Clinical Care/Education/Nutrition B R I E F
R E P O R T
Screening for Gestational Diabetes Mellitus A decision and cost-effectiveness analysis of four screening strategies WANDA K. NICHOLSON, MD, MPH1,2 LEE A. FLEISHER, MD3
HAROLD E. FOX, MD, MHS1 NEIL R. POWE MD, MPH, MBA4,5,6
G
sequential strategy using a decision model.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
standard cutoff value of 140 mg/dl for the 50-g glucose challenge test (sensitivity 80%, specificity 86%) (6) and the World Health Organization 2-h threshold value (ⱖ140 mg/dl) for the 75-g GTT (7). Because of limited data on the efficacy of the 75-g GTT in pregnancy, we used a baseline sensitivity and specificity of 80 and 86%, respectively, but varied sensitivity and specificity to 100% (8). Modified threshold levels were used for the 100-g GTT (fasting ⬍95 mg/dl, 1 h 180 mg/dl, 2 h 155 mg/dl, 3 h 145 mg/dl), and a sensitivity and specificity of 100% was assumed (9). Costs for maternal and infant care were estimated using the 2003 Medicare resource-based relative value units (10) and the 1997–2001 Maryland Health Care Commission Database (11). This database contains clinical and cost information on hospital discharges for maternity care within the state. Hospital charges were converted to costs using cost-tocharge ratios (12). Physician and hospital costs were adjusted to 2003 dollars based on the annual medical care component of the Consumer Price Index (14). The model incorporates three maternal health states: perfect health (utility: 1.0), perfect health following hysterectomy (0.9), and maternal death (0). There were three neonatal health states: none/ mild morbidity (utility: 1), moderate morbidity (0.7), and severe morbidity/ neonatal death (0) (13,14). QALYs for both mothers and infants were then calculated by combining the utility estimates with the 2003 life table data from the National Center for Health Statistics. Costs and utilities were discounted 3% per year (5). Indirect costs, including lost productivity and wages for an average U.S. female worker, were estimated using the Bureau of Labor Statistics (15). In the base-case analysis, the costeffectiveness of screening a 30-year-old pregnant woman between 24 and 28 weeks gestation was estimated with the incremental cost-effectiveness ratio, expressed as QALYs gained. A strategy was considered “dominated” if it was both
1482
DIABETES CARE, VOLUME 28, NUMBER 6, JUNE 2005
estational diabetes mellitus (GDM), defined as carbohydrate intolerance of variable degree with onset or first recognition during pregnancy, is the most common medical condition of pregnancy. GDM affects ⬃190,000 (4 –5%) of the ⬎4 million births occurring annually in the U.S. and is associated with several maternal and infant complications (1). Worldwide, the three primary screening strategies for GDM are the sequential strategy (initial 50-g glucose challenge test followed by, in those who test positive, a 100-g glucose tolerance test [GTT]), the 75-g GTT strategy, and the 100-g GTT strategy (2). The efficacy of these strategies, however, is debated. There are few randomized trials on the effectiveness of GDM screening (3). Yet, the majority of U.S. obstetricians provide universal screening for GDM (4). We conducted a cost-effectiveness analysis to compare four screening strategies for universal screening of GDM, including the sequential strategy, the 75and 100-g GTT, and a no-screening strategy. We assessed the relative cost and effectiveness (quality-adjusted life-years [QALYs]) of each strategy relative to the
RESEARCH DESIGN AND METHODS — We performed a decision and cost-effectiveness analysis from a societal perspective that incorporates all health effects and medical resources consumed regardless of who pays (5). Separate cost-effectiveness ratios were estimated for mothers and infants. Maternal outcomes were hypertensive disease, polyhydramnios, cesarean or vaginal delivery, and the potential complications of cesarean and vaginal delivery. Complications included operative injury, endometritis, deep vein thrombosis, severe hemorrhage requiring blood transfusion, and hysterectomy. Neonatal outcomes were mild hypoglycemia (ⱕ50 g/dl requiring intravenous fluid and observation), macrosomia (birth weight ⬎4,500 g), respiratory distress syndrome, shoulder dystocia, none/mild morbidity, moderate morbidity, and severe morbidity/infant death. Test effectiveness (sensitivity, specificity), GDM prevalence, clinical probabilities, and quality of life measures were obtained from the literature. We used the
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
From the 1Department of Gynecology and Obstetrics, The Johns Hopkins School of Medicine, Baltimore, Maryland; the 2Department of Population and Family Health Sciences, The Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland; the 3Department of Anesthesiology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; the 4Department of Medicine, The Johns Hopkins School of Medicine, Baltimore, Maryland; the 5Welch Center for Prevention, Epidemiology, and Clinical Research, The Johns Hopkins School of Medicine, Baltimore, Maryland; the 6Department of Epidemiology, The Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland. Address correspondence and reprint requests to Wanda Nicholson, MD, MPH, Department of Gynecology and Obstetrics, The Johns Hopkins School of Medicine, 600 North Wolfe St., Phipps 247, Baltimore, MD 21287. E-mail:
[email protected]. Received for publication 17 December 2004 and accepted in revised form 21 February 2005. Abbreviations: GDM, gestational diabetes mellitus; GTT, glucose tolerance test; QALY, quality-adjusted life-year. A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion factors for many substances. © 2005 by the American Diabetes Association.
