Odibo, Anthony O. MD, FRCOG1,3; Stamilio, David M. MD, MSCE2; Macones, George A. MD, MSCE1; Polsky, Daniel PhD2,3
Preterm births affect 12.5% of pregnancies in the United States, contributing to more than 75% of perinatal mortality and morbidity.1 Despite many years of research and significant financial investments in efforts to prevent preterm delivery, no effective therapy has been reported, until recently. With more than 4 million deliveries occurring every year in the United States, the financial implications of preterm birth cannot be overemphasized.
Reports of studies from the 1970s and 1980s showed conflicting findings on the efficacy of using progestational agents to prevent preterm births.2 Two recent randomized-controlled trials have rekindled interest in the use of progestational agents for preventing preterm birth.3,4 Meis et al3 reported a significant reduction in the rate of recurrent preterm birth in a high-risk group of women randomly assigned to receiving 17α-hydroxyprogesterone caproate compared with the placebo group. Similarly, a smaller trial using progesterone vaginal suppository reported a significant reduction of recurrence of preterm delivery in the treatment group compared with placebo.4
In response to these recent reports, the American College of Obstetricians and Gynecology issued a Committee Opinion in support of the use of 17α-hydroxyprogesterone caproate.5 After the release of the committee opinion statement, many obstetricians now prescribe 17α-hydroxyprogesterone caproate. As with any new technology, some of the scenarios in which 17α-hydroxyprogesterone caproate is prescribed are outside the context of the two recent trials mentioned above. It also still remains to be shown that this therapy is cost-effective compared with a policy of “doing nothing.”
For several reasons, cost considerations are important before a universal policy of 17α-hydroxyprogesterone caproate for all women considered at risk for preterm delivery could be adopted. First, although preterm birth contributes significantly to the cost of obstetric and neonatal care, it is still uncertain to what extent 17α-hydroxyprogesterone caproate will affect the rate of preterm birth and the associated cost. Currently, 17α-hydroxyprogesterone caproate is produced by only few pharmaceutical companies, creating the potential for a monopolistic control of price and its long-term consequences. Finally, third-party reimbursement is still not universal, raising potential ethical issues of equity.
The aim of this study is to evaluate whether the use of 17α-hydroxyprogesterone caproate for the prevention of recurrent preterm delivery is cost-effective. We are also interested in determining the subgroups of women with prior preterm births in which 17α-hydroxyprogesterone caproate is cost-effective. This information could be useful in developing policies regarding the use of 17α-hydroxyprogesterone caproate in pregnancy.
MATERIALS AND METHODS
Decision-analysis models evaluating four distinct subgroups considered for the use of 17α-hydroxyprogesterone caproate for the prevention of preterm births were created. The subgroups evaluated were 1) history of early preterm birth (prior preterm delivery less than 32 weeks); 2) history of preterm birth 32–37 weeks (prior preterm delivery 32–37 weeks) and 3) prior term delivery (prior term delivery) and 4) no prior delivery (no prior delivery). For each of these subgroups, a strategy of 17α-hydroxyprogesterone caproate was compared with a “do-nothing” strategy (No 17α-hydroxyprogesterone caproate). Figure 1 shows the abbreviated model for the use of 17α-hydroxyprogesterone caproate in women with prior delivery less than 32 weeks. The full models can be obtained from the authors if desired. The analysis is from a health institution perspective.
The primary measures of effectiveness for each strategy were 1) the number of preterm deliveries less than 32 weeks prevented and 2) the number of preterm delivery less than 37 weeks prevented and the quality-adjusted life-years gained from preventing preterm delivery. Secondary measures included the number of neonatal deaths and neonatal morbidity secondary to preterm birth prevented. The baseline efficacy of 17α-hydroxyprogesterone caproate and probabilities for each strategy and for the outcomes were derived from a systematic review of the English literature, supplemented by bibliographic review. The following search terms were used: “preterm birth” “preterm delivery” “preterm prevention” “17α-hydroxyprogesterone” and combinations thereof and the databases queried include MEDLINE, PubMed, and Cochrane database. The probabilities included in the decision tree are shown in Table 1.
