The Edinburgh Postnatal Depression Scale (EPDS; Cox et al., 1987) is a 10-item self-report measure of depressive symptoms in women during pregnancy and the first year after birth. It is one of the most widely used screening tools for antenatal and postnatal depression.
Perinatal depression is one of the most common complications of pregnancy and childbirth. It is also frequently missed, because low mood around birth is often attributed to normal adjustment to caring for a newborn, and women often do not raise it themselves (Centre of Perinatal Excellence [COPE], 2023). In its more severe forms, perinatal depression is associated with impaired functioning, particularly in a woman’s capacity to care for her infant and in the formation of secure infant attachment, which may in turn be associated with poorer social, cognitive, and behavioural outcomes in the child (COPE, 2023).
Respondents rate how often they have experienced low mood, loss of interest, and anxiety over the past seven days. The EPDS is intended both as a screening tool, to help clinicians recognise perinatal women whose symptoms warrant further evaluation, and as a measurement-based care tool to monitor progress. On NovoPsych it is configured for women aged 18 and over; the original validation sample had a mean age of 26 years (Cox et al., 1987).
The EPDS contains mood and cognitive items, including one on thoughts of self-harm, and excludes somatic content such as fatigue, appetite change, and sleep disturbance, which commonly overlaps with the normal physical experiences of late pregnancy and new parenthood (Cox, 2019). Universal screening for perinatal depression and anxiety with a standardised, validated instrument such as the EPDS is recommended in both the United States and Australia (American College of Obstetricians and Gynecologists [ACOG], 2023; COPE, 2023), and screening programmes have been associated with a reduced risk of depression (O’Connor et al., 2016).
The EPDS is used across the full range of perinatal care settings: hospital and community midwifery, general practice, maternal and child health nursing, obstetric clinics, and specialist perinatal mental health services. It can be administered as a routine screener at antenatal and postnatal visits, and throughout ongoing care to monitor symptoms and any change in depression risk. In progress monitoring, it provides a standard method of tracking response to intervention. The suggested screening schedule covers early and late pregnancy, the early postnatal period, and again within the first year, with the EPDS completed before or at the start of each scheduled visit. Results guide the frequency of further screening and support clinical decisions about further evaluation, triage, and referral (COPE, 2023; Cox, 2017).
The total score is calculated by summing all 10 items, giving a range of 0 to 30. Higher scores indicate more frequent or more intense depressive symptoms over the previous seven days.
Two descriptor bands are applied to the total score, at the scale authors’ own cut-offs of 10 and 13. Each does a different job: 10 is set low to catch as many cases as possible, and 13 is set high to flag the women most likely to have depression. Scores of 9 or below fall beneath both and carry no descriptor. The report gives the number without a label, and clinical judgement guides follow-up.
| Total score | Descriptor | What it means |
|---|---|---|
| 0 to 9 | No descriptor | Below both published cut-offs; the report gives the score without a label |
| 10 to 12 | Elevated | May reflect a depressive illness; further enquiry now, and repeat the EPDS within two to four weeks |
| 13 to 30 | High | Probable depressive illness; further mental health assessment is warranted |
Elevated (10 to 12): a more sensitive screen, in other words a wider net that catches more cases but includes more non-cases. Scores of 10 or more identify 85% of women with major depression (Levis et al., 2020). A score in this range may reflect a depressive illness and warrants further enquiry now, with the EPDS repeated within two to four weeks (COPE, 2023).
High (13 to 30): probable depressive illness. Most women without major depression (95%) score below 13, so scores in this range are seldom seen without it. The trade-off is that about a third of women with major depression also score below 13, which is why the Elevated range sits beneath it (Levis et al., 2020). A score in this range is likely to reflect a depressive illness and warrants further mental health assessment.
How likely a High score is to reflect depression depends on how common depression is in the population being screened, so the score identifies who needs a closer look rather than confirming a diagnosis. Responses may also shift when the questionnaire is completed in front of family members, or where disclosure feels risky.
