1 Introduction Ghubash and Abou-Saleh, 1997; Patel et al., 2002; Aydin et al., 2005; Cooper et al., 1999 Rahman et al., 2004 Goodman, 2004 Weich and Araya, 2004 Rahman et al., 2003a,b 2 Method 2.1 Study area, subjects and sampling n Fig. 1 2.2 Data collection World Health Organization, 1994a Rahman et al., 2003a,b κ World Health Organization, 1994b VonKorff et al., 1996 VonKorff et al., 1996 Socio-demographic variables (age, education, employment, family structure and composition) were assessed at T1 by the same interviewers using a specially designed Personal Information Questionnaire (PIQ). Education was categorised into no education versus at least four years primary education. Four years of schooling was chosen as a cut-off because many Pakistani female children attend primary school for four years, after which many children from low-income families stop attending. Family structure was categorised into nuclear family (parents and children only) or extended family (three generations, or one or both parents with married sons, their wives and children). Socioeconomic status was assessed at T1 by inquiring if the household was in debt and by asking Lady Health Workers, who lived in the same locality and had intimate knowledge of the families being studied, to rate the household on a 5-point Likert scale ranging from 1 (richest) to 5 (poorest). A single dichotomous variable of ‘poverty’ was created by combining these 2 measures, i.e., subjects who were both in debt and rated below 3 on the socioeconomic 5-point Likert scale were classified as being poor. Maternal financial empowerment within the household was measured at T1 by asking the mothers if they were given a lump-sum amount of money for day-to-day household expenses, and whether they could take independent decisions about its use. Mothers who answered ‘yes’ to both questions were classified as financially empowered within the household. Social support was assessed by inquiring if the woman received any support during pregnancy from relatives or friends. Brown and Harris, 1989 Husain et al., 2000 2.3 Statistical analysis StataCorp., 2001 T p Table 1 The study was approved by the ethics committees of Rawalpindi Medical College, Pakistan and University of Manchester, UK. 3 Results 3.1 Sample characteristics and Nine percent were primigravid, 16% already had one child, 21% two children, 17% three children and the remaining 37% had three or more. Thirty-seven percent lived in nuclear families (parents and children only) while the remaining lived in extended families (three generations, or one or both parents with married sons, their wives and children). Fifty-seven percent delivered at home with a traditional birth attendant, 98% without any reported complication. The gender of newborns was equally distributed. Of the 129 women, 121 (94%) were depressed at 3 months, 98 (76%) at 6 months and 80 (62%) at 12 months. Eighty out of 129 (62%) mothers depressed during the third trimester of pregnancy were still depressed at 12 months postnatally but 7 of these had not been depressed at 6 months; thus 73 (57%) were depressed at all time points. Table 1 Tables 2 and 3 p p Table 4 4 Discussion Patel et al., 2002; Rahman et al., 2003a,b Rahman et al., 2004; Patel et al., 2004 Grace et al., 2003; O'Connor et al., 2002 O'Hara et al. (1984) Campbell and Cohn (1997) Beeghly et al. (2002) Rubertsson et al. (2005) Najman et al., 2000 Horowitz and Goodman (2004) Bernazzani et al. (1997) Yonkers et al. (2001) Patel et al., 2002 Rahman et al., 2003a,b While only physically healthy mothers were included in the study, we did not include any physical measures to screen mothers for common physical problems such as anaemia or iodine-deficiency. However, chronic malnutrition in the mothers indicated by a low body-mass index was not significantly associated with persistent depression. Other limitations include a relatively small sample size and the fact that all the women came from one sub-district of Rawalpindi. Husain et al., 2000; Mirza and Jenkins, 2004 O'Hara and Swain, 1996; Beck, 2001 Bernazzani et al., 1997; Verkerk et al., 2003