A history of personal violence and postpartum depression: is there a link

Please download to get full document.

View again

of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Background: A link between violence and depression has been shown, but not a link between violence and postpartum depression. This study sought to determine if there is an association between a history of abuse (physical, sexual, emotional as a child
Document Share
Document Tags
Document Transcript
  Arch Womens Ment Health (2002) 4: 83–92 Original contribution A history of personal violence and postpartum depression: is there a link? M. M. Cohen 1,2 , B. Schei 1,3 , D. Ansara 1,2 , R. Gallop 4 , N. Stuckless 5 , and D. E. Stewart 6 1 Centre for Research in Women’s Health, Toronto, Canada 2 Department of Health Administration, University of Toronto and The Clinical Epidemiology & Health Care Research Program,University of Toronto, Toronto, Canada 3 Department of Community Medicine and General Practice, Norwegian University of Science and Technology, Trondheim, Norway 4 Faculty of Nursing, University of Toronto, Toronto, Canada 5 Society, Women and Health Program, Sunnybrook and Women’s College Health Sciences Centre, the Centre for Addiction andMental Health, and Department of Psychiatry, University of Toronto, Toronto, Canada 6 University Health Network, Women’s Health Program, Toronto Canada and the Departments of Psychiatry, Obstetrics andGynaecology, and Medicine, University of Toronto, Canada Introduction The United Nations definition of gender-based vio-lence includes any act “that results in, or is likely toresult in, physical, sexual, or psychological harm orsuffering to women, including threats of such acts,coercion or arbitrary deprivations of liberty, whetheroccurring in public or private life” (Fischbach andHerbert, 1997). Violence against women is common.For example, a Statistics Canada survey of 12,300women over age 18, found that in the 12 monthsprior, 10% experienced violence but 51% hadexperienced physical or sexual assault since age 16(Statistics Canada, 1993).The mental health consequences of violence aresignificant. Several studies now confirm the relation-ship between abuse and poor mental health espe-cially depression (e.g. Mullen et al., 1988; Moellerand Bachmann, 1993). As well, studies have nowdemonstrated that particularly for women, childhoodsexual abuse is associated with long term conse-quences such as depression as an adult (Bifulco et al.,1991; Cutler and Nolen-Hoeksema, 1991; Laws,1993). Since a history of exposure to violence hasbeen shown to be associated with depression, severalauthors have suggested that it is not unreasonablethat there may be a relationship between violenceand post partum depression (Buist, 1998; Campbell,1998; Stewart, 1994). Summary Background: A link between violence and depression has beenshown, but not a link between violence and postpartum depres-sion. This study sought to determine if there is an associationbetween a history of abuse (physical, sexual, emotional as a childor adult) and postpartum depression (PPD). Method: 200 postpartum women were recruited from 6 hospitals.At 8–10 weeks postpartum, a telephone interviewer asked womenabout physical, emotional or sexual abuse as an adult or child andsociodemographic, obstetrical and personal medical history. PPDwas assessed using the Edinburgh Postnatal Depression Scale(EPDS, score of  12). Abuse was determined by the ConflictTactics Scale or the Abuse Assessment Screen. Chi-square andlogistic regression were used to determine the relationship be-tween violence and PPD. Results: 11% of women had EPDS scores of  12. Rates of child-hood (6.5%), or adult (6.5%) physical abuse; and childhood (13%)or adult (14%) sexual abuse were reported by respondents. Emo-tional abuse in the current relationship (29.6%) exceeded that of childhood abuse (3.5%). Overall 43.2% of respondents had atleast one form of abuse. Having a history of depression (OR   3.3(95% CI, 1.3–8.7)), panic attack during pregnancy (OR   5.4 (1.6–19.0)), maternal complications (OR   5.0 (1.7–15.1)), low socialsupport (OR   3.3 (1.3–8.7)) and emotional abuse (OR   2.8 (1.1–7.4) were associated with PPD. Conclusion: Emotional abuse but not physical or sexual abusewas found to be associated with PPD. A possible explanation forthis relationship may be that being in an abusive situation puts oneat risk for depression and in turn, postpartum depression. Keywords:  Postpartum depression; emotional abuse; physicalabuse; sexual abuse; depression during pregnancy.  