoc J 1994;151: 1601-1604

8. Campbell SB, Cohn JF: Prevalence and correlates of post-partum depression in first-time mothers. J Abnorm Psychol1991; 100: 594-599

9. O'Hara MW, Zekowski EM, Philipps LH et al: Controlledprospective study of postpartum mood disorders: compari-son of childbearing and nonchildbearing women. J AbnormPsychol 1990; 99: 3-15

10. Kumar R, Robson KM: A prospective study of emotionaldisorders in childbearing women. Br J Psychiatry 1984; 144:35-47

11. Philipps LHC, O'Hara MW: Prospective study of postpar-tum depression: 4 A-year follow-up of women and children.JAbnorm Psychol 1991; 100: 151-155

12. Cogill SR, Caplan HL, Alexandra H et al: Impact of mater-nal postnatal depression on cognitive development of young

children. BMJ 1986; 292: 1165-1167

13. Belsky J, Spanier G, Rovine M: Stability and change in mar-riage across the transition to parenthood. J Marriage Fam1983; 8: 567-577

14. Miller BC, Sollie DL: Normal stresses during the transitionto parenthood. Fam Relat 1980; 29: 459-465

15. Russell C: Transition to parenthood: problems and gratifica-tions. J Marriage Fam 1974; 4: 294-301

16. Waldron H, Routh D: The effect of the first child on themarital relationship. J Marriage Fam 1981; 43: 785-788

17. Beautrais AL, Fergusson DM, Shannon FT: Life events andchildhood morbidity: a prospective study. Pediatrics 1982;70: 935-940

18. Aagaard J, Husfeldt P, Husfeldt V: Social factors and lifeevents as predictors for children's health: a one-yearprospective study within a general practice. Acta PaediatrScand 1983, 72: 275-281

19. Canadian Task Force on the Periodic Health Examination:The periodic health examination. Can Med Assoc J 1979; 121:1193-1254

20. Department of Clinical Epidemiology and Biostatistics, Mc-Master University Health Sciences Centre: How to readclinical journals: 111. To learn the clinical course and progno-sis of disease. Can Med Assoc J 1981; 124: 869-872

21. Oxman AD, Guyatt GH: Guidelines for reading literaturereviews. CanMedAssocJ 1988; 138: 697-703

22. Murphy S, Orkow B, Nicola RM: Prenatal prediction ofchild abuse and neglect. Child Abuse Negl 1985; 9: 225-235

23. Hotaling GT, Sugarman DB: An analysis of risk markers inhusband to wife violence: the current state of knowledge,Violence Vict 1986; 1: 101-124

24. Telch C, Lindquist C: Violent versus nonviolent coiuples: acomparison of patterns. Psychotherapy 1984; 21: 242-248

25. O'Hara MW, Schlechte JA, Lewis DA et al: Controlledprospective study of postpartum mood disorders: psycho-logical, environmental, and hormonal variables. J AbnormPsychol 1991; 100: 63-73

26. Whiffen VE: Vulnerability to postpartum depression: aprospective multivariate study. J Abnorm Psychol 1988; 97:467-474

27. O'Hara MW, Rehm LP, Campbell SB: Postpartum depres-sion: a role for social network and life stress variables. J NervMentDis 1983; 171: 336-341

28. Paykel ES, Emms EM, Fletcher J et al: Life events and socialsupport in puerperal depression. Br J Psychiatry 1980; 136:339-346

29. Watson JP, Elliott SA, Rugg AJ et al: Psychiatric disorder inpregnancy and the first postnatal year. Br J Psycbiatry 1984;144: 453-462

30. O'Hara MW: Social support, life events and depression dur-ing pregnancy and the puerperium. Arch G^ Psychiatry 1986,43: 569-573


31. Boyce P, Hickie 1, Parker G: Parents, partners or personal-ity? Risk factors for post-natal depression. J Affect Disord1991; 21: 245-255

32. Cox JL, Connor Y, Kendell RE: Prospective study of thepsychiatric disorders of childbirth. Br J Psychiatry 1982; 140:111-117

33. Lewis J: The transition to parenthood: 1. The rating of pre-natal marital competence. Fam Process 1988; 27: 149-165

34. Tomlinson PS: Spousal differences in marital satisfactionduring transition to parenthood. Nurs Res 1987; 36: 239-243

35. Grossman FK, Eichler LS, Winickoff SA: Pregnancy, Birth andParenthood, Josey-Bass, San Francisco, 1980

36. Milner J, Wimberley R: An inventory for the identificationof child abusers. J Clin Psychol 1979; 35: 95-100

37. Anderson CL: Assessing parenting potential for child abuserisk. Pediatr Nurs 1987; 13: 323-327

38. Strauss MA: Stress and child abuse. In Kempe CH, HelferRE (eds): The Battered Child, University of Chicago Press,Chicago, 1980: 86-102

