An important advance in prevention of perinatal human immunodeficiency virus (HIV) transmission was the finding by AIDS Clinical Trials Group protocol 076 that administration of zidovudine to women antepartum and intrapartum and to infants postpartum reduced the rate of perinatal transmission from 22.6% to 7.6%.1,2 Additional studies have been performed to determine whether shorter courses of zidovudine3,4 or if another antiretroviral therapy5 would be as effective or more effective in preventing perinatal transmission.
Prospective cohort studies have shown a decreased likelihood of perinatal HIV transmission with elective cesarean delivery6,7 and an additive protective effect with zidovudine therapy.8,9 A recent meta-analysis of observational data of 8500 mother–infant pairs showed that HIV-infected women receiving zidovudine who underwent elective cesarean delivery had a significant reduction in the rate of perinatal transmission from 7.3% to 2.0% compared with other delivery methods.10 These results were consistent with those of a recent European randomized clinical trial in which cesarean delivery in HIV-infected women receiving zidovudine was associated with an insignificant decrease in the risk of perinatal transmission from 4.3% to 0.8% compared with vaginal delivery.11 The ACOG reviewed this body of research and recommended that HIV-infected women, regardless of whether they are taking antiretroviral therapy, be offered elective cesarean delivery.12
The apparent benefits of elective cesarean delivery to reduce perinatal HIV transmission must be weighed against the risk of complications and increased cost of the procedure. Among women without HIV infection, those who undergo cesarean deliveries have increased maternal morbidity and mortality compared with those who have vaginal delivery.13,14 However, elective cesarean delivery is associated with a lower risk of maternal complications than nonelective cesarean delivery.15 Information on the relationship between method of delivery and subsequent maternal morbidity in HIV-infected women is limited but suggests that HIV-infected women who undergo cesarean delivery may be at increased risk for morbidity associated with surgical delivery compared with non–HIV-infected women.16,17
Researchers from the United Kingdom recently evaluated the cost-effectiveness of various strategies, including elective cesarean delivery, to reduce perinatal transmission of HIV.18 However, the analysis was based on different practice styles and costs than those in the United States and was done before publication of more exact estimates of perinatal transmission rates associated with elective cesarean delivery.
We used a decision-analysis model from the third-party payer perspective to evaluate the cost-effectiveness of two delivery strategies to reduce perinatal HIV transmission in the United States: 1) usual care, in which HIV-infected women receiving zidovudine undergo the method of delivery consistent with obstetric indications, and 2) the recommendation that HIV-infected women receiving zidovudine undergo elective cesarean delivery. We conducted threshold analyses to determine when the two strategies would be cost neutral.
Materials and Methods
We studied a hypothetical cohort of 7000 HIV-infected women in the United States. The size of the cohort was based on an estimate of the number of HIV-infected women giving birth each year in the United States.19 For this hypothetical cohort, we assumed that the HIV-infected women received zidovudine therapy, were in prenatal care by 36 weeks, did not have an obstetric indication or contraindication for delivery by cesarean delivery, and did not breastfeed.
The usual care strategy is shown in Figure 1. Human immunodeficiency virus–infected women receiving zidovudine therapy may undergo one of three different methods of delivery: vaginal delivery (instrumented or noninstrumented), nonelective cesarean delivery (in which labor has occurred or membranes have ruptured), and elective cesarean delivery (before onset of labor or ruptured membranes). Women may or may not experience complications as a result of the method of delivery. Women will give birth to either noninfected or HIV-infected infants.
The elective cesarean delivery strategy is shown in Figure 2. Human immunodeficiency virus–infected women receiving zidovudine therapy will be counseled about the risks and benefits of elective cesarean delivery. Of the women who accept the recommendation, most will undergo elective cesarean delivery at 38 weeks as planned. Some will enter labor or their membranes will rupture before the planned cesarean delivery; these women undergo either nonelective cesarean delivery or vaginal delivery. Women who do not accept the recommendation will undergo the method of delivery consistent with obstetric indications: vaginal delivery, nonelective cesarean delivery, or elective cesarean delivery. Regardless of the delivery method, women may or may not experience complications and give birth to either noninfected or HIV-infected infants.
We conducted the analysis from the third-party payer perspective. We programmed the model on a Microsoft Excel 4.0 spreadsheet (Microsoft Corp., Redmond, WA).
