B. Model input parameters
Resource use: Detailed finance data was only available for the lead study site, and these data have been used for all participants. Costs have been estimated inclusive of nursing / midwifery, medical and allied health staff costs, equipment and consumables, pharmacy, pathology, radiology, hotel services and business overheads. These have been calculated as a cost per hour of admission to antenatal ward, birth suite, theatre, postnatal ward and neonatal nursery. Cervical priming cost (PGE2 gel, PGE2 tape, cervical ripening balloon) was available at the patient level, whereas the hourly costs for each inpatient clinical unit have been calculated by dividing total cost by total clinical hours. The precise durations of care in each clinical unit was recorded for all 448 trial participants who were induced.
Health outcomes: Health status was measured using the generic EQ-5D-3L questionnaire and comprised (a) five domains: mobility, self-care, usual activities, pain/discomfort and anxiety/depression with 3 response levels each (no problems, some problems and extreme problems); and (b) a self-reported summary measure of health, using a visual analogue score (EQVAS), ranging from 0 (worst imaginable health state) to 100 (perfect health) (17). These health state assessments were available for 359 women who completed their assessment. The health utility index ranged from 0 (worst possible health) to 1 (best possible health) and values represented the utility scores used in the model according to Australian population norms (18).
Transition probabilities: Transition probabilities between model pathways were derived by calculations from observed events in the actual trial data (see Supplementary material, Figure S2). As no deaths were observed we did not include any risk of mortality associated with either method of IOL.