We assumed that 2 million pregnant women with singletons per year will present for prenatal care within the first 4 months of gestation (by 16 weeks), with 8.7% of these being multiparous with a prior preterm birth (less than 37 weeks) and 1.3% with a past history of early preterm delivery (less than 32 weeks).13 The proportion of primigravid and women with prior term delivery were derived using estimates from the most recent National Vital Statistics.1 In addition, 75% of recurrent preterm births were assumed to be spontaneous. The natural history for women with a prior preterm delivery and their performance in subsequent pregnancies were derived from the results of the Maternal Fetal Medicine Units network study13 and from the report by Petrini et al14 using data from the States of Missouri and New Jersey. For deliveries occurring less than 32 weeks, probability of neonatal outcomes for a median gestational age of 28 weeks were used, and for deliveries less than 37 weeks, the median gestational age of 34 weeks was used for probabilities.
The probability of neonatal death and morbidity were obtained from a study of the use of antenatal steroids to prevent respiratory distress syndrome and related morbidity and mortality in premature infants.12 Because respiratory distress syndrome is one of the most common neonatal complications from preterm birth, we used the probability for respiratory distress syndrome as a surrogate for neonatal morbidity. Neonatal deaths were assumed to occur within the first week of delivery. For deliveries less than 32 weeks an average cost for a 28-day stay in the neonatal intensive care unit (NICU) was used, and for deliveries after 32 weeks an average of 2 weeks NICU stay was used.
The projected numbers of preterm delivery prevented by 17α-hydroxyprogesterone caproate were derived by using the odds ratios and relative risks reported in the literature for the efficacy of 17α-hydroxyprogesterone caproate and applying these risk reduction estimates to the expected prevalence of preterm delivery for each strategy.3,6–11,13 There has been no significant complication from 17α-hydroxyprogesterone caproate injections (with the exception of pain) reported in the literature. Therefore our analysis does not include any complications from 17α-hydroxyprogesterone caproate administration.
The costs included in the decision tree were derived from published and unpublished sources as listed in Table 2. Only direct medical costs were included. All costs were adjusted for 2005 U.S. dollars.14 For the cost of 17α-hydroxyprogesterone caproate in women delivering before 32 weeks, we used an average cost for a 10-week course, and an average cost for a 14-week course was used for deliveries between 32 and 37 weeks. For term deliveries an average for a 20-week course was used. These costs were varied widely (see Table 2) in our sensitivity analysis. We included the nursing charge for only one visit as is the current practice in our unit. A nurse visits the patient to teach and observe them administer the 17α-hydroxyprogesterone caproate. Following this visit, the patients self-administer the medication. For hospital admissions secondary to preterm labor we used an average between hospital charges for an average length of stay of 3 days and values from Nicholson et al15 based on Maryland State discharge data, including costs for a 48-hour course of magnesium sulfate and dexamethasone. These charges were multiplied by a cost-charge ratio of 0.6 as an approximation to third party reimbursements.18 The cost for NICU stay was prorated according to the expected average duration of admission. All costs were discounted at a rate of 3%.
The utilities for neonatal morbidity were derived from the study by Teng and Wallace.17 For neonatal death we used a utility of 0 and for a term or normal (neonate with no complication from preterm delivery) infant a utility score of 1 was used. Deliveries occurring less than 32 weeks were assumed to incur severe morbidities with a utility score of 0.61 and those between 32 weeks and 37 weeks mild or moderate morbidity with a utility score of 0.82 (Table 2).17 These were converted to quality-adjusted life years by using average life expectancy values for the United States, discounted at a 3% rate.
In the baseline analysis, pathway probabilities were used to calculate the costs, the number of preterm deliveries less than 32 weeks and less than 37 weeks prevented, and the number of neonatal deaths and morbidity prevented by each strategy, accounting for different gestational ages (less than 32 weeks and less than 37 weeks) of delivering.
A cost-utility analysis was also performed and reported as cost-effectiveness ratios and incremental cost-effectiveness ratios. The incremental cost-effectiveness ratios is the additional cost of moving from a policy of no treatment to using 17α-hydroxyprogesterone caproate.
Due to the considerable uncertainty regarding several of our baseline assumptions, we performed one-way and multi-way sensitivity analyses (effect of varying the baseline estimates on the results) commensurate with the degree of uncertainty for each point estimate. The ranges of values used for the sensitivity analyses are shown in Tables 1 and 2. All computations were performed using a commercially available decision-analysis software package (TreeAge Pro 2005, Tree Age Inc, Williamstown, MA) and STATA 8.0 (StataCorp LP, College Station, TX).
