- Email: [email protected]
SURGERY OR ASSISTED REPRODUCTION? A DECISION ANALYSIS OF TREATMENT COSTS IN MALE INFERTILITY
0022-5347/05/1745-1926/0 THE JOURNAL OF UROLOGY® Copyright © 2005 by AMERICAN UROLOGICAL ASSOCIATION
Vol. 174, 1926 –1931, November 2005 Printed in U.S.A.
DOI: 10.1097/01.ju.0000176736.74328.1a
SURGERY OR ASSISTED REPRODUCTION? A DECISION ANALYSIS OF TREATMENT COSTS IN MALE INFERTILITY MAXWELL V. MENG,* KIRSTEN L. GREENE
AND
PAUL J. TUREK
From the Department of Urology, University of California-San Francisco, San Francisco, California
ABSTRACT
Purpose: Assisted reproductive technology (ART), including in vitro fertilization and intracytoplasmic sperm injection, is routinely used to treat male factor infertility. Because of the success of ART, the optimal method to achieve pregnancy with male infertility is controversial. Two examples in which ART competes with traditional male infertility treatments are varicocelectomy and vasectomy reversal. We used formal decision analysis to estimate and compare the cost-effectiveness of surgical therapy and ART for varicocele and vasectomy reversal. Materials and Methods: Decision analysis models were created for infertile men seeking paternity with varicocele and with post-vasectomy obstruction. Outcome probabilities applied to the model were derived from institutional and published sources. Costs of interventions were calculated from institutional data. Sensitivity analyses determined which elements were most important and, thus, were used to calculate threshold values. Results: Vasectomy reversal is as cost-effective as ART if bilateral vasovasostomy can be performed. However, if unilateral or bilateral vasoepididymostomy is required, sperm retrieval/ intracytoplasmic sperm injection may be more cost-effective due to lower patency rates. Vasectomy reversal is more cost-effective across all pregnancy rates provided that patency rates are greater than 79%. Surgical repair of varicocele is more cost-effective when the postoperative pregnancy rate is greater than 14% in men with a preoperative total motile sperm count of less than 10 million sperm and greater than 45% in men with greater than 10 million total motile sperm. Conclusions: A decision analysis based comparison of ART and classic surgical therapy suggests that varicocelectomy and vasectomy reversal are the most economical treatments in many cases of infertility due to these lesions. Tailoring the decision models to individual centers permits more accurate comparisons using specific costs as well as the surgical outcomes and results of ART. KEY WORDS: infertility, male; decision support techniques; reproductive techniques, assisted
Infertility affects between 10% and 15% of couples in the United States and male infertility underlies half of the cases. The advent of assisted reproductive technology (ART) has revolutionized infertility treatment by providing a viable alternative to classic male infertility treatment. Specifically in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) allow us to overcome even the most severe defects in spermatogenesis in which only a few sperm are available. Although they are effective, these interventions are expensive and a discussion of costs must be included in the care provided to infertile couples. Few cost-benefit studies are available to help guide physicians and patients in making appropriate decisions regarding infertility management.1, 2 Two clinical conditions exemplify male factor diagnoses amenable to surgical and ART treatment, namely varicocele and obstruction due to vasectomy. Varicocele occurs in 35% of infertile men and treatment with surgical ligation or embolization is an established method for improving sperm quality and pregnancy rates in infertile couples.3, 4 However, not all patients who undergo varicocele repair conceive, making ART a viable alternative strategy for family building. Vasectomy is a common birth control method in the United States but 2% to 6% of men seek future fertility after the procedure. Fertility after vasectomy is possible with surgical reconstruction or sperm retrieval and ART.
Although ART offers a successful pregnancy alternative to varicocelectomy and vasectomy reversal, the optimal method for achieving pregnancy remains controversial. Many reproductive endocrinologists routinely apply ART in men with vasectomy and varicocele associated infertility. However, in studies that compared the costs of microsurgical reconstruction and varicocelectomy to those of sperm retrieval/ICSI classic surgical therapy was viewed favorably.2, 5 With such strong opinions and scarce data the ability of couples to intelligently navigate their reproductive options is limited. Therefore, we clarified the cost-effectiveness of management options for vasectomy and varicocele associated infertility using decision analysis modeling. Decision models are constructed with predefined assumptions and they serve as useful tools for estimating outcomes when multiple complex treatments are available. In urology decision analysis modeling has been applied to prostate cancer treatment to clarify the critical variables involved in treatment decisions.6 In this study we applied formal decision analyses to calculate the cost of pregnancy for initial surgical or initial ART treatment in men with infertility due to varicocele and vasectomy. MATERIALS AND METHODS
Our analyses addressed the initial decision of whether to intervene surgically or use ART in cases of vasectomy reversal or varicocelectomy. Two algorithms were created to model treatment decisions and their outcomes for each urological
Submitted for publication December 30, 2004. * Correspondence: Department of Urology, University of California San Francisco, 1600 Divisadero St., Room A633, San Francisco, California 94115-1695. 1926
SURGERY OR ASSISTED REPRODUCTION DECISION ANALYSIS
1927
FIG. 1. Treatment algorithm for pregnancy after vasectomy. Decision is choice between initial surgical reconstruction and ART. IC, intercourse.
condition (figs. 1 and 2). For the purpose of analysis we included intrauterine insemination (IUI) along with sperm retrieval/ICSI as a component of ART treatment. Decision modeling for vasectomy and varicocele associated infertility. In a patient seeking fertility after vasectomy the first decision is the choice between vasectomy reversal and ART (fig. 1). In this decision model the cost per pregnancy was the end point for comparing the outcomes of the 2 branches. This was calculated using the cumulative costs of each outcome state in the specific branch and dividing this value by the number of pregnancies achieved in that branch. Several important assumptions were made in the model, namely 1) fecundity depended on female age but was independent of treatment modality, 2) pregnancy rates were stable and did not decrease with subsequent treatments and 3) the fraction of couples that pursued additional therapy after initial treatment failure was the same in each decision algorithm arm. A similar approach was used to create the decision model in men with varicocele associated infertility. The first decision was the choice between varicocelectomy and ART (fig. 2). Again, the cost per pregnancy was the end point for comparing the outcomes of the 2 branches of the model. Assumptions in this model were 1) fecundity depended on female age and not on treatment modality, 2) sperm retrieval/ICSI would be used for patients with a total motile sperm count (TMC) (calculated as volume ⫻ concentration ⫻ motile fraction) of less than 10 million sperm and 3) IUI would be used in couples with a TMC of greater than 10 million sperm. The decision models were analyzed using Decision Analyses, Data 2.5 computer software (TreeAge, Williamstown, Massachusetts). The cost per pregnancy and the pregnancy rate in each arm were calculated and compared. The more cost-effective method for treating vasectomy and varicocele associated infertility was the method that yielded the lower cost per pregnancy. The value at which variation in a parameter altered the outcome of the decision analysis, ie a lower cost per pregnancy, was termed the threshold value. Probability and utility data for decision modeling. In the
vasectomy reversal model data were gathered from the literature regarding vasectomy reversal patency and pregnancy rates after IUI and sperm retrieval/ICSI.5, 7–9 Table 1 lists the derived probability values. For the purpose of initial decision modeling we chose reasonable success rates but not necessarily the highest published rates from the literature. Thus, the vasectomy reversal pregnancy and patency rates were simplified and estimated to be 30% and 80%, respec-
FIG. 2. Treatment algorithm for pregnancy with male factor infertility due to varicocele. Decision is choice between surgical varicocelectomy and ART.