84 M. M. Cohen et al. Although postpartum depression (PPD) is widelyrecognised, its causes and correlates are still poorlyunderstood (Weissman and Olfson, 1995). PPD iscommon with the estimated rate in the three monthsafter delivery being about 10 – 15% of new mothers.PPD is known to be associated with signi fi cant seque-lae for mothers, infants and families (Murray andCooper, 1997). Known risks for PPD include biologi-cal/physical health, social and psychological factors.One factor that has not been widely examined inrelationship to PPD is a prior history of violence. Wefound only one study; Stewart (1994) found womenabused in pregnancy had an increased risk of physicalabuse in the postpartum period that was associatedwith PPD in over 50% of these women.Prior studies have shown that being in an unstablerelationship (Schaper et al., 1994), poor relationshipwith partner (Hickey et al., 1997) or marital dishar-mony (Murray et al., 1995) were found to be relatedto PPD. These fi ndings are consistent with an abusiverelationship but to the best of our knowledge, nostudy of PPD to date has asked women speci fi callyabout abuse. Given that many prior studies havefound that poor marital relationships are associatedwith PPD, it is possible that an important part of the association found may been mediated throughabuse.Thus we designed a pilot study with the primaryobjective being to determine if there is an associationbetween a history of abuse (physical, sexual, emo-tional abuse as an adult or as a child) and postpartumdepression. Patients After receiving university and institutional ethics committee ap-provals, a pilot study was carried out at six hospitals in the Torontoarea (three teaching and three community hospitals). Participantswere eligible if they were 18 years of age or greater, spoke andunderstood English, delivered a full-term singleton infant and con-sented to participate. Exclusion criteria were: not understandingEnglish, unable to give consent (language or other reasons),women who could not be easily contacted for the postpartuminterview (e.g. no telephone), and women who were giving up thechild for adoption. As well, women who had delivered prematureor multiple infants, infants with major congenital anomalies orneeding neonatal intensive care, still births or early neonataldeaths were also excluded as we felt that it would not be appropri-ate to ask these women to participate. Methods After delivery, eligible women were approached in the hospital onthe postpartum ward and were invited to participate in the study.With the help of the postpartum ward personnel, the researchassistant identi fi ed eligible women and those who did not meet theinclusion criteria. As many women as possible were approachedfor participation, but this varied due to length of hospital post-partum stay, presence of visitors, or other factors.Women were approached on the day after delivery andwere asked if they would participate in a study about women ’ shealth after pregnancy. Women were told that the study wouldinclude sensitive questions, but were not speci fi cally told thatthe study was about postpartum depression and a history of abuse. The study protocol was explained to them and a consentform and information sheet was provided. A brief interview onsociodemographic information (education, employment and fam-ily income) was undertaken at this time. Women were then askedif they would consent to a telephone interview at 8 to 10 weekspostpartum. This time frame was used for a number of reasons.We wanted to allow enough time for postpartum depressionto manifest (  4 weeks). We also wanted to reach women athome before they went back to work. As well, we thought thatwomen would not be able to remember the hospital recruitmentvisit if the time period of the interview was too long after therecruitment. A period of 8 to 10 weeks seemed to be the bestcompromise.We attempted to contact each woman by telephone to set upa convenient time when a private interview could be carriedout. Up to 8 telephone calls were tried for each participant beforeshe was considered a non-respondent. For those who were stillinterested in the study, a private telephone interview of lastingbetween 25 to 30 minutes was undertaken in which questionsabout depression and a history of abuse were included. Anywoman whose score on the postpartum depression screen washigh (over 12) or who revealed a recent history of abuse, wasoffered a referral to a psychiatrist, a social worker or anothersocial service.  