Probability estimates for the variables in the model were obtained from published reports and are shown in Table 1. The baseline probabilities were varied in sensitivity analyses according to the degree of uncertainty.
Results from a randomized clinical trial11 were used to generate probability estimates of method of delivery in the usual care and elective cesarean delivery strategies. The proportion of HIV-infected women who refused the elective cesarean delivery strategy was estimated from the number of women who refused randomization in a randomized clinical trial.20 Of the women who refused randomization in the clinical trial, more than half did so because they wanted a vaginal delivery.
The complication rate for vaginal deliveries was found to be 3.7% in a retrospective study that included more than 29,000 single vaginal vertex deliveries.21 The complication rates of 24.2% for nonelective cesarean delivery and 4.7% for elective cesarean were obtained from a prospective study of 1300 women who had a cesarean delivery.22 For baseline estimates, we increased the above complication rates three times in accordance with a study showing that HIV-infected women who underwent cesarean delivery had a threefold increase in complications compared with non–HIV-infected women.16
The perinatal transmission rate for HIV-infected women receiving zidovudine therapy ranged from 4.8% to 7.6%.2,23,24 However, these studies did not take into account method of delivery. We used a perinatal transmission rate of 4.3% for vaginal delivery and 0.8% for elective cesarean delivery; these rates were obtained from a randomized clinical trial.11 For the perinatal transmission rate associated with nonelective cesarean delivery, we used 4.3% as the baseline estimate.25 We obtained values for sensitivity analyses from various other studies.8–10
We summarized cost estimates for the variables in the model in Table 2. As appropriate, we used yearly inflation factors derived from the medical consumer price index to convert the published results to 1998 United States dollars and used a discount rate of 5% per year26 to change undiscounted published results to discounted United States dollars.
Baseline estimates for the cost of the delivery procedures with and without complications were obtained from the University Healthsystem Consortium Database.27 The database contains quarterly observed and expected mean costs of Health Care Financing Administration diagnosis-related groups for 93 university hospitals in the United States. Seventy-six university hospitals reported expected mean costs in the year 1998 for diagnosis-related group 370 (cesarean delivery with complications), 371 (cesarean delivery without complications), 372 (vaginal delivery with complications), and 373 (vaginal delivery without complications). Baseline estimates for the cost of the delivery procedures with and without complications were obtained by averaging the expected mean costs reported from the 76 university hospitals for each diagnosis-related group. The lowest and highest expected mean costs were used in sensitivity analyses.
Review of the literature revealed wide variation in estimates of the lifetime medical care cost of pediatric HIV infection. Using the methodology of Havens et al,28 we estimated the lifetime medical care cost to be $335,809. This value was comparable to $312,589 found in another study.29 Using the same yearly data for health care charges,30 various researchers have estimated the lifetime care cost to be between $86,13031,32 and $220,403.33 We chose $86,130 as our baseline estimate because it was the lowest that we could find in the literature and would probably bias the results against the elective cesarean delivery strategy.
We expressed outcomes of the decision analysis model as the absolute number of each delivery method (with and without complications) and the absolute number of noninfected and HIV-infected infants resulting from each strategy. The overall cost of each strategy included the cost of delivery method and the lifetime medical care cost saved from cases of perinatal HIV transmission prevented. The incremental cost of strategy was defined as the difference between the overall cost of strategy incurred in the elective cesarean delivery strategy and those incurred in the usual care strategy. We calculated incremental cost-effectiveness as the added cost of preventing a case of perinatal HIV transmission with the elective cesarean delivery strategy that was not prevented with the usual care strategy. By convention, a negative cost-effectiveness ratio indicates cost savings. We performed a series of univariate sensitivity analyses and one multivariate sensitivity analysis for variables about which uncertainty was greatest. When available, we took the lowest and highest values directly from the range of values reported in the literature or in the database.
Threshold analyses were performed to calculate the values of variables at which the two strategies were cost neutral.
Using baseline estimates, the two strategies were applied to a hypothetical cohort of 7000 HIV-infected pregnant women receiving zidovudine therapy (Table 3). In the usual care strategy, 4554 women had a vaginal delivery without complications, 569 had a vaginal delivery with complications, 1017 had a cesarean delivery without complications, and 860 had a cesarean delivery with complications. The cost associated with the usual care strategy was $21.4 million. In the usual care strategy, 271 infants were infected with HIV. The discounted lifetime cost of medical care for these infants totaled $23.3 million.
