CONTROVERSY: INSURANCE COVERAGE

Transcription

1 CONTROVERSY: INSURANCE COVERAGE FERTILITY AND STERILITY VOL. 80, NO. 1, JULY 2003 Copyright 2003 American Society for Reproductive Medicine Published by Elsevier Inc. Printed on acid-free paper in U.S.A. Does insurance coverage decrease the risk for multiple births associated with assisted reproductive technology? Meredith A. Reynolds, Ph.D., a Laura A. Schieve, Ph.D., a Gary Jeng, Ph.D., a and Herbert B. Peterson, M.D. b Centers for Disease Control and Prevention, Atlanta, Georgia Received September 3, 2002; revised and accepted December 2, Reprint requests: Meredith A. Reynolds, Ph.D., Division of Reproductive Health, CDC Mailstop K- 34, 4770 Buford Highway NE, Atlanta, Georgia (FAX: ; mtr6@cdc.gov). a Division of Reproductive Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention. b Department of Reproductive Health and Research, World Health Organization, Geneva, Switzerland /03/$30.00 doi: /s (03) Objective: To determine whether insurance coverage for ART is associated with transfer of fewer embryos and decreased risk of multiple births. Design: Retrospective cohort study of a population-based sample of IVF procedures performed in six U.S. states during Setting: Three states with mandated insurance coverage (Illinois, Massachusetts, and Rhode Island) and three states without coverage (Indiana, Michigan, and New Jersey). Participant(s): Seven thousand, five hundred sixty-one IVF transfer procedures in patients 35 years of age. Main Outcome Measure(s): Number of embryos transferred, multiple-birth rate, triplet or higher order birth rate, and triplet or higher order gestation rate. Result(s): A smaller proportion of procedures included transfer of three or more embryos in Massachusetts (64%) and Rhode Island (74%) than in the noninsurance states (82%). The multiple-birth rate in Massachusetts (38%) was less than in the noninsurance states (43%). The insurance states all had protective odds ratios for triplet or higher order births, but only the odds ratio (0.2) for Massachusetts was significant. This decreased risk in Massachusetts resulted from several factors, including a smaller proportion of patients with three or more embryos transferred, lower implantation rates when three or more embryos were transferred, and greater rates of fetal loss among triplet or higher order gestations. Conclusion(s): Insurance appears to affect embryo transfer practices. Whether this translates into decreased multiple birth risk is less clear. (Fertil Steril 2003;80: by American Society for Reproductive Medicine.) Key Words: In vitro fertilization, embryo transfer, multiple birth offspring, pregnancy, multiple, insurance Whether insurance coverage for infertility diagnosis and treatment should be mandated has been debated at both federal and state levels. Although no federal law currently mandates insurers to cover infertility services, multiple bills requiring coverage were introduced into the U.S. House of Representatives and the Senate during the 107th Congress (1 4). In addition, 14 states have enacted some type of infertility insurance law (5). Several of these laws cover some form of ART, including IVF. The debate on mandating coverage for infertility services has been argued from multiple legal, ethical, and cost benefit perspectives (6 13). One argument for mandated coverage posits that the presence of insurance will lower the high risk for multiple birth currently associated with ART and thereby reduce concomitant risks for adverse health consequences among women and infants. A reduction in multiple births and their sequelae would, in turn, yield economic savings to insurance companies. Multiple gestations and births are arguably one of the most important public health issues associated with use of ART. Increased use of infertility treatments, and specifically the widespread practice of transferring multiple embryos during ART treatments, has led to significant increases in triplet and higher order multiple-birth rates in the United States (14, 15). Multiple births are associated with increased infant mortality and morbidity (e.g., prematurity and low birth weight) and long- 16

2 term disability among survivors and lead to increased pregnancy and delivery complications (16 22), all of which increase health care costs. Projected estimates of the health care costs per family of multiple-birth deliveries conceived by ART in 2000 have ranged from $58,865 for twins to $281,698 for quadruplets (23). Goldfarb et al. (24) estimated the cost per woman delivered of pregnancies from IVF to be approximately $39,000 for singleton or twin pregnancies but about $340,000 for triplet and quadruplet pregnancies. State laws mandating insurance coverage of infertility treatments that include ART typically do not limit the number of embryos that may be transferred in a given procedure. Nonetheless, many believe that having insurance is associated with transfer of fewer embryos and thus with a decreased risk of multiple-gestation pregnancy and births. The assumption is that patients who pay out of pocket for ART are less willing to risk a failed procedure and are therefore more likely to elect to transfer a higher number of embryos so as to maximize their chances of a live birth (25 27). Little evidence is available to support the premise that insurance coverage for ART will result in a reduction in higher-order embryo transfers (i.e., more than 2 embryos) and lower risk of multiple birth. Two studies have attempted to evaluate this premise