The purpose of this pilot study was to perform a cost-identification analysis of care for gestational diabetes mellitus (GDM) by determining the direct costs of the diagnostic procedures and treatment used for the outpatient management of GDM (program input costs) and the direct costs of maternal hospitalization after diagnosis of GDM, delivery of the baby, and newborn care (outcome costs). Reimbursed average charges in the Northern California (NoCal) managed care market in 1996 were used to establish the direct costs, and the direct costs were then applied to the elements of care and pregnancy outcomes of three GDM management programs in NoCal, Southern California (SoCal), and New England (NewEng), using prospectively collected data. Reimbursed amounts for the detailed elements of GDM management (program input costs) are presented in the categories of diagnosis of GDM, diabetes treatment supplies, doctor's office visits, office visits to ancillary providers, and fetal surveillance. Program input costs per patient were $817 for diet-treated and $1,838 for insulin-treated women in NoCal, and were estimated to be $882 for diet-treated and $1,425 for insulin-treated women in NewEng. Program input costs for women requiring insulin treatment who were randomized to premeal or postprandial blood glucose testing (N Engl J Med 333:1237, 1995) in SoCal were estimated to be $3,596 per patient for the premeal group and $3,770 per patient for the postprandial group. Reimbursed amounts for health care expenditures related to pregnancy outcomes are detailed in the categories of hospital and physician charges for maternal antepartum hospitalization ($1,864 for 2 days), vaginal delivery with 50% use of epidural anesthesia ($4,050), cesarean section ($5,932), and neonatal intensive care ($9,130 for 4 days). Outcome costs per patient were $5,792 for diet-treated and $6,462 for insulin-treated women in NoCal. Outcome costs per patient were estimated to be $6,096 for diet-treated and $11,216 for insulin-treated women in NewEng, and $8,013 for the premeal blood glucose group and $7,495 for the postprandial blood glucose group in SoCal (both groups required insulin treatment). Incremental cost-effectiveness of postprandial monitoring in the SoCal controlled trial was $35 per patient in input costs per cesarean section averted and $25 per patient in input costs per neonatal intensive care unit day prevented. The benefit-to-cost ratio of the difference in input and outcome costs was 2.98 in favor of postprandial monitoring in the SoCal study. Cost analysis should be included in clinical trials of the management of GDM.