Psychiatric History and Postpartum Depression: The Mediating Role of Obstetric Complications Authors: Mette‐Marie Zacher Kjeldsen, Katrine Holde, Line Bager, Veerle Bergink, Emely Ek Blæhr, Janne Tidselbak Larsen, Kathrine Bang Madsen, Merete Lund Mægbæk, Liselotte Vogdrup Petersen, Trine Munk‐Olsen Published: 2026 ABSTRACT ABSTRACTIntroductionPsychiatric history is the strongest risk factor for postpartum depression (PPD). Obstetric complications, more prevalent among women with a psychiatric history, are also independent risk factors. However, the mechanisms linking these factors to PPD remain unclear. We examined whether obstetric complications mediate the association between psychiatric history and PPD.MethodsThis cohort study utilized Danish nationwide Edinburgh Postnatal Depression Scale (EPDS) screenings (2015–2021) linked with register data. Psychiatric history was defined as psychiatric diagnoses (ICD‐10: F00‐99) or filled psychotropic prescriptions (ATC: N05A, N05BE01, N06A, N06BA) from 1995 until conception. Complications were defined as a composite measure of complications occurring between conception and delivery. PPD symptoms were defined as a positive EPDS score (≥ 11), and PPD diagnosis was defined as a depression diagnosis (ICD‐10: F32‐33) or antidepressant prescription fill (ATC: N06A indicated for depression).ResultsOf 170,218 mothers (163,326 in analyses), 23.9% had a psychiatric history. These mothers had higher levels of PPD symptoms (13.4% vs. 6.1%), PPD diagnosis (7.0% vs. 0.4%), and complications (34.1% vs. 28.5%) compared to those without. (A) Psychiatric history (PPD symptoms: OR = 2.32 [95% CI, 2.22; 2.41]; PPD diagnosis: OR = 5.09 [95% CI, 4.48; 5.79]) and complications (PPD symptoms: OR = 1.16 [95% CI, 1.11; 1.21]; PPD diagnosis: OR = 1.18 [95% CI, 1.04; 1.34]) were independently associated with PPD. (B) Psychiatric history did not modify the association between complications and PPD (PPD symptoms: OR = 1.20 [95% CI, 1.14; 1.26] vs. 1.09 [95% CI, 1.02; 1.17]; PPD diagnosis: OR = 1.22 [95% CI, 1.00; 1.49] vs. 1.15 [95% CI, 0.97; 1.36]). (C) Complications mediated only a small fraction of the association between psychiatric history and PPD (proportion mediated: PPD symptoms = 0.68% [95% CI, 0.50%; 1.00%], PPD diagnosis = 0.42% [95% CI, 0.14%; 0.79%]).ConclusionsPsychiatric history and complications are independently associated with PPD, but complications explain only a negligible portion. These findings suggest that the link between psychiatric vulnerability and PPD is primarily driven by direct mechanisms rather than mediation through complications. Introduction Postpartum depression (PPD) affects one in six new mothers [1, 2]. Despite well‐documented risk factors, PPD remains underdiagnosed and undertreated [3, 4]. The complex interplay between risk factors likely contributes to challenges in identifying at‐risk women, underscoring the need for a deeper understanding of underlying disease mechanisms. A personal psychiatric history is the strongest predictor of PPD [5, 6, 7, 8, 9, 10, 11]. More specifically, risk of PPD is elevated across a broad range of psychiatric disorders, including major depressive disorder (MDD), bipolar disorder (BD), anxiety and panic disorders, posttraumatic stress disorder (PTSD), eating disorders, attention deficit hyperactivity disorder (ADHD), and obsessive–compulsive disorder (OCD) [12, 13, 14, 15]. Importantly, women with a psychiatric history are also at heightened risk of obstetric complications, particularly when having recent or persistent psychiatric episodes [16, 17]. These obstetric complications, which are also independently associated with an increased PPD risk, include preeclampsia [18], gestational diabetes [9, 19], gestational hypertension [10], hyperemesis gravidarum [20], postpartum hemorrhage [21], Cesarean section (C‐section) [22], and preterm birth [23, 24]. However, whether these obstetric complications mediate the association between personal psychiatric history and PPD remains unexplored. Examining this provides important insights to uncover mechanisms underlying PPD and advancing personalized prevention and intervention. Aims of the Study This study aimed to estimate the extent to which obstetric complications