Under the baseline assumptions, the model favors the use of 17α-hydroxyprogesterone caproate in women with a history of preterm delivery less than 32 weeks and between 32 weeks and 37 weeks as the most cost-effective compared with the other strategies. The costs per quality-adjusted life-year gained from each strategy are shown in Table 3. The use of 17α-hydroxyprogesterone caproate was associated with cost savings in women with prior preterm delivery. The cost saved per quality-adjusted life-year gained secondary to using 17α-hydroxyprogesterone caproate were $3,090 and $2,963 in women with prior preterm delivery less than 32 weeks and between 32 weeks and 37 weeks, respectively. The use of 17α-hydroxyprogesterone caproate in women with prior term, or no previous delivery was also found to be cost-effective, but not cost-saving, with cost per quality-adjusted life-year gained of $1,730 and $2,353, respectively.
The cost per additional preterm delivery less than 37 weeks averted using 17α-hydroxyprogesterone caproate were $35,319 and $36,093 for the strategies administering the medication to women with prior deliveries less than 32 weeks and between 32 weeks and 37 weeks, respectively, compared with $130,649 and $129,991 for women with no prior delivery or prior term deliveries, respectively. When considering preterm delivery less than 32 weeks averted, the cost-effectiveness ratio ($95,402) was more favorable in the strategy using 17α-hydroxyprogesterone caproate in women with prior delivery less than 32 weeks, compared with the other strategies.
The costs per neonatal outcomes averted are shown in Table 4. The cost per neonatal death avoided was cheapest with the use of 17α-hydroxyprogesterone caproate for prior preterm delivery less than 32 weeks, whereas the cost per neonatal morbidity avoided favors using 17α-hydroxyprogesterone caproate for women with prior deliveries between 32 weeks and 37 weeks.
The following measures were subjected to one-way and two-way sensitivity analyses using the range of values in tables 1 and 2: the probabilities for preterm delivery given prior delivery at less than 32 weeks, 32–37 weeks, term delivery, and no prior delivery; the effectiveness of 17α-hydroxyprogesterone caproate in preventing preterm delivery; the probability of neonatal death and morbidity given the gestational age at delivery; the cost of 17α-hydroxyprogesterone caproate; cost of hospitalization for preterm labor and NICU admissions; the utilities in the models and the discount rates. The model was robust to almost all parameters evaluated using both one-way and two-way sensitivity analyses. The analysis was universally sensitive to the efficacy of 17α-hydroxyprogesterone caproate in preventing preterm delivery less than 32 weeks in all models evaluated. When the probability of preterm delivery less than 32 weeks was greater than 31.5% in women using 17α-hydroxyprogesterone caproate, the “No 17α-hydroxyprogesterone caproate ” strategy became more cost-effective (Fig. 2).
The analysis shows the use of 17α-hydroxyprogesterone caproate to be most cost-effective for the prevention of preterm delivery in women with prior delivery at less than 32 weeks or between 32 weeks and 37 weeks using the commonly accepted threshold of $50,000 dollars per quality-adjusted life-year gained. These strategies were actually associated with cost-savings. The use of 17α-hydroxyprogesterone caproate in women with previous term or no prior delivery was also found to be cost-effective using the same threshold of $50,000 dollars per quality-adjusted life-year gained. The latter should, however, be interpreted with caution, because the model assumptions for these two strategies are yet to be confirmed clinically. The findings are not shocking given the cheap cost of 17α-hydroxyprogesterone caproate, relative to the cost and consequences of preterm delivery. The results of the cost per number of preterm delivery less than 32 weeks or 37 weeks or the number of neonatal complications prevented, are difficult to interpret because there is no a priori established willingness-to-pay threshold for the prevention of preterm delivery or the associated neonatal complications. Depending on how much the society or health care payers are willing to pay to prevent one preterm birth (with its associated complications), the use of 17α-hydroxyprogesterone caproate in women with prior delivery less than 32 weeks and between 32 weeks and 37 weeks could be considered cost-effective for all outcomes evaluated. Considering the significant contribution of preterm birth to neonatal mortality and morbidity, it is expected that the willingness-to-pay thresholds for these outcomes should be quite high.
