OBJECTIVE: To evaluate the cost-effectiveness of thromboprophylaxis at cesarean delivery with intermittent pneumatic compression.
METHODS: A decision tree model using Markov analysis was developed to compare two approaches to perioperative care at the time of cesarean delivery: 1) no use of perioperative thromboprophylaxis and 2) the use of intermittent pneumatic compression for thromboprophylaxis at the time of cesarean delivery. Postcesarean deep venous thrombosis was estimated to occur in 0.7% of patients (75% of whom were asymptomatic), and result in a 9% chance of postthrombotic syndrome. Mechanical prophylaxis was assumed to decrease the risk of deep venous thrombosis by 70% and to cost $120. Probability of morbidity and mortality of venous thromboembolism as well as anticoagulation and the costs and utilities for different health state were derived from published studies. Sensitivity analysis was performed over a wide range of variable estimates.
RESULTS: Using the assumptions in our base case, routine thromboprophylaxis for cesarean delivery cost $39,545 per quality-adjusted life year. One-way sensitivity analysis revealed that as long as the incidence of postcesarean deep venous thrombosis was at least 0.68%, intermittent pneumatic compression reduced the incidence of deep venous thrombosis by at least 50%, or the cost of intermittent pneumatic compression was less than $180, the cost-effectiveness of mechanical prophylaxis did not exceed $50,000 per quality-adjusted life year.
CONCLUSION: Mechanical thromboprophylaxis is estimated to be a cost-effective strategy under a wide range of circumstances.
LEVEL OF EVIDENCE: III
Under certain conditions, intermittent pneumatic compression is estimated to be a cost-effective intervention for the prevention of venous thromboembolism in women undergoing cesarean delivery.
From the 1Department of Obstetrics and Gynecology, Section of Maternal–Fetal Medicine, Northwestern University Medical School, Chicago, Illinois.
Presented at the Twenty-Fifth Annual Meeting of the Society for Maternal–Fetal Medicine, Reno, Nevada, February 7–12, 2005.
Corresponding author: William A. Grobman, 333 East Superior Street, Suite 410, Chicago, IL 60611; e-mail: email@example.com.
All aspects of Virchow’s triad of hypercoagulability, venous stasis, and endothelial damage are enhanced during pregnancy. Correspondingly, the risk of venous thromboembolism is six-fold higher in pregnant women than in age-matched nonpregnant women.1,2 Depending on the diagnostic criteria used to establish the diagnosis, venous thromboembolism in pregnancy is reported to complicate 0.5 to 3 of every 1,000 deliveries. A cesarean delivery further increases the risk for stasis and endothelial damage and is thought to increase the risk for venous thromboembolism 3- to 10-fold over women delivering vaginally.2–4 When a venous thromboembolism does occur, both the short-term and long-term consequences can be substantial due to the increased risk of fatal pulmonary embolism, recurrent venous thromboembolism, and postthrombotic syndrome.5–7
Given the health burden associated with venous thromboembolism, women at particularly high risk for this complication, such as those who have had a prior venous thromboembolism and have a thrombophilia, are recommended to receive anticoagulation during pregnancy. Yet, there remains no consensus with regard to which women who are not anticoagulated during pregnancy should receive perioperative venous thromboembolism prevention at the time of cesarean. Expert panels in the United Kingdom have recently advocated universal thromboprophylaxis with heparin or low-molecular-weight heparin for all patients undergoing cesarean delivery after observing a lack of compliance with the risk-based guidelines that they had promoted.4 Due to concerns about cost and the risk associated with pharmacologic prophylaxis, one group in North America has been reluctant to recommend universal prophylaxis and instead advocate prophylaxis only for at-risk groups.8
Several prospective randomized trials have shown that intermittent pneumatic compression has comparable efficacy to pharmacologic prophylaxis in preventing perioperative deep venous thrombosis (DVT) in patients undergoing general surgery and gynecologic surgery without the risk of major and minor bleeding that would be associated with unfractionated heparin or low-molecular-weight heparins.9–12 Moreover, a recent analysis of risks and benefits of methods of perioperative prophylaxis for cesarean delivery has demonstrated that prophylaxis with intermittent pneumatic compression is preferable to prophylaxis with pharmacologic treatments.13 Yet, this analysis did not evaluate whether, and under what circumstances, intermittent pneumatic compression would be cost-effective. In the present study, we have created a model to examine whether thromboprophylaxis at cesarean delivery with intermittent pneumatic compression is a cost-effective strategy.
