Clapp Mark A, Castro Victor M, Verhaak Pilar, McCoy Thomas H, Shook Lydia L, Edlow Andrea G, Perlis Roy H
Department of Obstetrics and Gynecology (Clapp, Shook, Edlow) and Center for Quantitative Health and Department of Psychiatry (Castro, Verhaak, McCoy, Perlis), Massachusetts General Hospital and Harvard Medical School, Boston; Research Information Science and Computing, Mass General Brigham, Somerville, MA (Castro).
Am J Psychiatry. 2025 Jun 1;182(6):551-559. doi: 10.1176/appi.ajp.20240381. Epub 2025 May 19.
Postpartum depression (PPD) is a major contributor to postpartum morbidity and mortality. Beyond efforts at routine screening, risk stratification models could enable more targeted interventions in settings with limited resources. The authors sought to develop and estimate the performance of a generalizable risk stratification model for PPD in patients without a history of depression, using information collected as part of routine clinical care.
The authors conducted a retrospective cohort study of all individuals who delivered between 2017 and 2022 in one of two large academic medical centers and six community hospitals. An elastic net model was constructed and externally validated to predict PPD, defined as having a mood disorder, an antidepressant prescription, or a positive screen on the postpartum Edinburgh Postnatal Depression Scale. Predictors used included sociodemographic factors, medical history, and prenatal depression screening information, all of which were known before discharge from the delivery hospitalization.
The cohort included 29,168 individuals; 2,696 (9.2%) met at least one criterion for postpartum depression in the 6 months following delivery. In the external validation data, the model had good discrimination and remained well calibrated: the area under the receiver operating characteristic curve was 0.721 (95% CI=0.709, 0.736), and the Brier calibration score was 0.087 (95% CI=0.083, 0.091). At a specificity of 90%, the positive predictive value was 28.8% (95% CI=26.7, 30.8), and the negative predictive value was 92.2% (95% CI=91.8, 92.7).
These findings demonstrate that a simple machine-learning model can be used to stratify the risk for PPD before delivery hospitalization discharge. This tool could help identify patients within a practice at the highest risk and facilitate individualized postpartum care planning for the prevention of, screening for, and management of PPD at the start of the postpartum period and potentially the onset of symptoms.
产后抑郁症(PPD)是导致产后发病和死亡的主要因素。除了进行常规筛查外,风险分层模型能够在资源有限的情况下实现更具针对性的干预。作者试图利用作为常规临床护理一部分收集的信息,开发并评估一种适用于无抑郁症病史患者的PPD通用风险分层模型的性能。
作者对2017年至2022年期间在两家大型学术医疗中心和六家社区医院之一分娩的所有个体进行了一项回顾性队列研究。构建了一个弹性网络模型并进行外部验证,以预测PPD,PPD定义为患有情绪障碍、开具抗抑郁药处方或产后爱丁堡产后抑郁量表筛查呈阳性。使用的预测因素包括社会人口统计学因素、病史和产前抑郁筛查信息,所有这些在分娩住院出院前都是已知的。
该队列包括29168名个体;2696名(9.2%)在分娩后的6个月内符合至少一项产后抑郁症标准。在外部验证数据中,该模型具有良好的区分度且校准良好:受试者工作特征曲线下面积为0.721(95%CI=0.709,0.736),Brier校准分数为0.087(95%CI=0.083,0.091)。在特异性为90%时,阳性预测值为28.8%(95%CI=26.7,30.8),阴性预测值为92.2%(95%CI=91.8,92.7)。
这些发现表明,一个简单的机器学习模型可用于在分娩住院出院前对PPD风险进行分层。该工具可帮助识别医疗机构中风险最高的患者,并促进个性化的产后护理计划,以便在产后初期及可能出现症状时预防、筛查和管理PPD。