by using published summary data from the Centers for Disease Control and Prevention s annual ART Success Rate Reports (28, 29). The limited, cliniclevel data on mean number of embryos transferred and multiple birth and gestation rates published in these reports substantially constrained the analyses and precluded any assessment of important confounding variables. In addition, both studies combined the data for the insurance states as one group, thus potentially masking important differences in results among these states. We used data from the complete population-based surveillance dataset of ART procedures performed in the United States during 1998 to determine whether insurance coverage was associated with embryo transfer practices and therefore affected the risk of multiple birth. MATERIALS AND METHODS The Fertility Clinic Success Rate and Certification Act of 1992 (30) mandates every clinic in the United States that performs ART procedures to report data related to its pregnancy success rates annually to the Centers for Disease Control and Prevention. Each year, the Society for Assisted Reproductive Technology (SART) creates a database of ART procedures performed in U.S. clinics and, per contract, shares these data with Centers for Disease Control and Prevention. Selected national and clinic-level summary statistics from this database are published annually. A more detailed description of this database has been published (31). We estimate that the annual database includes 95% of ART procedures performed in the U.S. each year. For the current study, we selected procedures initiated in We further selected procedures performed in the three states with the most comprehensive infertility insurance laws that year (Illinois, Massachusetts, and Rhode Island) and in three states with no insurance mandate (Indiana, Michigan, and New Jersey). The latter states were chosen on the basis of several factors, including their geographical proximity to the three insurance states, completeness of residency data, proportion of procedures performed in in-state residents, and overall number of procedures performed. Because ART encompasses heterogeneous treatments, we limited this analysis to IVF procedures using fresh embryos derived from the patients own eggs. Thus, excluded procedures included those that involved eggs or embryos donated from another woman, use of a gestational surrogate to carry the pregnancy, or transfer of eggs or embryos into the fallopian tubes rather than the uterus. In addition, we restricted our analyses to patients 35 years of age or younger, because these women are at higher risk of multiple birth and have the highest overall success rates. The final sample consisted of 7,561 IVF procedures: 1,814 in Illinois (21 clinics), 2,369 in Massachusetts (9 clinics), 370 in Rhode Island (1 clinic), and 3,008 in the noninsurance states (38 clinics). By 1998, the insurance mandates of Illinois, Massachusetts, and Rhode Island had been in effect for at least 7 years. These three states require all insurance policies providing pregnancy-related benefits to cover comprehensive diagnosis and treatment of infertility, including IVF. Each state defines infertility as the inability to conceive during a 1-year period. Illinois limits the number of ART procedures covered to four, with an additional two covered if a live birth results from a procedure. Illinois also exempts businesses with 25 or fewer employees from providing coverage and exempts religious organizations from covering procedures that violate their teachings and beliefs. The Rhode Island law covers only married persons and limits copayment to 20%. The Massachusetts law is the most comprehensive, with no unique exclusions. We evaluated four main outcomes: number of embryos transferred, proportion of live births that were multiple, proportion of live births that were triplet or higher order, and proportion of pregnancies that were triplet or higher order. We examined several measures of embryos transferred average number transferred, proportion of procedures with three or more embryos transferred, and proportion of procedures with four or more embryos transferred. Because relative comparisons were similar for each of these measures, we focus this report on the proportion of procedures with three or more embryos transferred. Research indicates that this outcome is probably the most meaningful practice indicator for evaluating multiple-birth risk among patients younger than 36 years of age (30). FERTILITY & STERILITY 17

3 Multiple birth was defined as a delivery of two or more infants in which at least one was live born. Triplet or higher order live-birth deliveries (i.e., delivery of three or more infants in which at least one was live born) were considered separately. Because some triplet or higher order live births are avoided due to spontaneous and medical fetal reductions of triplet or higher order gestations resulting in the delivery of singletons or twins, we also examined the percentage of triplet or higher order pregnancies as the total potential for triplet or higher order births. Triplet or higher order pregnancies were defined as three or more fetal hearts on sonography. Differences between the insurance states and the noninsurance states were evaluated by using 2 tests. Because the pattern of results was not consistent among the three insurance states, findings are presented separately by insurance state (Illinois, Massachusetts, and Rhode Island). The three noninsurance states, however, were grouped for