Digital and AI-assisted approaches across the preeclampsia care pathway: a scoping review

To map peer-reviewed evidence across four linked functions of digital and AI-assisted preeclampsia care-risk prediction, patient education, remote blood pressure monitoring, and clinician-directed escalation-and determine whether these functions have been evaluated together as an integrated clinical service.

PubMed, Scopus, and ScienceDirect were searched for records published from January 2010 to May 11, 2026.

Peer-reviewed original empirical and technical studies of artificial intelligence, machine learning, clinical decision support, telemedicine, mobile health, remote monitoring, digital education, or care escalation for preeclampsia or hypertensive disorders of pregnancy were eligible. Reviews, protocols, conference abstracts without a full report, editorials, and opinion articles were excluded.

The review followed PRISMA-ScR and Joanna Briggs Institute guidance. Two reviewers independently screened records. One reviewer charted data and a second independently verified them. Findings were organized by the four predefined functions and synthesized thematically. Formal critical appraisal was not performed; the synthesis therefore describes the evidence map rather than certainty or implementation readiness.

The searches identified 1396 records. After 313 duplicates were removed, 1083 records were screened, and 320 full-text articles were assessed; 286 full-text articles were excluded, and 34 studies conducted in diverse single- and multicountry settings were included. Prediction models reported favorable discrimination estimates, but calibration, independent validation, and equity assessment were inconsistent. Remote monitoring findings were mixed: BUMP 1 did not show earlier clinic-recorded detection of hypertension, whereas some replacement-care models reduced visits or admissions without an observed increase in adverse outcomes. Education studies mainly examined hypothetical responses or communication gaps rather than measured behavior change. Adoption surveys and qualitative implementation studies did not establish clinical effectiveness. No study prospectively evaluated all four functions as a coordinated service; this was treated as a descriptive observation, not evidence of effectiveness or novelty.

Component evidence supports staged study, not routine implementation of an integrated pathway. Searching three electronic information sources, excluding gray literature, and not performing formal appraisal may have omitted relevant implementation evidence. Future codesigned prospective studies should evaluate calibration, safety, workload, equity, and maternal and perinatal outcomes.