Individual participant data meta-analysis to compare EPDS accuracy to detect major depression with and without the self-harm item

Qiu, X, Wu, Y, Sun, Y, Levis, B, Tian, J, Boruff, JT, Cuijpers, P, Ioannidis, JPA, Markham, S, Ziegelstein, RC, Vigod, SN, Benedetti, A, Thombs, BD ORCID: https://orcid.org/0000-0002-5644-8432, He, C, Krishnan, A, Bhandari, PM, Neupane, D, Negeri, Z, Imran, M, Rice, DB, Azar, M, Chiovitti, MJ, Gilbody, S, Kloda, LA, Patten, SB, Mitchell, ND, Alvarado, R, Barnes, J, Beck, CT, Bindt, C, Correa, H, Couto, TCE, Chorwe-Sungani, G, Eapen, V, Favez, N, Felice, E, Fellmeth, G, Fernandes, M, Field, S, Figueiredo, B, Fisher, JRW, Green, EP, Honikman, S, Howard, LM, Kettunen, PA, Kohlhoff, J, Kozinszky, Z, Leonardou, AA, Maes, M, Martínez, P, Radoš, SN, Nishi, D, Pawlby, SJ, Rochat, TJ ORCID: https://orcid.org/0000-0002-5223-770X, Rowe, HJ, Sharp, DJ, Skalkidou, A, Smith-Nielsen, J, Stein, A, Su, KP, Sundström-Poromaa, I, Tadinac, M, Tandon, SD, Tendais, I, Töreki, A, Tran, TD, Trevillion, K, Turner, K, Væver, MS, van Heyningen, T, Vega-Dienstmaier, JM, Wynter, K and Yonkers, KA (2023) Individual participant data meta-analysis to compare EPDS accuracy to detect major depression with and without the self-harm item. Scientific Reports, 13 (1). 4026. ISSN 2045-2322

Preview

Published Version Available under License Creative Commons Attribution. Download (1MB) | Preview

Abstract

Item 10 of the Edinburgh Postnatal Depression Scale (EPDS) is intended to assess thoughts of intentional self-harm but may also elicit concerns about accidental self-harm. It does not specifically address suicide ideation but, nonetheless, is sometimes used as an indicator of suicidality. The 9-item version of the EPDS (EPDS-9), which omits item 10, is sometimes used in research due to concern about positive endorsements of item 10 and necessary follow-up. We assessed the equivalence of total score correlations and screening accuracy to detect major depression using the EPDS-9 versus full EPDS among pregnant and postpartum women. We searched Medline, Medline In-Process and Other Non-Indexed Citations, PsycINFO, and Web of Science from database inception to October 3, 2018 for studies that administered the EPDS and conducted diagnostic classification for major depression based on a validated semi-structured or fully structured interview among women aged 18 or older during pregnancy or within 12 months of giving birth. We conducted an individual participant data meta-analysis. We calculated Pearson correlations with 95% prediction interval (PI) between EPDS-9 and full EPDS total scores using a random effects model. Bivariate random-effects models were fitted to assess screening accuracy. Equivalence tests were done by comparing the confidence intervals (CIs) around the pooled sensitivity and specificity differences to the equivalence margin of δ = 0.05. Individual participant data were obtained from 41 eligible studies (10,906 participants, 1407 major depression cases). The correlation between EPDS-9 and full EPDS scores was 0.998 (95% PI 0.991, 0.999). For sensitivity, the EPDS-9 and full EPDS were equivalent for cut-offs 7–12 (difference range − 0.02, 0.01) and the equivalence was indeterminate for cut-offs 13–15 (all differences − 0.04). For specificity, the EPDS-9 and full EPDS were equivalent for all cut-offs (difference range 0.00, 0.01). The EPDS-9 performs similarly to the full EPDS and can be used when there are concerns about the implications of administering EPDS item 10. Trial registration: The original IPDMA was registered in PROSPERO (CRD42015024785).