MATERIALS AND METHODS
A decision tree model using Markov analysis was developed to compare two approaches to perioperative care at the time of cesarean delivery: 1) no use of perioperative venous thromboembolism prophylaxis and 2) the use of intermittent pneumatic compression for venous thromboembolism prophylaxis at the time of cesarean delivery.14 This approach uses cost-effectiveness analysis, in which both the costs and health outcomes of the different strategies are incorporated into the outcome measure. The Markov methodology was used to allow the model the capacity to follow women as they transition through the different health states considered in the model. The women in this model are those who have not been anticoagulated during their pregnancy. The analytic decision model was created using Data Pro 4.0 (TreeAge Software, Inc., Williamstown, MA).
In the baseline model, women enter the Markov cycle at 30 years of age. One possibility is that they undergo a cesarean without an associated lower extremity DVT, in which case they continued to be in good health with a chance of dying only due to their age-specific background risk. The age-specific background risk was derived from a life table in which the overall life expectancy is 76 years. Alternatively, a woman could have a delivery complicated by either an asymptomatic or symptomatic DVT. An asymptomatic DVT, given that it is not discovered, correspondingly remains untreated unless it manifests further pathology such as a pulmonary embolism, at which time it would be treated with short-term heparin-based therapy and 6-month warfarin treatment. Conversely, a symptomatic DVT is treated with short-term heparin-based therapy and 3-month warfarin treatment. A woman who develops a DVT has a risk of developing other pathology, and these risks are incorporated into the Markov analysis. Specifically, a woman with a DVT can develop a pulmonary embolism, an event that has a corresponding risk of death. Because the magnitude of both these risks is dependent upon whether a DVT has been treated, the difference in this magnitude is incorporated into the analysis. Even if a woman does not acutely develop a pulmonary embolism, longer-term complications of a thrombotic event include an increased risk for postthrombotic syndrome, a potentially debilitating condition that can continue to affect a woman throughout her life. Also, while on anticoagulation, women can experience minor bleeding complications as well as major bleeding complications (such as a cerebrovascular accident). Once anticoagulation regimen is complete, there is a risk of recurrent DVT. If a recurrence occurs, a woman must again hazard the types of DVT-related complications that have already been noted, although the chance of postthrombotic syndrome is greater than after a single DVT occurrence. The estimates for the probabilities of all events were derived from the published literature and are presented in Table 1.
The two strategies differ only in that women receiving the prophylaxis strategy receive intermittent pneumatic compression during and after their cesarean delivery until they are ambulating. This intermittent compression decreases the risk of a DVT by a magnitude that is predicated on observed efficacy in the literature.
All health states in our model were assigned a utility as a measure of effectiveness. All utilities are assigned a value from 0 to1, with 0 defined as no quality of life (death) and 1 defined as full quality of life (perfect health). During the time that a woman remains on warfarin, her quality of life is reduced due to the more frequent physician visits and testing that she must undergo, as well as the restrictions on activity to which she should adhere. Additional decrements in life quality were not accorded to minor and transient bleeding complications (eg, nosebleed), although major bleeding complications generated a further reduced quality of life. Lastly, postthrombotic syndrome leads to a decrement in quality of life. These utility values were derived from the published literature and are presented in Table 1. These utilities are summated throughout a woman’s remaining life to yield the number of quality-adjusted life years that she will have. The use of quality-adjusted life years allows both the length of a woman’s survival in a given health state as well as the quality of life during that survival to be incorporated into the model.
