Journal of Affective Disorders 117 (2009) 42–47

Contents lists available at ScienceDirect

Journal of Affective Disorders

j ourna l homepage: www.e lsev ie r.com/ locate / j ad

Research report

Young, single and not depressed: Prevalence of depressive disorder amongyoung women in rural Pakistan

Atif Rahman a,⁎, Mansoor Ahmed b, Siham Sikander b, Abid Malik b,Barbara Tomenson c, Francis Creed c

a University of Liverpool, Alder Hey Children's NHS Foundation Trust, Eaton Road, Liverpool, L12 2AP, UKb Human Development Research Foundation, No1, Street 64, Sector F-7/3, Islamabad 34000, Pakistanc Department of Psychological Medicine, University of Manchester, UK

a r t i c l e i n f o

⁎ Corresponding author. Tel.: +44 151 252 5509; faE-mail address: Atif.Rahman@liverpool.ac.uk (A. R

0165-0327/$ – see front matter © 2008 Elsevier B.V.doi:10.1016/j.jad.2008.12.008

a b s t r a c t

Article history:Received 9 July 2008Received in revised form 7 December 2008Accepted 8 December 2008Available online 8 January 2009

Background: The prevalence of depression is very high among adult women in Pakistan but it isnot knownwhether such a high prevalence occurs in younger women. We aimed to assess theprevalence and correlates of depression in 16 to 18-year old unmarried women in Pakistan.

Method: Population-based survey of all 16 to 18-year old unmarried women in one ruralcommunity in Rawalpindi District, Punjab, Pakistan. Depressive disorder and psychologicaldistress were assessed using the Structured Clinical Interview for DSM-IV Disorders (SCID) andSelf-Reporting Questionnaire (SRQ) respectively.

Results: 337 eligible women were identified of whom 321 (95%) were interviewed. Fourteen(4.4%) had depressive disorder; one third scored 9 or more on SRQ. On multivariate analysis ahigh SRQ score was associated with childhood experience of poverty, father's education,stressful life events, disturbed family relationships and mother's depression.

Limitations: The sample was derived from one rural community only and the results should begeneralised with caution.

Conclusion: Depressive disorder is not common in young women in rural Pakistan thoughdistress appears common and is associated with early and recent adversity and familydifficulties. These results suggest future work might aim to understand onset and preventchronic depression.

© 2008 Elsevier B.V. All rights reserved.

Keywords:Depressive disorderAdolescentWomenDeveloping countrySouth AsiaMental health

1. Introduction

The global burden of depression, especially in women, isnowwidely recognised (Lopez et al., 2006; Prince et al., 2007).The prevalence of depression among women in Pakistan isparticularly high (45.5%) and it has been associated withpoverty, lack of education, relationship problems and socialstatus of women (Mirza and Jenkins, 2004). Evidence fromWestern studies suggests that the peak age of onset ofdepression is between late adolescence and early adulthood(Jaffee et al., 2002; Oldehinkel et al.,1999). Studies from SouthAsia suggest a different pattern, with a lower prevalence in

x: +44 151 252 5285ahman).

All rights reserved.

.

adolescence that rises steeply in adulthood (Pillai et al., 2008).We sought to determine whether high rates of depression inolder Pakistani women were evident in unmarried lateteenaged women and whether the correlates of such depres-sionwere similar. In a cross-sectional population-based studyin rural Pakistan we assessed the prevalence of depression in16 to 18-year old unmarried women and the social factorscorrelated with such depression.