39. Milner JS, Wimberley RC: Prediction and explanation ofchild abuse. J Clin Psychol 1980; 36: 875-885

40. Schneider CJ: The Michigan screening profile of parenting.In Starr RH (ed): Child Abuse Prediction: Policy Implications,Ballinger, Cambridge, Mass, 1982; 6: 157-174

41. Zuravin S: Child maltreatment and teenage first births: a re-lationship mediated by chronic sociodemographic stress? AmOrthopsychiatry Assoc J 1988; 58: 91-103

42. Kotelchuk M: Child abuse and neglect - prediction andmisclassification. In Helfer RE, Kempe CH (eds): Child Abuseand Neglect: the Family and the Community, Ballinger, Cambridge,Mass, 1976: 67-103

43. Schneider C, Hoffmeister JK, Helfer RE: A predictivescreening questionnaire for potential problems in mother-child interaction. In Helfer RE, Kempe CH (eds): Child Abuseand Neglect: the Family and the Community, Ballinger, Cambridge,Mass, 1976: 393-407

44. Bergman B, Larsson G, Brismar B et al: Aetiological and pre-cipitating factors in wife battering: a psychosocial study ofbattered wives. Acta Psychiatr Scand 1988; 77: 338-345

45. Lewis B: The wife abuse inventory: a screening device forthe identification of abused women. Soc Work 1985; Jan-Feb32-35

46. Amaro H, Fried LE, Cabral H et al: Violence during preg-nancy and substance abuse. Am J Public Health 1990; 80:575-579

47. Campbell JC, Poland ML, Ager J: Correlates of batteringduring pregnancy. Res Nurs Health 1992; 15: 219-226

48. Martin GJ, Brown GW, Goldberg DP et al: Psychosocialstress and puerperal depression. J Affect Disord 1989; 16:283-293

49. Cutrona CE, Troutman BR: Social support, infant tempera-ment, and parenting self-efficacy: a mediational model ofpostpartum depression. Child Dev 1986; 57: 1507-1518

50. Cutrona CE: Social support and stress in the transition toparenthood. J Abnorm Psychol 1984; 93: 378-390

51. Stein A, Cooper PJ, Campbell EA et al: Social adversity andperinatal complications: their relation to post natal depres-sion. BMJ 1989; 298: 1073-1074

52. Kantor GK, Straus MA: Substance abuse as a precipitant ofwife abuse victimizations. Am J Drug Alcohol Abuse 1989; 15:173-189

53. LaRossa R, LaRossa M: Transition to Parenthood, Sage, BeverlyHills, Calif, 1981

54. Belsky J, Lang M, Huston T: Sex typing and division of la-bor as determinants of marital change across the transitionto parenthood. J Pers Soc Psychol 1986; 50: 517-522

55. Egan TG, Monaghan SM, Muir RC et al: Prenatal screeningof pregnant nothers for parenting difficulties: final resultsfrom the Queen Mary Child Care Unit. Soc Sci Med 1990;30: 289-295

56. Muir RC, Monaghan SM, Gilmore RJ et al: Predicting childabuse and neglect in New Zealand. Aust N Z J Psychiatry1989; 23: 255-260

57. Altemeier WA, O'Connor S, Vietze P et al: Prediction ofchild abuse: a prospective study of feasibility. Child AbuseNegi 1984; 8: 393-400

58. Bullock L, McFarlane J, Bateman LH et al: The prevalenceand characteristics of battered women in a primary care set-ting. Nurs Pract 1989; 14: 47-55

59. Jaffe P, Wolfe D, Wilson S et al: Emotional and physicalhealth problems of battered women. Can J Psychiatry 1986;31: 625-629

60. MacEwen KE, Barling J: Multiple stressors, violence in thefamily of origin and marital aggression: a longitudinal inves-tigation. JFam Viol 1988; 3 (1): 73-87

61. O'Hara MW, Rehm LP, Campbell SB: Predicting depressivesymptomatology: cognitive-behavioural models and post-partum depression. J Abnorm Psychol 1982; 91: 457-461

62. Hilberman E, Munson K: Sixty battered women. Victimol IntJ 19778; 2: 460-470

63. Gondolf EW, Foster RA: Wife assault among VA alcohol re-habilitation patients. Hosp Community Psychiatry 1991; 42:74-78