Nicholson and Associates
Table 1—Cost-effectiveness of screening strategies for GDM*
Screening strategies Maternal model Sequential (50-g GCT ⫹ 100-g GTT)‡ 100-g GTT 75-g GTT No screening Neonatal model Sequential (50-g GCT ⫹ 100-g GTT) 100-g GTT 75-g GTT No screening
Cost of strategy in 2003 (U.S. $)
Incremental cost ($)
Effectiveness of strategy (QALY)
Incremental effectiveness (QALY)
Cost-effectiveness ratio ($ per QALY saved)†
$2,836 $2,874 $2,895 $2,995
Referent ⫹36.00 ⫹59.00 ⫹159
25.9219 25.9231 25.91557 25.9201
Referent ⫹0.001 ⫺0.007 ⫺0.002
Referent $32,374 Dominated‡ Dominated
$77 $89 $91 $80
Referent ⫹11 ⫹13 ⫹3
29.9177 29.9190 29.9108 29.8985
Referent ⫹0.001 ⫺0.007 ⫺0.02
Referent $8,252 Dominated‡ Dominated
*Data are from the base-case analysis using a GDM prevalence of 4%. †Cost-effectiveness ratio of the 75- and 100-g strategies compared with the sequential strategy (reference strategy). ‡Dominance occurs when one clinical strategy is both more costly and less effective compared with the reference strategy. The 75-g GTT strategy has both higher cost (maternal outcomes 关$2,895 vs. $2,836兴; infant outcomes 关$91 vs. $77兴) and less effectiveness (maternal outcomes 关$25.91557 vs. 25.9219兴; infant outcomes 关29.9108 vs. 29.9177兴) than the sequential strategy. ⫹, QALYs gained; ⫺, QALYs lost. GCT, glucose challenge test.
more costly and less effective than its comparative strategy. One-way sensitivity analyses were performed in which model parameters (e.g., GDM prevalence, probabilities, utilities, costs) were changed individually to bias them in favor or against screening with the sequential strategy. Cost-effectiveness ratios of ⬍$50,000/ QALY were considered highly favorable toward a particular screening strategy (5). RESULTS — With a GDM prevalence of 4%, the sequential strategy (reference strategy) was the least costly strategy for mothers and infants (Table 1). Absolute cost and effectiveness of the sequential strategy was $2,836.00 and 25.9219 QALYs for maternal outcomes and $77.00 and 29.9177 QALYs for infant outcomes. The cost-effectiveness ratio for the 100-g strategy compared with the sequential strategy was highly favorable for both maternal outcomes ($32,374 per QALYs saved) and neonatal outcomes ($8,251 per QALYs saved), respectively. The 75-g strategy was more costly and less effective (dominated) compared with the sequential strategy for maternal and neonatal outcomes. The no-screening strategy was less effective (⫺0.002) and more costly ($158.80) relative to the sequential strategy, probably due to the additional cost and disutility in unrecognized cases (40 per 1,000) of GDM. CONCLUSIONS — The sequential screening method was the most costDIABETES CARE, VOLUME 28, NUMBER 6, JUNE 2005
effective strategy. Because the 100-g strategy had a highly favorable costeffectiveness ratio, it may be useful in populations where GDM is more prevalent (e.g., Hispanics) to avoid the need for subsequent confirmatory testing. Future screening guidelines might incorporate the prevalence of GDM in different racial groups. While the no-screening strategy has no upfront screening costs, the initial cost savings is not sufficient to overcome the cost and disutility of potential obstetrical complications. There are limitations to this study, most notably the lack of well-established test parameters for the 75-g GTT in pregnancy and the paucity of data on newborn health utilities. However, the costeffectiveness ratios for screening strategies were not highly influenced at the extremes of our estimates in sensitivity analyses. We recognize that the probability of maternal or neonatal outcomes is dependent on the severity of glucose intolerance (16). Pregnancies with mild GDM controlled by diet alone or falsenegative cases with mild glucose intolerance may have outcomes similar to normal pregnancy. Finally, the analysis does not include potential effects of the intrauterine environment on the longterm health of the offspring. We conclude that conventional use of the sequential strategy is the preferred strategy followed by the 100-g GTT. The 75-g GTT and the no-screening strategy are not currently viable screening methods.