2. Methods

2.1. Study population and sample

The study was performed in one rural community, KooriDulal Union Council in district Rawalpindi. Koori Dulal, has a

43A. Rahman et al. / Journal of Affective Disorders 117 (2009) 42–47

population of about 18,000 living in 2100 households and isserved by 13 local village-based community health workers(Lady Health Workers) who provide basic health educationandmaternal-child health care to every household. Each LHWis responsible for about a hundred households. They arerequired by the Health Department to keep accurate recordsof the individuals in each household. From November 2006 toFebruary 2007, we engaged them to carry out a survey of allfemales aged 16 to 18 years to determine their marital status.370 womenwere identified; 16 (4.5%) were ineligible as theywere married and 17 (4.5%) were excluded as they had aphysical illness or learning disability. Of the remaining 337eligible women, 321 (95%) agreed to be interviewed.

2.2. Measurements

All the questionnaires were culturally adapted using astandardised procedure (Rahman et al., 2003b). We took careto ensure that sensitive personal information was obtainedconfidentially by an experienced, culturally-aware andtrained female interviewer, who was blind to the mentalhealth assessment. Mental health assessments were carriedout by a trained psychiatrist who was blind to the othermeasurements.

2.3. Mental health assessments

Current Major Depressive Episode was diagnosed usingthe Structured Clinical Interview for DSM-IV Disorders (SCID)(First et al., 1994) The SCID is a semi-structured diagnosticinterview which has been widely used in cross-nationalepidemiological and treatment studies (Gorman et al., 2004;Weissman et al., 1996). The interview consists of standardiseddiagnostic questions arranged in modules corresponding toeach DSM-IV Axis I disorder.

Psychological distress was measured using the Self-Reporting Questionnaire (SRQ-20) (World Health Organiza-tion, 1994), a brief self-assessment or interviewer adminis-tered scale. It consists of twenty items, each with a yes/noanswer and a maximum score of 20. It has been used widelyin developing countries and has good psychometric proper-ties. We have validated it in Pakistan and SRQ score of 9 ormore is best cut off for probable psychological problem.(Rahman et al., 2005).

2.4. Early and recent adversity

Early childhood physical abuse was measured using theSelf-Report of Childhood Abuse Physical (SRCAP), (Widomand Shepard, 1996) designed to enable self-report ofchildhood physical abuse up to the end of secondary school.The questionnaire has 6 statements describing variousscenarios of serious physical abuse (e.g., up to the timeyou finished secondary school, did anyone inside or outsideof your family beat or really hurt you by hitting you with abare hand or fist; hurt you badly enough so that you neededa doctor or other medical treatment; beat you when youdidn't deserve it). The respondent is asked if she experi-enced any such abuse. A positive response to any of theSRCAP items would be considered an indication of physicalabuse.

Experience of childhood poverty was established througha semi-structured interview with the mother. This was ratedpositive if the child had experienced a period where thefamily had no money to buy everyday necessities and hadexperienced hunger due to lack of food.

Life events and difficulties in the previous year wereassessed with a checklist derived from the Life Events andDifficulties Schedule (LEDS) used in a previous study in thestudy area (Rahman et al., 2003a). We included the eventswhich are most often severely stressful in Pakistan (financial,housing, work, health, family arguments) and events thatoccur between the ages of 16 and 20 years (leaving school,starting a job, marriage plans, and marriage/moving out ofthe parental home). A cut-off of 4 or more life events in theprevious year was used to indicate difficulties in this area(Kendler et al., 2003).

2.5. Family factors

The Index of Family Relations (IFR) (Hudson, 1990) wasused to measure the magnitude of problems of personaland social functioning in the area of family adjustment. The18-item IFR characterizes the severity of family relationshipproblems in a global manner and can be regarded as ameasure of intra-familial stress. The IFR has been translatedinto Urdu and has been used in a previous study in Pakistan(Deeba, 2004). We used the established cut off score of 37to indicate high levels of family stress.

To explore if there were any intergenerational associa-tions, we assessed all mothers of the young women for majordepressive disorder using the SCID.

We also enquired about family structure, dichotomisedinto joint vs nuclear.