All costs in the model are direct, derived from the published literature, and presented in Table 2 in 2004 United States dollars. Costs in the literature before 2004 were adjusted to 2004 dollars using the medical care component of the Consumer Price Index. All costs are direct and take into account costs of all medical consequences of a given health event. For example, the costs of a DVT include those from pharmacologic therapy, inpatient stays, physician visits, and laboratory monitoring. This analysis was from the perspective of the medical care system.
In this model, both costs and utilities are discounted at a 3% annual rate. One-way sensitivity analyses were performed on all model variables and two-way sensitivity analyses were performed on variables of interest. The most stringent criterion used to define a cost-effective strategy was a cost-effectiveness ratio of less than $50,000 per quality-adjusted life year. However, given the controversy regarding the most appropriate ratio to use as a cost-effective threshold, cost-effective thresholds of up to $100,000 per quality-adjusted life year were also explored in the sensitivity analysis.29
The results for the base-case model are presented in Table 3. Our model predicts that prophylaxis is more costly but also more effective than no prophylaxis. Specifically, for any given woman, prophylaxis costs $104 more but gains .00263 quality-adjusted life years. Thus, given the baseline estimates, the incremental cost-effectiveness (the difference between the costs of the two strategies divided by the difference between the effectiveness of the two strategies) of routine intermittent pneumatic compression during cesarean is $39,545 per quality-adjusted life year.
We performed a one-way sensitivity analysis to evaluate the effect of changing the probability and cost variables on the cost-effectiveness ratio. All probability, utility, and cost variables were changed from the lowest to highest estimate. Altering the variables had little effect upon the result, with three exceptions. Specifically, the cost of intermittent pneumatic compression, the risk reduction with intermittent pneumatic compression, and the probability of having a DVT after cesarean were the three variables that, when changed, altered the cost-effectiveness ratio substantially. The change in cost-effectiveness ratio for these three variables as each is changed across its entire range is presented in Figure 1. As long as the incidence of postcesarean venous thromboembolism (both symptomatic and asymptomatic) was at least 0.68%, intermittent pneumatic compression reduced the chance of venous thromboembolism by at least 50%, or the cost of intermittent pneumatic compression was less than $180, the cost-effectiveness of mechanical prophylaxis did not exceed $50,000 per quality-adjusted life year.
Given the dramatic changes caused in the outcomes due to changes in the probability of having a DVT as well as the risk reduction due to intermittent pneumatic compression, we further performed a two-way sensitivity analysis on these variables. The results of this analysis are presented in Figure 2. In this figure, each line represents a different cost-effective threshold for combinations of the two variables of interest. As can be seen, even a relatively small risk reduction (25%) due to intermittent pneumatic compression prophylaxis yields a cost-effectiveness ratio of between $50,000 per quality-adjusted life year and $100,000 per quality-adjusted life year if the risk of perioperative DVT in a given population is approximately between 1.5% and 1%, respectively. Alternatively, when intermittent pneumatic compression is responsible for a large risk reduction in DVT risk, such as 90%, the risk of a perioperative DVT needs to be between 0.2% and 0.45% for the result to continue in the same cost-effective range.
Routine thromboprophylaxis at cesarean delivery is not standard practice in the United States. Although guidelines in the United Kingdom have called for universal prophylaxis and the American College of Chest Physicians has advocated a risk-based system for thromboprophylaxis, these guidelines are based solely on expert opinion. To date, there has been no prospective randomized trial of adequate sample size that has allowed the development of clear consensus with regard to proper clinical management.30 Similarly, there has not been an investigation that has provided a well-accepted system whereby individual risk factors for venous thromboembolism can be used to arrive at an accurate risk of venous thromboembolism associated with cesarean delivery for a given individual. Because pharmacologic prophylaxis carries the potential for risks of bleeding as well as the less common but more serious potential for heparin-induced thrombocytopenia, many practitioners have been reluctant to adopt non–evidence-based pharmacologic strategies for thromboprophylaxis for women for whom the probability of venous thromboembolism is unclear.
A strategy of prophylaxis with intermittent pneumatic compression avoids the potential complications of pharmacologic prophylaxis. Indeed, a recent analysis revealed that if any prophylaxis were to be used, mechanical compression is the most rational choice.13 Nevertheless, it is not clear that this strategy should be routinely employed, because it would require the use of resources for more than 1 million women annually for a disease process that is uncommon and most often able to be treated without severe acute morbidity or mortality.