This analysis was sensitive to the probability of delivering less than 32 weeks while on 17α-hydroxyprogesterone caproate. The critical threshold was 31.5%, which is close to our baseline probability of preterm delivery less than 32 weeks (30%). The probability of 31.5% can be reached only if 17α-hydroxyprogesterone caproate is not effective in preventing preterm delivery in any of the subgroups evaluated.
When compared with other preventive programs that have been adopted, such as cholesterol reduction using cholestyramine resins ($132,000 per year of life saved)19 or colorectal cancer screening in elderly ($45,000 per year of life saved),20 the use of 17α-hydroxyprogesterone caproate in women with prior preterm delivery (both less than 32 weeks and between 32 and 37 weeks) would seem not only to be cost-effective but also to be cost-saving for the outcomes of preventing preterm birth less than 32 weeks and less than 37 weeks.
Although the current indication for 17α-hydroxyprogesterone caproate is in women with prior preterm delivery, we sought to determine whether the medication may be cost-effective in other patient populations, such as those with no prior delivery or with previous term delivery. Our findings suggest that the use of 17α-hydroxyprogesterone caproate in these subgroups is cost-effective using the traditional threshold for cost-effectiveness per quality-adjusted life-year and within our baseline assumptions.
Our analysis is the first to consider the economic implications of the use of 17α-hydroxyprogesterone caproate. A previous review of the effect of the use of 17α-hydroxyprogesterone caproate on the rate of preterm delivery suggested only a limited effect given the small group of at-risk women who are candidates for the therapy.14 The latter report however, did not consider the cost implications of the use of 17α-hydroxyprogesterone caproate. Given the inexpensive and safe nature of 17α-hydroxyprogesterone caproate, further studies evaluating new subgroups of women who could benefit from 17α-hydroxyprogesterone caproate are needed. Although some secondary analyses using the original data by Maternal Fetal Medicine Units network 17α-hydroxyprogesterone caproate study have been published with the aim of identifying potential candidates for 17α-hydroxyprogesterone caproate,21,22 more studies using different patient populations will be useful. Other studies evaluating the economic burden of preterm births have been limited to costs of hospitalization in the prenatal and neonatal period.16,23–25 For example, Petrou23 evaluated the economic consequences of preterm birth in the first year of life. The analysis was limited to the cost of hospitalization and inpatient services. No considerations of quality-adjusted life-years were included. Other authors have focused on the economic consequences of very low birth weight infants, a complication of preterm birth.24,25
This analysis has several limitations. We considered direct medical costs only and did not include indirect costs such as costs from work time loss or loss of future earnings. This is due to lack of reliable data on such costs. The probabilities were derived from limited sources due to few studies evaluating the consequences of preterm births considered in this analysis. However, despite wide variations of these probabilities in our sensitivity analyses, the models remained robust. The utilities used were derived from quality-adjusted life-year weights from evaluation of very low birth-weight neonates.17 Although these quality-adjusted life-year weights may be close approximations to those from preterm births, future studies are needed to confirm this assumption. Such studies should address the utility of outcomes from preterm births both from the parental perspective as well as the perspective of providers.
In conclusion, despite the limitations of this analysis, the use of 17α-hydroxyprogesterone caproate was associated with cost-savings with respect to prevention of preterm birth less than 32 weeks and 32–37 weeks and may be cost-effective in the other subgroups. Determinations of the cost-effectiveness of 17α-hydroxyprogesterone caproate for the other outcomes evaluated (cost per preterm delivery prevented and neonatal complications avoided) will depend on what society is willing to pay to avoid these complications.
1. Hamilton BE, Martin JA, Ventura SJ, Sutton PD, Menacker F. Births: preliminary data for 2004. Natl Vital Stat Rep 2005;54:1–17.
2. Keirse MJ. Progestogen administration in pregnancy may prevent preterm delivery. Br J Obstet Gynaecol 1990;97:149–54.
3. Meis PJ, Klebanoff M, Thom E, Dombrowski MP, Sibai B, Moawad AH, et al. Prevention of recurrent preterm delivery by 17α-hydroxyprogesterone caproate. N Engl J Med 2003;348:2379–85.