1928
SURGERY OR ASSISTED REPRODUCTION DECISION ANALYSIS
TABLE 1. Derived probabilities for vasectomy reversal and pregnancy rates used in decision model Variable
% Estimated Probability
Patency: Bilat vasovasostomy Vasovasostomy/vasoepididymostomy Bilat vasoepididymostomy Pregnancy rate: After vasectomy reversal After 1 ICSI cycle After 4 IUI cycles Second treatment sought
87 70 65 30 30 32 50
TABLE 2. Institutional intervention costs Procedure
Cost (US $)
Microsurgical vasovasostomy Microsurgical varicocelectomy 1 IVF-ICSI cycle 1 IUI cycle
10,000 4,500 10,000 500
TABLE 3. Pregnancy rate after varicocelectomy10 Preop TMC (1 ⫻ 106)
No.
% Pregnancy Rate
0–1.5 1.5–5 5–10 10–20 20–40 Greater than 40
132 62 58 68 60 65
16.5 30.6 36.2 39.7 58.3 61.5
445
36.6
Total
tively. In addition, the fraction of couples pursuing additional therapy after primary treatment failure was assumed to be 50%. Table 2 lists the cost of each treatment using data from estimated total charges at our institution. In the varicocele model we applied published institutional data to generate probability values.10 Table 3 lists pregnancy rates in the varicocelectomy cohort stratified by preoperative TMC. Of all patients 36.6% of couples achieved spontaneous pregnancy after surgery. The pregnancy rate for a single ICSI cycle and 4 IUI cycles was assumed to be 30% and 32%, respectively.11, 12 RESULTS
Post-vasectomy infertility. Figure 1 shows the decision model comparing vasectomy reversal and sperm retrieval/ ICSI. Table 4 shows the calculated cost per pregnancy for vasectomy reversal and sperm retrieval/ICSI for reversal patency rates between 40% and 100%, and female fecundity
rates between 10% and 50%. For each cell in the table the calculated cost per pregnancy for the specific intervention is listed for direct comparison. The baseline decision analysis favored initial vasectomy reversal rather than sperm retrieval/ICSI if bilateral vasovasostomy was performed at a cost per pregnancy of $38,983 vs $39,506, which corresponds to column 4 (80% patency) and row 3 (30% pregnancy rate) (table 4). If a 100% patency rate were achievable, this approach would be uniformly favored over sperm retrieval/ICSI across all pregnancy rates. However, sperm retrieval/ICSI was initially more cost-effective if surgical reconstruction required unilateral or bilateral vasoepididymostomy due to lower associated patency rates (columns to the left of 80% reversal patency) (table 4). Figure 3, A represents 1-way sensitivity analysis in which costs were compared for a range of reversal patency rates while holding the pregnancy rate constant at 40%. In this scenario vasectomy reversal was more cost-effective than sperm retrieval/ICSI when the patency rate was greater than 79%, the threshold value. Similarly figure 3, B shows 1-way sensitivity analysis of the pregnancy rate when the patency rate was held constant at 60%. There is no threshold value in this graph since sperm retrieval/ICSI was always more cost-effective in this scenario with lower patency rates. This information was summarized in the 2-way sensitivity analysis, in which reversal patency and pregnancy rates were varied (fig. 3, C). In this analysis vasectomy reversal was more cost-effective when the patency rate was greater than 78%, independent of female fecundity estimates (ie the pregnancy rate). Varicocele associated infertility. Figure 2 shows the decision model comparing initial varicocele ligation and ART for varicocele associated infertility. Table 5 shows the cost per pregnancy of surgical intervention and ART in the entire group and the cost by preoperative TMC. Overall initial surgical repair was more cost-effective than ART. However, IUI yielded a lower cost per pregnancy than varicocelectomy in men with a preoperative TMC of between 10 and 20 million sperm ($9,000 vs $11,333, respectively). In men with lower sperm counts (TMC less than 10 million) who would qualify for sperm retrieval/ICSI instead of IUI we compared the cost-effectiveness of sperm retrieval/ ICSI and varicocele repair after varying the postvaricocelectomy pregnancy rate. In these men due to the high cost of sperm retrieval/ICSI varicocelectomy was the more cost-effective treatment when the postoperative pregnancy rate was greater than the 14% threshold value (fig. 4, A). In men with higher sperm counts (TMC greater than 10 million) who might qualify for IUI instead of sperm retrieval/ICSI varicocelectomy was more cost-effective than IUI only when the pregnancy rate was greater than the 45% threshold value (fig. 4, B).
TABLE 4. Cost per pregnancy for ART and vasovasostomy by pregnancy and patency rates % Pregnancy Rate 10: Reversal ICSI 20: Reversal ICSI 30: Reversal ICSI 40: Reversal ICSI 50: Reversal ICSI
40% Reversal Patency (US $)
60% Reversal Patency (US $)
80% Reversal Patency (US $)
100% Reversal Patency (US $)
168,182 117,241
137,383 117,241
115,873 117,241
100,000 117,241
84,884 58,929
69,231 58,929
58,197 58,929
50,000 58,929
57,143 39,506
46,535 39,506
38,983 39,506
33,333 39,506
43,293 29,808
35,204 29,808
29,386 29,808
25,000 29,808
35,000 24,000
28,421 24,000
23,636 24,000
20,000 24,000
SURGERY OR ASSISTED REPRODUCTION DECISION ANALYSIS
1929
FIG. 3. Sensitivity analysis. A, comparison of vasectomy reversal patency costs for range of reversal patency rates when pregnancy rate was constant at 40%. B, 1-way analysis of pregnancy rate in vasectomy associated male infertility with vasectomy reversal patency constant at 60%. C, 2-way analysis of pregnancy rate and reversal patency in vasectomy associated male infertility comparing cost-effectiveness of surgical reconstruction and ICSI when varying reversal patency and pregnancy rates. Shaded area indicates most cost-effective modality.
