Validating the Edinburgh Postnatal Depression Scale as a screening tool for postpartum depression in Kathmandu, Nepal

Postpartum depression is defined as an episode of non-psychotic depression according to standardized diagnostic criteria with onset within 1 year of childbirth [1]. Approximately 50–80 % of women suffered from maternal sadness in the puerperal period, with about 20 % of those women develop postpartum depression [2]. However, there is a wide variation in prevalence of postpartum depression from almost affecting 8–50 % of postnatal mothers [3‐7] in different countries. The prevalence of PPD is estimated to be higher in low and middle income countries than for high income countries primarily as many women don’t seek help or tell others about their feelings [8‐12]. British Medical Bulletin, 2012 reported that low income countries such as Nepal and Malaysia had prevalence of PPD below 13 % and in other low income countries it ranges from 13 to 50 % [7]. The studies conducted in Nepal showed the prevalence of PPD ranging from 4.9 to 19.4 % [13‐15]. The differences in the magnitude of prevalence may be due to transcultural variations and interpretation of symptoms and also socio-economic variables [16].

Postpartum Depression can have several consequences which includes maternal death due to suicide [17], the mother infant relationship [18], psychological development of child [19], infant nutrition [20] and infant growth [21]. Postpartum depression is infrequently diagnosed and indeed treated, though its significant incidence and morbidity [22]. Depressive symptoms are dismissed as they mimic normal physiologic changes [23‐25] and formerly used to be ignored in health care services [26]. The effects of postnatal depression on the mother, her marital relationship, and her children make it an important condition to diagnose, treat and prevent [1].

The routine use of screening scales for the purpose of identifying symptoms of depression is an effective, simple and economical way to identify women at risk [27]. There are some instruments designed to measure PPD which includes: The Edinburgh Postnatal Depression Scale [28], Beck Depression Scale (BDI) [29], Zung Self Rating Scale [30], the Kessler Psychological Distress Scale (K10) [31] and the Self-Reporting Questionnaire (SRQ20) [32]. In the postpartum period, the EPDS has been the most widely used scale to identify postpartum depression [4, 33‐35]. Cox, Halden and Sagovsky, in 1987 identified EPDS as fully acceptable tool for detection of PPD with simple method of scoring. It is a 10-item self-report questionnaire in which women are asked to rate how they have felt in the previous 7 days and each question is scored 0–3 resulting in total score range of 0–30. The tool has satisfactory sensitivity and specificity and was also sensitive to change in severity of depression over time [36].

With the development of EPDS, it has been validated and used in many countries [4, 34, 35, 37, 38]. The sensitivity observed in the validation studies presented variations ranging from 65 to 100 %, while the specificity with 49–100 % in range with different cut-off points [4, 34, 35, 37, 39, 40]. In the validation studies done in Nepal there is also wide range of variation in sensitivity with 68–100 % whereas less variation in specificity 93–94 % [13, 41]. The great variability of results among the different studies was due to variations in methodology used, diagnostic criteria and different weeks of postpartum [4, 34, 35, 37, 39‐41].

Validation of EPDS is an important endeavor that provides reliability of the instrument when next version other than original English is used. As there is wide variation in sensitivity of EPDS in various validation studies conducted in Nepal, this study aimed to validate original English version of EPDS into Nepali version in order to measure the accuracy of the EPDS to screen postpartum depressive symptoms.

The present study showed the overall prevalence of 17.1 % with cutoff point of 12/13.The prevalence of the study was lower than the prevalence found by Budathoki (19.4 % at 6 weeks and 22.2 % at 10 weeks of postpartum) which was done in Kathmandu Medical College Teaching Hospital and Dhading District Hospital with cutoff point of >13 [15], whereas, the prevalence was higher than in the study done by Regmi (12 %) in tertiary level hospital [13]. The findings lie in between these two studies. The minor alteration in prevalence might be due to differences in the study methodology and diagnostic criteria and also due to variation in the sample size. The prevalence not essentially contradict with that of the other countries which includes Thailand (10 %) [35], New Zealand (17.1 %) [40], Pakistan (16.4 %) [42] and United Kingdom (10–15 %) [43]. The prevalence of this study was not that far with the studies of other countries though there is variation in social structure, different educational status, social support, economic status, availability of health services, parenting methods and more. However, several studies have found high rate of depression in low and middle income countries [8‐12].

The sensitivity, specificity, positive predictive value and negative predictive value of the Nepalese version EPDS found by this study was 92, 95.6, 77 and 99.3 % respectively at threshold level 12/13. The original English version of EPDS by Cox, identified sensitivity and specificity of 86 and 78 % respectively with threshold level above 12/13 [36]. The study done for validation of Nepalese version of EPDS by Regmi had shown sensitivity of 100 %, specificity of 92.6 %, positive predictive value of 41.6 % and negative predictive value of 100 % with cutoff point 12/13[13]. Also, an another study done in Nepal in 1999, had detected sensitivity of 68.3 %, specificity of 93.8 %, positive predictive value of 65 % and negative predictive value of 94.64 % at cut off point of 12/13 [41]. The hospital based study done in India, revealed the sensitivity of 92 % and specificity of 85 % [44] while the study done in urban slum of Karachi, Pakistan had shown sensitivity of 79 % and specificity of 74 % [42]. This study showed better psychometric properties than in the original English version EPDS by Cox et al., earlier Nepali language EPDS validation studies, studies from neighboring countries like India and Pakistan. This might be due to difference in methodology, translation process, sample size, diagnostic interview and timing of testing. The participants who visit health facilities of urban area are well educated and can have good understanding of translated version. In our study, more than two-third of participants were educated up to secondary level and above, so may have the greater emotional literacy and this might be the reason for good psychometric properties. However, the systemic study of Shrestha and colleagues found psychometric properties from low-and middle-income countries to be lower than that for the original English version of EPDS [45].

Having well defined cut off point appropriate at local setting is essential for accurate detection and estimation of PPD burden in the country. The cutoff point of Nepalese version of EPDS identified by this study was 12/13 which is similar with the previous studies done in Nepal by Regmi et al. and Nepal et al. [13, 41]. The cutoff score was also somewhat equivalent with the study from low and lower-middle-income countries like India (11/12) [44], Pakistan (13/14) [42], Mongolia (12/13) [46], Ghana (10/11) [47] and Zimbabwe (10/11) [48]. But the cutoff point was higher than the studies done in Ethiopia (5/6) [49], Nigeria (8/9) [50], Vietnam (3/4) [51] and Malawi (4/5) [52]. The variation in cutoff point score seen in different validation studies may be due to social and economic diversity, cultural norms and heterogeneities in sample characteristics. The higher EPDS cutoff score ensures the screening test has the best performance by identifying most cases so that a diagnostic or/and therapeutic intervention could be offered.

Culture sensitive translation, empirical validation, better psychometric properties and consistency of findings with national and international literature suggested that the Nepalese version of EPDS is valid instrument to screen the postpartum depressive symptoms. On the other hand, the study exhibited some limitations to be used in all settings of Nepal. As the sample was drawn from only one tertiary level hospital, the findings from this study may be limited to clinical setting and might have less generalizability in the culturally and socially diverse context.

From our result and those obtained from other studies it can be concluded that Nepalese version of the EPDS can be successfully used to screen the postpartum depression in Nepalese women. These findings might be the inferences for the future planning and implementation of maternal mental health programs and policies in low and lower-middle-income countries.