In the elective cesarean delivery strategy, 1456 women had a vaginal delivery without complications, 182 had a vaginal delivery with complications, 4352 had a cesarean delivery without complications, and 1011 had a cesarean delivery with complications. The cost associated with the elective cesarean delivery strategy was $28.3 million. In the elective cesarean delivery strategy, 129 infants were infected with HIV. The discounted lifetime cost of medical care for these infants totaled $11 million.
An incremental cost of $7.0 million resulted in 142 cases of perinatal HIV transmission prevented in 1 year. The incremental cost savings from those prevented cases of perinatal HIV transmission was $12.3 million. The incremental overall cost savings would be $5.3 million. The incremental cost-effectiveness ratio was $37,284 saved per case of perinatal transmission prevented.
A series of univariate sensitivity analyses was performed on selected variables (Table 4). The value for a variable was varied while all other variables were maintained at baseline estimates. Incremental cost-effectiveness ratios (added cost per case of perinatal transmission prevented by elective cesarean delivery strategy) were calculated; all were found to be cost saving. The range of cost savings per additional infection prevented varied from a low of $7800 to a high of $287,000.
Since costs for diagnosis-related groups 370 (cesarean delivery with complications) and 371 (cesarean delivery without complications) did not differentiate between elective and nonelective cesarean deliveries, we performed a multivariate sensitivity analysis in which the cost of nonelective cesarean delivery with and without complications was twice that of the baseline estimate. The result was $58,694 saved per additional infection prevented.
We performed threshold analyses on selected variables to determine when the elective cesarean delivery strategy would fail to be cost saving. Elective cesarean delivery was not cost saving when the perinatal HIV transmission rates associated with vaginal delivery, nonelective cesarean delivery, and elective cesarean delivery were all reduced by 43.3%; the cost of uncomplicated vaginal delivery was less than $556; the cost of uncomplicated cesarean delivery was greater than $5907; or the discounted lifetime cost of medical care for pediatric HIV infection was less than $48,846.
This analysis suggests that the elective cesarean delivery strategy in HIV-infected women receiving zidovudine therapy would both reduce the number of cases of perinatal HIV transmission and be cost saving. The outcomes were robust, and the elective cesarean delivery strategy was cost saving over a wide range of clinical and cost estimates. However, the elective cesarean delivery strategy was not cost saving when the baseline perinatal HIV transmission rates for vaginal delivery of 4.3%, nonelective cesarean delivery of 4.3%, and elective cesarean delivery of 0.8% were all reduced by at least 50%.
The baseline assumption that all HIV-infected women in the United States are receiving zidovudine therapy enabled us to use perinatal HIV transmission rates from peer-reviewed studies. A recent medical record review of HIV-infected women delivering at 66 sites that belonged to Pediatric AIDS Clinical Trials Group or were supported by Ryan White Title IV suggested that only 26% of HIV-infected women received either monotherapy or no therapy. The rest received combination antiretroviral therapy (Tuomala R, Shapiro D, Samelson R, Burchett G, Ciupak J, Mc-Namara H, et al. Antepartum antiretroviral therapy and viral load in 949 HIV-infected women in 1998–1999 [PACTG 367] [abstract]. Society for Maternal Fetal Medicine, Miami, FL, 2000). If the hypothetical cohort size in the model were changed to reflect this finding, the elective cesarean delivery strategy would still be cost saving per case of perinatal transmission prevented. However, only 37 cases of perinatal HIV transmission would be prevented, and the incremental overall cost savings would be reduced to $1.4 million.
