PurposeWe aimed to determine the incidence of mental health diagnoses and associated health and social risk factors among perinatal people in three different COVID-19 phases.MethodsWe conducted a population-based, retrospective cohort study using linked administrative datasets. We included persons with live, in-hospital births in Ontario, Canada from January 1 to March 31 in 2019, 2021, or 2022 (three phases relative to COVID-19 with different public health policy measures). We excluded people with prior mental health diagnoses. We used diagnostic codes to identify new onset of depression, anxiety, or adjustment disorder in the antenatal and postpartum period. We developed multivariable, modified Poisson models to examine associations between sociodemographic and clinical factors and new mental health diagnoses in each phase.ResultsThere were 72,242 people in our cohort. Antenatal mental health diagnoses were significantly higher in 2021 (aRR = 1.32; CI = 1.20–1.46) and 2022 (aRR = 1.22; CI = 1.11–1.35) versus 2019. Postpartum diagnoses were significantly greater in 2021 (aRR = 1.16; CI = 1.08–1.25) versus 2019. Antenatal diagnoses were associated with birth year, previous stillbirth, pre-existing hypertension, multiparity, residential instability, and ethnocultural diversity. Postpartum diagnoses were associated with birth year, maternal age, multiparity, care provider profession, assisted reproductive technology, birthing mode, pre-existing hypertension, intensive care admission, hospital readmission, residential instability, and ethnocultural diversity. Family physicians increasingly made mental health diagnoses in 2021 and 2022.ConclusionIncreased incidence of perinatal mental health diagnoses during COVID-19 suggests complex dynamics involving pandemic and health and social risk factors.RegistrationThis study was registered with Clinicaltrials.gov (NCT05663762) on December 21, 2022.