reference. Because the three insurance states differed from the noninsurance states in background characteristics, we conducted stratified and multivariable logistic regressions to control for confounding and to investigate effect modification between insurance status and the following variables: patient age in years, infertility diagnosis, previous ART procedure, previous live birth, cryopreservation of nonused embryos (a proxy measure of embryo availability), and two IVF procedural techniques (use of ICSI and assisted hatching). Using the state in which ART was performed as an indirect measure of patients insurance coverage is complicated by the fact that some patients undergo ART in clinics outside of their state of residence. Therefore, we took available residency data into account in examining our findings. Because the percentage of missing data for patient residence was high in some states, we conducted sensitivity analyses for all of our statistics. First, we considered the total number of procedures regardless of residency (n 7,561). We then limited analyses to known in-state residents only (n 5,091), and finally, we performed analyses including both known in-state residents as well as patients with missing residency data who were reclassified as in-state residents (n 1,654). We present bivariable analyses for these three samples. Our final logistic regression models are presented for known in-state residents only, although findings were similar for all three samples. The known in-state resident sample provides the best test for assessing whether insurance coverage is associated with embryo transfer practices and multiple birth risk, because this sample comprises the patients most likely to be covered by insurance if they resided in an insurance state and the patients least likely to be covered by insurance if they resided in a noninsurance state. However, if practice patterns of ART are influenced by insurance coverage, one might argue that such an effect would occur globally, in that practitioners would not develop different practice patterns for their patients with and without insurance coverage. Thus, we considered both procedures on in-state residents and total procedures performed in a state in our analyses. Sample sizes for the logistic regression models varied depending on the outcome. The first model included transfer of at least three embryos (yes/no) as the dependent variable and included all 5,091 IVF transfer procedures of known in-state residents. The second and third models examined multiple and triplet or higher order live-birth deliveries (yes/ no), respectively, as the dependent variable and included the 1,850 IVF procedures that had resulted in a live birth. The fourth model examined triplet or higher order gestations (yes/no) for all IVF procedures that had resulted in a pregnancy (n 2,183). Finally, since the goal of all IVF procedures is a live birth, we conducted a fifth model for all IVF transfer procedures (n 5,091) to determine whether insurance affected the likelihood of a live birth (yes/no). Differences between insurance and noninsurance states in number of embryos transferred would be expected to result in differences in multiple birth rates. However, other factors affecting multiple-birth risk may also differ systematically between insurance and noninsurance states. Insurance and noninsurance states may differ in the types of patients who typically receive higher-order embryo transfers. Similarly, patients from insurance and noninsurance states may have differing rates of fetal loss, reflecting differences in patients who receive ART. To investigate these factors, we examined triplet or higher order live-birth rates in more detail by holding number of embryos transferred constant and comparing embryo implantation rates by state. In addition, we compared fetal loss rates among the outcomes of triplet or higher order gestations. This study was approved by the Institutional Review Board at the Centers for Disease Control and Prevention. RESULTS Patients in Illinois, Massachusetts, and Rhode Island differed from patients in noninsurance states in terms of several variables (Table 1). Patients in Massachusetts and Rhode Island were somewhat older than those in noninsurance states. Patients in all three insurance states were more likely to have undergone a previous ART procedure and to have had a previous pregnancy, and those in Illinois and Massachusetts were more likely to have had a previous live birth. Although various differences were found in terms of infertility diagnosis, female factor (tubal factor, endometriosis, ovulatory dysfunction, or uterine factor) was always the most common diagnosis, followed by male factor and unexplained infertility. Compared with the noninsurance states, procedures in Massachusetts and Rhode Island were less likely to include ICSI and assisted hatching, but those in 18 Reynolds et al. Insurance and IVF multiple birth risk Vol. 80, No. 1, July 2003