Item 10 asks directly about thoughts of self-harm. Any endorsement generates a risk alert in the report, whatever the total score, and Australian guidance recommends immediate further mental health assessment (COPE, 2023). The item does not ask specifically about suicidal ideation, so responses should be explored further and read both on their own and alongside the total score.
The thresholds work the same with or without item 10: total scores with and without it correlate at .998 (Qiu et al., 2023). A respondent may score in the Elevated or High range while answering Never to item 10, or may endorse item 10 while the total score stays below 10. The second pattern does not on its own indicate depression, but it does indicate a risk of self-harm that warrants attention in its own right.
The report opens with the total score. At 10 or above it names the band and lists the three highest-scoring items with the response given to each, so the clinician can see which symptoms drive the score. Where items 3, 4 and 5, which are commonly associated with an anxious presentation, stand out from the rest of the profile, the report says so. No anxiety score is produced.
Where EPDS scores are available from more than one administration, change between administrations supports progress monitoring. A change of 4 or more points is treated as a reliable change, meaning a change large enough to be unlikely to reflect measurement error alone (Matthey, 2004).
On a first administration, a single vertical bar chart displays the total score against the full 0 to 30 range.
Where two or more administrations are available, a line graph shows the total score across administration dates instead, covering up to the 60 most recent administrations and split into two graphs, oldest first, where there are more than 30. Shading on both charts is a visual guide to the descriptor bands: unshaded below 10, grey from 10 to 12 (Elevated), and a red tint from 13 to 30 (High), with 13 marked on the vertical axis.
The EPDS was developed and validated against Research Diagnostic Criteria for depressive illness, obtained from Goldberg’s Standardised Psychiatric Interview, in a sample of 84 mothers (Cox et al., 1987). Items were drawn from a pool derived from existing depression and anxiety scales, and refined with input from women and primary care workers on content, wording, acceptability, and likelihood of detecting postnatal depression. Two thresholds were proposed. A cut-off of 13 or more was reported as optimal for identifying probable depressive illness, while a lower cut-off of 10 or more was proposed to reduce false negatives, with the authors noting that failed detection could be reduced to under 10% at that threshold (Cox et al., 1987). Cox (2019) confirmed both cut-offs, with the caveat that users should establish the sensitivity, specificity, positive predictive value, and optimal cut-off appropriate to their own clinical or research purpose. The scale has been translated into more than 60 languages (Cox, 2019). NovoPsych administers the original English version, and findings from translated versions are identified as such below.
Internal consistency is strong across the English version and its translations. The original validation reported a standardised α of .87 and split-half reliability of .88, both derived from a subsample of 60 of the 84 mothers (Cox et al., 1987). Subsequent studies have replicated this, with α of .84 to .86 in pregnant and postpartum women presenting to psychiatric services (N = 185; Flynn et al., 2011), α = .87 for the Brazilian Portuguese version in a postnatal sample (N = 245; Figueira et al., 2009), and α above .87 for the total score of the Spanish version in a longitudinal study of 2,994 perinatal women (Domínguez-Salas et al., 2026).
Short-term test-retest reliability was assessed in 118 Australian mothers of infants under 12 months, who completed the EPDS a second time an average of 2.82 days after the first administration (Kernot et al., 2015). The intraclass correlation for the total score was strong (ICC = .92), with 90.7% agreement on risk category, supporting the stability of scores over short retest intervals.