mediate the association between personal psychiatric history and PPD. We hypothesized that these obstetric complications partially mediate the association, meaning women with a psychiatric history are at increased risk of PPD both directly and indirectly through a higher likelihood of experiencing obstetric complications during pregnancy and delivery. Materials and Methods This study followed the AGReMA guideline [25]. A preregistered protocol is available at Open Science Framework (https://osf.io/x85d9/). Data Sources and Study Population This cohort study utilized the HOPE cohort, comprising 170,218 maternal Edinburgh Postnatal Depression Scale (EPDS) screenings conducted within 12 weeks postpartum (January 1, 2015, to December 31, 2021), linked with Danish population‐wide registers [26, 27]. The cohort is broadly representative of the Danish background population; details are available elsewhere [26]. To ensure that the outcome, PPD, was not a continuation of a pre‐existing depression, a wash‐out period was implemented. All mothers with a recorded depression diagnosis (ICD‐10: F32‐33) or a redeemed prescription for antidepressants (ATC: N06A) from conception until delivery were excluded. Data on ICD‐10 diagnoses and ATC codes were retrieved from the Psychiatric Central Research Register, the National Patient Register, and the National Prescription Register [28, 29, 30]. See Figure S1 for details on variables and temporal order. Exposure Personal psychiatric history was the exposure of interest. It was defined as either a redeemed psychotropic prescription (ATC: N05A (antipsychotics), N05BE01 (buspirone), N06A (antidepressants), N06BA (psychostimulants)) or a psychiatric diagnosis (ICD‐10: F00‐99) recorded from 1995 until conception. It was categorized as “Yes” or “No” and sourced from the Psychiatric Central Research Register, the National Patient Register, and the National Prescription Register [28, 29, 30]. Outcome PPD was the outcome of interest and was operationalized using two separate measures, both dichotomized (“Yes” or “No”): PPD symptoms were defined as a positive EPDS screening (cut‐off ≥ 11) within 12 weeks postpartum [31]. PPD diagnosis was defined as a redeemed antidepressant prescription (ATC: N06A) with indication for depression (indication codes: 168 and 270), or an in‐ or outpatient hospital contact with a depression diagnosis (ICD‐10: F32‐33) within 6 months postpartum. Data were sourced from the National Prescription Register, the National Patient Register, and the Psychiatric Central Research Register [28, 29, 30]. PPD symptoms were defined as a positive EPDS screening (cut‐off ≥ 11) within 12 weeks postpartum [31]. PPD symptoms were defined as a positive EPDS screening (cut‐off ≥ 11) within 12 weeks postpartum [31]. PPD diagnosis was defined as a redeemed antidepressant prescription (ATC: N06A) with indication for depression (indication codes: 168 and 270), or an in‐ or outpatient hospital contact with a depression diagnosis (ICD‐10: F32‐33) within 6 months postpartum. Data were sourced from the National Prescription Register, the National Patient Register, and the Psychiatric Central Research Register [28, 29, 30]. PPD diagnosis was defined as a redeemed antidepressant prescription (ATC: N06A) with indication for depression (indication codes: 168 and 270), or an in‐ or outpatient hospital contact with a depression diagnosis (ICD‐10: F32‐33) within 6 months postpartum. Data were sourced from the National Prescription Register, the National Patient Register, and the Psychiatric Central Research Register [28, 29, 30]. Previous research has shown that the overlap between PPD symptoms and PPD diagnosis in the HOPE cohort is only 0.8%, indicating that these are two distinct definitions of PPD [26]. Mediators Obstetric complications were the potential mediators of interest. Obstetric complications included preeclampsia/eclampsia (ICD‐10: O14‐15), gestational hypertension (ICD‐10: O13), gestational diabetes (ICD‐10: O24), hyperemesis gravidarum (ICD‐10: O21), postpartum hemorrhage (ICD‐10: O72 code VPH ≥ 1000) [32], acute C‐section (ICD‐10: DO821, DO843, DO829, DO842, procedure codes: KMCA10A and KMCA10E), and preterm birth (delivery < 37 weeks of gestation). Each obstetric