According to the United States Public Health Service, combination antiretroviral therapy should be offered to HIV-infected pregnant women.34 The applicability of elective cesarean delivery for prevention of perinatal HIV transmission to HIV-infected women receiving combination antiretroviral therapy is unknown. Four groups of investigators reporting on single-site or non-randomized trials have found one case of perinatal transmission among 210 HIV-infected pregnant women receiving combination antiretroviral therapy35 (Beckerman KP, Benson M, Dahud S, Shannon M, Wara D. Control of maternal HIV-1 disease during pregnancy [abstract]. Twelfth World AIDS Conference, Geneva, Switzerland, 1998; Morris A, Zorilla C, Vajaranant M. A review of protease inhibitors use in 89 pregnancies [abstract]. Sixth Conference on Retroviruses and Opportunistic Infections, Chicago, IL, 1999; Stek A, Khoury M, Kramer F, Homans J, Rother C, Kovacs A. Maternal and infant outcomes with highly active antiretroviral therapy during pregnancy [abstract]. Sixth Conference on Retroviruses and Opportunistic Infections, Chicago, IL, 1999). Analysis of data from the medical record review at 66 sites showed that the perinatal transmission rate of 188 women on combination therapy including a protease inhibitor was 1.1% (confidence interval 0.1, 3.8) (Samelson R, Shapiro D, Tuomala R, Burchett S, Ciupak G, McNamara J, et al. HIV vertical transmission rates according to antiretroviral therapy and viral load during pregnancy in 526 mother-child pairs, 1998–99 [PACTG 367] [abstract]. Society for Maternal Fetal Medicine, Miami, FL, 2000). This value approaches the threshold limit found in our study, in which the elective cesarean delivery strategy is not cost saving when baseline rates of perinatal HIV transmission are all reduced by at least 50%.
Elective cesarean delivery may not be equally efficacious in prevention of perinatal transmission among all HIV-infected women receiving therapy. Two recent studies suggested the importance of viral load in perinatal HIV transmission. In a report by Mofenson et al,25 HIV-infected women receiving zidovudine who had undetectable viral loads (less than 500 copies of HIV RNA per mL) at delivery did not transmit HIV to their children, whereas women with detectable viral loads had a transmission rate of 6.1%. Garcia et al36 found that viral loads in HIV-infected pregnant women are directly related to transmission and that no cases of transmission occurred when viral loads were less than 1000 copies of HIV RNA per mL. None of the studies that suggested a benefit of elective cesarean delivery in preventing perinatal HIV transmission analyzed the effect of viral load on this benefit. Whether elective cesarean delivery provides additional benefit in HIV-infected women with viral loads below the limit of detection, who seem to already have very low risk for perinatal HIV transmission, remains to be seen. A recent survey of the Obstetrical Subcommittee of the Pediatric AIDS Clinical Trial Group, which includes 20 obstetricians practicing at largely tertiary centers, revealed that the majority of these physicians would discourage elective cesarean delivery for women with undetectable viral loads (Personal communication). Again, our analysis suggests that if a strategy can reduce the baseline perinatal transmission rates for vaginal delivery of 4.3%, nonelective cesarean delivery of 4.3%, and elective cesarean delivery of 0.8% by at least 50%, the elective cesarean delivery strategy will not be cost saving.
We did not represent in the model the risk of HIV transmission from HIV-infected women to health care providers during delivery. Prospective data suggest that the risk associated with a single percutaneous exposure is less than 0.5%.37 The risk is not greater during cesarean delivery than vaginal delivery.12 However, more health care personnel are present during a cesarean delivery than a vaginal delivery, and more people are therefore at risk for HIV transmission. This slight increase in risk may result in costs that would select against the elective cesarean delivery strategy. Appropriate precautions against skin-penetrating injuries should be taken,38 and postexposure prophylaxis should be offered when injuries occur.39
Elective cesarean delivery is one management strategy for prevention of perinatal HIV transmission. In properly selected cases, elective cesarean delivery is likely to be cost saving per case of perinatal transmission prevented. The cost savings per case of perinatal transmission prevented in the elective cesarean delivery strategy may decrease with successful efforts at increased prenatal identification and appropriate use of combination antiretroviral therapy for control of viral load.
1. Connor EM, Sperling RS, Gelber R, Kiselev P, Scott G, O'Sullivan MJ, et al. Reduction of maternal-infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. Pediatric AIDS Clinical Trials Group Protocol 076 Study Group. N Engl J Med 1994;331:1173–80.
2. Sperling RS, Shapiro DE, Coombs RW, Todd JA, Herman SA, McSherry GD, et al. Maternal viral load, zidovudine treatment, and the risk of transmission of human immunodeficiency virus type 1 from mother to infant. N Engl J Med 1996;335:1621–9.