Across factor-analytic studies, a general depression factor is the most consistently supported and accounts for the most variance (Small et al., 2007; Ratislavová et al., 2025), and the total score is the unit against which thresholds and accuracy data are benchmarked. The initial factor analysis identified a depression factor accounting for 46% of the variance, and a second factor loading on three items, which were removed to produce the 10-item scale (Cox et al., 2014). Multi-factor solutions have since been studied with inconsistent results. Several studies support three correlated factors (anxiety, depression, and anhedonia) over a one-factor solution. Matsumura and colleagues (2020; N = 91,063 mothers, Japanese version) reported adequate fit for three factors (AGFI = .972, RMSEA = .054, CFI = .974, TLI = .961, SRMR = .033) where a one-factor model met none of the same criteria (AGFI = .774, RMSEA = .149, CFI = .766, TLI = .700, SRMR = .083), as did Domínguez-Salas and colleagues (2026; N = 2,994 women, Spanish version) in pregnancy (CFI = .995, TLI = .993, RMSEA = .047) and postpartum (CFI = .996, TLI = .994, RMSEA = .039). The composition of the anxiety factor varies, however, spanning items 3 to 5 in some samples (van der Zee-van den Berg et al., 2019, Dutch version) and items 3 to 6 in others (Matsumura et al., 2020), and Gollan and colleagues (2017; N = 15,172) tabulated roughly twenty prior analyses with heterogeneous item assignment. The scale’s author describes the total score as a measure of depressive symptoms rather than of anxiety (Cox, 2019).
In the original validation, against Research Diagnostic Criteria for major and minor depressive illness, a threshold of 13 or more gave a sensitivity of 86%, specificity of 78%, and positive predictive value of 73% (Cox et al., 1987). The strongest criterion-validity evidence since comes from an individual participant data meta-analysis of 58 datasets comprising 15,557 pregnant and postpartum women, of whom 2,069 met criteria for major depression (Levis et al., 2020). Among the 36 studies using a semi-structured diagnostic interview as the reference standard (n = 9,066, 1,330 with major depression), sensitivity and specificity were .85 and .84 at a cut-off of 10 or more, .81 and .88 at 11 or more, and .66 and .95 at 13 or more. Combined sensitivity and specificity were maximised at 11 or more, and accuracy was similar across reference standards and subgroups, including pregnant and postpartum women considered separately.
Positive predictive value depends heavily on how common depression is in the population being screened. Across major-depression prevalence of 5% to 25%, positive predictive value at a cut-off of 11 or more ranges from 26% to 69%, while negative predictive value ranges from 93% to 99% (Levis et al., 2020).
Convergent validity is supported by strong correlations with established measures of depression and anxiety: .76 with the General Health Questionnaire and .62 with the Hamilton Depression Rating Scale in a review of studies of the Persian versions (Rahmani Ivary et al., 2019), and above .50 with the GAD-7 for the Spanish version (Domínguez-Salas et al., 2026).
Predictive validity is supported across repeated administrations during pregnancy and after birth. In 53,032 Australian women receiving public perinatal care, an antenatal score of 10 to 12 was associated with higher odds of scoring 13 or more postnatally (adjusted OR = 4.5, 95% CI 3.4 to 5.9), and an antenatal score of 13 or more was associated with higher odds of that elevation persisting into the postnatal period (adjusted OR = 13.0, 95% CI 10.32 to 16.47; Khanlari et al., 2019). Both outcomes were defined by a later EPDS score rather than by a diagnostic interview. Taken together, the EPDS identifies who warrants further assessment, and the likelihood that a positive screen reflects depression depends on how common depression is in the setting being screened.
The EPDS captures change in scores over repeated administrations. In the original validation study, mothers who met criteria for depression at the first interview but not at follow-up showed a significant reduction in average scores (n = 16; 15.8 to 9.8; t = 3.72, p = .002), while those who continued to meet criteria showed no significant change (n = 15; 16.5 to 15.38; Cox et al., 1987). Treatment responsiveness has since been demonstrated in independent intervention research: a meta-analysis of 11 studies comprising 20 trials of exercise interventions in postpartum women scoring 10 or more found a small but significant reduction in EPDS scores relative to control conditions (standardised mean difference = −0.28, 95% CI −0.38 to −0.18; Norhamazi et al., 2022). Score change also tracks change on concurrent symptom measures, correlating with change on the PHQ-9 (r = .55) and the GAD-7 (r = .54) in 117 pregnant women (Mao et al., 2021). Together, these findings support the use of the EPDS to monitor progress across repeated administrations.
