Validation of the Hindi version of the Edinburgh postnatal depression scale (EPDS): Translation, cross-cultural adaptation, and psychometric evaluation in postpartum mothers

ABSTRACT

Background:

Postpartum depression (PPD) is a significant global public health concern, with a high prevalence in India. While the EPDS is widely used, there is a lack of rigorous validation studies specifically for Hindi-speaking postpartum mothers.

Aim:

This study aimed to translate, culturally adapt, and evaluate the psychometric properties of the Hindi EPDS in postpartum mothers against a clinical diagnosis.

Methods:

An observational study was conducted at a tertiary care hospital involving 181 mothers, 48 hours to 7 days after childbirth. The original English EPDS underwent a five-stage translation and adaptation process, including independent forward and back-translations and expert committee reviews. Validity was assessed against a clinical diagnosis made by a psychiatrist. Statistical analysis included Cronbach’s alpha for internal consistency, Principal Axis Factoring with Promax rotation for construct validity, and receiver operating characteristic (ROC) curve analysis to determine the optimal cut-off score.

Results:

The prevalence of depression in the cohort was 26.0% (47/181). Hindi EPDS showed good internal consistency (Cronbach’s α =0.859) and demonstrated good diagnostic accuracy with an Area Under the Curve (AUC) of 0.912. At an optimal cut-off score of 9, the scale showed a sensitivity of 70.2% and a specificity of 91.8%. Factor analysis revealed a three-factor structure: Anxiety, Depressed Affect, and Anhedonia, which accounted for 62% of the total variance.

Conclusions:

The Hindi version of the EPDS is a valid, reliable, and culturally sensitive tool for screening postpartum depression in Indian tertiary care settings. A threshold score of nine is recommended for identifying at-risk mothers, enabling clinicians to apply cultural formulations for mothers facing significant neonatal stressors or breastfeeding difficulties.

INTRODUCTION

Postpartum depression (PPD) is a common and debilitating mental health condition affecting women in the period following childbirth, with significant consequences for maternal well-being, infant development, and family functioning.[1] Despite growing attention to perinatal mental health, PPD remains under-recognized and under-treated in many low- and middle-income countries, including India, where linguistic diversity and cultural factors further complicate detection and care. The Edinburgh Postnatal Depression Scale (EPDS) is one of the most widely used self-report tools for screening depressive symptoms in the perinatal period. translated and validated in multiple languages and cultural contexts worldwide due to ease of use, and established psychometric properties.[2,3] Its utility extends across diverse populations and settings, making it a preferred instrument for both clinical screening and research. However, screening tools must undergo rigorous translation and validation to ensure semantic, conceptual, and cultural equivalence, particularly when used in populations different from those in which they were initially developed.

In the Indian context, several Indian-language versions of the EPDS (e.g., Marathi, Konkani, and Bengali) have undergone translation and diagnostic validation against clinical reference standards. However, comprehensive psychometric evaluation of these versions, including assessment of factor structure, construct validity, and measurement properties, remains limited and inconsistent across languages and settings.[4] Although the Hindi EPDS is validated for antenatal depression, postpartum depression has distinct symptoms, necessitating validation during the 48-hour to 7-day postpartum period.[5] Hindi is one of the most widely spoken languages in India and the world, and a valid, culturally adapted Hindi EPDS for postpartum women would fill a crucial gap in perinatal mental health assessment in this population. Given that postpartum depression and antenatal depression, while related, may present with different symptom profiles and clinical implications, it is not sufficient to assume that a tool validated in the antenatal period will perform equivalently in the postpartum period.[6,7,8] Moreover, a systematic review of Indian validation studies of perinatal screening instruments has highlighted the importance of language-specific validation studies to ensure diagnostic accuracy and clinical utility.[3]

The present study, therefore, aimed to translate and culturally adapt the EPDS into Hindi using a standardized cross-cultural adaptation protocol and to evaluate its psychometric properties, including criterion-related validity, construct validity, known-group validity, and diagnostic accuracy, in a sample of postpartum mothers. To the best of our knowledge, this is the first study in India to validate the Hindi EPDS in a postpartum population, addressing a significant gap in screening and early identification of PPD in Hindi-speaking settings.