3. Wiktor SZ, Ekpini E, Karon JM, Nkengasong J, Maurice C, Severin ST, et al. Short-course oral zidovudine for prevention of mother-to-child transmission of HIV-1 in Abidjan, Cote d'Ivoire: A randomised trial. Lancet 1999;353:781–5.
4. Shaffer N, Chuachoowong R, Mock PA, Bhadrakom C, Siriwasin W, Young NL, et al. Short-course zidovudine for perinatal HIV-1 transmission in Bangkok, Thailand: A randomised controlled trial. Bangkok Collaborative Perinatal HIV Transmission Study Group. Lancet 1999;353:773–80.
5. Guay LA, Musoke P, Fleming T, Bagenda D, Allen M, Nakabiito C, et al. Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: HIVNET 012 randomised trial. Lancet 1999;354:795–802.
6. Caesarean section and risk of vertical transmission of HIV-1 infection. The European Collaborative Study. Lancet 1994;343:1464–7.
7. Kind C. Mother-to-child transmission of human immunodeficiency virus type 1: Influence of parity and mode of delivery. Paediatric AIDS Group of Switzerland. Eur J Pediatr 1995;154:542–5.
8. Kind C, Rudin C, Siegrist CA, Wyler CA, Biedermann K, Lauper U, et al. Prevention of vertical HIV transmission: Additive protective effect of elective Cesarean section and zidovudine prophylaxis. Swiss Neonatal HIV Study Group. AIDS 1998;12:205–10.
9. Mandelbrot L, Le Chenadec J, Berrebi A, Bongain A, Benifla JL, Delfraissy JF, et al. Perinatal HIV-1 transmission: Interaction between zidovudine prophylaxis and mode of delivery in the French Perinatal Cohort. JAMA 1998;280:55–60.
10. Read JS. The mode of delivery and the risk of vertical transmission of human immunodeficiency virus type 1—a meta-analysis of 15 prospective cohort studies. The International Perinatal HIV Group. N Engl J Med 1999;340:977–87.
11. Elective caesarean-section versus vaginal delivery in prevention of vertical HIV-1 transmission: A randomised clinical trial. The European Mode of Delivery Collaboration. Lancet 1999;353:1035–9.
12. American College of Obstetricians and Gynecologists. Scheduled cesarean delivery and the prevention of vertical transmission of HIV infection. ACOG Committee Opinion no. 219. Washington, D.C.: The College, 1999.
13. Miller JM Jr. Maternal and neonatal morbidity and mortality in cesarean section. Obstet Gynecol Clin North Am 1988;15:629–38.
14. Petitti DB. Maternal mortality and morbidity in cesarean section. Clin Obstet Gynecol 1985;28:763–9.
15. van Ham MA, van Dongen PW, Mulder J. Maternal consequences of caesarean section. A retrospective study of intraoperative and postoperative maternal complications of caesarean section during a 10-year period. Eur J Obstet Gynecol Reprod Biol 1997;74:1–6.
16. Grubert TA, Reindell D, Kastner R, Lutz-Friedrich R, Belohradsky BH, Dathe O. Complications after caesarean section in HIV-1-infected women not taking antiretroviral treatment. Lancet 1999; 354:1612–3.
17. Semprini AE, Castagna C, Ravizza M, Fiore S, Savasi V, Muggiasca ML, et al. The incidence of complications after caesarean section in 156 HIV-positive women. AIDS 1995;9:913–7.
18. Ratcliffe J, Ades AE, Gibb D, Sculpher MJ, Briggs AH. Prevention of mother-to-child transmission of HIV-1 infection: Alternative strategies and their cost-effectiveness. AIDS 1998;12:1381–8.
19. Byers RH Jr, Caldwell MB, Davis S, Gwinn M, Lindegren ML. Projection of AIDS and HIV incidence among children born infected with HIV. Stat Med 1998;17:169–81.
20. Newell ML, Parazzini F, Mandelbrot L, Peckham C, Semprini A, Bazin B, et al. A randomised trial of mode of delivery in women infected with the human immunodeficiency virus. Br J Obstet Gynaecol 1998;105:281–5.
21. Nyirjesy I, Pierce WE. Perinatal mortality and maternal morbidity in spontaneous and forceps vaginal delivery. Am J Obstet Gynecol 1964;89:568–78.