2.6. Socioeconomic factors

Current socioeconomic status was assessed in 3 ways:Firstly we asked the Lady Health Workers (who live in thesame community and have intimate knowledge of thefamilies being studied) to rate each family's socio-economicsituation, relative to overall prosperity in the sub-district, ona five-point Likert scale ranging from 1 (richest) to 5(poorest). We have used this method in previous studies(Rahman et al., 2004). Secondly, we asked the mother if thefamily was currently in debt, and thirdly we recorded thehousehold income. Other sociodemographic variablesincluded education, parental education and the women'sbody mass index (BMI) as an indicator of their nutritionalstatus.

2.7. Statistical analysis

Univariate analysis using Fisher's Exact test assessed thecorrelates of depressive disorder but since so fewwomen hadthis disorder we divided the sample into 3 tertiles accordingto SRQ score. The correlates of these tertiles were analysedusing the χ2 test for dichotomous data and ANOVA forcontinuous data. Multiple regression analysis was performedwith SRQ score as the dependent variable and all of thevariables listed in Tables 2 and 3 were included as indepen-dent variables.

Table 1Correlates of major depressive disorder.

Factor Nondepressed

Depressed Relative risk(95% CI)

Fisher'sexactp valueN=307 N=14

n (%) n (%)

Sociodemographic factorsIncome less than Rs 5000 96 (31%) 6 (43%) 1.6 (0.6 to 4.5) 0.39Family in debt 97 (32%) 6 (43%) 1.6 (0.6 to 4.5) 0.39Rated 4 or 5 by 5 point

scale for socioeconomicstatus

78 (25%) 2 (14%) 0.5 (0.1 to 2.2) 0.53

≤4 years schooling 18 (6%) 3 (21%) 3.9 (1.2 to 12.9) 0.055Mother uneducated 229 (75%) 8 (57%) 0.5 (0.2 to 1.3) 0.21Father uneducated 54 (18%) 3 (21%) 1.3 (0.4 to 4.4) 0.72Joint family structure 210 (68%) 12 (86%) 2.7 (0.6 to 11.7) 0.24

Early and recent adversityExperience of poverty 26 (8%) 4 (29%) 3.9 (1.3 to 11.6) 0.032Physical abuse 30 (10%) 2 (14%) 1.5 (0.4 to 6.4) 0.644 or more recent

life events136 (44%) 13 (93%) 15.0 (2.0 to 113.4) b0.001

Family factorsHigh score on Index of

Family Relations (N36) 152 (17%) 6 (43%) 3.4 (1.2 to 9.4) 0.025

Mother depressed 44 (14%) 9 (64%) 9.1 (3.2 to 26.1) b0.001

1 Missing data for 1 subject in the non-depressed group.

Table 2Comparison of 3 SRQ tertiles with risk factors as categorical variables.

SRQscore 0–4

SRQscore 5–8

SRQscore 9–20

χ2 p

(n=110) (n=106) (n=105)

No (%) No (%) No (%)

Sociodemographic factorsSubject's age:16 45 (41%) 40 (38%) 39 (37%) 1.7 0.7917 30 (27%) 35 (33%) 37 (35%)18 35 (32%) 31 (29%) 29 (28%)

Income less than Rs 5000 34 (31%) 29 (27%) 39 (37%) 2.4 0.30Family in debt 31 (28%) 35 (33%) 37 (35%) 1.3 0.52Rated4or 5 by 5 point scalefor socioeconomic status

29 (26%) 23 (22%) 28 (27%) 0.9 0.64

≤4 years schooling 6 (6%) 4 (4%) 11 (11%) 4.2 0.12Mother uneducated 78 (71%) 78 (74%) 81 (77%) 1.1 0.58Father uneducated 12 (11%) 16 (15%) 29 (28%) 11.0 0.004 a

Family structure: joint 82 (74%) 67 (63%) 73 (70%) 3.3 0.20Nuclear or multiple 28 (26%) 39 (37%) 32 (30%)

Childhood adversity/recent life eventsExperience of poverty 8 (7%) 5 (5%) 17 (16%) 9.0 0.011 b