TABLE 5. Cost per pregnancy of male infertility due to varicocele Preop TMC (1 ⫻ 106) 0–1.5 1.5–5 5–10 10–20 20–40 Greater than 40
Cost/Pregnancy (US $) Varicocelectomy
ART
28,286 14,684 12,429 11,333 7,714 7,313
33,333 33,333 33,333 9,000 9,000 9,000
DISCUSSION
The success of ART for treating female and male infertility has led to its empirical application in cases of surgically treatable diagnoses in male infertility. In the last decade several concerns regarding ART treatment have been realized. 1) It is expensive.13–15 2) It has associated complications. 3) It has implications for the health of offspring, including low birth weight, prematurity, chromosomal abnormalities and an increased incidence of birth defects.16, 17 Surprisingly despite its widespread application, to our knowledge no published study has documented that ART is more cost-effective than classic urological treatments for male infertility. In this study we further clarified the costeffectiveness issues that surround the management options for vasectomy and varicocele associated infertility through decision analysis modeling. Our decision analysis suggests that vasectomy reversal is more cost-effective than sperm retrieval and IVF-ICSI. Moreover, other observations became apparent as variables in the model were altered to determine the effect on outcomes. A threshold patency rate of 78% was identified below which sperm retrieval/ICSI was more cost-effective than vasectomy reversal. This benchmark value for the reversal patency rate implies that vasectomy reversal surgery must be performed
at a high standard of quality to be more cost-effective than sperm retrieval/ICSI. In general vasectomy reversal is more likely to be cost-effective in men with a shorter time from vasectomy, in whom bilateral vasovasostomy can typically be performed. Conversely men with a longer interval and greater potential for requiring vasoepididymostomy may benefit from initial sperm retrieval/ICSI. In addition, the cost-effectiveness of vasectomy reversal treatment depends primarily on male and not on female fertility factors. Although it is true that female fecundity alters the actual costs of surgical reconstruction and ART, the relative costs remain unchanged. Varying reversal patency rates dramatically altered cost-effectiveness outcomes in the model. This implies that despite decreased pregnancy rates in couples with an older female partner vasectomy reversal can remain the more cost-effective approach if the patency rate exceeds the threshold value. This concept is substantiated by a review of the literature. In a series of 29 patients with vasectomy reversal Deck and Berger found a lower cost per newborn with vasectomy reversal compared with sperm retrieval/ICSI at a median female age of 40 years (range 38 to 48).18 Interestingly the patency rate achieved in that series was greater than 90%, which is a value consistent with vasectomy reversal cost-effectiveness in our decision model. The decision model provides a framework for the individual surgeon to counsel patients. Although efficacy data are available regarding postoperative patency rates after vasectomy reversal from the multicenter Vasovasostomy Study Group,7 this information reflects the practice of experienced and dedicated microsurgeons and not necessarily that of the individual community based surgeon. Conveniently this decision analysis model can be tailored to be accurate for an individual surgical practice by simply including data from the vasectomy reversal experience of that surgeon. In addition, costs vary greatly among individual practices and geographic locations, and the availability of the models permits
1930
SURGERY OR ASSISTED REPRODUCTION DECISION ANALYSIS
FIG. 4. A, cost-effectiveness of varicocelectomy vs IVF-ICSI in men with varicocele associated male infertility and TMC less than 10 million at various post-varicocelectomy pregnancy rates. B, sensitivity analysis in men with varicocele associated infertility with TMC greater than 10 million comparing cost of varicocele and ART with IUI across pregnancy rates.
customization of the calculations and comparisons at a particular center. We are in the process of translating the decision models to web based tools, similar to currently popular nomograms and models providing outcome estimates for individuals with prostate cancer, allowing one to input the costs and estimated success of interventions with the resulting calculation of the cost per pregnancy and relative costeffectiveness. The second decision model was created to address the cost-effectiveness of varicocele associated infertility treatment. In a previous study Schlegel compared sperm retrieval/ ICSI and varicocelectomy pregnancy rates, and concluded that primary treatment with varicocele repair was more costeffective than ART, while providing comparable delivery rates.2 In our study we refined their cost-benefit analysis by 1) applying decision modeling, 2) stratifying the degree of semen impairment according to motile sperm counts and 3) including IUI as a viable treatment option in the subset of men with varicocele associated infertility. We observed that surgical treatment is more cost-effective than ART in certain cases of varicocele associated infertility. If in a surgeon’s experience men with a less severely impaired preoperative motile sperm count (greater than 10 million motile sperm) cannot achieve a 45% pregnancy rate after surgery, ART can be more cost-effective. This is largely due to the relatively inexpensive IUI procedure for which such couples may qualify. On the contrary, if a surgeon can achieve a greater than 14% pregnancy rate in men with more severe impairment in semen quality (less than 10 million motile sperm), varicocele repair is more cost-effective since these couples generally require sperm retrieval/ICSI instead of IUI to overcome infertility.19
Decision analysis models have inherent limitations. In our modeling we did not consider factors such as the number of children desired, the social and monetary costs of intervention in the female partner, the cost of time lost from work or the costs attributable to procedural complications and multiple gestations. However, in our estimation incorporating these variables would likely support the argument for initial surgical treatment rather than ART for the 2 types of male factor infertility. Other intangible elements, such as the potential for future natural conception, favor surgical approaches to treatment. Conversely our model did not incorporate the complex issue of the effects of age on maternal reproductive potential. Inclusion of this variable into the model would likely favor ART over surgery in a higher proportion of affected couples because the mean time to pregnancy after vasectomy reversal and varicocelectomy (12 and 8 months, respectively)7, 20 may be longer than that needed for a pregnancy with ART. Overall the quality of our analysis is supported by the fact that the calculated cost per pregnancy is reasonable in each model and they are within the range of values that have been published for these treatments.1, 5, 19 The information provided by these decision models serves as a framework for future analyses and discussion. Further refinements may account for more accurate assessment of variables, such as female fecundity and improvements in IVF-ICSI. In addition, customization of the decision analysis can address concerns of differences in costs and estimations of treatment efficacy. These data provide important information for all parties involved in the treatment of reproductive impairments, including physicians, infertile couples and insurance carriers, of whom each has specific concerns and desired end points. CONCLUSIONS