complication was categorized as “Yes” or “No.” Based on all obstetric complications, a composite variable was created (“Any complications” or “No complications”). Data were collected from the Medical Birth Register and the National Patient Register [30, 33]. Covariates As outlined in the Directed Acyclic Graph (DAG) (Figure S2), analyses were adjusted for potential confounders in the associations between exposure, mediator, and outcome. These included parity, sourced from the Medical Birth Register, categorized as “1 child,” “2 children,” or “3+ children” [33]. Calendar year and maternal age at delivery, sourced from the Civil Registration System, were included as continuous variables and modeled linearly [34]. Education level, sourced from Statistics Denmark's Register of Education, categorized as “mandatory” (primary education), “short” (upper secondary and vocational training), “medium” (bachelor's degree), or “high” (master's and doctoral degrees) [35, 36]. And lastly, maternal origin of birth, sourced from the Civil Registration System, categorized as “Denmark” or “Not Denmark” [34]. Statistical Analyses Descriptive characteristics of the study population were stratified by exposure status, comparing mothers with and without psychiatric history. The statistical analyses followed a three‐step approach (Figure 1) using logistic regression models with robust standard errors to estimate odds ratios (OR) with 95% confidence intervals (CIs). Robust standard errors accounted for some mothers being included more than once in the HOPE cohort. First, associations between the exposure and outcome (adjusted for parity, calendar year, educational level, and origin of birth), the mediator and outcome (adjusted for personal psychiatric history, parity, calendar year, maternal age, educational level, origin of birth), and the exposure and mediator (adjusted for parity, calendar year, maternal age, educational level, and country of origin) were estimated to confirm independent associations. Second, the association between the mediator (obstetric complications) and outcome (PPD), stratified by the exposure (psychiatric history), was estimated using fully adjusted models. Due to minimal variation across strata, the mediation models were specified without an exposure‐mediator interaction term. Third, causal mediation analysis within the counterfactual framework was applied to examine the mediating role of obstetric complications in the association between psychiatric history and PPD [37]. This approach decomposed the total effect of psychiatric history on PPD into the natural direct effect (NDE)—the effect of psychiatric history on PPD independent of complications—and the natural indirect effect (NIE)—the effect mediated through obstetric complications. The proportion mediated was then estimated as the NIETotal effect[38]. The proportion mediated quantifies the proportion of the total effect explained by obstetric complications. The primary analysis focused on the mediator variable indicating the presence or absence of any obstetric complications, while individual obstetric complications were analyzed separately. A secondary mediation analysis stratified psychiatric history by timing: “≥ 3 years prior to conception” and “< 3 years prior to conception.” As missing data were < 5%, adjusted analyses were conducted using complete case analysis. CIs for the NDE and NIE were estimated using a quasi‐Bayesian approximation with 10 Monte Carlo draws. Analytical approach. Given the binary outcome, causal mediation analysis offers advantages over traditional methods due to its flexibility, of not being restricted to parametric models [39]. However, this approach comes with the following assumptions: (1) no unmeasured confounding of the exposure‐outcome relationship, (2) no unmeasured confounding of the mediator‐outcome relationship, (3) no unmeasured confounding of the exposure‐mediator relationship, and (4) no confounders of the mediator‐outcome effect that are affected by the exposure (cross‐world independence assumption) [39]. These assumptions were considered in the a priori DAG (Figure S2) and evaluated through a sensitivity analysis. All analyses were performed in R version 4.4.1, using the Mediation (version 