The Course of Postpartum Psychiatric Disorders in … · The Course of Postpartum Psychiatric Disorders in Women ... participate who scored at or above the cut point of ... low-up

The Course of Postpartum Psychiatric Disorders in Women

and Their Partners

PHYLLIS ZELKOWITZ, ED.D., and TAMARA H. MILET, M.D.1

This study examined the course of postpartum psychiatric disorders in a com-munity sample of mothers and their partners to determine whether sociodemo-graphic variables, life stress, and psychiatric history were related to persistence ofmental health problems. At 6 months postpartum, 48 index couples where the wifehad a psychiatric disorder at 2 months postpartum and 50 control couples with nosuch diagnosis underwent diagnostic interviews and completed questionnaires onpsychological symptoms, life stress, and treatment history. The results indicate thatat follow-up, 54% of the index mothers still had a psychiatric diagnosis, as did 60%of their partners who had had a psychiatric diagnosis at 2 months postpartum.Socioeconomic status, country of origin, and life stress were related to persistence,as were diagnosis and timing of onset of the disorder. About a third of the parentswere referred for treatment. It is concluded that for many families, postpartumpsychiatric disorders are not a transient phenomenon.

—J Nerv Ment Dis 189:575–582, 2001

Postpartum psychiatric disorders pose a signifi-cant mental health problem in the community be-cause of their prevalence and their impact on par-ent-infant and couple relationships (Campbell et al.,1992; Gotlib et al., 1991; Murray and Cooper, 1997;Weinberg and Tronick, 1998). Rates of postpartumdepressive disorders vary with the method of assess-ment and the timing of the evaluation; on average,approximately 13% of childbearing women are af-fected (O’Hara and Swain, 1996). There is also anelevated risk for anxiety disorders during the post-partum period (Altshuler et al., 1998). The partnersof women with postpartum psychiatric disordersare themselves at risk for mental health problems(Ballard and Davies, 1996; Fawcett and York, 1986;Lovestone and Kumar, 1993; Raskin et al., 1990;Soliday et al., 1999; Zelkowitz and Milet, 1996). Asimilar set of risk factors is associated with postpar-tum psychiatric disorders in both mothers and fa-thers: a history of mental health problems (duringpregnancy or earlier), stressful life events, a lack of

social support, and poor marital adjustment (Areiaset al., 1996b; O’Hara and Swain, 1996; Swendsen andMazure, 2000; Zelkowitz and Milet, 1996, 1997). De-mographic variables, such as socioeconomic status(SES) and family size, have not been consistentlyfound to be related to postpartum psychiatric disor-ders in women (O’Hara, 1995); however, immigrantstatus may place women at risk for mental healthproblems after childbirth (Fisch et al., 1997; Glasseret al., 1998; Zelkowitz, 1996; Zelkowitz and Milet,1995).

There are limited data on the course of postpar-tum psychiatric disorders. A significant number ofwomen, perhaps as many as one third of those withpostpartum depression in the first 3 months afterbirth, continue to be symptomatic at 6 months post-partum (Areias et al., 1996a; Campbell and Cohn,1997; Milgrom and McCloud, 1996). Few data areavailable on the course of postpartum disorders infathers, with one study reporting that 3 of 10 fatherswho were symptomatic at 6 weeks postpartum con-tinued to have elevated levels of depressive symp-toms at 6 months postpartum (Ballard et al., 1994).There is little information on factors associated withchronicity, though past psychiatric history and fam-ily history of mental illness were not found to berelated to a chronic course in women (Campbell andCohn, 1997). This literature is also limited by the factthat most studies include only primiparous women;it is important to consider the effects of parity on thecourse of the disorder, because responsibility forseveral young children is a risk factor for depression

1 Department of Psychiatry, Sir Mortimer B. Davis-JewishGeneral Hospital and McGill University. Send reprint requests toDr. Phyllis Zelkowitz, Institute of Community and Family Psy-chiatry, Sir Mortimer B. Davis-Jewish General Hospital, 4333Cote Ste. Catherine Road, Montreal, Quebec, Canada H3T 1E4.

This research was supported by a grant from the Conseilquébecois de la recherche sociale (RS-1893 1 092).