Instruments Postpartum depression The study used the Edinburgh Postnatal Depression Scale (EPDS)(Cox et al., 1987). This instrument was developed for the postpar-tum period since other general depression scales contain severalitems relating to somatic symptoms such as fatigue and insomniawhich are common for new mothers. The 10-item EPDS instru-ment has been widely used in the study of postpartum depressionin various settings and among different populations (Zelkowitzand Milet, 1995; Schaper et al., 1994; Warner et al., 1996; Wickbergand Hwang, 1996; Stamp and Crowther, 1994; Jadresic et al.,1995). History of violence Since this was a pilot study, we tested two versions to assess ahistory of personal violence. (Details are given in the Appendix).In brief, for interviewing women at the fi rst three hospitals, amodi fi ed Con fl ict Tactics Scale (Straus, 1996) was used for adultphysical and sexual abuse and selected questions from Tolman(1989) were used to assess adult emotional abuse. For thelast three hospitals, we used the Abuse Assessment Screen(MacFarlane, 1992) for physical and sexual abuse and the ques-tions from the Violence Against Women survey to assess emo-tional abuse as an adult. Since we found that the statisticalrelationship between violence and PPD was very similar for thetwo versions when analyzed separately, we combined the fi ndingsfor presentation here.  Violence and postpartum depression 85 Childhood sexual and emotional abuse For both versions, we asked two questions on childhood sexualabuse and childhood physical abuse. These questions were modi- fi ed from Briere (1988) and are those being used at the Society,Women and Health therapy program at the Centre for Addictionand Mental Health, Toronto Ontario. These involved asking aboutunwanted sexual experiences before the age of 14 and about physi-cal abuse (hit, punched, left with bruises, cuts or bleeding) beforethe age of 14. For emotional abuse, women were asked: “ were youneglected as a child? ” In addition to the items associated with a history of violence,women were asked a number of questions related to socio-demographic characteristics, a history of depression (family his-tory of depression, depression prior to this pregnancy, depressionin a previous pregnancy, and depression during this pregnancy),the pregnancy, their experience of labor and delivery, socialsupport, and their health in the post-partum period.  Analysis We de fi ned probable PPD as a score of  12 on the EPDS, thecut-off value used in most studies. Since some studies have used acut off of 9/10 (Carpiniello et al., 1997; Jadresic, 1995), we catego-rized the EPDS scores into three groups:  12, 9 to 11, and 8 orless.Physical abuse as an adult was de fi ned as having answered “ yes ”  to at least one of the physical abuse questions. Sexual abuseas a child was de fi ned as having answered “ yes ”  to the questionabout childhood sexual abuse. Similarly, physical abuse as a childwas determined by the number of respondents answering “ yes ”  tothese questions. Sexual abuse as an adult was considered if theparticipants answered “ yes ”  to questions about sexual abuse as anadult (version 2) or “ yes ”  to any sexual abuse or sexual coercionquestions from version 1. Emotional abuse was considered to bepresent if respondents had given a rating of “ sometimes, often, orvery often ”  to three or more questions (version 1) or if respond-ents answered “ yes ”  to 3 or more questions about emotional abusein version 2.We examined the variable “ violence ”  in a number of ways. Weexamined a history of childhood or adult sexual abuse, adult orchild physical abuse, and adult or child emotional abuse as well asany abuse (physical, sexual or emotional) or any adult abuse