NovoPsych applies two thresholds, at 10 and 13, following the current Australian national guideline (COPE, 2023) and the author’s recommendation that cut-offs reflect the specific clinical application (Cox, 2019).
The higher threshold of 13 or more is the point at which further assessment is recommended. The original authors identified it as optimal for detecting depressive illness (Cox et al., 1987), and the Australian guideline recommends further assessment at this level (COPE, 2023). It prioritises specificity, which is .95 at 13 or more against a semi-structured interview reference standard (36 studies, n = 9,066; Levis et al., 2020).
The lower threshold of 10 or more marks symptom elevation that warrants monitoring (COPE, 2023). Cox and colleagues (1987) proposed it as a first-stage screen, or for settings where potential cases must not be missed, and at 10 or more the scale detects 85% of major depression cases (Levis et al., 2020). Australian guidance recommends monitoring and rescreening within two to four weeks for scores of 10 to 12 (COPE, 2023), and US guidance uses 10 or more as the positive screening threshold (ACOG, 2023).
Percentiles need a reference sample that represents the women being screened, with a full score distribution from that sample. The two largest Australian EPDS datasets, together covering more than 65,000 women, report the proportion of women above each threshold but not the underlying distribution (Buist et al., 2008; Khanlari et al., 2019). Scores are also heavily concentrated at the low end, with around 86 percent of women in the Sydney cohort scoring 9 or below, so a percentile says little about most respondents. The two thresholds, by contrast, rest on strong evidence, so the report places each score against them: 10, a more sensitive screen, and 13, which indicates probable depressive illness.
It indicates probable depressive illness, and further mental health assessment is warranted. It is a screening result, not a diagnosis. Most women without major depression (95 percent) score below 13 (Levis et al., 2020), but how likely a particular positive screen is to reflect depression depends on how common depression is in the population being screened.
A screening measure serves two purposes. A lower threshold catches more true cases at the cost of more false alarms; a higher threshold gives fewer false alarms but misses more cases. The scale’s authors proposed both a lower and a higher cut-off for exactly this reason (Cox et al., 1987), and the largest accuracy meta-analysis reaches the same conclusion (Levis et al., 2020). Scores of 10 or more pick up 85 percent of women with major depression; about a third of women with major depression score below 13.
Not in place of a dedicated anxiety measure. Some items ask about anxious symptoms, and the report comments on them when they stand out from the rest of the profile, but the scale was built to detect depression and its evidence base is about depression. Where anxiety is the clinical question, treat an elevated EPDS as a prompt to ask about it directly, or administer a measure designed for the purpose such as the GAD-7. Anxiety and depression often co-occur in the perinatal period, so an elevated score in a woman whose presentation seems mainly anxious is worth exploring.
Not necessarily that treatment is failing. The postnatal months bring changes in sleep, support and circumstance that can move mood independently of any intervention, and a rise around a known stressor reads differently from a rise with no obvious trigger. The report separates changes of 4 points or more, which are large enough to be unlikely to reflect measurement error alone, from smaller movements, so a modest increase is not reported as deterioration. The more useful question is usually what has changed for the woman since the last administration.
Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression: Development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry, 150(6), 782–786. https://doi.org/10.1192/bjp.150.6.782
Royal College of Psychiatrists (RCPsych), 1987, as publisher of the British Journal of Psychiatry. Scale authored by J. L. Cox, J. M. Holden and R. Sagovsky.
