2022 · IEEE journal of biomedical and health informatics · paywalled
AI Aided Analysis on Saliva Crystallization of Pregnant Women for Accurate Estimation of Delivery Date and Fetal Status
Li et al.
The finding, in our words
AI analysis of salivary crystallisation patterns from self-collected saliva predicted delivery dates with high accuracy in a small cohort, offering a virtually painless home monitoring option for pregnant women that could support decentralised antenatal care.
A paraphrase to the Library’s standard, never the abstract. The source is one link away and is always the authority.
An intensive protocol combining daily at-home saliva collection for ovarian hormones with smartphone-based ecological momentary assessment achieved 85% compliance in 50 female smokers, revealing substantial within- and between-subject heterogeneity in hormone levels, anxiety and nicotine craving. This demonstrates feasibility of decentralised, patient-centric data collection for studying complex comorbidities.
In a Phase 1 crossover study of etrasimod, self-collected blood microsamples yielded pharmacokinetic exposures comparable to conventional venous sampling, both with and without participant practice sessions. Continuous wearable sensors concurrently enabled real-time remote monitoring of safety vital signs and electrocardiograms, demonstrating the viability of hybrid decentralised designs in early-phase clinical trials.
This review assesses wearable sensors for the continuous biochemical monitoring of body fluids such as sweat, saliva and interstitial fluid, confirming their potential for decentralised healthcare through pilot trials. It emphasises that successful clinical translation depends on large-scale validation and the integration of ethical and sociocultural considerations.
Daily remote urine hormone monitoring of LH and PdG, together with age and cycle start day, can identify menstrual cycle phase and day with 95% confidence; this supports precise, patient-centric cycle tracking outside clinic settings.
Decentralised monitoring of people with type 2 diabetes using continuous glucose monitoring and a hybrid smartwatch was feasible, with high participant retention and successful collection of glucose, heart rate, activity, sleep and GPS data via a study app.