The authors acknowledge the work of Nettie Harris andEugenia Silva in recruiting research participants and JacquelineBeaulne, Cheryl-Lynn Rogers, and Camille Surprenant in collect-ing the data. We are grateful for the cooperation of RuthBresnen, Shari Hecht, Monique Lapointe, and Zubeida Ali of thecommunity health centers involved in this study.

0022-3018/01/1899–575 Vol. 189, No. 9THE JOURNAL OF NERVOUS AND MENTAL DISEASE Printed in U.S.A.

Copyright © 2001 by Lippincott Williams & Wilkins

575

in women (Bebbington et al., 1991; McGrath et al.,1990).

In the present study, we report a 6-month fol-low-up of a community sample of women who wereevaluated for psychiatric disorders at 2 months post-partum. We also examined continuing mental healthproblems in their partners. A further purpose of thestudy was to identify factors associated with persis-tent mental health problems in childbearing womenand their partners, including sociodemographic vari-ables, life stress, psychiatric and treatment history.

Methods

Participants

The sample for this 6-month follow-up study in-cludes 48 of 50 women who had participated in astudy of postpartum psychological adjustment andwho had been diagnosed with a psychiatric disorderat 2 months postpartum. In addition, all 50 womenwho had been recruited as controls participated inthe 6-month follow-up. All the women were re-cruited through two community health centers thatmake routine postpartum contacts with all child-bearing women in their catchment areas. The sam-ple included English- and French-speaking womenwho were married or living in a stable relationshipand had a singleton birth with no significant perina-tal complications.

Details of the recruitment procedures can befound in Zelkowitz and Milet (1996). In brief, therewas a two-stage approach for identifying womenwith postpartum psychiatric disorders at 2 monthspostpartum. Women were screened over the tele-phone by using the Edinburgh Postnatal DepressionScale (EPDS; Cox et al., 1987). Those agreeing toparticipate who scored at or above the cut point of10, which is recommended for community screeningwith this measure, were then interviewed in theirhomes by a clinical psychologist to determinewhether a diagnosis was present. Of the 119 womenscreened who had scores above the cut point, 61(51%) agreed to participate. Eleven women weresubsequently excluded because they had no currentdiagnosis upon interview. There was no differencebetween participants and nonparticipants in age,parity, gender of infant, or SES.

The index group consisted of women who metcriteria for a current axis I disorder; cases wereidentified in this manner because postpartum psy-chiatric illness can be manifested through a varietyof symptoms and syndromes, including depression,panic disorder, and obsessive-compulsive disorder(cf. Metz et al., 1988; Sichel et al., 1993). At the

2-month assessment, 20 women (42%) met criteriafor major depression, and 20 (42%) met criteria foradjustment disorder with depressed mood. Eightwomen (17%) met criteria for anxiety disorders.

The control group was comprised of women whoscored below 7 on the EPDS and had no currentdiagnosis on clinical interview. Partners of womenin both groups also participated in the research.Control couples were matched to index couples onthe basis of parity and SES. For half the couples, thecurrent pregnancy was their first; the other coupleswere having second or third children.

Two index couples refused to participate in the6-month follow up. In addition, two other couples(one index and one control) had separated by 6months postpartum; both mothers and the controlfather completed the assessment, whereas the indexfather did not. Thus, the final sample at 6 monthspostpartum consisted of 98 women and 97 men.Demographic characteristics of the sample are pre-sented in Table 1.

Measures

Edinburgh Postnatal Depression Scale (EPDS).This is a 10-item scale specifically designed toscreen for postpartum depression in communitysamples (Cox et al., 1987). This measure has alsobeen used to screen for postpartum anxiety disor-ders (Ballard et al., 1993). The items are rated on ascale from zero to three and refer to depressedmood, anhedonia, guilt, anxiety, and suicidal ide-ation, as experienced in the past 7 days. The scale isinternally consistent and has adequate sensitivity

TABLE 1Demographic Characteristics of the Sample

Index(N 5 48)

Control(N 5 50)

Mean maternal age (yr) 29.9 30.3Mean paternal age (yr) 32.0 33.8Mean maternal education (yr) 14.9 15.8Mean paternal education (yr) 15.8 15.0Mothers working outside the home (%) 56 64Fathers working outside the home (%) 83 74Maternal occupational status: (%)

Professional 21 36Managerial/white collar 27 30Blue collar/unskilled 17 4No occupation 35 30

Paternal occupational status: (%)Professional 33 32Managerial/white collar 44 42Blue collar/unskilled 15 16No occupation 8 10

Male children (%) 54 60

ZELKOWITZ AND MILET576

and specificity, as compared with a psychiatricdiagnosis of major depression (Cox et al., 1987;Murray and Carothers, 1990). The authors recom-mend the use of a cut point of 9/10 to increasesensitivity for the purposes of community screening.Sensitivity in our study population was found to be91% (Zelkowitz and Milet, 1995).