Physical abuse 6 (6%) 7 (7%) 19 (18%) 11.6 0.003 a, b

4 or more life events 32 (29%) 48 (45%) 69 (66%) 29.1 b0.001 a, b

Family factorsIFR total score 37 or more c 10 (9%) 11 (10%) 37 (35%) 30.9 b0.001 a, b

Mother depressed 10 (9%) 20 (19%) 23 (22%) 7.0 0.030 a

a Group 3 significantly greater than group 1.b Group 3 significantly greater than group 2.c Missing data for 1 subject in the lowest SRQ group.

44 A. Rahman et al. / Journal of Affective Disorders 117 (2009) 42–47

Ethical approvalwas obtained from the Institutional ReviewBoardofHumanDevelopmentResearchFoundation, Islamabad.The study conformed to good practice and to the Declaration ofHelsinki. Participants were interviewed after obtaining writtenvalid formal consent. They had the option of opting out of thestudy at any point during the interview and assessment.

3. Results

3.1. Prevalence of DSM IV depressive disorder and psychologicaldistress

Major depressive disorder, diagnosed using the SCID, wasfound in 14 out of the 321 (4.4%) young women. Age-wiseprevalence was 4/124 (3%) in 16 year olds, 4/102 (4%) in17 year olds and 6/95 (6%) in 18 year olds. 53/321 (17%) of themothers had depressive disorder.

The number of women scoring 9 or above on the SRQwas105 (33%). In order to examine the correlates of psycholo-gical distress in more details, we divided our sample intotertiles according to SRQ score. The top tertile (n=105)included women with SRQ score of 9 or above, the middletertile (n=106) included women with SRQ score of 5–8,while the bottom tertile (n=110) had women with scoresof 0–4. The mean (sd) SRQ scores for the tertiles were: 2.5(1.2), 6.5 (1.1) and 12.2 (2.9). The 14 young women withmajor depressive disorder had a mean SRQ score of 14.4(sd=3.8) compared to 6.7 (4.1) for the remainder (t=6.9,pb0.0005). All but one of the women with depressivedisorder had an SRQ score of 9 or more. Sixty-three women(20%) responded yes to the item: ‘Has the thought of endingyour life been on your mind?’ This was reported affirma-tively by 4/110 (3.6%) of the bottom SRQ tertile, 15/106 ofthe middle SRQ tertile (14.1%) and 44/105 (42%) of the topSRQ tertile.

3.2. Correlates of depressive disorder and psychological distress

The correlates of DSM IV depressive disorder are shown inTable 1. Univariate analysis shows that depressive disorderwas significantly associated with experience of childhoodpoverty, 4 or more recent life events and a high score on indexof family relations and maternal depression.

Tables 2 and 3 show the univariate analysis of the threetertiles of SRQ score with the risk factors as dichotomous andcontinuous variables, respectively. Significant correlates werefather's level of education, experience of childhood poverty,physical abuse, 4 or more recent life events and a high scoreon index of family relations and maternal depression.

Due to the low number of women with a DSM IVdepressive disorder, multiple regression analysis was carriedout with total SRQ score as the dependent variable, and all thefactors listed in Tables 1 and 2, and subject's BMI asindependent variables. This showed that the variables thatwere significantly and independently associated with totalSRQ score were: 4 or more life events, father's education,IFRN=37, experience of childhood poverty and mother'sdepression. This result remained unaltered when additionalindependent variables mother's age and BMI, were alsoincluded as independent variables, neither of which wassignificant. Therewas a high Variance Inflation Factor (VIF) forsocio-economic score and low income, due to the very highlevel of association between the two variables. The multipleregression analyses were repeated without socio-economicscore, and the same set of variables as above were found to be

Table 3Comparison of 3 SRQ tertiles with risk factors as continuous variables.