64. Marshall LL, Rose P: Premarital violence: the impact of fam-ily of origin violence, stress, and reciprocity. Violence Vict1990; 5: 51-64

65. Hilliard PA: Physical abuse in pregnancy. Obstet Gynecol1985; 66: 185-190

66. Goldberg WG, Tomlanovick MC: Domestic violence vic-tims in the emergency department: new findings. JAMA1984; 251: 3259-3264


67. Stuart EP, Campbell JC: Assessment of patterns of danger-ousness with battered women. Issues Ment Health Nurs 1989;10: 245-260

68. Stewart Helton A: Protocol of Carefor the Battered Woman: Preven-tion of Battering During Pregnancy, March of Dimes Birth De-fects Foundation, New York, 1987

69. Poteat GM, Grossnickle WF, Cope JG et al: Psychometricproperties of the Wife Abuse Inventory. J Clin Psychol 1990;46: 828-834

70. Bland R, Orn H: Family violence and psychiatric disorder.Can JPsycbiatry 1988; 31: 129-1 37

71. Hapgood CC, Elkind CS, Wright JJ: Maternal blues: phe-nomena and relationship to later postpartum depression.AustNZ J Psychiatry 1988; 22: 299-306

72. O'Neill T, Murphy P, Greem VT: Postnatal depressionaetiological factors. IrMed J 1990; 83: 17-18

73. O'Hara MW, Neunaber DJ, Zekowski EM: Prospectivestudy of postpartum depression: prevalence, course and pre-dictive factors. J Abnorm Psychol 1984; 93: 158-171

74. Gotlib IH, Whiffen VE, Mount JH et al: Prevalence ratesand demographic characteristics associated with depressionin pregnancy and the postpartum. J Consult Clin Psychol 1989;57: 269-274

75. McFarlane J, Parker B, Soeken K et at: Assessing for abuseduring pregnancy. JAMA 1992; 267: 3176-3178

76. Shuerger JM, Reigle N: Personality and biographic data thatcharacterize men who abuse their wives. J Clin Psychol 1988;44: 75-81

77. Fagan RW, Barnett OW, Patton JB: Reasons for alcohol usein maritally violent men. Am J Drug Alcohol Abuse 1988; 14:371-392

78. Berensen AB, Stiglich NJ, Wilkinson GS et al: Drug abuseand other risk factors for physical abuse in pregnancyamong white non-Hispanic, black and Hispanic women. AmJ Obstet Gynecol 1991; 164: 1491-1499

79. Miller BA, Downs WR, Gondoll DM: Spousal violenceamong alcoholic women as compared to a random house-hold sample of women. J Stud Alcohol 1989; 50: 53 3-540

.so ...................9 .Ml' .9,1,12 .01-1

------ 0 ,1......amosi 1 W.'

Factors in the marriage or familyPoor marital adjustment or satisfactionDivorce soughtReligionReligious incompatibilityUnmarried statusDuration of marriageEducational or occupational incompatibilitySocial isolationLow family income or social classFrequency of argumentsPoor sex lifeInfrequent sexual intercourseProblems caused by pregnancyRecent stressful life eventsFamily problemsDeath of close relativeAddition of third adult to familyParenting problemsAbsent fatherOvercrowded or chaotic living arrangementsNumber of children residing in the householdHarsh discipline of children

Factors affecting the motherWitnessed violence as a child or adolescentExperienced violence as a childPast or present abusePast or present psychiatric problemPrevious suicide attemptLow socioeconomic statusAlcohol or drug abuseLow education levelSomatic complaints, conversion symptomsLow self-esteemEmotional needs unmetDistressLow intelligenceCriminal recordPoor health habitsFrequent pregnanciesBeginning to work outside the homePoor healthRigidity in attitudesAgeUnwanted pregnancy (after 20 weeks)

Lack of attendance at prenatal classesWorking during pregnancyChange of role associated with giving up workAntepartum depressionLack of prenatal careParityNumber of previous pregnanciesEmploymentHistory of menstrual problemsPrevious therapeutic abortionPrevious miscarriagePrevious gynecologic problemCountry of originRaceRelationship with parentsFrequent medical visitsPrevious physical illness requiring admissionto hospital

Current illnessDissatisfaction with leisure activitiesSmoking at end of pregnancyChronic illnesses in mother or familyPrevious abnormal deliveryAnxiety during pregnancyHostility during pregnancyHigher than normal emotional lability duringpregnancy