Acknowledgments — W.K.N. is supported by the Harold Amos Medical Faculty Development Award of the Robert Wood Johnson Foundation and a Patient-Oriented Mentored Research Scientist Award (1K23DK067944) from the National Institute of Diabetes and Digestive and Kidney Diseases. N.R.P. is supported in part by K24-DK-02643 from the National Institute of Diabetes and Digestive and Kidney Diseases.
References 1. Gabbe SG, Graves C: Management of diabetes mellitus complicating pregnancy. Obstet Gynecol 102:857– 868, 2003 2. American Diabetes Association: Gestational diabetes mellitus (Position Statement). Diabetes Care 27 (Suppl. 1):S88 – S90, 2004 3. Brody SC, Harris R, Lohr K: Screening for gestational diabetes: a summary of the evidence for the U.S. Preventive Services Task Force. Obstet Gynecol 101:380 –392, 2003 4. Gabbe SG, Hill L, Schmidt L, Schulkin J: Management of diabetes by obstetriciangynecologist. Obstet Gynecol 91:643– 647, 1998 5. Russell L, Gold M, Siegel J, Daniels N, Winstein M: The role of cost-effectiveness analysis in health and medicine: panel on cost-effectiveness in health and medicine. JAMA 276:1172–1177, 1996 6. O’Sullivan JB, Mahan C: Criteria for the oral glucose tolerance test in pregnancy. Diabetes 13:278 –285, 1964 7. World Health Organization: Definition, Diagnosis, and Classification of Diabetes Mellitus and its Complications: Report of a WHO Consultation. Part 1: Diagnosis and Classification of Diabetes Mellitus. Geneva,
1483
Cost-effectiveness of GDM screening
World Health Org., 1999 (WHO/NCD/ NCS/99.2) 8. Poncet B, Touzet S, Rocher L, Berland M, Orgiazzi J, Colin C: Cost-effectiveness analysis of gestational diabetes mellitus screening in France. Eur J Obstet Gynecol Reprod Biol 103:122–129, 2002 9. Carpenter MW, Coustan D: Criteria for screening tests for gestational diabetes. Am J Obstet Gynecol 144:768 –773, 1982 10. Health Care Financing Administration OACT, Office of National Statistics, New York, DRI/McGraw-Hill HCC, 2003
1484
11. Maryland Health Care Commission Hospital Discharge Data, 1997–2001. Baltimore, MD, Department of Health and Mental Hygiene, 2003 12. Health Services Cost Review Commission Annual Cost Filings and Approved Rate Orders, 1997–2001. Baltimore, MD, Maryland Health Care Commission, 2003 13. Chung A, Macario A, El-Sayed Y, Riley E, Duncan G, Druzin M: Cost-effectiveness of a trial of labor after previous cesarean. Obstet Gynecol 97:932–941, 2000 14. Waitzman NJ, Romano P, Scheffler RM,
Harris JA: Economic costs of birth defects and cerebral palsy-1992. MMWR 44:694 – 699, 1995 15. Bureau of the Census: Summary of population and wage earnings: Maryland, Washington, DC [article online], 2003. Available from http://www.bls.gov. Accessed 4 October 2004 16. Langer O, Anyaegbunam A, Brustman L, Divon M: Management of women with one abnormal oral glucose tolerance test value reduces adverse outcome in pregnancy. Am J Obstet Gynecol 161:593–599, 1989
DIABETES CARE, VOLUME 28, NUMBER 6, JUNE 2005
