Physical Activity During Pregnancy in Women With Pre-existing Diabetes

This is a multi-centre, mixed-methods observational study designed to examine physical activity patterns across pregnancy in women with pre-existing Type 1 diabetes (T1D) or Type 2 diabetes (T2D), and to investigate associations with glucose control and diabetes-related emotional wellbeing. The study integrates longitudinal quantitative assessments with qualitative insights into participant experiences. The study will be conducted across NHS sites in the South West of England. Following informed consent, participants will be assigned a unique study identification number. The study is observational and does not involve any intervention or alteration to routine clinical care. Quantitative Data Collection Physical activity will be assessed using objective and self-reported measures. Device-measured physical activity will be collected using a GENEActiv wrist-worn accelerometer worn continuously for 7 days during each trimester (8-13, 18-22, and 26-31 weeks' gestation). Accelerometer data will be processed using validated analytical software to derive summary metrics including daily minutes of moderate-to-vigorous physical activity (MVPA), total activity, sedentary time, and sleep. Structured exercise will be recorded using a study diary completed during each monitoring period. The diary captures activity type, timing, and perceived intensity, and includes prospective and retrospective components to characterise habitual behaviour. Dietary intake will be assessed using the Remote Food Photography Method (RFPM). Participants will submit photographs of meals and snacks via secure digital platforms. Images will be linked to study data using the assigned study identification number. Continuous glucose monitoring (CGM) data corresponding to each monitoring period will be obtained from routine clinical data-sharing platforms used in standard care. De-identified glucose data will be securely transferred for analysis. Glucose metrics will be derived according to international consensus standards, including time in range, mean glucose, and glycaemic variability. Diabetes-related emotional wellbeing will be assessed at each monitoring period using the 20-item Problem Areas in Diabetes (PAID) questionnaire administered electronically. Qualitative Component A purposive sub-sample of \~16 participants will be invited to participate in online focus groups. Discussions will explore experiences of physical activity during pregnancy, perceived barriers and facilitators, interactions between activity and diabetes self-management, and perceived support needs. Sessions will be audio-recorded, transcribed verbatim, anonymised, and analysed using reflexive thematic analysis. Data Management and Quality Assurance All participants will be assigned a unique study identification number. Identifiable information will be stored separately from research data. Data transfers between sites and the coordinating centre will use secure, approved electronic systems. Data will be stored on secure institutional servers with restricted access. Accelerometer and CGM data will undergo predefined processing and validation procedures to ensure completeness and plausibility. Range and consistency checks will be applied to quantitative variables prior to analysis. Qualitative transcripts will be reviewed against audio recordings to ensure transcription accuracy. Sample Size Considerations The target sample size is 175 participants (75 with T1D and 100 with T2D). The study is powered to detect a clinically meaningful change in daily MVPA across pregnancy, accounting for repeated measures and potential clustering by site. Statistical Analysis Changes in MVPA across pregnancy will be analysed using multilevel linear modelling to account for repeated measures within participants and clustering by site. Pregnancy stage will be modelled both categorically (trimester) and continuously (gestational week). Associations between physical activity, glucose metrics, and emotional wellbeing will be examined using regression-based approaches adjusted for relevant covariates. Missing quantitative data will be examined to determine patterns of missingness. Multilevel models inherently accommodate unbalanced repeated measures; additional sensitivity analyses will be conducted if appropriate. Qualitative data will be analysed using Braun and Clarke's reflexive thematic analysis to identify themes related to behavioural capability, opportunity, motivation, and contextual influences. The study does not include an interventional component and involves no modification to usual clinical management.