Social adjustmentPrevious experience with childrenLate menarcheLack of sexual enjoymentLack of self-confidenceSelf-image as immature and dependentBereavementSelf-control behavioursPoor coping skillsInterpersQnal sensitivityInterpersonal eventsDysfunctional attitudeLow locus of controlPrepartum attributionsPersonality factorsDepression in first-degree relativeMental disorder in second-degree relativeDisturbed family relationships during childhoodUnhappy childhood

Problems in present relationship with motherPoor relationship with mother during childhoodLoss of motherPoor relationship with father during childhoodLoss of fatherProblems in present relationship with fatherPoor relationship with mother-in-law

Factors affecting pregnancy and childbirthStressful events related to pregnancy, labour and

deliveryUnplanned pregnancyDuration of pregnancyLength of time to conceptionPrevious consideration of therapeutic abortionMother's feelings on discovering pregnancyMother troubled by "old wives' tales"Complications during pregnancyLength of labourSubjective impact of labourAbnormal delivery (cesarean section,

instrumental delivery)Complications of deliveryMother's satisfaction with careMother's degree of participation in careSex of baby (baby of unpreferred sex)Breast-feedingMother's expectation of difficulty with infant

Factors affecting mother's partnerAgeInterest in pregnancyPresence at deliverySexual aggressiveness toward motherViolence toward childrenWitnessed violence in family of originExperienced violence as a child or adolescentNeed for power or dominance but lowassertiveness

Criminal arrest recordExtreme jealousyPsychopathologyAlcohol abuseLow education levelLow incomeLow self-esteem


Risk factor Definitions used in articles reviewed

Poor marital adjustment or satisfaction Marital problems22Attempted separation or divorce24Husband rates marriage as a less important part of his life than does his

wife38Woman feels that partner does not love her27Poor communication28Score on Dyadic Adjustment Scale'`Score on Locke-Wallace Marital Adjustment Test5"

Lack of social support Social isolation40Lack of social, instrumental or emotional support from spouse27

Traditional, rigid sex-role expectations Woman feels emotionally, financially or materially dependent on herpartner44

Inequitable division of labour34

Recent stressful life events Life stressors experienced during previous year (e.g., financial problems,illness of parent, legal problems, move, work problems, death of friend orfamily member)38

Score on Life Events Inventory"OScore on Social Readjustment Rating Scale28Score on Pilkonis Life Events Schedule'0

Mother or her partner experienced or Sexual exploitation in childhood"witnessed violence or neglect in Severe discipline,37 deprivation22 or "unfair" severe punishment inchildhood childhood

Mother placed in foster care during childhood517

Current or past emotional, physical, Woman is hit, slapped, pushed, has objects thrown at her,68 choked,sexual or verbal abuse of the woman by assaulted with weapon, burned,67 tied up or whipped'her partner Man controls, criticizes and denigrates partner68

Partner suspected of child abuse in the Official report of mild or severe abusepast or of harsh discipline of children Child placed in foster care

Death of childCurrent frequent spankings or use of beltPhysical punishment of baby before crawlingSadistic or dangerous punishment22Children show evidence of punishment37Belief that physical punishment is acceptable38Cursing at childCalling children names37

Poor relationship of mother with her Lack of closeness to mother4"parents Feeling that parents were displeased with her in childhood'7

Low self-esteem in mother Mother does not feel good about herself37Mother feels that she is "usually unsuccessful" in life'7

Past or present psychiatric disorder inmother

Pregnancy unwanted by mother (after20 weeks)

Antepartum depression

Prenatal care not started until thirdtrimester................................................................................................................................................

Mother refused or quit prenatal classes

Alcohol or drug abuse by mother or herpartner

Current psychosisChronic psychiatric problemsChronic depression22

~~~~~~~~~~~~~~~~~~..

Child unwanted22Mother changed decision concerning relinquishment of child for adoption"5

Score on Beck Depression Inventory6'Score on Centre for Epidemiologic Studies - Depression scaleScore on Hamilton Rating Scale for Depression73Result of Goldberg's Standardized Psychiatric Interview30

Self-explanatory

Self-explanatory

Heavy alcohol use (may include self-reports, black-outs, need for an "eye-opener", loss of control, dependency on alcohol, drinking more than heor she should have during the last year, hallucinations or deliriumtremens in the abstinence phase)

Use of illicit drugs (e.g., marijuana, cocaine or opiates) as determined byurine assay or self-report46

Daily or periodic use of sedatives or hypnotics (usually benzodiazepines) bymother to induce sleep

Continuous abuse of sedatives and hypnotics by partner"


Documents

[systematic review * examen critique systematique