In this analysis, we have attempted to discern whether prophylaxis with intermittent pneumatic compression would ever be a cost-effective intervention, and if so, under what circumstances that cost-effectiveness would exist. Based on this Markov model, we can draw several conclusions. It is conceivable that prophylaxis could be cost-effective, even for all women undergoing cesarean. This possibility, however, is dependent on the values of three variables. To some degree, the variable cost of the compression boots affects the cost-effectiveness of the intervention. Yet, as evidenced in Figure 1, the variables that most affect the cost-effectiveness are the risk reduction for DVT from intermittent pneumatic compression and the prevalence of perioperative DVT. From data that is available from surgical and gynecologic patients, intermittent compression is more than 50% effective in the reduction of DVT, and thus is likely to provide a risk reduction that would allow intermittent pneumatic compression to be a cost-effective strategy for obstetric patients.9–13 Ultimately, the possibility of cost-effectiveness will depend upon perioperative DVT prevalence. Wide variations in the prevalence of DVT, ranging from 0.5 in 1,000 to 3 in 1,000, have been reported in obstetric populations, and even these frequencies may have been affected by ascertainment bias and are not specifically the perioperative prevalence. Further investigations into the frequency of perioperative DVT, and possibly into the frequencies for different subpopulations of women, will provide data that will, in conjunction with our analysis, further clarify whether intermittent pneumatic compression is cost-effective, at the very least for selected populations with an elevated risk of perioperative DVT.
As with any cost-effectiveness analysis, the results of the model are dependent upon the variables that are used. In most cases, the choice of the values for the baseline estimate was not highly relevant, because with three exceptions, any reasonable estimate that was used did not significantly affect the results. The reduction in DVT risk from intermittent pneumatic compression, a value that could materially alter the result, was derived from the nonobstetric literature, and the generalizability of its efficacy to an obstetric population is not well known. And, as noted above, the perioperative risk for cesarean-associated DVT remains somewhat imprecise.
Nevertheless, across a wide range of possible scenarios, it seems that prophylaxis with intermittent pneumatic compression devices is cost-effective for the prevention of venous thromboembolism. Although a prospective randomized trial remains the “gold standard” in attempting to determine the true risks and benefits of interventions, until such a trial can be completed, our analysis provides information that may be useful in considering the best strategy for thromboprophylaxis for cesarean delivery.
1. Greer IA, Thomson AJ. Management of venous thromboembolism in pregnancy. Best Pract Res Clin Obstet Gynaecol 2001;15:583–603.
2. Andres RL, Miles A. Venous thromboembolism and pregnancy. Obstet Gynecol Clin North Am 2001;28:613–30.
3. Bonnar J. Can more be done in obstetric and gynecologic practice to reduce morbidity and mortality associated with venous thromboembolism? Am J Obstet Gynecol 1999;180:784–91.
4. Drife J, Lewis G. Why mothers die 1997-1999: the fifth report of the confidential enquiries into maternal deaths in the United Kingdom. London (UK):RCOG Press; 2001.
5. Kearon C. Natural history of venous thromboembolism. Circulation 2003;107:I22–30.
6. Lindhagen A, Bergqvist A, Bergqvist D, Hallbrook T. Late venous function in the leg after deep venous thrombosis occurring in relation to pregnancy. Br J Obstet Gynecol 1986;93:348–52.
7. Greer IA. Prevention and management of venous thromboembolism in pregnancy. Clin Chest Med 2003;24:123–37.
8. Bates SM, Greer IA, Hirsch J, Ginsberg JS. Use of antithrombotic agents during pregnancy: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2004;126:627S–44S.
9. Tsapogas MJ, Goussous H, Peabody RA, Karmody AM, Eckert C. Postoperative venous thrombosis and the effectiveness of prophylactic measures. Arch Surg 1971;103:561–7.
10. Clarke-Pearson DL. Prevention of venous thromboembolism in gynecologic surgery patients. Curr Opin Obstet Gynecol 1993;5:73–9.