4.5.0) package. Sensitivity Analyses Three sensitivity analyses were conducted to evaluate the robustness of the findings. First, to address potential unmeasured mediator‐outcome confounding, a sensitivity analysis inspired by Ding and VanderWeele was performed [40]. This analysis quantified the degree of unmeasured confounding required to nullify the point estimate and the lower CIs of the NIE and NDE. Second, to account for dependency between observations, a sensitivity analysis was performed restricting the sample to a random pregnancy per woman. Third, to ensure the mediator, obstetric complications, preceded the outcome, PPD, women with no psychiatric history who were diagnosed with a psychiatric disorder (ICD‐10: F00‐99) or received psychotropic medication (ATC: N05‐06) during pregnancy for the first time were excluded. Ethics The study was approved by the Danish Data Protection Agency (through local registration at Aarhus University, journal number 2016‐051‐000001, serial number 2304). All data were anonymized and analyzed on a secure platform at Statistics Denmark. Informed consent was not required in accordance with Danish law. Results Of 170,218 mothers included in the study population, 40,707 (23.9%) had a personal psychiatric history (Table 1). Compared to their counterparts without a personal psychiatric history, they had a significantly higher risk of developing both PPD symptoms (13.4% vs. 6.1%) and PPD diagnosis (7.0% vs. 0.4%), had lower educational levels (mandatory: 19.1% vs. 6.6%), slightly higher parity (3+ children: 15.4% vs. 11.6%), were more often of Danish origin (91.6% vs. 84.9%), and experienced more complications (34.1% vs. 28.5%). See Table S1 for characteristics of women classified based on prescriptions alone, diagnosis alone, or both. For the analysis population, a washout period excluded 5028 women with a depression diagnosis or antidepressant prescription fills during pregnancy (Figure S3). The proportion of excluded women was mainly among women having a personal psychiatric history (11.7% vs. 0.2%). Additionally, complete case analysis was used, excluding women with missing data on covariates (n excluded = 1864), ending up with an analysis population of 163,326. The proportion of exclusions was comparable between the two groups (education: 0.2% vs. 0.4%, parity: 0.8% vs. 0.8%, preterm birth: 0.8% vs. 0.7%). TABLE 1Characteristics of the study population separated on personal psychiatric history.CharacteristicsPersonal psychiatric historyNoYes N = 170,218129,511 (76.1%)40,707 (23.9%)PPD symptomsNo121,641 (93.9%)35,251 (86.6%)Yes7870 (6.1%)5456 (13.4%)PPD diagnosisNo128,987 (99.6%)37,845 (93.0%)Yes524 (0.4%)2862 (7.0%)Age at deliveryMean (SD)30.8 (4.7)31.1 (5.0)EducationMandatory8590 (6.6%)7764 (19.1%)Short36,596 (28.3%)14,378 (35.3%)Medium7001 (5.4%)1841 (4.5%)High76,788 (59.3%)16,635 (40.9%)Missing536 (0.4%)89 (0.2%)Parity1 child68,388 (52.8%)20,100 (49.4%)2 children45,063 (34.8%)13,996 (34.4%)3+ children15,081 (11.6%)6280 (15.4%)Missing979 (0.8%)331 (0.8%)Calendar year of delivery2014 a 2505 (1.9%)744 (1.8%)201516,172 (12.5%)4877 (12.0%)201618,415 (14.2%)5790 (14.2%)201719,613 (15.1%)6205 (15.2%)201820,867 (16.1%)6431 (15.8%)201920,924 (16.2%)6576 (16.2%)202015,863 (12.2%)5174 (12.7%)202115,152 (11.7%)4910 (12.1%)Origin of birthDenmark109,929 (84.9%)37,306 (91.6%)Not Denmark19,582 (15.1%)3401 (8.4%)Washout (ATC: N06A or ICD: F32‐33 during pregnancy)No129,253 (99.8%)35,937 (88.3%)Yes258 (0.2%)4770 (11.7%)Any complicationNo91,773 (70.9%)26,538 (65.2%)Yes36,870 (28.5%)13,877 (34.1%)Missing868 (0.7%)292 (0.7%)Preeclampsia/eclampsiaNo125,286 (96.7%)39,048 (95.9%)Yes4225 (3.3%)1659 (4.1%)Gestational hypertensionNo126,387 (97.6%)39,530 (97.1%)Yes3124 (2.4%)1177 (2.9%)Gestational diabetesNo123,734 (95.5%)38,160 (93.7%)Yes5777 (4.5%)2547 (6.3%)Hyperemesis gravidarumNo126,672 (97.8%)39,065 (96.0%)Yes2839 (2.2%)1642 (4.0%)Postpartum hemorrhageNo119,807 (92.5%)37,376 (91.8%)Yes9704 (7.5%)3331 (8.2%)C‐sectionNo111,519 (86.1%)33,976 (83.5%)Yes17,992 (13.9%)6731 (16.5%)Preterm birthNo124,862 (96.4%)38,901 (95.6%)Yes3724 (2.9%)1491 (3.7%)Missing925 (0.7%)315 (0.8%) a Some women gave birth in 2014; they are included because the EPDS screening was performed within 12 weeks