Factor SRQ score0-4

SRQ score5-8

SRQ score9-20

ANOVAP value

(n=110) (n=106) (n=105)

Mean (SD) Mean (SD) Mean (SD)

Subject's education (years) 8.9 (2.7) 9.0 (2.5) 8.2 (3.1) 0.065Subject's Body MassIndex (BMI)

19.8 (3.7) 19.5 (3.4) 19.6 (3.1) 0.82

Life events total score 2.6 (2.4) 3.4 (2.5) 4.5 (2.5) b0.001 a, b

Index of Family Relationstotal score

28.6 (6.6) 28.0 (7.0) 33.7 (11.6) b0.001 a, b

a Group 3 significantly greater than group 2.b Group 3 significantly greater than group 1.

45A. Rahman et al. / Journal of Affective Disorders 117 (2009) 42–47

significantly and independently associated with the total SRQscore (Table 4).

4. Discussion

Toourknowledge this is thefirst study tohave examined theprevalence of depression in a population-based sample ofwomen in this age group in a developing country. Strengths ofthis study include the high response rate from a population-based sample, the use of a well-recognised interview scheduleadministered by trained psychiatrists and the concurrentcollection of data from the women's mothers regarding earlyadversity, current financial problems and her current mentalstate. Limitations of the study include its cross-sectional nature,which prevents clear statements about causality and therelatively small sample from one rural locality only.

We found a relatively low prevalence of depressivedisorder (4.4%) in unmarried women aged 16–18 yearswhen compared to the prevalence (25%) we found in ourprevious study of married pregnant women (mean age26 years) living in the same area (Rahman et al., 2003a). Asystematic review of Pakistani studies suggests the rate ofdepression in non-pregnant older womenmay even be higher(Mirza and Jenkins, 2004). In a more recent study conductedin the same area by our group (Rahman, 2006), theprevalence of depressive disorder was 20% in pregnantwomen below 20 years of age (personal communication).

Table 4Multiple regression analysis with SRQ total score and all risk factors.

Factor Unstandardised coefficien

B SE

Age in years −0.016 0.2Body Mass Index (BMI) −0.001 0.0Income less than Rs 5000 −0.608 0.5Family in debt −0.061 0.4≤4 years schooling 0.788 0.9Mother uneducated −0.308 0.5Father uneducated 1.837 0.6Joint family structure −0.340 0.4Experience of poverty 1.837 0.8Physical abuse 1.028 0.74 or more recent life events 1.775 0.4High score on Index of Family Relations (N36) 3.364 0.6Mother depressed 1.245 0.6

Themaximum Variance Inflation Factor is 1.44. Socio economic status rated by LHW i3.8)—see text.

Why was there a five-fold difference in the rate ofdepressive disorder between unmarried young women andolder married women? It may be a measurement artefact, butthis is unlikely, given that we used the same interviewschedule, the same interviewers, and the same study area inboth studies. A similar steep rise in the rate of depressionbetween late adolescence and early adulthood has beenobserved in studies from India (Patel et al., 2006; Pillai et al.,2008). Such a pattern may be peculiar to South Asian culturewhere women lead a protected life till early adulthood, whenthere is a rapid role transition accompanied by severalpotentially difficult life events clustered closely togetherincluding marriage, moving in with the husband's family,pregnancy, and childbirth (Pillai et al., 2008). In contrast, inWestern cultures, the transition from childhood to adulthoodmay be more gradual and some type of stresses may appearearlier than in South Asian cultures (for example, in thesphere of intimate relationships).

The chronic nature of depressive disorder in developingcountries could also contribute to the higher prevalencefound in older age groups. In a follow-up study of 160 womendepressed during pregnancy, we found that over half weredepressed at 6 months and 1 year postnatal (Rahman andCreed, 2007). It could be hypothesized that events followingmarriage or childbirth trigger the onset of the condition, andmaintaining factors, over which the woman has little control,allow the condition to persist.

