A decision tree was used to assess the costs and outcomes of the two interventions, by combining data from published studies. The time horizon was 75 years. The authors stated that the perspective was societal.
The effectiveness data were from a literature review in PubMed. This review identified studies in English, using several key terms. Studies without a control group and reviews were excluded. Each model parameter was the mean of selected studies, weighted by sample size, and its range was the lowest and highest reported values. The main estimate of effectiveness was the relative risk of cerebral palsy and death with versus without treatment with magnesium. This was derived from four published trials, in which magnesium was primarily given for the prevention of cerebral palsy.
Monetary benefit and utility valuations:
The utility estimates were from two published studies.
Measure of benefit:
The benefit measure was quality-adjusted life-years (QALYs) which were discounted at an annual rate of 3%.
The direct costs included magnesium treatment, magnesium reactions, cerebral palsy, neonatal death, and neonatal survival. These estimates were from the published literature or Medicare reimbursement rates. Charges were multiplied by a cost-to-charge ratio of 0.6. The costs were discounted using an annual rate of 3% and reported in US $.
Analysis of uncertainty:
One-way and multiway sensitivity analyses were performed to assess how variations in the model parameters changed the results. A probabilistic sensitivity analysis was undertaken, by sampling the point estimates from distributions around the model parameters, using Monte Carlo simulation. Two scenario analyses were performed. In one, magnesium was only offered to women with pre-term premature rupture of the membranes. In the other, it was offered only to women at risk of delivery at a gestational age of less than 28 weeks.