Methods

This study used a cross-sectional design from January 2023 to December 2023 to evaluate the Hindi version of the EPDS among mothers 48 hours to 7 days after childbirth at a tertiary care hospital. Participants were recruited via a consecutive sampling technique from the obstetrics ward, with written informed consent obtained for the research. Ethical approval was granted by the Institutional Human Ethics Committee (IHEC vide-IM0264), and all collected data were securely stored in the psychiatry records section, with no provision for data sharing. The inclusion criteria targeted mothers aged 18 to 45 years with a minimum literacy level of primary education, and encompassed only inpatient settings. The study included mothers regardless of delivery nature, including both spontaneous vaginal and caesarean sections, and considered cases of both viable children and stillbirths. Mothers with intellectual disability, impairment in comprehension of the research consent due to hearing or visual impairment were excluded from the study. The sample size was based on established psychometric validation recommendations. For the 10-item EPDS, guidelines suggest 5–10 participants per item, with 100–200 participants generally considered adequate for factor analysis and reliability assessment.

The translation and cross-cultural adaptation phase focused on ensuring linguistic equivalence while maintaining the scale’s sensitivity to local nuances. This process evaluated cross-language equivalence using intra-class correlation coefficients. Statistical analyses were performed using SPSS version 27, employing Cronbach’s alpha for internal consistency and kappa values, along with intra-class correlation coefficients, for reliability. Validity was rigorously assessed through clinical diagnosis. Principal component analysis and factor analysis were utilized to extract factors of postnatal depression psychiatric morbidity. Finally, a detailed ROC curve analysis was conducted to determine the optimal cut-off and establish a recommended threshold score for this population in North India. To ensure linguistic and conceptual equivalence of the EPDS for use in an Indian Hindi-speaking postpartum population, we followed a standardized cross-cultural adaptation process consistent with best practice guidelines for scale translation and validation. This included forward translation, synthesis, backward translation, expert committee review, and pre-testing.[9]

We followed a multidisciplinary, five-stage process to translate and cross-culturally adapt the English EPDS into Hindi, ensuring conceptual and linguistic fidelity for postpartum mothers. Initially, two independent forward translations were produced by a clinician (T1: DD, MD) and a Hindi literature expert (T2: SFH, PhD), prioritizing culturally appropriate phrasing and conceptual equivalence over literal word-for-word translation. These versions were synthesized into a reconciled Hindi version (T-12) through a joint review by the translators and an expert committee comprising a methodologist (DD), clinician (DB), language expert (SMH), and an additional member (JC) to resolve discrepancies and ensure conceptual integrity. To verify the accuracy of the Hindi phrasing, two translators blind to the original instrument (B1: SMH; B2: RS, MD) performed independent backward translations into English.[9] This was followed by an expert committee’s comparison of the original scale to address semantic soundness, ensuring the tool can capture the cultural context relevant to the target population. For illiterate participants who consented, items were read aloud verbatim by trained research staff in a standardized manner, and responses were recorded without interpretation or guidance. Finally, the pre-final Hindi EPDS underwent pretesting with a ten-representative sample of postpartum mothers to confirm item clarity and comprehensibility. The translation process is followed as per Figure 1.

Figure 1.

Translation process of the Hindi version of EPDS

The concurrent validity of the Hindi EPDS was established by comparing participants’ self-reported scores with a clinical diagnosis of depression, which was established by the evaluating psychiatrist through a comprehensive psychiatric assessment. This clinical assessment used a method that categorized mothers into “diagnosed” and “non-diagnosed” groups as per the ICD-10 diagnosis of Depression. To evaluate the scale’s diagnostic accuracy, an ROC curve analysis was performed, and the AUC was calculated to measure the scale’s ability to discriminate between depressed and non-depressed mothers. Concurrent validity was further refined by calculating sensitivity and specificity at various thresholds. We adhered to the STARD 2015 checklist,[10] as applicable, for the validation of EPDS provided in the supplementary section.

RESULTS

Based on data from the 181 mothers who provided consent for the study, the mean age was 27.6 years (SD = 3.5), and 70.7% (n = 128) resided in urban areas, while 29.3% (n = 53) resided in rural areas. The religious distribution was predominantly Hindu (94.5%, n = 171), and the educational background showed a majority of graduates (38.1%, n = 69) or post-graduates (24.3%, n = 44). Occupationally, 75.1% (n = 136) were homemakers, while obstetric data revealed a balanced distribution between primigravida (49.7%, n = 90) and multigravida (50.3%, n = 91) status. Furthermore, 51.4% (n = 93) of deliveries were conducted via cesarean section, with a final depression prevalence of 26.0% (n = 47) established for the total sample, as shown in Table 1.