American College of Obstetricians and Gynecologists. (2023). Screening and diagnosis of mental health conditions during pregnancy and postpartum. Clinical Practice Guideline No. 4. Obstetrics & Gynecology, 141(6), 1232–1261. https://doi.org/10.1097/AOG.0000000000005200
Buist, A. E., Austin, M.-P. V., Hayes, B. A., Speelman, C., Bilszta, J. L. C., Gemmill, A. W., Brooks, J., Ellwood, D., & Milgrom, J. (2008). Postnatal mental health of women giving birth in Australia 2002–2004: Findings from the beyondblue National Postnatal Depression Program. Australian and New Zealand Journal of Psychiatry, 42(1), 66–73. https://doi.org/10.1080/00048670701732749
Centre of Perinatal Excellence. (2023). Effective mental health care in the perinatal period: Australian clinical practice guideline. COPE. https://www.cope.org.au/health-professionals/national-perinatal-mental-health-guideline/new-2023-national-perinatal-mental-health-guideline
Cox, J. (2017). Use and misuse of the Edinburgh Postnatal Depression Scale (EPDS): A ten point ‘survival analysis’. Archives of Women’s Mental Health, 20(6), 789–790. https://doi.org/10.1007/s00737-017-0789-7
Cox, J. (2019). Thirty years with the Edinburgh Postnatal Depression Scale: Voices from the past and recommendations for the future. British Journal of Psychiatry, 214(3), 127–129. https://doi.org/10.1192/bjp.2018.245
Cox, J., Holden, J., & Henshaw, C. (2014). Perinatal mental health: The Edinburgh Postnatal Depression Scale (EPDS) manual (2nd ed.). Royal College of Psychiatrists. https://doi.org/10.1017/9781108617574
Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression: Development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry, 150, 782–786. https://doi.org/10.1192/bjp.150.6.782
Domínguez-Salas, S., Gómez-Gómez, I., Rodríguez-Muñoz, M. F., Rodríguez-Domínguez, C., Motrico, E., & Gómez-Baya, D. (2026). Psychometric properties of the Edinburgh Postnatal Depression Scale (EPDS) for use in digital perinatal mental health research: A longitudinal study among Spanish pregnant and postpartum women. Human Behavior and Emerging Technologies, 2026, 8792831. https://doi.org/10.1155/hbe2/8792831
Figueira, P., Corrêa, H., Malloy-Diniz, L., & Romano-Silva, M. A. (2009). Edinburgh Postnatal Depression Scale for screening in the public health system. Revista de Saúde Pública, 43(Suppl 1), 79–84. https://doi.org/10.1590/S0034-89102009000800012
Flynn, H. A., Sexton, M., Ratliff, S., Porter, K., & Zivin, K. (2011). Comparative performance of the Edinburgh Postnatal Depression Scale and the Patient Health Questionnaire-9 in pregnant and postpartum women seeking psychiatric services. Psychiatry Research, 187(1–2), 130–134. https://doi.org/10.1016/j.psychres.2010.10.022
Gollan, J. K., Wisniewski, S. R., Luther, J. F., Eng, H. F., Dills, J. L., Sit, D., Ciolino, J. D., & Wisner, K. L. (2017). Generating an efficient version of the Edinburgh Postnatal Depression Scale in an urban obstetrical population. Journal of Affective Disorders, 208, 615–620. https://doi.org/10.1016/j.jad.2016.10.013
Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59(1), 12–19. https://doi.org/10.1037/0022-006X.59.1.12
Kernot, J., Olds, T., Lewis, L. K., & Maher, C. (2015). Test-retest reliability of the English version of the Edinburgh Postnatal Depression Scale. Archives of Women’s Mental Health, 18(2), 255–257. https://doi.org/10.1007/s00737-014-0461-4