postpartum, between January 1, 2015, and December 31, 2022. Characteristics of the study population separated on personal psychiatric history. Some women gave birth in 2014; they are included because the EPDS screening was performed within 12 weeks postpartum, between January 1, 2015, and December 31, 2022. Figure 2 displays the associations between the exposure and outcome, the mediator and outcome, and the exposure and mediator (Step 1, Figure 1). Women with a psychiatric history had more than twofold higher odds (OR = 2.32 [95% CI, 2.22; 2.41]) of PPD symptoms and a more than fivefold higher odds (OR = 5.09 [95% CI, 4.48; 5.79]) of PPD diagnosis compared to those without a psychiatric history. Women with obstetric complications had an OR of 1.16 (95% CI, 1.11; 1.21) for PPD symptoms and an OR of 1.18 (95% CI, 1.04; 1.34) for PPD diagnosis compared to those without. Women with personal psychiatric history had an OR of 1.19 (95% CI, 1.16; 1.23) for obstetric complications. For specific complications, see Supporting Information (Table S2a,b). Logistic regressions estimating OR (95% CI) for the association separately between the exposure, mediator, and outcome. Association between psychiatric history and PPD was adjusted for parity, calendar year, educational level, and country of origin. Association between psychiatric history and complications was adjusted for parity, calendar year, maternal age, educational level, and country of origin. Association between complications and PPD was adjusted for personal psychiatric history, parity, calendar year, maternal age, educational level, and country of origin. Table S3 shows the association between the mediator and outcome, stratified by the exposure (Step 2, Figure 1). Most point estimates across strata fall within each other's 95% CI, suggesting limited evidence of effect modification by psychiatric history. Table 2 displays the direct, indirect, and total mediating effects of obstetric complications in the association between psychiatric history and PPD, along with the proportion mediated (Step 3, Figure 1). See Supporting Information for specific complications (Table S4). Obstetric complications mediated 0.68% (0.50%; 1.00%) and 0.42% (0.14%; 0.79%) of the association between psychiatric history and PPD symptoms or diagnosis, respectively. The proportion mediated was similar for past episodes and recent episodes for PPD symptoms (0.62% [95% CI, 0.44%; 0.82%] vs. 0.73% [95% CI, 0.53%; 0.93%]) and slightly higher for past compared to recent episodes for PPD diagnosis (0.61% [95% CI, 0.14%; 1.09%] vs. 0.40% [95% CI, 0.04%; 0.81%]) (Table S5). TABLE 2Mediation analysis estimating OR (95% CI) and proportion mediated with personal psychiatric history as the exposure, complications as the mediator, and PPD as the outcome.Direct effectIndirect effectTotal effectProportion mediatedPPD symptoms1.07 (1.07; 1.07)1.00 (1.00; 1.00)1.07 (1.07; 1.07)0.68% (0.50%; 1.00%)PPD diagnosis1.01 (1.01; 1.02)1.00 (1.00; 1.00)1.01 (1.01; 1.02)0.42% (0.14%; 0.79%) Mediation analysis estimating OR (95% CI) and proportion mediated with personal psychiatric history as the exposure, complications as the mediator, and PPD as the outcome. Sensitivity analyses (Tables S6 and S7) showed similar trends. The sensitivity analysis inspired by VanderWheele and Ding indicated that, in the absence of any observed mediational effect, only a minimal level of unmeasured confounding would be required to explain the observed association (Tables S8 and S9). Discussion For this study, we aimed to disentangle the complexity behind how major risk factors directly and indirectly influence PPD risk by applying mediation analysis to explore potential underlying mechanisms. Specifically, we sought to investigate whether the strongest PPD risk factor—personal psychiatric history—partly exerts its effect through other established risk factors, namely obstetric complications. These obstetric complications are themselves more prevalent among women with a psychiatric history, raising the question of whether they represent a pathway through which psychiatric vulnerability increases PPD risk. Using a stepwise approach, we first confirmed independent associations between psychiatric history, obstetric complications, and