Although the prevalenceofmajordepressive disorder is low,a third of the women had high levels of psychological distressmeasured by the SRQ. This high level of psychological distresswas associated with early adversity (poverty and physicalabuse), family stress and adverse recent life events. The recentlife events most strongly associated with psychological distress(data not shown) included loss of livelihood in themain bread-earner, educational problems (many girls have to stop theireducation after high school to prepare for a more domesticrole), marriage plans (lack of a suitor or rejection) and fallingoutwith significant familymembers. These domains are similarto those seen in other Pakistani studies of adultwomen (Husainet al., 2006; Rahman et al., 2003a).

However, it is notable that in our study, current indicatorsof low socioeconomic status were not associated with

ts Standardised coefficients t P

Beta

76 −0.003 −0.06 0.9566 −0.001 −0.02 0.9958 −0.065 −1.09 0.2895 −0.006 −0.12 0.9032 0.044 0.85 0.4005 −0.031 −0.61 0.5414 0.160 2.99 0.00390 −0.036 −0.69 0.4933 0.122 2.20 0.02886 0.070 1.31 0.1955 0.202 3.90 b0.00108 0.295 5.54 b0.00138 0.105 1.98 0.048

s omitted from this analysis because of multicollinearity with income (3.4 and

46 A. Rahman et al. / Journal of Affective Disorders 117 (2009) 42–47

depression or distress but childhood experience of povertywas associated with both. Low socioeconomic status, on theother hand, is associated with depression in older women(Rahman et al., 2003a). Although the association betweenpoverty and depression is well-established in developingcountries, the exact nature of the relationship is unclear (Pateland Kleinman, 2003). It may be that young people areprotected from the stress of financial insecurity while theyliving with their parents, but this protection is lost aftermarriage. Of all socioeconomic indicators measured, the onlyone independently associated with a high SRQ score washaving an uneducated father. Rural Pakistani society is highlypatriarchal in nature and it may be that educationalawareness in the head of the household has importantbearings on the manner in which women are treated in thehousehold. The other recent change within rural Punjab hasbeen the government's drive to increase levels of girls'education. In this sample, only 6.5% of the subjects (averageage 17) had less than 4 years of education whereas in ourprevious study conducted in 2003, 47% of the subjects(average age 26) were uneducated. More studies with largersamples are required to clarify the relationship of depressionwith socioeconomic status and education, both of which aredynamic rather than static in rural Pakistan.

In our study, living in a nuclear family was not associatedwith increased risk of depression or distress, as has beenpreviously reported (Mumford et al., 1997). It has beensuggested that the strong traditional support provided by theextended family contributes strongly to the low prevalence ofdepression seen in Indian adolescents (Pillai et al., 2008).While traditional support systems, such as those afforded byjoint family systems and close knit agrarian communities mayprovide comprehensive social support to young women,intra- or inter-family tensions might lead to abuse or othertraumas that continue for a long time within a closed familysystem. In the absence of any external helping agencies ayoung woman could become trapped in such a situation. Suchwomen might be more likely to passively accept abuse in hernew home after marriage.

Depression and distress in the young women were bothassociated with depressive disorder in their mothers. Thisraises questions about the family environment; it may reflectthe degree of disturbed family relationships. Since itremained an independent correlate in our multiple regressionanalysis, however, it could also raise the possibility of an inter-generational, possibly genetic, association.

There are likely to be interactions between childhoodadversity, high levels of family distress, recent life difficulties,and genetic propensity, and all of these probably combine toconfer vulnerability for depression. It may be hypothesizedthat women with high levels of psychological distress andpossible ‘sub-threshold’ depression, and possibly a greatergenetic predisposition, are more likely to become depressedfollowing the potentially difficult life events immediatelyfollowing marriage. This notion requires testing in a pro-spective study. If supported, it would raise the possibility of anintervention early in married life, prior to pregnancy, whichaims to prevent depressive disorder completely or prevent itbecoming a chronic disorder. Even earlier interventionsfocussing on preventing abuse and improving family func-tioning could have a role in reducing distress in youngwomen

and preventing depression. Designing such interventionsrequires a more complete understanding of the timing andreasons for onset of depressive disorder in this population.