Structured Clinical Interview for DSM-III-R,

Non-Patient Version (SCID-NP; Spitzer et al., 1990).This semistructured interview was used to makecurrent as well as lifetime axis I diagnoses at 2months postpartum. Interrater reliability in thisstudy was established by having an observer presentduring 12 interviews. A kappa of .77 between thetwo raters on the presence or absence of specificdiagnoses was obtained.

Longitudinal Interval Follow-up Evaluation

(LIFE; Keller et al., 1987). This is a semistructuredinterview designed to track the course of diagnosesmade with the SCID and of symptoms that may havearisen during the follow-up interval. This instrumenthas demonstrated interrater reliability for weeklychanges in symptomatology, with intraclass correla-tions in the .90 range. For the purposes of this study,the interviewer reviewed the period from the initialassessment at 2 months postpartum until the fol-low-up assessment at 6 months postpartum to deter-mine the presence of symptoms on a weekly basis.Participants who continued to meet diagnostic cri-teria at 6 months were considered to be persistentcases, whereas those who no longer met diagnosticcriteria were classified as “in remission.”

Symptom Checklist 90 - R (SCL-90-R; Derogatis,1983). The somatization, depression, and anxietysubscales of this self-report inventory were used toassess psychological symptoms in both spouses.This questionnaire was completed at 2 and 6 monthspostpartum.

The Life Stress Scale of the Parenting Stress Index(Abidin, 1990) is a list of 19 stressful life events,relating to friends and family, work and schooling,income, and health. Respondents indicate whetheror not each event has occurred in the past 12months. The total number of stressors was calcu-lated for each respondent, and weighted accordingto the scoring manual for the measure.

Treatment History. Respondents were askedwhether they had been referred for treatment formental health problems during the period betweenthe 2- and 6-month assessments and whether theyhad sought treatment. Those who had done so wereasked to indicate the type of treatment provider theyhad seen.

Procedures

Both partners in all participant couples were in-terviewed individually by a clinical psychologist.Interviews took place in the couple’s home, at ap-proximately 6 months postpartum. While one part-ner was being interviewed, the other completed theself-report measures.

Statistical Analyses

The chi-square statistic was used to comparegroups on categorical variables, including remission,type of diagnosis, and sociodemographic variables,such as employment status and country of origin;logistic regression was employed for multivariateanalysis of the variables associated with remissionrates. Student’s t statistic was used to comparegroups on continuous variables such as parental ageand years of education. Repeated measures analysisof variance was employed to examine differences inpsychological symptoms related to time of assess-ment and diagnostic group. Correlational analysiswas used to determine the relationship betweenstressful life events and psychological symptoms.

Results

Factors Associated with Psychiatric Diagnosis in

Mothers at 6 Months Postpartum

Of the 48 women in the index group, 26 (54%)continued to be symptomatic at 6 months postpar-tum. Sixteen (62%) women had depressive disor-ders, four (15%) had adjustment disorders with de-pressed mood, and six (23%) had anxiety disorders.Among those with major depression, two had co-morbid diagnoses of dysthymia, and five had comor-bid diagnoses of generalized anxiety disorder. Onewoman in the control group was diagnosed with ananxiety disorder at 6 months postpartum.

A past history of mental health problems was notrelated to remission, with approximately two-thirdsof the women in both the remitted and symptomaticgroups having reported a history of such problemswhen they were assessed at 2 months postpartum.Timing of the onset of the disorder that had beendiagnosed at 2 months postpartum was significantlyassociated with remission: 73% of the women had apostpartum onset, 15% had an onset during preg-nancy, and 12% indicated that the onset was prior tothe pregnancy. Among women with a postpartumonset, 46% were still symptomatic at 6 months, ascompared with 77% of those for whom the onset hadbeen during or before the pregnancy (x2 5 3.72, p 5.05). Furthermore, the nature of the diagnosis was

THE COURSE OF POSTPARTUM PSYCHIATRIC DISORDERS 577

also related to remission. Only 20% of the womenwho were diagnosed with a depressive disorderwere in remission at 6 months, whereas 44% of thosewith anxiety disorders and 74% of those with adjust-ment disorders had remitted by 6 months postpar-tum (x2[2] 5 11.3, p , .01).

