The resulting four-profile solution separated participants into a low-symptom profile, a profile marked by diminished positive affect, a profile with prominent self-blame and sleep disturbance, and a profile combining high symptoms with elevated self-harm-related thoughts. Because latent profile analysis identifies statistical groupings, the profiles describe patterns within this survey rather than established clinical categories. The low-symptom profile was the largest, containing 427 participants, or 42.2% of the sample.

A further 203 workers, representing 20.1%, were assigned to the diminished-positive-affect profile. The self-blame and sleep-disturbance profile included 260 participants, or 25.7%, while 122 workers, or 12.1%, were placed in the high-symptom profile with elevated self-harm-related thoughts. Together, the three profiles beyond the low-symptom group accounted for 57.9% of participants.

The model’s entropy was 0.845, and average posterior probabilities for the profiles ranged from 0.869 to 0.945. Multinomial logistic regression was then used to examine whether sociodemographic, occupational, familial and psychological factors were associated with profile membership. Candidate factors were selected using unadjusted omnibus between-profile tests with P values below 0.05.

After false discovery rate correction, two factors remained associated with profile membership: career motivation and participants’ reports of a history of depression in their mother. These were statistical associations in the model, not evidence that either factor caused a symptom profile. The analysis therefore links occupational and family-related characteristics to differences in profile membership without establishing a direction of effect or temporal sequence.

The results show that total EPDS screening results can conceal variation in symptom composition among women in this occupational sample. Some participants had comparatively low symptom levels, whereas others showed reduced positive affect, pronounced self-blame and disturbed sleep, or a broader high-symptom pattern that included elevated self-harm-related thoughts. Several features limit the findings.

The anonymous online survey was cross-sectional, so it could not establish whether symptoms or associated factors changed over time or whether the profiles persisted after childbirth. Participants were at different stages within the 24-month postpartum window, and the sample came from one geographic area and one occupational group. The symptom profiles and associated factors were also described as exploratory, while the 61.5% screening-positive proportion applies only to the surveyed sample and is not an estimate of diagnostic prevalence in the wider population.

Subject of Research: Latent profiles of postpartum depressive symptoms among female healthcare workers in northeastern Sichuan, China

Article Title: Psychological symptom profiles of postpartum depressive symptoms among female healthcare workers: a cross-sectional latent profile analysis

Article References: Su, L., Chen, Y., Wen, G., Tan, M., Wei, X., Wang, M., Li, F., & Ke, X. (2026). Psychological symptom profiles of postpartum depressive symptoms among female healthcare workers: a cross-sectional latent profile analysis. BMC Psychology. https://doi.org/10.1186/s40359-026-05577-0

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05577-0

Keywords: postpartum depression, female healthcare workers, EPDS, latent profile analysis, symptom profiles, self-blame, sleep disturbance, self-harm-related thoughts, cross-sectional study, China, Psychological, symptom