Khanlari, S., Eastwood, J., Barnett, B., Naz, S., & Ogbo, F. A. (2019). Psychosocial and obstetric determinants of women signalling distress during Edinburgh Postnatal Depression Scale (EPDS) screening in Sydney, Australia. BMC Pregnancy and Childbirth, 19, 407. https://doi.org/10.1186/s12884-019-2565-3
Levis, B., Negeri, Z., Sun, Y., Benedetti, A., & Thombs, B. D. (2020). Accuracy of the Edinburgh Postnatal Depression Scale (EPDS) for screening to detect major depression among pregnant and postpartum women: Systematic review and meta-analysis of individual participant data. BMJ, 371, m4022. https://doi.org/10.1136/bmj.m4022
Mao, F., Sun, Y., Wang, J., Huang, Y., Lu, Y., & Cao, F. (2021). Sensitivity to change and minimal clinically important difference of Edinburgh Postnatal Depression Scale. Asian Journal of Psychiatry, 66, 102873. https://doi.org/10.1016/j.ajp.2021.102873
Matsumura, K., Hamazaki, K., Tsuchida, A., Kasamatsu, H., & Inadera, H. (2020). Factor structure of the Edinburgh Postnatal Depression Scale in the Japan Environment and Children’s Study. Scientific Reports, 10, 11647. https://doi.org/10.1038/s41598-020-67321-x
Matthey, S. (2004). Calculating clinically significant change in postnatal depression studies using the Edinburgh Postnatal Depression Scale. Journal of Affective Disorders, 78(3), 269–272. https://doi.org/10.1016/S0165-0327(02)00313-0
Norhamazi, I., Linoby, A., Abd Latif, R., Mursyid, S., Naimah, M., Nor Haslan, A., Mohd Noor, M. A., & Zamri, F. N. S. (2022). Exercise intervention up to 6 months postpartum improves the outcome of Edinburgh postnatal depression scale. Journal of Physical Education and Sport, 22(9), 2127–2133. https://doi.org/10.7752/jpes.2022.09272
O’Connor, E., Rossom, R. C., Henninger, M., Groom, H. C., & Burda, B. U. (2016). Primary care screening for and treatment of depression in pregnant and postpartum women: Evidence report and systematic review for the US Preventive Services Task Force. JAMA, 315(4), 388–406. https://doi.org/10.1001/jama.2015.18948
Qiu, X., Wu, Y., Sun, Y., Levis, B., Tian, J., Boruff, J. T., Cuijpers, P., Ioannidis, J. P. A., Markham, S., Ziegelstein, R. C., Vigod, S. N., Benedetti, A., & Thombs, B. D. (2023). Individual participant data meta-analysis to compare EPDS accuracy to detect major depression with and without the self-harm item. Scientific Reports, 13, 4026. https://doi.org/10.1038/s41598-023-29114-w
Rahmani Ivary, F., Fanaei, S., Ghahremani, S., Falah Ardizi, N., Abdollahpour, N., Khorasani, F., & Ghazanfarpour, M. (2019). A systematic review of psychometric properties of the Edinburgh Postnatal Depression Scale in Iranian population. International Journal of Pediatrics, 7(5), 9497–9505. https://doi.org/10.22038/ijp.2019.36747.3202
Ratislavová, K., Hendrych Lorenzová, E., Lochmannová, A., & Martin, C. R. (2025). Multidimensionality within the Edinburgh Postnatal Depression Scale: Application issues of specific structure. Journal of Reproductive and Infant Psychology, 43(4), 1036–1049. https://doi.org/10.1080/02646838.2023.2285837
Small, R., Lumley, J., Yelland, J., & Brown, S. (2007). The performance of the Edinburgh Postnatal Depression Scale in English speaking and non-English speaking populations in Australia. Social Psychiatry and Psychiatric Epidemiology, 42(1), 70–78. https://doi.org/10.1007/s00127-006-0134-3
van der Zee-van den Berg, A. I., Boere-Boonekamp, M. M., Groothuis-Oudshoorn, C. G. M., & Reijneveld, S. A. (2019). The Edinburgh Postpartum Depression Scale: Stable structure but subscale of limited value to detect anxiety. PLOS ONE, 14(9), e0221894. https://doi.org/10.1371/journal.pone.0221894