Table 1.

Sociodemographic characteristics of the sample

Category	Sub-category	Frequency (n)	Percentage (%)
Residence	Urban	128	70.7%
	Rural	53	29.3%
Religion	Hindu	171	94.5%
	Muslim	7	3.9%
	Other	3	1.7%
Education	Illiterate	8	4.4%
	Primary	15	8.3%
	Secondary	45	24.9%
	Graduate	69	38.1%
	Post-graduate	44	24.3%
Occupation	Homemaker	136	75.1%
	Employed	45	24.9%
Parity	Primigravida	90	49.7%
	Multigravida	91	50.3%
Delivery	Caesarean	93	51.4%
	Spontaneous	88	48.6%

Translation outcome

Emotional terms were adapted using semantic nuance and cultural idioms to ensure lay comprehension and conceptual fidelity. Accessible scoring and reverse-coding instructions supported a validated screening threshold of 10 for the identified cases in the sample. Phrasing refinements from pilot feedback ensured item consistency across participants with diverse obstetric histories and magico-religious practices. The pre-final Hindi version of the EPDS was pilot-tested among 10 postpartum mothers in both English and Hindi. Cross-language equivalence was assessed using the Intraclass Correlation Coefficient (ICC) based on total EPDS scores. The Hindi version demonstrated excellent agreement with the original English version (ICC = 0.92). The mean EPDS scores obtained from the English and Hindi versions were comparable (18.3 ± 3.0 vs. 18.9 ± 2.1), respectively.

Reliability, Internal consistency

The 10-item EPDS-Hindi demonstrated good internal consistency, with a Cronbach’s alpha of 0.859. Item-deletion analyses indicated that removal of no individual item resulted in a meaningful improvement in the overall internal consistency, supporting retention of all 10 items.

Concurrent validity of the Hindi EPDS

A total of 181 postpartum mothers were evaluated to examine the validity of the Hindi version of the EPDS relative to the clinical diagnosis. Consistent with the prescribed scoring instructions, items E1, E2, and E4 were scored on a 0–3 direct scale, and items E3, E5–E10 were reverse-coded from 3–0. EPDS total scores demonstrated clear discrimination between the two groups with and without the ICD-10 diagnosis of Depression, as shown in Table 2.

Table 2.

Concurrent validity among the depressed and nondepressed samples

Group	Number of Mothers	Percentage (%)	Mean EPDS Hindi Total Score
Depressed	47	26.0%	12.30
Not Depressed	134	74.0%	3.51
Total Sample	181	100%	5.79

Among the sample, 47 mothers (26.0%) were diagnosed with meeting ICD-10 diagnostic criteria for depression. The mean EPDS total score in this diagnosed group was 12.30, whereas the mean in the non-diagnosed group (n = 134; 74.0%) was 3.51. The total sample mean EPDS score was 5.79, evidencing a substantial overall range in symptom severity. Individual score observations further highlight this distinction. For example, a diagnosed participant scored 24 on the EPDS, whereas a non-diagnosed participant scored 0.

ROC analysis and diagnostic performance

To quantify the diagnostic performance of the Hindi EPDS, a ROC analysis was conducted, with EPDS total scores as the test variable and clinical diagnosis as the criterion standard. The ROC curve plots sensitivity against 1—specificity across all possible EPDS thresholds.

The AUC was 0.912, indicating excellent diagnostic accuracy of the Hindi EPDS in distinguishing depressed from non-depressed mothers. An EPDS cut-off score of ≥ 9 was identified as the optimal pragmatic threshold, balancing sensitivity and specificity in this sample [Figure 2]. At this cut-off, sensitivity was 70.2%, reflecting the proportion of depressed mothers correctly identified by the scale, and specificity was 91.8%, indicating a high rate of correct classification among non-symptomatic mothers.

Figure 2.