orany childhood abuse.Frequencies of demographic factors, outcome and independ-ent variables were determined for all participants. Using chisquare tests, we then determined the relationship of each inde-pendent variable with the outcome measure, as de fi ned by thethree categories of EPDS. Odds ratios and 95% con fi dence inter-vals were also calculated with EDPS as two categories (  12 versus0 – 11).For the multivariable analysis, variables which were signi fi cantat the p   0.15 were entered into a stepwise logistic regressionmodel where the outcome variable was dichotomized as probablepostpartum depression, EPDS   12 versus score of 11 or less.Independent variables were tested for multicollinearity and forthose variables showing a high intercorrelation, only one variablewas used in the fi nal model (e.g. a history of prior depressionversus depression in this pregnancy). Two fi nal models were run.The fi rst model consisted of those variables which remained sig-ni fi cant in the stepwise regression plus the variable “ emotionalabuse ”  which was “ forced ”  in. The second model was similarexcept all variables related to depression or anxiety were notincluded. Results Two hundred and fi fty three women were recruitedfrom six hospitals. Of those approached for partici-pation, 76% consented ranging from 60.2% to 87%across the six hospitals. Of the 253 women, 53 womenwere lost to follow up at 8 weeks (20.8%) leaving 200women in the study. Those lost to follow up weredifferent from the interviewed women in that theyhad less education, had lower income and were morelikely to be born outside of Canada.The 200 interviewed women re fl ected the ethnicand economic mix of the city of Toronto in that about58% of women were born in Canada (Table 1). Mostwomen were well educated with only 24.5% havinghigh school or less. The women were mainly em-ployed prior to and during the pregnancy and wereliving with a partner at the time of the interview;most of the women ’ s partners were also employed(95.4%). Regarding household income, 13.4% of thewomen had income less than the Low Income Cutoff for the City of Toronto. On the other hand, themajority of women were well off; 28.9% had incomesin excess of $100,000 per annum. Most of the womenwere aged 30 – 34 years (41.2%).Among the participants, 11.1% had EPDS scoresof  12 (probable PPD), 10.6% had scores rangingfrom 9 to 11 and 78.4% had scores of 8 or less (Table1). Regarding violence, we found a history of child-hood or adult physical abuse among 11.0% of par-ticipating women. Thirteen percent of interviewedwomen gave a history of adult sexual abuse and14.1% gave a history of childhood sexual abuse.Adult emotional abuse was reported in 29.6% of women and a history of emotional abuse as a child in3.5%. Overall, 86 women (43.2%) told of a history of at least one form of abuse.Several variables (sociodemographic, personalhistory of depression, pregnancy-related, labor anddelivery, and postpartum) were assessed in associa-tion with the risk of postpartum depression (Table 2).The variables signi fi cantly related to an increasedrisk of post-partum depression (score of 12 or more)were: “ a history of severe/very severe PMS ”  (oddsratio 4.41, 95% CI 1.33 – 14.65) “ depression prior tothis pregnancy ”  (odds ratio 5.78, 95% CI 2.2.5 – 14.84), “ depression during this pregnancy ”  (oddsratio, 3.74, 95% CI 1.41 – 9.94), “ panic attack duringthis pregnancy ”  (odds ratio 3.35, 95%CI 1.22 – 9.17),having had a caesarian section (odds ratio 2.80, 95%CI 1.00 – 7.82), “ maternal complications ”  (odds ratio  86 M. M. Cohen et al. 6.65, 95% CI 2.38 – 18.65), and “ social support ”  (oddsratio 3.63, 95% CI 1.47 – 8.97).When examining a possible relationship betweenthe abuse variables and postpartum depression(Table 3), only “ adult emotional abuse ”  was signi- fi cantly associated with a higher risk of postpartumdepression (odds ratio for adult emotional abuse3.39, 95% CI 1.37, 8.38).All independent variables that were signi fi cantlyassociated with PPD in the bivariate analysis (p  0.15 in Table 2) were entered into a multivariablelogistic regression (Table 4). Since the variables “ having a history of depression prior to this preg-nancy, and depression in this pregnancy were highlycorrelated (p   0.0001), the variable having a historyof depression prior to this pregnancy was used in the fi nal model.For the fi rst model, having a history of depressionprior to this pregnancy, having a panic attack duringpregnancy, two or more maternal complicationsduring or after delivery and low social support wererisk factors for postpartum depression. Since thedepression-related variables were signi fi cantly cor-related with the variable “ adult emotional abuse ” , Table 1. Characteristics of participants (n       200) CharacteristicN%Born in Canadayes11658.0Highest level of educationhigh school or less4924.5technical or community college4824.0undergraduate degree7437.0graduate or professional degree2914.5Employment in last 12 monthsemployed/self employed15979.5student73.5unemploymentinsurance/disability/welfare63.0homemaker2512.5other31.5Currently living with partneryes19497.0Partner employmentemployed/self employed18595.4other94.6Household income 1999 Cdn $   $21,000126.4$21 – 32,000 2 137.0$32 – 55,000 3 3418.2$55 – 80,0004725.1$80 – 100,0002714.4  $100,0005428.9Age (yrs)18 – 24189.625 – 293920.930 – 347841.235 – 394524.140  84.3EPDS score 1  122211.19 – 112110.60 – 815678.4Adult physical abuse136.5Childhood physical abuse136.5Any physical abuse2211.0Adult sexual abuse2613.0Childhood sexual abuse2814.1Any sexual abuse4723.5Adult emotional abuse5929.6Childhood emotional abuse73.5Any emotional abuse6331.7Any history of abuse8643.2 1 One participant did not complete the EPDS. 2 Low income cutoff for family of 4, City of Toronto, 1996. 3 Average Canadian income for cities over 500,000 population, 1996.  Violence and postpartum depression 87 Table 2. Relationship between EPDS categories and risk factors for postpartum depression (proportions and odds rations) VariableN0 – 8EPDS12  Odds ratio and 95%(N   156)9 – 11(N   22)con fi dence intervals%(N   21)%(0 – 11 vs. 12  )% Socio-demographic Age (yrs)   30*5729.445.023.81.0030 – 347741.835.042.91.38 (0.44 – 4.35)35  5328.820.033.31.58 (0.47 – 5.33)Income   $32,0002513.64.822.22.35 (0.61 – 9.10)$32 – 80,0008142.947.644.41.35 (0.45 – 4.09)  $80,000*8043.547.633.31.00Canadian bornyes*11558.361.950.01.00no8441.738.150.01.42 (0.59 – 3.46)Education   high school4925.019.927.31.64 (0.54 – 5.02)technical school4821.828.636.42.35 (0.82 – 6.70)university*10253.252.436.41.00Partner education   high school4121.823.815.80.56 (0.15 – 2.07)technical school4524.523.815.80.50 (0.14 – 1.87)university*10553.652.468.41.00Employedyes*15881.466.777.31.00no4118.633.322.71.15 (0.40 – 3.33)Partner employedyes*18494.1100.0100.01.00no95.90.00.0 –  (cell count 0) Personal history PMS  severe/very severe187.74.822.74.41 (1.33 – 14.65)moderate4321.914.327.31.86 (0.64 – 5.36)mild/none*13770.381.050.01.00Depression prior to this pregnancy 5 yes4518.123.857.15.78 (2.25 – 14.84)no*15281.976.242.91.00Family history of depressionyes3117.09.514.30.87 (0.24 – 3.15)no*16483.090.585.71.00Depression in previous pregnancy 2 yes3024.341.753.83.20 (0.97 – 10.51)no*6975.758.346.21.00Depression during this pregnancy 4 yes3312.823.838.13.74 (1.41 – 9.94)no*16587. Pregnancy Paritynulliparous9854.740.042.90.67 (0.27 – 1.66)multiparous*9145.360.057.11.00Sick leave on doctor ’ s ordersyes2912.823.818.21.35 (0.42 – 4.32)no*17087. plannedyes*14475.066.759.11.00no5525.033.340.91.97 (0.79 – 4.92)Ever smoked during pregnancyyes2512.214.313.61.11 (0.30 – 4.07)no*17487.885.786.41.00Ever used alcohol during pregnancyyes6332.033.327.30.79 (0.29 – 2.12)no*13668.066.772.71.00Panic attack during pregnancy 3 yes3012.814.333.33.35 (1.22 – 9.17)no*16887.285.766.71.00 Labour and delivery Length of labour (hrs)12 or more4032.420.023.50.69 (0.21 – 2.27)less than 12*9467.680.076.51.00Mode of deliveryspontaneous*9754. delivery3620.310.019.01.61 (0.44 – 5.86)c-section5625.740.047.62.80 (1.00 – 7.82)Emergency c-sectionyes2814. (0.44 – 4.56)no*16185.885.081.01.00Level of painnone/mild*3719.214.318.21.00moderate3818.09.536.42.20 (0.60 – 8.06)severe/very severe12462.876.245.40.72 (0.21 – 2.46)Length of stay in hospital (days)2 or less*11259.647.640.91.00more than 28740.452.459.12.01 (0.82 – 4.95)Maternal complications 5 2 or more227.79.536.46.65 (2.38 – 18.56)less than 2*17792.390.563.61.00
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks