Effect of mode of administration on Edinburgh Postnatal Depression Scale in the South Indian population: A comparative study on self-administered and interviewer-administered scores.

BACKGROUND: Edinburgh Postnatal Depression Scale (EPDS) is a validated screening tool widely used to assess perinatal depression (PND). However, due to stigma associated with PND, respondents could answer sensitive questions differently depending on the mode of administration, especially in culturally and linguistically diverse country like India. The present study explored longitudinal differences in EPDS scores between self-administered and interviewer-administered modes. METHODS: 177 women from rural South India were administered EPDS, self-administration followed by interviewer-administered for four visits, twice each during prenatal and postnatal visits. EPDS scores were compared between the two modes descriptively, graphically and by repeated mixed measure models. Classification of antenatal depression (AD), postnatal depression (PD) and PND based on the two modes were compared by McNemar Chi-square test. Clinical and psychosocial characteristics were examined to identify factors associated with differences in the scoring modes. Concordance rates and Goodman Kruskal's Gamma coefficients were measured for individual EPDS items. RESULTS: Longitudinal EPDS scores and rates of AD, PD and PND were significantly higher in self-administered mode. Recent adverse life events were the only factor observed to be significantly associated with the differences between the two modes. Rank correlation and concordance rates suggested stronger association for EPDS items relating to anhedonia subscale and moderate/weaker association for EPDS items relating to anxiety/depression subscales. CONCLUSION: Our study findings suggest that the effect of mode of administration should be taken into account while using PND screening tools such as EPDS, especially in countries such as India with higher levels of illiteracy.

Identification of clinical and psychosocial characteristics associated with perinatal depression in the south Indian population.

BACKGROUND: Longitudinal perinatal depression (PND) data is sparsely available in the Indian population. We have employed Edinburgh Postnatal Depression Scale (EPDS) to assess the prevalence and identify characteristics associated with PND in the south Indian population. PND was assessed longitudinally using EPDS scores with traditional cut-off approach as well as a novel method of latent class mixture modeling (LCMM). The LCMM method, to the best of our knowledge, has been used for the first time in the Indian population. METHODS: Three hundred and forty seven women, predominantly from economically-weaker sections of rural and urban South India were longitudinally assessed for antenatal depression (AD) and postnatal depression (PD) using EPDS cutoff-scores ≥13 and ≥10, respectively. Uni/multivariable analyses were used to identify PND associated characteristics. LCMM was then implemented, followed by risk characteristics identification. RESULTS: PND prevalence from traditional approach was 24.50 % (12.68 % AD; 18.16% PD). Characteristics associated with PND were urban-site and recent adverse life events. Irregular menstrual history and chronic health issues were associated with AD and PD, respectively. Three distinct PND trajectories were observed from LCMM-analysis: low-risk (76.08%), medium-risk (19.89%) and high-risk (4.04%). Urban-site, recent adverse life events, irregular menstrual history and pregnancy complications were associated with medium-risk/high-risk trajectories. LIMITATIONS: EPDS is a screening tool and not a diagnostic tool for depression. Since the study population included women from economically-weaker sections, the results need verification in other socio-economic groups. CONCLUSIONS: Both the traditional cut-off-based approach and LCMM provided very similar conclusions regarding the prevalence of PND and characteristics associated with it. Higher PND prevalence was observed in urban women compared to rural women. In low-income countries, identifying risk characteristics associated with PND is a critical component in designing prevention strategies for PND related conditions because of the limited access to mental health resources.