ROC curve for Hindi EPDS at a pragmatic cut-off of 9

Construct validity: Factor structure

Exploratory factor analysis was conducted to assess the underlying psychological constructs measured by the Hindi EPDS. Principal Axis Factoring (PAF) with oblique (Promax) rotation was chosen, given theoretical expectations of correlated affective symptom dimensions. The analysis yielded a three-factor solution accounting for approximately 62% of total variance, with factor 1 as ‘Anxiety’ construct comprising of items E3 (self-blame), E4 (anxious or worried), and E5 (frightened or panicky) loaded strongly, consistent with anxiety symptoms, factor 2 as ‘Depressed Affect’ construct comprising of items E7, E8, E9, and E10 relating to low mood and factor 3 as ‘Anhedonia’ construct comprising of items E1 and E2 (inability to laugh and difficulty looking forward to enjoying things) loaded together, representing anhedonia as a features distinct from anxiety and depressed affect.

Promax rotation allowed these factors to correlate, reflecting the clinical reality that anxious and depressive symptoms commonly co-occur in postpartum populations. For example, participants with elevated anhedonia also tended to score high on anxiety items, supporting the use of oblique rather than orthogonal rotation. Most items show high communality (above 0.60), indicating that the three-factor model explains a substantial amount of variance for each item in this Hindi-speaking sample [see Table 3].

Table 3.

Rotated factor structure (pattern matrix) of the Hindi EPDS

Item	EPDS Item Content (Hindi Context)	Factor 1: Anxiety	Factor 2: Depressed Affect	Factor 3: Anhedonia	Communality
E4	Anxious or worried for no good reason	0.85	0.10	0.04	0.73
E5	Felt frightened or panicky	0.82	0.07	0.12	0.69
E3	Blaming oneself unnecessarily when things went wrong	0.78	0.14	-0.05	0.62
E6	Things have been getting on top of me	0.69	0.22	0.08	0.56
E8	Felt sad or miserable	0.12	0.89	0.03	0.81
E9	So unhappy, I have been crying	0.08	0.84	0.05	0.71
E7	So unhappy, I have had difficulty sleeping	0.21	0.75	0.10	0.64
E10	Thought of harming myself	-0.06	0.66	0.09	0.45
E1	Able to laugh and see the funny side of things	0.05	0.06	0.90	0.82
E2	Looked forward to things with enjoyment	0.10	0.04	0.86	0.75
Eigenvalue		3.21	1.84	1.15
% of Variance	(Total: 62.0%)	32.1%	18.4%	11.5%

Extraction Method: Principal Axis Factoring. Rotation Method: Promax with Kaiser Normalization. Loadings > 0.40 are bolded

Known-group Validity

To further establish the known-group validity, we compared EPDS total scores between depressed and non-depressed mothers. The difference in mean EPDS scores between the two groups was highly significant (t = 12.45, P < 0.001), indicating that mothers with a diagnosis of postpartum depression scored substantially higher on the Hindi EPDS than those without such a diagnosis. Diagnostic accuracy was evaluated using ROC analysis. To evaluate the diagnostic performance of the Hindi EPDS, a ROC curve analysis was conducted using clinically diagnosed postpartum depression as the reference standard. It demonstrated excellent ability in distinguishing depressed from non-depressed mothers, with an AUC of 0.912 (95% CI: 0.868–0.956, P < 0.001).

Examination of the ROC coordinates indicated that a threshold score of ≥9 provided the optimal balance between sensitivity and specificity, yielding 70.2% sensitivity and 91.8% specificity. A threshold score of ≥10 yielded a sensitivity of 59.6% and specificity of 92.5%, while a threshold of ≥11 increased specificity to 94.8% at the cost of reduced sensitivity (55.3%). These findings support the utility of the Hindi EPDS as an effective screening instrument for postpartum depression in Hindi-speaking postpartum mothers.

DISCUSSION

This study is a comprehensive translation and validation of the Hindi version of EPDS for postpartum mothers in India, addressing a critical gap in the lack of locally validated tools for screening postpartum mental health problems. The methodology, incorporating forward and back translations, captured the unique linguistic and cultural contexts of the Hindi-speaking cohort. It allowed the tool to factorize into a three-factor solution of anxiety, anhedonia, and depressed mood, which accounted for approximately 62% of the total variance. This structure demonstrates that the Hindi version is susceptible to the multi-dimensional nature of postnatal distress, particularly how anhedonia remains a distinct psychometric feature among Indian mothers while facing the changes in postpartum bio-psycho-social dynamics.