11. Turpie AG, Hirsh J, Gent M, Julian D, Johnson J. Prevention of deep vein thrombosis in potential neurosurgical patients. A randomized trial comparing graduated compression stockings alone or graduated compression stockings plus intermittent pneumatic compression with control. Arch Intern Med 1989;149:679–81.
12. Jeffery PC, Nicolaides AN. Graduated compression stockings in the prevention of postoperative deep vein thrombosis. Br J Surg 1990;77:380–3.
13. Quinones JN, James DN, Stamilio DM, Cleary KL, Macones GA. Thromboprophylaxis after cesarean delivery: a division analysis. Obstet Gynecol 2005;106:733–40.
14. Sonnenberg FA, Beck JR. Markov models in medical decision making: a practical guide. Med Decis Making 1993;13:322–38.
15. Barbour LA, Pickard J. Controversies in thromboembolic disease during pregnancy: a critical review. Obstet Gynecol 1995;86:621–33.
16. Clarke-Pearson DL, Synan IS, Colemen RE, Hinshaw W, Creasman WT. The natural history of postoperative venous thromboembolism in gynecologic oncology: a prospective study of 382 patients. Am J Obstet Gynecol 1984;148:1051–4.
17. Wessler S. Medical management of venous thrombosis. Annu Rev Med 1976;27:313–9.
18. Kelly J, Hunt BJ. Do anticoagulants improve survival in patients presenting with venous thromboembolism? J Intern Med 2003;254:527–39.
19. Clagett GP, Reisch JS. Prevention of venous thromboembolism in general surgical patients. Results of meta-analysis. Ann Surg 1988;208:227–40.
20. Villasanta U. Thromboembolic disease in pregnancy. Am J Obstet Gynecol 1965;93:142–60.
21. Marchetti M, Pistorio A, Barone M, Serafini S, Barosi G. Low molecular weight heparin versus warfarin for secondary prophylaxis of venous thromboembolism: a cost-effectiveness analysis. Am J Med 2001;111:130–9.
22. Maxwell GL, Myers ER, Clarke-Pearson KL. Cost-effectiveness of deep venous thrombosis prophylaxis in gynecologic oncology surgery. Obstet Gynecol 2000;95:206–14.
23. Pradoni P, Lensing AW, Cogo A, Cuppini S, Villalta S, Carta M, et al. The long-term clinical course of acute deep venous thrombosis. Ann Intern Med 1996;125:1–7.
24. Gage BF, Cardinalli AB, Owens DK. The effect of stroke and stroke prophylaxis with aspirin or warfarin on quality of life. Arch Intern Med 1996;156:1829–36.
25. Dainty L, Maxwell GL, Clarke-Pearson DL, Myers ER. Cost-effectiveness of combination thromboembolism prophylaxis in gynecologic oncology surgery. Gynecol Oncol 2004;93:366–73.
26. Marchetti M, Quaglini S, Barosi G. Cost-effectiveness of screening and extended anticoagulation for carriers of both factor V Leiden and prothrombin G20210A. QJM 2001;94:365–72.
27. Botteman MF, Caprini J, Stephens JM, Nadipelli V, Bell CF, Pashos CL, et al. Results of an economic model to assess the cost-effectiveness of enoxaparin, a low-molecular-weight heparin, versus warfarin for the prophylaxis of deep vein thrombosis and associated long-term complication in total hip replacement surgery in the United States. Clin Ther 2002;24:1960–86.
28. Schulman K, Burke J, Drummond M, Davies L, Carlsson P, Gruger J, et al. Resource costing for multinational neurologic clinical trials: methods and results. Health Econ 1998;7:629–38.
29. Ubel PA, Hirth RA, Chernew ME, Fendrick AM. What is the price of life and why doesn’t it increase at the rate of inflation? Arch Intern Med 2003;163:1637–41.
30. Gates S, Brocklehurst P, Davis LJ. Prophylaxis for venous thromboembolic disease and the early postnatal period. Cochrane Database Syst Rev 2002: CD001689.