PPD. Second, the association between obstetric complications and PPD was similar in strata with and without psychiatric history. And third, given that psychiatric history is associated with adverse pregnancy outcomes, we hypothesized that obstetric complications would partially mediate the association between psychiatric history and PPD [17]. However, our findings did not support this hypothesis, as only 0.68% (95% CI, 0.50%; 1.00%, PPD symptoms) and 0.42% (95% CI, 0.14%; 0.79%, PPD diagnosis) of the observed association was explained by complications. We confirmed that psychiatric history and obstetric complications were independently associated with PPD (Step 1, Figure 1). Our findings align with existing evidence, confirming the well‐established association between psychiatric history and both PPD symptoms and diagnosis, with a twofold and fivefold increase, respectively [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]. Compared to existing literature, our risk estimates of PPD diagnosis were somewhat higher than those reported on specific psychiatric disorders in an American study based on insurance claims. This divergence could possibly be due to population differences, as our study was population‐based within a universal healthcare system [12]. On the other hand, our risk estimate was much lower than in a Swedish population‐based study, which reported an almost twentyfold increase for PPD diagnosis. This may potentially be due to the washout period applied by us to ensure incident PPD cases [41]. Additionally, prior Danish research reported a low risk of PPD among mothers without a psychiatric history (0.6%), which we replicated (0.4%) [14]. Similarly, we confirmed that a wide range of obstetric complications were independently associated with PPD [9, 10, 18, 19, 20, 21, 22, 23, 24]. We showed that the association between obstetric complications and PPD was not modified by psychiatric history (Step 2, Figure 1), suggesting that obstetric complications increase the risk of PPD similarly among women with and without a psychiatric history. This contributes to a sparsely studied area, as most PPD research treats established risk factors as independent predictors, without examining whether they interact. As highlighted by Yim et al., interactive effects are rarely explored, and many studies include a broad set of predictors without explicitly examining how these factors may operate together or modify one another's effects [42]. Our results therefore add nuance to the evidence by providing results that two of the most well‐established PPD risk factors appear to operate independently, rather than synergistically. We did not find that obstetric complications mediate the association between psychiatric history and PPD (Step 3, Figure 1). Despite a recent study establishing a clear link between maternal psychiatric history and adverse neonatal outcomes, our hypothesis that the association between psychiatric history and PPD would be partially mediated through obstetric complications was not supported [17]. This suggests that while obstetric complications contribute to PPD risk, they do not explain the pathway through which psychiatric history influences PPD development. PPD affects a substantial number of mothers, yet it remains challenging to predict who will develop the disorder. We aimed to enhance the understanding of mechanisms underlying the disorder, as it may inform targeted strategies, ideally aiming to prevent PPD from developing. Our study contributes to this etiological understanding by examining whether obstetric complications mediate the association between psychiatric history and PPD. The finding that obstetric complications do not explain any part of this association suggests that these risk factors operate through distinct pathways. Although psychiatric history is not a modifiable risk factor per se and thus cannot be targeted directly to prevent PPD, it remains a critical predictor. Gaining deeper insights into how psychiatric vulnerability contributes to PPD development is essential for improving early identification and design of supportive interventions tailored to women at risk. Our findings highlight the need for further research into alternative mechanisms, such as