Formal decision analysis provides a structured means of assessing treatment choices in male factor infertility. According to models constructed for varicocele and vasectomy related male infertility surgery is often more cost-effective than ART. However, threshold values for surgical success rates are needed to maintain the cost-effectiveness of urological surgery over ART and provide a form of quality standard for the practicing surgeon. Formal decision modeling can also serve as a web based instrument for comparing treatment costs for individual urologists using practice specific success rates and, thereby, it can more accurately present the complex treatment decisions facing couples seeking infertility care. REFERENCES
1. Pavlovich, C. P. and Schlegel, P. N.: Fertility options after vasectomy: a cost-effectiveness analysis. Fertil Steril, 67: 133, 1997 2. Schlegel, P. N.: Is assisted reproduction the optimal treatment for varicocele-associated male infertility? A cost-effectiveness analysis. Urology, 49: 83, 1997 3. Evers, J. L. and Collins, J. A.: Surgery or embolisation for varicocele in subfertile men. Cochrane Database Syst Rev, 3: CD000479, 2004 4. Madgar, I., Weissenberg, R., Lunenfeld, B., Karasik, A. and Goldwasser, B.: Controlled trial of high spermatic vein ligation for varicocele in infertile men. Fertil Steril, 63: 120, 1995 5. Kolettis, P. N. and Thomas, A. J., Jr.: Vasoepididymostomy for vasectomy reversal: a critical assessment in the era of intracytoplasmic sperm injection. J Urol, 158: 467, 1997 6. Simpson, K. N.: Problems and perspectives on the use of decision-analysis models for prostate cancer. J Urol, 152: 1888, 1994 7. Belker, A. M., Thomas, A., Jr., Fuchs, E. F., Konnak, J. W. and Sharlip, I. D.: Results of 1,469 microsurgical vasectomy reversals by the Vasovasostomy Study Group. J Urol Nurs, 11: 93, 1992
SURGERY OR ASSISTED REPRODUCTION DECISION ANALYSIS 8. Kolettis, P. N., Sabanegh, E. S., Nalesnik, J. G., D’Amico, A. M., Box, L. C. and Burns, J. R.: Pregnancy outcomes after vasectomy reversal for female partners 35 years old or older. J Urol, 169: 2250, 2003 9. Nalesnik, J. G. and Sabanegh, E. S., Jr.: Vasovasostomy: multiple children and long-term pregnancy rates. Curr Surg, 60: 348, 2003 ¨ ¸ 10. Cayan, S., Erdemir, F., Ozbey, I., Turek, P. J., Kadiog˘lu, A. and Tellalog˘lu, S.: Can varicocelectomy significantly change the way couples use assisted reproductive technologies? J Urol, 167: 1749, 2002 11. Guzick, D. S., Carson, S. A., Coutifaris, C., Overstreet, J. W., Factor-Litvak, P., Steinkampf, M. P. et al: Efficacy of superovulation and intrauterine insemination in the treatment of infertility. National Cooperative Reproductive Medicine Network. N Engl J Med, 340: 177, 1999 12. Jee, B. C., Ku, S. Y., Suh, C. S., Choi, Y. M., Kim, J. G., Moon, S. Y. and Kim, S. H.: Cumulative ongoing pregnancy rate in intracytoplasmic sperm injection cycles. J Obstet Gynaecol Res, 30: 372, 2004 13. Donovan, J. F., Jr., DiBaise, M., Sparks, A. E., Kessler, J. and Sandlow, J. I.: Comparison of microscopic epididymal sperm aspiration and intracytoplasmic sperm injection/in-vitro fertilization with repeat microscopic reconstruction following vasectomy: is second attempt vas reversal worth the effort? Hum Reprod, 13: 387, 1998 14. Van Voorhis, B. J., Sparks, A. E., Allen, B. D., Stovall, D. W., Syrop, C. H. and Chapler, F. K.: Cost-effectiveness of infertility treatments: a cohort study. Fertil Steril, 67: 830, 1997 15. Neumann, P. J., Gharib, S. D. and Weinstein, M. C.: The cost of a successful delivery with in vitro fertilization. N Engl J Med, 331: 239, 1994 16. Hansen, M., Kurinczuk, J. J., Bower, C. and Webb, S.: The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization. N Engl J Med, 346: 725, 2002 17. Schieve, L. A., Meikle, S. F., Ferre, C., Peterson, H. B., Jeng, G. and Wilcox, L. S.: Low and very low birth weight in infants conceived with use of assisted reproductive technology. N Engl J Med, 346: 731, 2002 18. Deck, A. J. and Berger, R. E.: Should vasectomy reversal be performed in men with older female partners? J Urol, 163: 105, 2000 19. Garceau, L., Henderson, J., Davis, L. J., Petrou, S., Henderson, L. R., McVeigh, E. et al: Economic implications of assisted reproductive techniques: a systematic review. Hum Reprod, 17: 3090, 2002 20. Cayan, S., Lee, D., Black, L. D., Reijo Pera, R. A. and Turek, P. J.: Response to varicocelectomy in oligospermic men with and without defined genetic infertility. Urology, 57: 530, 2001 EDITORIAL COMMENT There are many different ways to evaluate the value of a treatment. An approach that takes into account the cost as well as the outcomes is a cost-effectiveness evaluation. As with most models, the devil of such an evaluation is in the details. How are the data derived? What assumptions were made? These are critical points to help determine whether the modeled data have any value to another practitioner. Cost-effectiveness is of particular value for determining the role of a fertility treatment because patients may bear more of
1931
the cost of such medical care out of pocket, since insurers may not cover such treatments or may provide only partial insurance coverage. Prior reports have used national (references 1, 2, 15 and 19 in article), and local/single institution costs and results (references 5 and 14 in article). Various costs can be included beyond the direct costs of medical care. The resulting costs of multiple gestations is important to consider. Although twins appear to be a blessing for the couple, the arrival of triplets almost always engenders prematurity, which can result in substantive risks for the child (cerebral palsy and blindness) and massive costs for the health care system (reference 15 in article.) Such costs were not considered in this analysis, although 17% to 54% of pregnancies after IVF cycles at the University of California-San Francisco center are multiple gestations.1 Of interest is that the success rate for IVF/ICSI quoted for the University of California-San Francisco center is 30% in the report, whereas the Centers for Disease Control website quotes a 28.5% delivery rate per IVF attempt in women 40 years younger,1 suggesting that the data in the article may overstate the relative benefits of assisted reproduction. The article provides a guide and an example from 1 center of how to evaluate the relative value of specific male fertility treatment with varicocelectomy and vasectomy reversal. Readers must consider the costs and specific pregnancy results obtained at their center to apply this information to the clinical care of patients. Peter N. Schlegel Department of Urology Weill Medical College of Cornell University New York Presbyterian Hospital New York, New York 1. 2002 Assisted Reproductive Technology (ART) Report: 2002 Fertility Clinic Report by state: Clinic Data. Department of Health and Human Services, Centers for Disease Control and Prevention. Available at http://apps.nccd.cdc.gov/ART2002/ clindata02.asp?Location⫽60. REPLY BY AUTHORS Part of the strength and usefulness of decision analytic models is that while disagreement may exist on a specific cost or pregnancy rate, these values are merely used as the starting point for analysis. The true value of decision models is the ability to change the numerous variables, either one at a time or simultaneously, and determine the impact on outcomes. Thus, quibbling about overstating outcomes (30% vs 28.5%) can be easily addressed using sensitivity analyses, as explicitly shown in figure 3 in the article. We appreciate the comments and acknowledge that our models, and decision analytic methods in general, have limitations. These assumptions and simplifications are clearly stated, and we discuss their implications overall, the issues raised (eg multiple gestations) only increase the costs of ART and further suggest that classic, surgical treatments of male factor infertility are cost-effective approaches. Ultimately, the usefulness of our cost analysis can be improved by incorporating a couple’s specific situation, using more accurate estimates of all variables including fecundity and hospital/physician charges. This individualized assessment of relative costs may help guide patients as well as physicians during the decision making process.