4. da Fonseca EB, Bittar RE, Carvalho MH, Zugaib M. Prophylactic administration of progesterone by vaginal suppository to reduce the incidence of spontaneous preterm birth in women at increased risk: a randomized placebo-controlled double-blind study. Am J Obstet Gynecol 2003;188:419–24.
5. American College of Obstetricians and Gynecologists. ACOG Committee Opinion. Use of progesterone to reduce preterm birth. Obstet Gynecol 2003;102:1115–6.
6. Hauth JC, Gilstrap LC, Brekken AL, Hauth JM. The effect of 17 [alpha]-hydroxyprogesterone caproate on pregnancy outcome in an active-duty military population. Am J Obstet Gynecol 1983;146:187–90.
7. Johnson JW, Austin KL, Jones GS, Davis G, King TM. Efficacy of 17 [alpha]-hydroxyprogesterone caproate in the prevention of premature labor. N Engl J Med 1975;293:675–80.
8. Yemini M, Borenstein R, Dreazen E, Apelman Z, Mogilner BM, Kessler I, et al. Prevention of premature labor by 17 [alpha]-hydroxyprogesterone caproate. Am J Obstet Gynecol 1985;151:574–7.
9. Hartikainen-Sorri AL, Kauppila A, Tuimala R. Inefficacy of 17 alpha-hydroxyprogesterone caproate in the prevention of prematurity in twin pregnancy. Obstet Gynecol 1980;56:692–5.
10. LeVine L. Habitual abortion: a controlled clinical study of progestational therapy. West J Surg Obstet Gynecol 1964;72:30–6.
11. Sanchez-Ramos L, Kaunitz AM, Delke I. Progestational agents to prevent preterm birth: a meta-analysis of randomized controlled trials. Obstet Gynecol 2005;105:273–9.
12. Simpson KN, Lynch SR. Cost savings from the use of antenatal steroids to prevent respiratory distress syndrome and related conditions in premature infants. Am J Obstet Gynecol 1995;173:316–21.
13. Mercer BM, Goldenberg RL, Moawad AH, Meis PJ, Iams JD, Das AF, et al. The preterm prediction study: effect of gestational age and cause of preterm birth on subsequent obstetric outcome. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol 1999;181:1216–21.
14. Petrini JR, Callaghan WM, Klebanoff M, Green NS, Lackritz EM, Howse JL, et al. Estimated effect of 17 alpha-hydroxyprogesterone caproate on preterm birth in the United States. Obstet Gynecol 2005;105:267–72.
15. Cost analysis. In: Drummond MF, Sculpher MJ, Torrance GW, O’Brien, Stoddart GL, editors. Methods for the economic evaluation of health care programmes. New York (NY): Oxford University Press; 2005. p. 77.
16. Nicholson WK, Frick KD, Powe NR. Economic burden of hospitalizations for preterm labor in the United States. Obstet Gynecol 2000;96:95–101.
17. Teng TO, Wallace A. One thousand health-related quality-of-life outcomes. Med Care 2000;38:583–637.
18. Medicare. US Department of Health and Human Services, National Limitations—Clinical Lab, Washington, DC (1995).
19. Kinosian BP, Eisenberg JM. Cutting into cholesterol: cost-effective alternatives for treating hypercholesterolemia. JAMA 1988;259:2249–54.
20. Wagner JL, Herdman RC, Wadhwa S. Cost-effectiveness of colorectal cancer screening in the elderly. Ann Intern Med 1991;115:807–17.
21. Spong CY, Meis PJ, Thom EA, Sibai B, Dombrowski MP, Moawad AH, et al. Progesterone for prevention of recurrent preterm birth: impact of gestational age at previous delivery. Am J Obstet Gynecol 2005;193:1127–31.
22. Meis PJ, Dombrowski MP, Leindecker S, Moawad AH, Northern A, Iams JD, et al. Does progesterone treatment influence risk factors for recurrent preterm delivery? Obstet Gynecol 2005;106:557–61.
23. Petrou S. The economic consequences of preterm birth during the first 10 years of life. BJOG 2005;112:10–5.
24. McCormick MC, Bernbaum JC, Eisenberg JM, Kustra SL, Finnegan E. Costs incurred by parents of very low birth weight infants after the initial neonatal hospitalization. Pediatrics 1991;88:533–41.
25. Rogowski J. Cost-effectiveness of care for very low birth weight infants. Pediatrics 1998;102:35–43.