Our findings lead us to conclude that the understandingabout aetiology of depression fromWestern studies cannot beentirely extrapolated to developing countries like Pakistanwhere the sociocultural milieu is very different. The chronicnature of the condition contributes heavily to years of lifelived with disability in a large number of women. There isstrong evidence that maternal depression is associated withlow birthweight, high rates of diarrhoea and poor growth ininfants (Rahman et al., 2004, 2007a,b). Such public healthimplications demand a greater understanding, through long-itudinal studies, of the origins of depression in these womenso that effective interventions can be designed.

Role of funding sourceThe project was funded by Human Development Research Foundation,

Pakistan. The funder had no further role in study design; in the collection,analysis and interpretation of data; in the writing of the report; and in thedecision to submit the paper for publication.

Conflict of interestAll the authors declare that they have no conflict of interest.

Acknowledgements

We are grateful to all the young women who took part inthe study. AR held a Wellcome Career Development Fellow-ship while this study was performed.

References

Deeba, F. The relationship between depression and gender discriminationagainst women in families. 2004. Quaid-e-Azam University, Islamabad,National Institute of Psychology. Ref Type: Thesis/Dissertation.

First, M.B., Spitzer, R.L., Gibbon, M., et al., 1994. Structured Clinical Interviewfor Axis I DSM-IV Disorders. Biometrics Research, NewYork. Ref Type:Serial (Book, Monograph).

Gorman, L.L., O'Hara, M.W., Figueiredo, B., Hayes, S., Jacquemain, F.,Kammerer, M.H., Klier, C.M., Rosi, S., Seneviratne, G., Sutter-Dallay, A.L.,2004. Adaptation of the structured clinical interview for DSM-IVdisorders for assessing depression in women during pregnancy andpost-partum across countries and cultures. Br. J. Psychiatr., Suppl. 46,s17–s23.

Hudson, W., 1990. The WALMYR Assessment Scales: Scoring Manual.WALMYR, Tempe, AZ.

Husain, N., Bevc, I., Husain, M., Chaudhry, I.B., Atif, N., Rahman, A., 2006.Prevalence and social correlates of postnatal depression in a low incomecountry. Arch. Women Ment. Health 9 (4), 197–202.

Jaffee, S.R., Moffitt, T.E., Caspi, A., Fombonne, E., Poulton, R., Martin, J., 2002.Differences in early childhood risk factors for juvenile-onset and adult-onset depression. Arch. Gen. Psychiatry 59 (3), 215–222.

Kendler, K.S., Hettema, J.M., Butera, F., Gardner, C.O., Prescott, C.A., 2003. Lifeevent dimensions of loss, humiliation, entrapment, and danger in theprediction of onsets of major depression and generalized anxiety. Arch.Gen. Psychiatry 60 (8), 789–796.

Lopez, A., Mathers, C., Ezzati, M., Jamison, D., Murray, C., 2006. Global Burdenof Disease and Risk Factors. World Bank, Washington.

Mirza, I., Jenkins, R., 2004. Risk factors, prevalence, and treatmentof anxiety anddepressive disorders in Pakistan: systematic review. BMJ 328 (7443), 794.

Mumford, D.B., Saeed, K., Ahmad, I., Latif, S., Mubbashar, M.H., 1997. Stress andpsychiatric disorder in rural Punjab: a community survey. Br. J. Psychiatry170, 473–478.

Oldehinkel, A.J., Wittchen, H.U., Schuster, P., 1999. Prevalence, 20-monthincidence and outcome of unipolar depressive disorders in a communitysample of adolescents. Psychol. Med. 29 (3), 655–668.