As can be seen in Table 2, maternal age, parity,country of origin, and infant gender were not relatedto maternal mental health. However, SES, in termsof maternal occupation and education, was associ-ated with remission, with higher status mothersmore likely to be in the remitted group. Althoughemployment status at the time of the birth of theinfant was not related to remission, women whowere still symptomatic at 6 months postpartum wereless likely to return to work. This was particularlytrue of primiparous women: only 18% of first-timemothers who were symptomatic at 6 months wereback at work, as compared with 75% of the primi-parous women in remission (x2 5 7.43, p , .01).

Logistic regression was used to evaluate the rela-tive contributions of the variables that were signifi-cantly associated with remission: diagnostic cate-gory, timing of onset of the disorder, and maternaloccupational status as an index of SES. Direct entryof these three variables yielded a model that fit thedata better than the model with only a constant(x2[5] 5 17. 16, p , .01). Diagnostic category, and inparticular major depression, was the only variable tobe significantly associated with remission in thismultivariate analysis; the b for major depression

was 22.26 (p , .01). The model correctly classified23 of 26 women who were still symptomatic (88.5%),and 14 of 22 women in remission (63.6%), for anoverall classification rate of 77%.

Psychological Symptoms in Mothers at 6 Months

Postpartum

We compared self-reported levels of psychologi-cal symptoms in the three groups of women: con-trols, those in remission, and those who continuedto be symptomatic. Scores on the SCL90-R Anxiety,Depression, and Somatic Symptoms scales contin-ued to be significantly elevated in the index group,whether or not they still met diagnostic criteria at 6months postpartum (see Table 3). Repeated mea-sures analysis of variance indicated significant maineffects of diagnostic group (F[6,186] 5 13.2, p ,.001) and time of assessment (2 and 6 months post-partum; F[3,93] 5 3.1, p , .05). Student-Newman-Keuls post-hoc comparisons indicated that the threegroups of women were significantly different fromone another in symptoms of anxiety and depres-sion. The remitted and symptomatic women didnot differ from each other in somatic symptoms,which were significantly elevated compared withcontrols. There was also a significant time 3 diag-nostic group interaction effect for the anxiety scale(F[2,95] 5 4.3, p , .05); women in remission showeda decrease in anxiety, whereas those who were stillsymptomatic did not.

Among the demographic variables, only maternaloccupational status and country of origin were re-lated to self-reported psychological symptoms onthe SCL90-R. There was a significant effect of occu-pation on anxiety (F[3,94] 5 4.4, p , .01) and so-matic symptoms (F[3,94] 5 5.7, p 5 .001). Student-Newman-Keuls post-hoc analyses indicated thatwomen working in professional or managerial/whitecollar occupations had lower anxiety scores thanthose in lower status jobs, whereas somatic symp-toms were elevated in women working in blue collaror unskilled jobs. Compared with native-bornwomen, foreign-born women reported higher levelsof depression (t[96] 5 2.2, p , .05) and somaticsymptoms (t[96] 5 3.0, p , .01).

Psychiatric Diagnoses in Fathers

In the original sample, 17 fathers (24%) were di-agnosed with a psychiatric disorder at 2 monthspostpartum. Two of these fathers were lost to fol-low-up. Of the 15 fathers, 9 (60%) remained symp-tomatic at 6 months postpartum. In addition, threenew cases were diagnosed at 6 months: two of these

TABLE 2Demographic Correlates of Remission in Mothers

InRemission(N 5 22)

StillSymptomatic

(N 5 26) Significance

Mean age (yr) 30.1 29.7 t(46) 5 20.2 nsMean education (yr) 15.8 14.1 t(46) 5 22.1*

% Working outside thehome at time ofinfant’s birth 55 58 x2 5 0.05 ns

% Working outside thehome at 6 monthspostpartum 59 23 x2 5 6.5*

Occupational status(%) x2 (3) 5 10.3*

Professional 41 4Managerial/white

collar 23 31Blue collar/unskilled 9 23No occupation 27 42

% Primiparous 55 42 x2 5 0.7 ns% Male infants 64 46 x2 5 1.5 ns% Foreign-born

women 56 33 x2 5 2.4 ns* p , .05.ns, not significant.


were the partners of index women who were stillsymptomatic at 6 months, and the third was in thecontrol group. In contrast, among the remitted fa-thers, five were the partners of women who werealso in remission, and the sixth was in the controlgroup. It should be noted that in 23% of the coupleswhere the woman had a diagnosis at 6 months post-partum, her partner also had a diagnosis; in contrast,only 8% of the women with no diagnosis had apartner with a diagnosis (x2 5 4.38, p , .05).

Four of the six fathers in remission had hadadjustment disorders, one an anxiety disorder, andone a major depressive disorder. The symptomaticgroup included three men with somatization disor-ders, three with mood disorders, two with general-ized anxiety disorder, and one with an adjustmentdisorder.

Remission was unrelated to psychiatric history.As with the mothers, timing of onset of the disorderwas significantly related to remission, with 83% ofthe fathers in remission reporting a postpartum on-set and only 11% with an onset during or before thepregnancy (x2 5 7.82, p , .01).

There were few demographic correlates of pater-nal mental health at 6 months postpartum. Infantgender, maternal parity, paternal age, education, andoccupational status were also not associated withremission rates. Only country of origin was relatedto paternal psychiatric diagnosis, with 24% of theforeign born and 6% of the native born fathershaving a current diagnosis at 6 months postpartum(x2 5 6.01, p , .05). Fully 83% of the fathers inremission were native-born Canadians, as comparedwith only 22% of the fathers who were still symp-tomatic at 6 months (x2 5 5.40, p , .05).

Psychological Symptoms in Fathers

Levels of psychological symptomatology differedaccording to the diagnostic group of the mothers(see Table 3). Repeated measures analysis of vari-

ance indicated a main between-subjects effect ofdiagnostic group (F[6,184] 5 2.3, p , .05). Specifi-cally, anxiety and depression scores were elevatedamong men in the index group, whether or not thewomen were in remission at 6 months. To determinewhether symptom elevation could be attributed tomen with a psychiatric diagnosis, we reanalyzed thedata using only those fathers who did not have acurrent diagnosis. The results were virtually identi-cal, indicating that even among fathers with no cur-rent diagnosis, those whose partners had a postpar-tum psychiatric disorder reported more symptoms.None of the demographic variables were related topsychological symptoms in the fathers at 6 months.

Relationship between Life Stress, Psychiatric

Diagnoses, and Psychological Symptoms

Mothers in remission reported more life stressthan either controls or mothers who were still symp-tomatic (see Table 4); the same was true for fathers.Life stress was found to be correlated (.26, p , .01)with maternal depressive symptoms on the SCL90-R,(.17, p , .05) with paternal depressive symptoms,and (.22, p , .05) with paternal somatic symptoms.These associations may help to explain why parentsin remission continued to report high levels of

TABLE 3Mean Psychological Symptom T-Scores in Mothers and Fathers by Maternal Diagnostic Group

Maternal Symptoms Paternal Symptoms

Control(N 5 50)

Remitted(N 5 22)

Symptomatic(N 5 26)

Control(N 5 50)

Remitted(N 5 22)

Symptomatic(N 5 25)

Anxiety 47.5 52.7 64.4 50.3 55.4 58.6(10.0)a (10.9) (9.1) (12.2) (11.1) (12.3)

Depression 51.4 59.2 65.9 52.2 59.2 60.1(9.9) (7.1) (5.6) (12.5) (9.4) (10.5)

Somatic symptoms 49.4 57.7 61.5 49.6 56.3 53.9(9.8) (10.7) (10.2) (13.6) (8.1) (10.6)

a Standard deviations in parentheses.

TABLE 4Mean Scores on the Life Stress Scale for Mothers and Fathers

by Diagnostic Groups

Maternal Diagnostic Group

Control Remitted Symptomatic

Mean life stressscores ofmothersb

7.5(8.2)a

13.0(7.6)

8.9(6.6)

Mean life stressscores offathersc

6.8(7.6)

13.8(8.0)

7.7(5.4)

a Standard deviations are in parentheses.b F(2,94) 5 3.9, p , .05.c F(2,91) 5 7.4, p , .001. Data are missing for 1 control father and

2 fathers in the symptomatic group.


symptoms, even though they no longer met diagnos-tic criteria for a psychiatric disorder. However, anal-ysis of covariance comparing women in the threediagnostic groups on depressive symptoms and con-trolling for life stress found that there continued tobe a significant main effect of diagnostic groupwhen life stress was taken into account (F[2,93] 524.22, p , .001).

Treatment for Mothers and Fathers with Postpartum

Psychiatric Disorders

Eighteen women and nine men had been referredfor treatment; of these, 14 women and 6 men hadactually received treatment. Thus, approximatelyone third of the parents with a postpartum psychi-atric diagnosis had been referred for treatment, andonly about a quarter had received treatment. All themen and 10 of the women were referred to either apsychologist, psychiatrist, or social worker. Otherreferrals for women were made to communityhealth centers, a family physician, and a self-helpgroup.

Demographic variables, including infant gender,parity, SES, and country of origin, were not relatedto either referrals or help-seeking among men andwomen. Diagnosis was not significantly associatedwith treatment, although referral was somewhatmore likely to occur among women with depressivedisorders (45%) than those with anxiety or adjust-ment disorders (22%). The same was true for men,with two thirds of those with mood disorders re-ferred for treatment, compared with 22% of thosewith anxiety or somatic disorders.

The relationship between remission and treat-ment was also not statistically significant: 14% of thewomen in remission had received treatment, as com-pared with 31% of those who were still symptomatic.Similarly, 33% of the men in remission and 22% ofthose who were still symptomatic had receivedtreatment. Given the small numbers of both men andwomen who had actually received treatment, wemay have lacked the power to detect significantassociations among variables relating to diagnosis,treatment, and remission.

Discussion

The results of this study of a community sample ofpostpartum women and their partners indicates thatmental health problems tended to persist for severalmonths after the birth of the infant. More than halfthe mothers and almost two thirds of the fatherswho had a diagnosis at 2 months postpartum con-tinued to meet diagnostic criteria at 6 months post-

partum. Furthermore, even those who were in re-mission continued to report elevated symptoms ofanxiety, depression, and somatic complaints.

Several factors were found to be related to per-sistent mental health problems among women in thepostpartum period, including SES, type of diagnosis,and timing of onset of the disorder. Women of higherSES, as indexed by maternal education and occupa-tional status, were more likely to be in remission at6 months postpartum, as were those with a postpar-tum onset of the disorder. However, in multivariateanalysis, only the nature of the diagnosis remainedsignificantly associated with remission. Depressivedisorders appeared to be particularly refractory, ascompared with anxiety and adjustment disorders.

Country of origin may be another important riskfactor for both men and women. Foreign-born womenreported higher levels of psychological symptoms at6 months postpartum, and foreign-born men hadlower remission rates. Life stress and a lack of socialsupport, which have been shown to be related topostpartum depression (cf. O’Hara, 1995) may beparticularly important risk factors for depression inchildbearing immigrant women (Zelkowitz et al.,2000).

The results of this study confirm other researchshowing that the partners of women with postpar-tum psychiatric disorders often exhibit mentalhealth problems (cf. Harvey and McGrath, 1988;Lovestone and Kumar, 1993). Many of the fathersappeared to be suffering from chronic mental healthproblems, which continued to affect them after thebirth of their children. However, even among fatherswith no psychiatric diagnosis, those whose partnershad a postpartum psychiatric disorder continued toreport relatively high levels of psychological symp-toms at 6 months postpartum. It is important to noteas a limitation of this study that its findings cannotbe generalized to the population of postpartum fa-thers, because the men were not screened directlyfor mental health problems, as were their partners.Routine screening of both partners in the postpar-tum period, and outreach to families suffering frommental health problems, is needed so that parentsobtain the help and support they may need.

Life stress is a risk factor for postpartum depres-sion (cf. O’Hara, 1995; Swendsen and Mazure, 2000);the results of this study indicate that stress was alsorelated to the course of the disorder. It is of interestthat parents in remission reported the highest levelsof stress, which may serve to maintain elevatedsymptom levels. It is possible that mothers who arefeeling better resume their regular activities, result-ing in greater stress. For example, women in remis-sion were more likely to have returned to work,


indicating that their functional status was moreadequate than women who were still symptomatic.However, the return to work, and the need tocombine family and job responsibilities, may be as-sociated with greater stress for mothers of younginfants.

The present study can only address the issue oftreatment in a very limited fashion, due to smallnumbers and therefore inadequate power. It appearsthat we still have much to learn about the types oftreatment that might be effective for postpartumpsychiatric disorders. As noted above, few parentsactually received treatment, and treatment was notfound to be related to remission. There is little re-search on either pharmacological or psychosocialtreatments (Nonacs and Cohen, 1998). There is aneed to undertake clinical trials of different modal-ities of treatment, including psychopharmacological,cognitive-behavioral, interpersonal, and couple ther-apies. Because breastfeeding women may be reluc-tant to take psychotropic medication, demonstrationof the efficacy of nonmedical treatments may affordservice providers and clients more treatment op-tions (Appleby et al., 1997).

This follow-up of a case-control study of postpar-tum women and their partners suggests that certainpsychosocial and clinical variables are associatedwith persistent mental health problems. Because ofthe heterogeneity of diagnoses in the index group, itwould be necessary to validate these findings inother cohorts, with larger numbers of women in thevarious diagnostic groups.

The results of this study indicate that for manyfamilies, postpartum psychiatric disorders are not atransient phenomenon. Evaluation, and even treat-ment, should include both partners, because themental health problems of one spouse may affectthe psychological well-being of the other. These arefamilies at risk, where vulnerable parents are re-sponsible for the care of the new infant, and in somecases other children as well. The extent to whichthese disorders affect family functioning needs to beaddressed in future research.

References

Altshuler LL, Hendrick V, Cohen LS (1998) Course of mood andanxiety disorders during pregnancy and the postpartum pe-riod. J Clin Psychiatry 59(Suppl 2):29–33.

Abidin RR (1990) Parenting Stress Index: Manual, 3rd Ed. Char-lottesville, VA: Pediatric Psychology Press.

Appleby L, Warner R, Whitton A, Faragher B (1997) A con-trolled study of fluoxetine and cognitive-behavioural coun-selling in the treatment of postnatal depression. Br Med J314:932–936.

Areias MEG, Kumar R, Barros H, Figueiredo E (1996a) Compar-ative incidence of depression in women and men, during preg-

nancy and after childbirth: Validation of the Edinburgh Post-natal Depression Scale in Portuguese mothers. Br J

Psychiatry 169:30–35.Areias MEG, Kumar R, Barros H, Figueiredo E (1996b) Corre-

lates of postnatal depression in mothers and fathers. Br J

Psychiatry 169:36–41.Ballard C, Davies R (1996) Postnatal depression in fathers. Int

Rev Psychiatry 8:65–71.Ballard C, Davis R, Cullen PC, Mohan RN, Dean C (1994) Prev-

alence of postnatal psychiatric morbidity in mothers and fa-thers. Br J Psychiatry 164:782–788.

Ballard C, Davis R, Handy S, Mohan RNC (1993) Postpartumanxiety in mothers and fathers. Eur J Psychiatry 7:117–121.

Bebbington PE, Dean C, Der G, Hurry J, Tennant C (1991)Gender, parity and the prevalence of minor affective disorder.Br J Psychiatry 158:40–45.

Campbell SB, Cohn JF (1997) The timing and chronicity ofpostpartum depression: Implications for infant development.In L Murray, PJ Cooper (Eds), Postpartum depression and

child development (pp 165–197). New York: Guilford.Campbell SB, Cohn JF, Flanagan C, Popper S, Meyers T (1992)

Course and correlates of postpartum depression during thetransition to parenthood. Dev Psychopathol 4:29–47.

Cox JL, Holden JM, Sagovsky R (1987) Detection of postnataldepression: Development of the 10-item Edinburgh PostnatalDepression Scale. Br J Psychiatry 150:782–786.

Derogatis L (1983) SCL- 90-R Manual-II. Towson, MD: ClinicalPsychometric Research.

Fawcett J, York R (1986) Spouses’ physical and psychologicalsymptoms during pregnancy and the postpartum. Nurs Res

35:144–148.Fisch RZ, Tadmor OP, Dankner R, Diamant YZ (1997) Postnatal

depression: A prospective study of its prevalence, incidenceand psychosocial determinants in an Israeli sample. J Obstet

Gynaecol Res 23:547–554.Glasser S, Barell V, Shoham A, Ziv A, Boyko V, Lusky A, Hart S

(1998) Prospective study of postpartum depression in an Is-raeli cohort: Prevalence, incidence and demographic risk fac-tors. J Psychosom Obstet Gynaecol 19:155–164.

Gotlib IH, Whiffen VE, Wallace PM, Mount JH (1991) Prospectiveinvestigation of postpartum depression: Factors involved inonset and recovery. J Abnorm Psychol 100:122–132.

Harvey L, McGrath G (1988) Psychiatric morbidity in spouses ofwomen admitted to a Mother and Baby Unit. Br J Psychiatry

152:505–510.Keller MB, Lavori PW, Friedman B, Nielsen E, Endicott J,

McDonald-Scott P, Andreasen NC (1987) The LongitudinalInterval Follow-up Evaluation: A comprehensive method forassessing outcome in prospective longitudinal studies. Arch

Gen Psychiatry 44:540–548.Lovestone S, Kumar R (1993) Postnatal psychiatric illness: The

impact on partners. Br J Psychiatry 163:210–216.Metz A, Sichel DA, Goff DC (1988) Postpartum panic disorder.

J Clin Psychiatry 49:278–279.McGrath E, Keita GP, Strickland BR, Russo NF (Eds) (1990)

Women and depression: Risk factors and treatment issues.Washington, DC: American Psychological Association.

Milgrom J, McCloud P (1996) Parenting stress and postnataldepression. Stress Med 12:177–186.

Murray L, Carothers AD (1990) The validation of the EdinburghPost-natal Depression Scale on a community sample. Br J

Psychiatry 157:288–290.Murray L, Cooper PJ (1997) The role of infant and maternal

factors in postpartum depression, mother-infant interactions,and infant outcomes. In L Murray, PJ Cooper (Eds), Postpar-

tum depression and child development (pp 111–135). NewYork: Guilford.

Nonacs R, Cohen LS (1998) Postpartum mood disorders: Diag-nosis and treatment guidelines. J Clin Psychiatry 59(Suppl2):34–40.


O’Hara MW (1995) Postpartum depression: Causes and conse-quences. New York: Springer-Verlag.

O’Hara MW, Swain AM (1996) Rates and risk of postpartumdepression: A meta-analysis. Int Rev Psychiatry 8:37–54.

Raskin VD, Richman JA, Gaines C (1990) Patterns in depressivesymptoms in expectant and new parents. Am J Psychiatry147:658–660.

Sichel DA, Cohen LS, Rosenbaum JF, Driscoll J (1993) Postpar-tum onset of obsessive-compulsive disorder. Psychosomatics34:277–279.

Soliday E, McCluskey-Fawcett K, O’Brien M (1999) Postpartumaffect and depressive symptoms in mothers and fathers. Am JOrthopsychiatry 69:30–38.

Spitzer RL, Williams JBW, Gibbon M, First MB (1990) User’sguide for the Structured Clinical Interview for DSM-III-R.Washington, DC: American Psychiatric Press.

Swendsen JD, Mazure JD (2000) Life stress as a risk factor forpostpartum depression: Current research and methodologicalissues Clin Psychol Sci Pract 7:17–31.

Weinberg MK, Tronick EZ (1998) The impact of maternal psy-chiatric illness on infant development. J Clin Psychiatry59(Suppl 2):53–61.

Zelkowitz P (1996) Childbearing and women’s mental health.Transcult Psychiatry Res Rev 33:391–412.

Zelkowitz P, Katofsky L, Saucier J-F, Valenzuela M, Westreich R,Dayan J, Schinazi J (2000) Screening for depression in preg-nant immigrant women. Poster presented at the 7th WorldCongress of the World Association for Infant Mental Health,July, 2000, Montreal, Quebec.

Zelkowitz P, Milet TH (1995) Screening for post-partum depres-sion in a community sample. Can J Psychiatry 40:80–86.

Zelkowitz P, Milet TH (1996) Postpartum psychiatric disor-ders: Their relationship to psychological adjustment andmarital satisfaction in the spouses. J Abnorm Psychol 105:281–285.

Zelkowitz P, Milet TH (1997) Stress and support as related topostpartum paternal mental health and perceptions of theinfant. Infant Mental Health J 18:424–435.


Documents

The Course of Postpartum Psychiatric Disorders in … · The Course of Postpartum Psychiatric Disorders in Women ... participate who scored at or above the cut point of ... low-up