biological, psychological, or social factors, to refine our understanding of PPD etiology and onset. Broadly, our findings suggest that both women with a psychiatric history and those with obstetric complications should receive close attention to enhance early identification and support. However, as our results indicate that these risk factors operate through separate pathways, clinical practice should consider tailored approaches. Women with a psychiatric history may benefit from early psychological support and monitoring throughout pregnancy and the postpartum. While those experiencing complications might require targeted interventions addressing the specific physical and emotional stressors associated with adverse obstetric outcomes, such as pain management, support during recovery, and reassurance around offspring health. Developing differentiated care pathways that reflect the distinct needs and mechanisms underlying each risk factor could improve intervention efforts. This approach also aligns with a more personalized model of perinatal mental healthcare, emphasizing risk stratification and timely, appropriate support based on individual risk profiles. Strengths and Limitations A key strength of this study is the use of the HOPE cohort, a population‐based cohort representative of the Danish population. By linking HOPE with population‐wide registers, this ensures a comprehensive coverage of the Danish source population on a wide range of information [26]. Additionally, HOPE provides a broad spectrum of PPD cases, including cases defined from EPDS scores, antidepressant use, and hospital diagnoses. We followed the AGReMA guideline to ensure transparency and robustness in the conduct and reporting of our mediation analysis. However, as recently highlighted, specific challenges frequently arise in psychiatric mediation studies [43]. Namely, lack of clear temporal ordering, insufficient control for confounding, and failure to consider exposure‐mediator interaction. We addressed these concerns in several ways. First, we leveraged longitudinal data to establish temporal ordering and further assessed the robustness of this ordering in a sensitivity analysis excluding mothers who received their first‐ever psychiatric diagnosis between conception and delivery. Second, we constructed an a priori DAG to clarify the assumed causal structure and guide confounder adjustment. Finally, we assessed interaction between exposure and mediator to avoid misspecification of the underlying model. Despite these strengths, our study also has limitations. First, the HOPE cohort is underrepresented in non‐Danish populations, limiting generalizability to immigrant populations [26]. Second, although we adjusted for key sociodemographic and reproductive factors (education, parity, maternal age at delivery, origin of birth, and calendar year of delivery), some potential confounders were not available in the registers, including physical activity [44], social support [45, 46], and unintended pregnancy [47]. Such unmeasured confounding could affect estimates of both direct and indirect effects and should be taken into account when interpreting the results of the mediation analysis (Figure S2). In addition, while our study focused on obstetric complications as mediators, we acknowledge that adverse neonatal outcomes could plausibly represent an additional pathway linking obstetric complications and PPD [17]. As adverse neonatal outcomes were not included in the mediation model, some indirect pathways may not be fully captured. Future studies could investigate this further. Third, while we adjusted for maternal education as a potential confounder, it is possible that early‐life psychiatric history influenced educational attainment for some women. Consequently, education may represent a pathway between exposure and outcome for certain individuals, rather than a pure confounder. Thus, adjusting for education may have resulted in some overadjustment, potentially attenuating the observed association. Finally, we dichotomized the EPDS score to define PPD symptoms using a cut‐off ≥ 11. While this approach identifies clinically relevant cases, it may result in some loss of information on the full range of symptom severity and