Vol. 174, 1926 –1931, November 2005 Printed in U.S.A.
DOI: 10.1097/01.ju.0000176736.74328.1a
SURGERY OR ASSISTED REPRODUCTION? A DECISION ANALYSIS OF TREATMENT COSTS IN MALE INFERTILITY MAXWELL V. MENG,* KIRSTEN L. GREENE
AND
PAUL J. TUREK
From the Department of Urology, University of California-San Francisco, San Francisco, California
ABSTRACT
Purpose: Assisted reproductive technology (ART), including in vitro fertilization and intracytoplasmic sperm injection, is routinely used to treat male factor infertility. Because of the success of ART, the optimal method to achieve pregnancy with male infertility is controversial. Two examples in which ART competes with traditional male infertility treatments are varicocelectomy and vasectomy reversal. We used formal decision analysis to estimate and compare the cost-effectiveness of surgical therapy and ART for varicocele and vasectomy reversal. Materials and Methods: Decision analysis models were created for infertile men seeking paternity with varicocele and with post-vasectomy obstruction. Outcome probabilities applied to the model were derived from institutional and published sources. Costs of interventions were calculated from institutional data. Sensitivity analyses determined which elements were most important and, thus, were used to calculate threshold values. Results: Vasectomy reversal is as cost-effective as ART if bilateral vasovasostomy can be performed. However, if unilateral or bilateral vasoepididymostomy is required, sperm retrieval/ intracytoplasmic sperm injection may be more cost-effective due to lower patency rates. Vasectomy reversal is more cost-effective across all pregnancy rates provided that patency rates are greater than 79%. Surgical repair of varicocele is more cost-effective when the postoperative pregnancy rate is greater than 14% in men with a preoperative total motile sperm count of less than 10 million sperm and greater than 45% in men with greater than 10 million total motile sperm. Conclusions: A decision analysis based comparison of ART and classic surgical therapy suggests that varicocelectomy and vasectomy reversal are the most economical treatments in many cases of infertility due to these lesions. Tailoring the decision models to individual centers permits more accurate comparisons using specific costs as well as the surgical outcomes and results of ART. KEY WORDS: infertility, male; decision support techniques; reproductive techniques, assisted
Infertility affects between 10% and 15% of couples in the United States and male infertility underlies half of the cases. The advent of assisted reproductive technology (ART) has revolutionized infertility treatment by providing a viable alternative to classic male infertility treatment. Specifically in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) allow us to overcome even the most severe defects in spermatogenesis in which only a few sperm are available. Although they are effective, these interventions are expensive and a discussion of costs must be included in the care provided to infertile couples. Few cost-benefit studies are available to help guide physicians and patients in making appropriate decisions regarding infertility management.1, 2 Two clinical conditions exemplify male factor diagnoses amenable to surgical and ART treatment, namely varicocele and obstruction due to vasectomy. Varicocele occurs in 35% of infertile men and treatment with surgical ligation or embolization is an established method for improving sperm quality and pregnancy rates in infertile couples.3, 4 However, not all patients who undergo varicocele repair conceive, making ART a viable alternative strategy for family building. Vasectomy is a common birth control method in the United States but 2% to 6% of men seek future fertility after the procedure. Fertility after vasectomy is possible with surgical reconstruction or sperm retrieval and ART.
Although ART offers a successful pregnancy alternative to varicocelectomy and vasectomy reversal, the optimal method for achieving pregnancy remains controversial. Many reproductive endocrinologists routinely apply ART in men with vasectomy and varicocele associated infertility. However, in studies that compared the costs of microsurgical reconstruction and varicocelectomy to those of sperm retrieval/ICSI classic surgical therapy was viewed favorably.2, 5 With such strong opinions and scarce data the ability of couples to intelligently navigate their reproductive options is limited. Therefore, we clarified the cost-effectiveness of management options for vasectomy and varicocele associated infertility using decision analysis modeling. Decision models are constructed with predefined assumptions and they serve as useful tools for estimating outcomes when multiple complex treatments are available. In urology decision analysis modeling has been applied to prostate cancer treatment to clarify the critical variables involved in treatment decisions.6 In this study we applied formal decision analyses to calculate the cost of pregnancy for initial surgical or initial ART treatment in men with infertility due to varicocele and vasectomy. MATERIALS AND METHODS
Our analyses addressed the initial decision of whether to intervene surgically or use ART in cases of vasectomy reversal or varicocelectomy. Two algorithms were created to model treatment decisions and their outcomes for each urological
Submitted for publication December 30, 2004. * Correspondence: Department of Urology, University of California San Francisco, 1600 Divisadero St., Room A633, San Francisco, California 94115-1695. 1926
SURGERY OR ASSISTED REPRODUCTION DECISION ANALYSIS
1927
FIG. 1. Treatment algorithm for pregnancy after vasectomy. Decision is choice between initial surgical reconstruction and ART. IC, intercourse.
condition (figs. 1 and 2). For the purpose of analysis we included intrauterine insemination (IUI) along with sperm retrieval/ICSI as a component of ART treatment. Decision modeling for vasectomy and varicocele associated infertility. In a patient seeking fertility after vasectomy the first decision is the choice between vasectomy reversal and ART (fig. 1). In this decision model the cost per pregnancy was the end point for comparing the outcomes of the 2 branches. This was calculated using the cumulative costs of each outcome state in the specific branch and dividing this value by the number of pregnancies achieved in that branch. Several important assumptions were made in the model, namely 1) fecundity depended on female age but was independent of treatment modality, 2) pregnancy rates were stable and did not decrease with subsequent treatments and 3) the fraction of couples that pursued additional therapy after initial treatment failure was the same in each decision algorithm arm. A similar approach was used to create the decision model in men with varicocele associated infertility. The first decision was the choice between varicocelectomy and ART (fig. 2). Again, the cost per pregnancy was the end point for comparing the outcomes of the 2 branches of the model. Assumptions in this model were 1) fecundity depended on female age and not on treatment modality, 2) sperm retrieval/ICSI would be used for patients with a total motile sperm count (TMC) (calculated as volume ⫻ concentration ⫻ motile fraction) of less than 10 million sperm and 3) IUI would be used in couples with a TMC of greater than 10 million sperm. The decision models were analyzed using Decision Analyses, Data 2.5 computer software (TreeAge, Williamstown, Massachusetts). The cost per pregnancy and the pregnancy rate in each arm were calculated and compared. The more cost-effective method for treating vasectomy and varicocele associated infertility was the method that yielded the lower cost per pregnancy. The value at which variation in a parameter altered the outcome of the decision analysis, ie a lower cost per pregnancy, was termed the threshold value. Probability and utility data for decision modeling. In the
vasectomy reversal model data were gathered from the literature regarding vasectomy reversal patency and pregnancy rates after IUI and sperm retrieval/ICSI.5, 7–9 Table 1 lists the derived probability values. For the purpose of initial decision modeling we chose reasonable success rates but not necessarily the highest published rates from the literature. Thus, the vasectomy reversal pregnancy and patency rates were simplified and estimated to be 30% and 80%, respec-
FIG. 2. Treatment algorithm for pregnancy with male factor infertility due to varicocele. Decision is choice between surgical varicocelectomy and ART.