Patel, V., Kleinman, A., 2003. Poverty and common mental disorders indeveloping countries. Bull. World Health Organ. 81 (8), 609–615.

47A. Rahman et al. / Journal of Affective Disorders 117 (2009) 42–47

Patel, V., Kirkwood, B., Pednekar, S., Pereira, B., Baros, B., Fernandes, J., Datta, J.,Pai, R., Weiss, H., Mabey, D., 2006. Gender disadvantage and reproductivehealth risk factors for common mental disorder in women: a communitysurvey in India. Arch. Gen. Psychiatry 63, 404–413.

Pillai, A., Patel, V., Cardozo, P., Goodman, R., Weiss, H.A., Andrew, G., 2008.Non-traditional lifestyles and prevalence of mental disorders inadolescents in Goa, India. Br. J. Psychiatry 192 (1), 45–51.

Prince, M., Patel, V., Saxena, S., Maj, M., Maselko, J., Phillips, M., Rahman, A.,2007. No health without mental health— a sloganwith substance. Lancet370 (9590), 859–877.

Rahman, A. Development and evaluation by randomised trial of a commu-nity-based Early Multimodal Intervention for depressed mothers andtheir infants in rural Rawalpindi, Pakistan. http://www.controlled-trials.com/ISRCTN65316374. 8-6-2006. 2-3-2007. Ref Type: ElectronicCitation.

Rahman, A., Creed, F., 2007. Outcome of prenatal depression and risk factorsassociated with persistence in the first postnatal year: prospective studyfrom Rawalpindi. Pakistan J. Affect. Disord. 100, 115–121.

Rahman, A., Iqbal, Z., Harrington, R., 2003a. Life events, social support anddepression in childbirth: perspectives from a rural community in thedeveloping world. Psychol. Med. 33 (7), 1161–1167.

Rahman, A., Iqbal, Z., Waheed, W., Hussain, N., 2003b. Translation andcultural adaptation of health questionnaires. J. Pak. Med. Assoc. 53 (4),142–147.

Rahman, A., Iqbal, Z., Bunn, J., Lovel, H., Harrington, R., 2004. Impact ofmaternal depression on infant nutritional status and illness: a cohortstudy. Arch. Gen. Psychiatry 61 (9), 946–952.

Rahman, A., Iqbal, Z., Lovel, H., Shah, M.A., 2005. Screening for postnataldepression in the developing world: a comparison of the WHO Self-Reporting Questionnaire (SRQ-20) and the Edinburgh Postnatal Depres-sion Screen (EPDS). J. Pakistan Psychiatr. Soc. 2, 69–72.

Rahman, A., Bunn, J., Lovel, H., Creed, F., 2007a. Association betweenantenatal depression and low birthweight in a developing country.Acta Psychiatr. Scand. 115 (6), 481–486.

Rahman, A., Bunn, J., Lovel, H., Creed, F., 2007b.Maternal depression increasesinfant risk of diarrhoeal illness — a cohort study. Arch. Dis. Child. 92 (1),24–28.

Weissman, M.M., Bland, R.C., Canino, G.J., Faravelli, C., Greenwald, S., Hwu, H.G.,Joyce, P.R., Karam, E.G., Lee, C.K., Lellouch, J., Lepine, J.P., Newman, S.C.,Rubio-Stipec, M., Wells, J.E., Wickramaratne, P.J., Wittchen, H., Yeh, E.K.,1996. Cross-national epidemiology of major depression and bipolardisorder. JAMA 276 (4), 293–299.

Widom, C.S., Shepard, R.L., 1996. Accuracy of adult recollections of childhoodvictimization: Part 1.Childhood physical abuse. Psychol. Assess. 8, 412–421.

World Health Organization, 1994. A User's Guide to the Self ReportingQuestionnaire (SRQ). WHO/MNH/PSF/94.8. World Health Organization,Geneva. Ref Type: Report.