precludes assessment of dose–response relationships between obstetric complications and depressive symptoms. However, separate analyses of individual obstetric complications were conducted (see Supporting Information), providing some insights into the differential associations of specific complications with PPD risk. Conclusion In this population‐based cohort study, we first confirmed that personal psychiatric history and obstetric complications independently were associated with PPD. Second, we found no evidence of interaction between psychiatric history and obstetric complications. And thirdly, obstetric complications did not mediate the association between personal psychiatric history and PPD, suggesting that these risk factors operate through distinct pathways. These findings highlight the importance of tailored approaches: women with a psychiatric history may benefit from proactive, continuous mental health care during the perinatal period, whereas those experiencing obstetric complications may require targeted screening and support in the immediate postpartum phase. Our results underscore the need for differentiated strategies that reflect the specific needs and risk profiles of distinct high‐risk groups, and they call for further research into the mechanisms linking psychiatric vulnerability to PPD to better inform such precision intervention efforts. Author Contributions Mette‐Marie Zacher Kjeldsen and Katrine Holde had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: All authors. Acquisition, analysis, or interpretation of data: Mette‐Marie Zacher Kjeldsen, Katrine Holde, Liselotte Vogdrup Petersen, and Trine Munk‐Olsen. Drafting of the manuscript: Mette‐Marie Zacher Kjeldsen. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: Katrine Holde. Obtained funding: Liselotte Vogdrup Petersen and Trine Munk‐Olsen. Administrative, technical, or material support: N/A. Supervision: Liselotte Vogdrup Petersen and Trine Munk‐Olsen. Funding M.M.Z.‐K. is supported by the Lundbeck Foundation (grant nos. 32691 and R433‐2023‐565). T.M.‐O. and M.L.M. are also supported by the Lundbeck Foundation (grant no. 32691). V.B., E.E.B., and T.M.‐O. are funded by the National Institute of Mental Health (NIMH, grant no. R01MH122869). The funding agencies had no role in the design, data collection, analysis, interpretation, writing the manuscript, or the decision to submit the manuscript for publication. Disclosure The corresponding author (M.‐M.Z.K.) confirms that the manuscript represents an honest, accurate, and transparent account of the study's finding. Conflicts of Interest K.B.M. has received a speaker's fee from Medice Nordic within the last 3 years. T.M.‐O. has received a speaker's fee from Lundbeck A/S within the last 3 years. The other authors declare no conflicts of interest. Supporting information Figure S1: Illustration of population, exposure, mediator, and outcome in relation to temporal order. Figure S2: Directed acyclic graph (DAG) on the confounding structure. Table S1: Characteristics of study population separated on prescriptions and diagnosis for personal psychiatric history. Figure S3: Flowchart from the study population to the analysis population. Table S2: (a) Logistic regressions estimating OR (95% CI) for the association between the exposure and outcome, as well as the mediator and outcome. (b) Logistic regressions estimating OR (95% CI) for the association between the exposure and mediator. Table S3: Logistic regressions estimating OR (95% CI) for the mediator stratified by the exposure. Table S4: Mediation analysis estimating OR (95% CI) and the proportion mediated for specific complications. Table S5: Mediation analysis estimating OR (95% CI) and the proportion mediated stratified by timing of the exposure (recent and past). Table S6: Sensitivity analysis of mediation analysis restricted to one random birth per mother in HOPE. Table S7: Sensitivity analysis of mediation analysis with exclusion of women with no prior psychiatric history who receive a psychiatric diagnosis or psychotropic medication from conception until delivery. Table S8: Sensitivity analysis estimating the E value for the indirect effect. Table S9: Sensitivity analysis estimating the E value for the direct effect.