1928
SURGERY OR ASSISTED REPRODUCTION DECISION ANALYSIS
TABLE 1. Derived probabilities for vasectomy reversal and pregnancy rates used in decision model Variable
% Estimated Probability
Patency: Bilat vasovasostomy Vasovasostomy/vasoepididymostomy Bilat vasoepididymostomy Pregnancy rate: After vasectomy reversal After 1 ICSI cycle After 4 IUI cycles Second treatment sought
87 70 65 30 30 32 50
TABLE 2. Institutional intervention costs Procedure
Cost (US $)
Microsurgical vasovasostomy Microsurgical varicocelectomy 1 IVF-ICSI cycle 1 IUI cycle
10,000 4,500 10,000 500
TABLE 3. Pregnancy rate after varicocelectomy10 Preop TMC (1 ⫻ 106)
No.
% Pregnancy Rate
0–1.5 1.5–5 5–10 10–20 20–40 Greater than 40
132 62 58 68 60 65
16.5 30.6 36.2 39.7 58.3 61.5
445
36.6
Total
tively. In addition, the fraction of couples pursuing additional therapy after primary treatment failure was assumed to be 50%. Table 2 lists the cost of each treatment using data from estimated total charges at our institution. In the varicocele model we applied published institutional data to generate probability values.10 Table 3 lists pregnancy rates in the varicocelectomy cohort stratified by preoperative TMC. Of all patients 36.6% of couples achieved spontaneous pregnancy after surgery. The pregnancy rate for a single ICSI cycle and 4 IUI cycles was assumed to be 30% and 32%, respectively.11, 12 RESULTS
Post-vasectomy infertility. Figure 1 shows the decision model comparing vasectomy reversal and sperm retrieval/ ICSI. Table 4 shows the calculated cost per pregnancy for vasectomy reversal and sperm retrieval/ICSI for reversal patency rates between 40% and 100%, and female fecundity
rates between 10% and 50%. For each cell in the table the calculated cost per pregnancy for the specific intervention is listed for direct comparison. The baseline decision analysis favored initial vasectomy reversal rather than sperm retrieval/ICSI if bilateral vasovasostomy was performed at a cost per pregnancy of $38,983 vs $39,506, which corresponds to column 4 (80% patency) and row 3 (30% pregnancy rate) (table 4). If a 100% patency rate were achievable, this approach would be uniformly favored over sperm retrieval/ICSI across all pregnancy rates. However, sperm retrieval/ICSI was initially more cost-effective if surgical reconstruction required unilateral or bilateral vasoepididymostomy due to lower associated patency rates (columns to the left of 80% reversal patency) (table 4). Figure 3, A represents 1-way sensitivity analysis in which costs were compared for a range of reversal patency rates while holding the pregnancy rate constant at 40%. In this scenario vasectomy reversal was more cost-effective than sperm retrieval/ICSI when the patency rate was greater than 79%, the threshold value. Similarly figure 3, B shows 1-way sensitivity analysis of the pregnancy rate when the patency rate was held constant at 60%. There is no threshold value in this graph since sperm retrieval/ICSI was always more cost-effective in this scenario with lower patency rates. This information was summarized in the 2-way sensitivity analysis, in which reversal patency and pregnancy rates were varied (fig. 3, C). In this analysis vasectomy reversal was more cost-effective when the patency rate was greater than 78%, independent of female fecundity estimates (ie the pregnancy rate). Varicocele associated infertility. Figure 2 shows the decision model comparing initial varicocele ligation and ART for varicocele associated infertility. Table 5 shows the cost per pregnancy of surgical intervention and ART in the entire group and the cost by preoperative TMC. Overall initial surgical repair was more cost-effective than ART. However, IUI yielded a lower cost per pregnancy than varicocelectomy in men with a preoperative TMC of between 10 and 20 million sperm ($9,000 vs $11,333, respectively). In men with lower sperm counts (TMC less than 10 million) who would qualify for sperm retrieval/ICSI instead of IUI we compared the cost-effectiveness of sperm retrieval/ ICSI and varicocele repair after varying the postvaricocelectomy pregnancy rate. In these men due to the high cost of sperm retrieval/ICSI varicocelectomy was the more cost-effective treatment when the postoperative pregnancy rate was greater than the 14% threshold value (fig. 4, A). In men with higher sperm counts (TMC greater than 10 million) who might qualify for IUI instead of sperm retrieval/ICSI varicocelectomy was more cost-effective than IUI only when the pregnancy rate was greater than the 45% threshold value (fig. 4, B).
TABLE 4. Cost per pregnancy for ART and vasovasostomy by pregnancy and patency rates % Pregnancy Rate 10: Reversal ICSI 20: Reversal ICSI 30: Reversal ICSI 40: Reversal ICSI 50: Reversal ICSI
40% Reversal Patency (US $)
60% Reversal Patency (US $)
80% Reversal Patency (US $)
100% Reversal Patency (US $)
168,182 117,241
137,383 117,241
115,873 117,241
100,000 117,241
84,884 58,929
69,231 58,929
58,197 58,929
50,000 58,929
57,143 39,506
46,535 39,506
38,983 39,506
33,333 39,506
43,293 29,808
35,204 29,808
29,386 29,808
25,000 29,808
35,000 24,000
28,421 24,000
23,636 24,000
20,000 24,000
SURGERY OR ASSISTED REPRODUCTION DECISION ANALYSIS
1929
FIG. 3. Sensitivity analysis. A, comparison of vasectomy reversal patency costs for range of reversal patency rates when pregnancy rate was constant at 40%. B, 1-way analysis of pregnancy rate in vasectomy associated male infertility with vasectomy reversal patency constant at 60%. C, 2-way analysis of pregnancy rate and reversal patency in vasectomy associated male infertility comparing cost-effectiveness of surgical reconstruction and ICSI when varying reversal patency and pregnancy rates. Shaded area indicates most cost-effective modality.