The existing literature has previously focused on validating the EPDS in antenatal populations. Still, this research establishes its performance in the immediate postnatal period in a general tertiary care hospital setting.[5] We intentionally utilized PAF with Promax (oblique) rotation, as this method accounts for the theoretical and empirical correlation between anxiety and depressive symptoms often seen in postpartum women.[11] By employing the clinical diagnosis, we were able to validate the scale against expert psychiatric diagnosis, identifying a depression prevalence of 26.0% among the 181 mothers. The high diagnostic accuracy, evidenced by an AUC of 0.912, confirms that the Hindi EPDS is a superior tool for distinguishing depression in an inpatient general hospital context.

The reported optimal cut-off scores have varied from 6 to 7 in a community-based Tamil study to 13 in Assamese and Bengali validations. Sensitivity values in Indian studies have ranged from 71.4% to 100%, while specificity has ranged from 77.7% to 98.0%. The Hindi EPDS demonstrated an AUC of 0.912, indicating excellent diagnostic discrimination comparable to that reported for validated Konkani, Tamil, Punjabi, Kannada, Gujarati, Assamese, Bengali, and Marathi versions. Notably, the optimal cut-off identified in the present study (≥9) is lower than those reported in several Indian studies but comparable to findings from some community and cross-cultural validations, where lower thresholds improved case detection.[3] The original EPDS validation recommended a cut-off of ≥13 for probable depression, although subsequent international studies have reported optimal thresholds ranging from 9 to 13 depending on the population, language, and diagnostic reference standard. Similarly, Indian language validations have identified varying cut-offs, including 11–12 for Konkani and 13 for Marathi and Bengali. Similar variability has been reported in the Vietnamese EPDS validation, which identified an optimal cut-off of 3/4, highlighting the influence of language, culture, and population characteristics on screening performance.[12]

Furthermore, though the scale has moderate sensitivity at the optimal cut-off of nine, it suggests that clinicians should remain vigilant with mothers who could be missed from screening for depression if a clinician relied solely on the numerical cut-off score. This gap highlights, in the Indian context, when a mother reports severe neonatal health complications or uses local idioms of distress to describe her mental state, even if her score is below the recommended threshold, as noted in our sample, which is beyond the scope of this article. A few mothers may express suffering through metaphorical descriptions of their struggles in culturally relevant expressions or by referencing an inability to perform traditional duties or rituals, which are not captured in the scale and should be supplemented by cultural formulation to unearth the conventional frameworks of health and spiritual well-being.[13,14] Some mothers expressed difficulty with infant bonding, and physical care can be a significant indicator of depression in clinical interviews, but they were underrepresented on the Hindi EPDS. Because the scale is particular to depressive symptoms but less sensitive to the full range of maternal distress, clinicians must use the score as a guide rather than a definitive diagnostic tool. Vigilance is especially required for those expressing their suffering through culturally relevant language or those facing significant newborn health crises, as these qualitative factors are strong predictors of a diagnosis, regardless of the numerical total.[15,16]

Despite its robust performance, the study is limited by its consecutive sampling technique and a single-center tertiary hospital design, which may limit the generalizability of the findings to more diverse rural populations. The Hindi EPDS demonstrated excellent specificity; its moderate sensitivity suggests that some milder cases may not be detected, emphasizing the importance of clinical follow-up assessment after screening. The performance of the Hindi EPDS in identifying subclinical or mild postpartum depressive symptoms could not be evaluated and warrants investigation in future community-based studies, where such presentations are more common and are often amenable to psychosocial interventions. Future research should focus on multi-center longitudinal studies to track the scale’s predictive validity across the first postpartum year. In conclusion, the Hindi EPDS is a valid and culturally appropriate instrument that provides clinicians with a reliable threshold for identifying at-risk mothers, thereby facilitating essential mental health interventions in the Indian context.[17] The Hindi EPDS appears suitable as a community screening tool when administered by trained frontline health workers. However, positive screens should be followed by psychosocial interventions and a formal clinical assessment.

CONCLUSION

The final Hindi EPDS retains the content validity, conceptual integrity, and clinical relevance of the original English instrument. It is linguistically appropriate for Hindi-speaking postpartum mothers and ready for application in Hindi-speaking areas after psychometric evaluation, including reliability testing and criterion-related validation against diagnosis among postpartum mothers. However, given the cultural context, we recommend supplementation with cultural formulation to unearth the traditional frameworks of health and spiritual well-being in postpartum mothers in India.

Declaration regarding the use of generative AI

The authors have used generative AI to improve the grammar and readability of the manuscript and assume full responsibility for its entire content, including parts generated by the AI tool.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.