22. Nielsen TF, Hokegard KH. Postoperative cesarean section morbidity: a prospective study. Am J Obstet Gynecol 1983;146:911–5.
23. Fiscus S, Adimora A, Schoenbach V, Lim W, McKinney R, Rupar D, et al. Perinatal HIV infection and the effect of zidovudine therapy on transmission in rural and urban counties. JAMA 1996;275:1483–8.
24. Stiehm ER, Lambert JS, Mofenson LM, Bethel J, Whitehouse J, Nugent R, et al. Efficacy of zidovudine and human immunodeficiency virus (HIV) hyperimmune immunoglobulin for reducing perinatal HIV transmission from HIV-infected women with advanced disease: Results of Pediatric AIDS Clinical Trials Group protocol 185. J Infect Dis 1999;179:567–75.
25. Mofenson LM, Lambert JS, Stiehm ER, Bethel J, Meyer WA 3d, Whitehouse J, et al. Risk factors for perinatal transmission of human immunodeficiency virus type 1 in women treated with zidovudine. Pediatric AIDS Clinical Trials Group Study 185 Team. N Engl J Med 1999;341:385–93.
26. Gold MR, Siegel JE, Russell LB, Weinstein MC. Cost-effectiveness in health and medicine. New York: Oxford University Press, 1996.
27. University HealthSystem Consortium. Cost summary by focus health care organization and comparison group for Diagnosis-Related Groups 370–373 for fiscal year 1998. Available on the Internet at www.uhc.edu
. Accessed 1999.
28. Havens PL, Cuene BE, Holtgrave DR. Lifetime cost of care for children with human immunodeficiency virus infection. Pediatr Infect Dis J 1997;16:607–10.
29. Postma MJ, Beck EJ, Mandalia S, Sherr L, Walters MD, Houweling H, et al. Universal HIV screening of pregnant women in England: Cost effectiveness analysis. BMJ 1999;318:1656–60.
30. Hsia DC, Fleishman JA, East JA, Hellinger FJ. Pediatric human immunodeficiency virus infection. Recent evidence on the utilization and costs of health services. Arch Pediatr Adolesc Med 1995;149:489–96.
31. Mauskopf JA, Paul JE, Wichman DS, White AD, Tilson HH. Economic impact of treatment of HIV-positive pregnant women and their newborns with zidovudine. JAMA 1996;276:132–8.
32. Stringer JSA, Rouse DJ. Rapid testing and zidovudine treatment to prevent vertical transmission of human immunodeficiency viurs in unregistered parturients: A cost-effectiveness analysis. Obstet Gynecol 1999;94:34–40.
33. Gorsky RD, Farnham PG, Straus WL, Caldwell B, Holtgrave DR, Simonds RJ, et al. Preventing perinatal transmission of HIV—costs and effectiveness of a recommended intervention. Public Health Rep 1996;111:335–41.
34. Public Health Service Task Force recommendations for the use of antiretroviral drugs in pregnant women infected with HIV-1 for maternal health and for reducing perinatal HIV-1 transmission in the United States. MMWR Morb Mortal Wkly Rep 1998;47:1–30.
35. McGowan JP, Crane M, Wiznia AA, Blum S. Combination antiretroviral therapy in human immunodeficiency virus-infected pregnant women. Obstet Gynecol 1999;94:641–6.
36. Garcia PM, Kalish LA, Pitt J, Minkoff H, Quinn TC, Burchett SK, et al. Maternal levels of plasma human immunodeficiency virus type 1 RNA and the risk of perinatal transmission. Women and Infants Transmission Study. N Engl J Med 1999;341:394–402.
37. Tokars JI, Marcus R, Culver DH, Schable CA, McKibben PS, Bandea CI, et al. Surveillance of HIV infection and zidovudine use among health care workers after occupational exposure to HIV-infected blood. The CDC Cooperative Needlestick Surveillance Group. Ann Intern Med 1993;118:913–9.
38. Update: Universal precautions for prevention of transmission of HIV, hepatitis B virus, and other bloodborne pathogens in health-care settings. MMWR Morb Mortal Wkly Rep 1988;37.
39. Public service guidelines for the management of health care workers exposed to HIV and recommendations for post-exposure prophylaxis. MMWR Morb Mortal Wkly Rep 1998;47:1–33.