TABLE 5. Cost per pregnancy of male infertility due to varicocele Preop TMC (1 ⫻ 106) 0–1.5 1.5–5 5–10 10–20 20–40 Greater than 40
Cost/Pregnancy (US $) Varicocelectomy
ART
28,286 14,684 12,429 11,333 7,714 7,313
33,333 33,333 33,333 9,000 9,000 9,000
DISCUSSION
The success of ART for treating female and male infertility has led to its empirical application in cases of surgically treatable diagnoses in male infertility. In the last decade several concerns regarding ART treatment have been realized. 1) It is expensive.13–15 2) It has associated complications. 3) It has implications for the health of offspring, including low birth weight, prematurity, chromosomal abnormalities and an increased incidence of birth defects.16, 17 Surprisingly despite its widespread application, to our knowledge no published study has documented that ART is more cost-effective than classic urological treatments for male infertility. In this study we further clarified the costeffectiveness issues that surround the management options for vasectomy and varicocele associated infertility through decision analysis modeling. Our decision analysis suggests that vasectomy reversal is more cost-effective than sperm retrieval and IVF-ICSI. Moreover, other observations became apparent as variables in the model were altered to determine the effect on outcomes. A threshold patency rate of 78% was identified below which sperm retrieval/ICSI was more cost-effective than vasectomy reversal. This benchmark value for the reversal patency rate implies that vasectomy reversal surgery must be performed
at a high standard of quality to be more cost-effective than sperm retrieval/ICSI. In general vasectomy reversal is more likely to be cost-effective in men with a shorter time from vasectomy, in whom bilateral vasovasostomy can typically be performed. Conversely men with a longer interval and greater potential for requiring vasoepididymostomy may benefit from initial sperm retrieval/ICSI. In addition, the cost-effectiveness of vasectomy reversal treatment depends primarily on male and not on female fertility factors. Although it is true that female fecundity alters the actual costs of surgical reconstruction and ART, the relative costs remain unchanged. Varying reversal patency rates dramatically altered cost-effectiveness outcomes in the model. This implies that despite decreased pregnancy rates in couples with an older female partner vasectomy reversal can remain the more cost-effective approach if the patency rate exceeds the threshold value. This concept is substantiated by a review of the literature. In a series of 29 patients with vasectomy reversal Deck and Berger found a lower cost per newborn with vasectomy reversal compared with sperm retrieval/ICSI at a median female age of 40 years (range 38 to 48).18 Interestingly the patency rate achieved in that series was greater than 90%, which is a value consistent with vasectomy reversal cost-effectiveness in our decision model. The decision model provides a framework for the individual surgeon to counsel patients. Although efficacy data are available regarding postoperative patency rates after vasectomy reversal from the multicenter Vasovasostomy Study Group,7 this information reflects the practice of experienced and dedicated microsurgeons and not necessarily that of the individual community based surgeon. Conveniently this decision analysis model can be tailored to be accurate for an individual surgical practice by simply including data from the vasectomy reversal experience of that surgeon. In addition, costs vary greatly among individual practices and geographic locations, and the availability of the models permits
1930
SURGERY OR ASSISTED REPRODUCTION DECISION ANALYSIS
FIG. 4. A, cost-effectiveness of varicocelectomy vs IVF-ICSI in men with varicocele associated male infertility and TMC less than 10 million at various post-varicocelectomy pregnancy rates. B, sensitivity analysis in men with varicocele associated infertility with TMC greater than 10 million comparing cost of varicocele and ART with IUI across pregnancy rates.
customization of the calculations and comparisons at a particular center. We are in the process of translating the decision models to web based tools, similar to currently popular nomograms and models providing outcome estimates for individuals with prostate cancer, allowing one to input the costs and estimated success of interventions with the resulting calculation of the cost per pregnancy and relative costeffectiveness. The second decision model was created to address the cost-effectiveness of varicocele associated infertility treatment. In a previous study Schlegel compared sperm retrieval/ ICSI and varicocelectomy pregnancy rates, and concluded that primary treatment with varicocele repair was more costeffective than ART, while providing comparable delivery rates.2 In our study we refined their cost-benefit analysis by 1) applying decision modeling, 2) stratifying the degree of semen impairment according to motile sperm counts and 3) including IUI as a viable treatment option in the subset of men with varicocele associated infertility. We observed that surgical treatment is more cost-effective than ART in certain cases of varicocele associated infertility. If in a surgeon’s experience men with a less severely impaired preoperative motile sperm count (greater than 10 million motile sperm) cannot achieve a 45% pregnancy rate after surgery, ART can be more cost-effective. This is largely due to the relatively inexpensive IUI procedure for which such couples may qualify. On the contrary, if a surgeon can achieve a greater than 14% pregnancy rate in men with more severe impairment in semen quality (less than 10 million motile sperm), varicocele repair is more cost-effective since these couples generally require sperm retrieval/ICSI instead of IUI to overcome infertility.19
Decision analysis models have inherent limitations. In our modeling we did not consider factors such as the number of children desired, the social and monetary costs of intervention in the female partner, the cost of time lost from work or the costs attributable to procedural complications and multiple gestations. However, in our estimation incorporating these variables would likely support the argument for initial surgical treatment rather than ART for the 2 types of male factor infertility. Other intangible elements, such as the potential for future natural conception, favor surgical approaches to treatment. Conversely our model did not incorporate the complex issue of the effects of age on maternal reproductive potential. Inclusion of this variable into the model would likely favor ART over surgery in a higher proportion of affected couples because the mean time to pregnancy after vasectomy reversal and varicocelectomy (12 and 8 months, respectively)7, 20 may be longer than that needed for a pregnancy with ART. Overall the quality of our analysis is supported by the fact that the calculated cost per pregnancy is reasonable in each model and they are within the range of values that have been published for these treatments.1, 5, 19 The information provided by these decision models serves as a framework for future analyses and discussion. Further refinements may account for more accurate assessment of variables, such as female fecundity and improvements in IVF-ICSI. In addition, customization of the decision analysis can address concerns of differences in costs and estimations of treatment efficacy. These data provide important information for all parties involved in the treatment of reproductive impairments, including physicians, infertile couples and insurance carriers, of whom each has specific concerns and desired end points. CONCLUSIONS
Formal decision analysis provides a structured means of assessing treatment choices in male factor infertility. According to models constructed for varicocele and vasectomy related male infertility surgery is often more cost-effective than ART. However, threshold values for surgical success rates are needed to maintain the cost-effectiveness of urological surgery over ART and provide a form of quality standard for the practicing surgeon. Formal decision modeling can also serve as a web based instrument for comparing treatment costs for individual urologists using practice specific success rates and, thereby, it can more accurately present the complex treatment decisions facing couples seeking infertility care. REFERENCES
1. Pavlovich, C. P. and Schlegel, P. N.: Fertility options after vasectomy: a cost-effectiveness analysis. Fertil Steril, 67: 133, 1997 2. Schlegel, P. N.: Is assisted reproduction the optimal treatment for varicocele-associated male infertility? A cost-effectiveness analysis. Urology, 49: 83, 1997 3. Evers, J. L. and Collins, J. A.: Surgery or embolisation for varicocele in subfertile men. Cochrane Database Syst Rev, 3: CD000479, 2004 4. Madgar, I., Weissenberg, R., Lunenfeld, B., Karasik, A. and Goldwasser, B.: Controlled trial of high spermatic vein ligation for varicocele in infertile men. Fertil Steril, 63: 120, 1995 5. Kolettis, P. N. and Thomas, A. J., Jr.: Vasoepididymostomy for vasectomy reversal: a critical assessment in the era of intracytoplasmic sperm injection. J Urol, 158: 467, 1997 6. Simpson, K. N.: Problems and perspectives on the use of decision-analysis models for prostate cancer. J Urol, 152: 1888, 1994 7. Belker, A. M., Thomas, A., Jr., Fuchs, E. F., Konnak, J. W. and Sharlip, I. D.: Results of 1,469 microsurgical vasectomy reversals by the Vasovasostomy Study Group. J Urol Nurs, 11: 93, 1992
SURGERY OR ASSISTED REPRODUCTION DECISION ANALYSIS 8. Kolettis, P. N., Sabanegh, E. S., Nalesnik, J. G., D’Amico, A. M., Box, L. C. and Burns, J. R.: Pregnancy outcomes after vasectomy reversal for female partners 35 years old or older. J Urol, 169: 2250, 2003 9. Nalesnik, J. G. and Sabanegh, E. S., Jr.: Vasovasostomy: multiple children and long-term pregnancy rates. Curr Surg, 60: 348, 2003 ¨ ¸ 10. Cayan, S., Erdemir, F., Ozbey, I., Turek, P. J., Kadiog˘lu, A. and Tellalog˘lu, S.: Can varicocelectomy significantly change the way couples use assisted reproductive technologies? J Urol, 167: 1749, 2002 11. Guzick, D. S., Carson, S. A., Coutifaris, C., Overstreet, J. W., Factor-Litvak, P., Steinkampf, M. P. et al: Efficacy of superovulation and intrauterine insemination in the treatment of infertility. National Cooperative Reproductive Medicine Network. N Engl J Med, 340: 177, 1999 12. Jee, B. C., Ku, S. Y., Suh, C. S., Choi, Y. M., Kim, J. G., Moon, S. Y. and Kim, S. H.: Cumulative ongoing pregnancy rate in intracytoplasmic sperm injection cycles. J Obstet Gynaecol Res, 30: 372, 2004 13. Donovan, J. F., Jr., DiBaise, M., Sparks, A. E., Kessler, J. and Sandlow, J. I.: Comparison of microscopic epididymal sperm aspiration and intracytoplasmic sperm injection/in-vitro fertilization with repeat microscopic reconstruction following vasectomy: is second attempt vas reversal worth the effort? Hum Reprod, 13: 387, 1998 14. Van Voorhis, B. J., Sparks, A. E., Allen, B. D., Stovall, D. W., Syrop, C. H. and Chapler, F. K.: Cost-effectiveness of infertility treatments: a cohort study. Fertil Steril, 67: 830, 1997 15. Neumann, P. J., Gharib, S. D. and Weinstein, M. C.: The cost of a successful delivery with in vitro fertilization. N Engl J Med, 331: 239, 1994 16. Hansen, M., Kurinczuk, J. J., Bower, C. and Webb, S.: The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization. N Engl J Med, 346: 725, 2002 17. Schieve, L. A., Meikle, S. F., Ferre, C., Peterson, H. B., Jeng, G. and Wilcox, L. S.: Low and very low birth weight in infants conceived with use of assisted reproductive technology. N Engl J Med, 346: 731, 2002 18. Deck, A. J. and Berger, R. E.: Should vasectomy reversal be performed in men with older female partners? J Urol, 163: 105, 2000 19. Garceau, L., Henderson, J., Davis, L. J., Petrou, S., Henderson, L. R., McVeigh, E. et al: Economic implications of assisted reproductive techniques: a systematic review. Hum Reprod, 17: 3090, 2002 20. Cayan, S., Lee, D., Black, L. D., Reijo Pera, R. A. and Turek, P. J.: Response to varicocelectomy in oligospermic men with and without defined genetic infertility. Urology, 57: 530, 2001 EDITORIAL COMMENT There are many different ways to evaluate the value of a treatment. An approach that takes into account the cost as well as the outcomes is a cost-effectiveness evaluation. As with most models, the devil of such an evaluation is in the details. How are the data derived? What assumptions were made? These are critical points to help determine whether the modeled data have any value to another practitioner. Cost-effectiveness is of particular value for determining the role of a fertility treatment because patients may bear more of
1931
the cost of such medical care out of pocket, since insurers may not cover such treatments or may provide only partial insurance coverage. Prior reports have used national (references 1, 2, 15 and 19 in article), and local/single institution costs and results (references 5 and 14 in article). Various costs can be included beyond the direct costs of medical care. The resulting costs of multiple gestations is important to consider. Although twins appear to be a blessing for the couple, the arrival of triplets almost always engenders prematurity, which can result in substantive risks for the child (cerebral palsy and blindness) and massive costs for the health care system (reference 15 in article.) Such costs were not considered in this analysis, although 17% to 54% of pregnancies after IVF cycles at the University of California-San Francisco center are multiple gestations.1 Of interest is that the success rate for IVF/ICSI quoted for the University of California-San Francisco center is 30% in the report, whereas the Centers for Disease Control website quotes a 28.5% delivery rate per IVF attempt in women 40 years younger,1 suggesting that the data in the article may overstate the relative benefits of assisted reproduction. The article provides a guide and an example from 1 center of how to evaluate the relative value of specific male fertility treatment with varicocelectomy and vasectomy reversal. Readers must consider the costs and specific pregnancy results obtained at their center to apply this information to the clinical care of patients. Peter N. Schlegel Department of Urology Weill Medical College of Cornell University New York Presbyterian Hospital New York, New York 1. 2002 Assisted Reproductive Technology (ART) Report: 2002 Fertility Clinic Report by state: Clinic Data. Department of Health and Human Services, Centers for Disease Control and Prevention. Available at http://apps.nccd.cdc.gov/ART2002/ clindata02.asp?Location⫽60. REPLY BY AUTHORS Part of the strength and usefulness of decision analytic models is that while disagreement may exist on a specific cost or pregnancy rate, these values are merely used as the starting point for analysis. The true value of decision models is the ability to change the numerous variables, either one at a time or simultaneously, and determine the impact on outcomes. Thus, quibbling about overstating outcomes (30% vs 28.5%) can be easily addressed using sensitivity analyses, as explicitly shown in figure 3 in the article. We appreciate the comments and acknowledge that our models, and decision analytic methods in general, have limitations. These assumptions and simplifications are clearly stated, and we discuss their implications overall, the issues raised (eg multiple gestations) only increase the costs of ART and further suggest that classic, surgical treatments of male factor infertility are cost-effective approaches. Ultimately, the usefulness of our cost analysis can be improved by incorporating a couple’s specific situation, using more accurate estimates of all variables including fecundity and hospital/physician charges. This individualized assessment of relative costs may help guide patients as well as physicians during the decision making process.