4 TABLE 1 Characteristics of in vitro fertilization procedures performed on women 35 years of age or younger in Percent distribution at start of case IVF procedure Characteristic Illinois (n 1,814) Massachusetts (n 2,369) Rhode Island (n 370) Noninsurance states (n 3,008) Maternal age (y) a 70 b 64 Previous pregnancy 45 a 48 a 49 a 39 Previous live births 19 c 19 c Previous ART procedures 47 a 48 a 55 a 34 Diagnosis Female factor a 49 a 62 Male factor 26 c b 29 Unexplained 12 a 18 a 15 a 9 Use of ICSI 63 a 36 a 44 c 49 Use of assisted hatching 57 a 25 a 6 a 51 Cryopreservation of extra embryos 31.2 a 30.5 a 48.9 a 37.1 Residency data complete 80 a 62 a 100 a 88 In-state resident d 96 a 89 a 74 b 81 a P.001 compared with noninsurance states. P.01 compared with noninsurance states. P.05 compared with noninsurance states. Among those for whom residency data were available. Reynolds. Insurance and IVF multiple birth risk. Fertil Steril Illinois were more likely to include these techniques. Massachusetts and Illinois had lower rates of cryopreservation of extra, nontransferred embryos compared with the noninsurance states. Completeness of residency data varied, but among procedures for which residency status was known, most were done for in-state residents. Table 2 shows the proportion of procedures that involved transfer of at least three embryos. Overall, these percentages were lower in Massachusetts and Rhode Island (66.4% and 73.5%) than in the noninsurance states (81.6%). Similar results were obtained when analyses were limited to known in-state residents and when those with missing residency were reclassified as in-state residents. Table 2 also shows the differences in overall multiplebirth and triplet or higher order births and gestations by insurance status. The proportion of live-birth deliveries that were multiple births was lower in Massachusetts than in the noninsurance states, but this difference reached significance only for known in-state residents. The proportion of livebirth deliveries and of pregnancies that were triplet or higher order was significantly lower in Massachusetts than in the noninsurance states. In contrast, the rates of multiple births, triplet or higher order births, and triplet or higher order gestations in Illinois and Rhode Island were similar to those in the noninsurance states. Findings were similar when analyses were limited to known in-state residents and when patients with unknown residence were recoded as in-state residents. Table 3 shows results from logistic regression models for procedures performed for known in-state residents. Compared with procedures in noninsurance states, procedures in Massachusetts (odds ratio, 0.4) and Rhode Island (odds ratio, 0.6) were less likely to include transfer of three or more embryos. For multiple births, only patients in Massachusetts (odds ratio, 0.7) were at significantly lower risk. All three insurance states had protective odds ratios for triplet or higher order multiple births, although this finding reached significance only in Massachusetts (odds ratio, 0.2). All three insurance states had a decreased chance of a live birth, but the effect was least pronounced for the Massachusetts sample. Because Massachusetts was the only insurance state to have a protective odds ratio for higher order embryo transfers, multiple births, triplet or higher order births, and triplet or higher order gestations, we further compared this state to the noninsurance states (Table 4). We sought to determine whether the factors underlying the state s lower risk for multiple births and gestations extended beyond the fact that a smaller proportion of procedures performed there involved transfer of three or more embryos. When the number of embryos transferred was held constant, Massachusetts pa- FERTILITY & STERILITY 19

5 TABLE 2 Outcome variables, by insurance status. Sample No. of transfers Transfer 3 embryos (%) No. of live-birth deliveries Overall Multiple-birth deliveries (%) Triplet or higher order No. of pregnancies Triplet or higher order Pregnancies s (%) No. of transfers Livebirth deliveries (%) Total sample Illinois a Massachusetts a a a a Rhode Island a a Noninsurance states a In-state residents (known residents) Illinois a Massachusetts a b 2.6 a a c Rhode Island c a Noninsurance states In-state residents (missing recoded as in-state) Illinois a Massachusetts a a a a Rhode Island a a Noninsurance states a P.001 compared with noninsurance states. b P.05 compared with noninsurance states. c P.01 compared with noninsurance states. Reynolds. Insurance and IVF multiple birth risk. Fertil Steril tients continued to have a lower triplet or higher order live-birth rate. For example, when three embryos were transferred, the triplet or higher order birth rate was 11.9% for the noninsurance states but 3.7% for Massachusetts. Examination of patients who became pregnant after at least three embryos were transferred revealed that those in Massachusetts were less likely to become pregnant with triplets or more (12.7% vs. 17.3%). Furthermore, among patients who were pregnant with triplet or higher order gestations, significantly fewer of these pregnancies progressed to a triplet or higher order live birth for residents of Massachusetts than for residents of noninsurance states (25.0% vs. 60.1%). The rates of complete and partial pregnancy loss among these women were significantly higher in Massachusetts than in noninsurance states (18.8% vs. 8.8% for complete and 56.2% vs. 31.1% for partial). For partial pregnancy losses, we could not differentiate between spontaneous and medical fetal reductions. TABLE 3 Odds of transferring at least three embryos, multiple birth, triplet or higher order live birth, triplet or higher order pregnancy, and live-birth delivery. a Location 3 embryos Multiple birth Triplet or higher order live birth Triplet or higher order pregnancy Live birth Noninsurance states Referent Referent Referent Referent Referent Illinois 1.0 ( ) 0.9 ( ) 0.8 ( ) 0.9 ( ) 0.7 ( ) Massachusetts 0.4 ( ) 0.7 ( ) 0.2 ( ) 0.5 ( ) 0.9 ( ) Rhode Island 0.6 ( ) 1.3 ( ) 0.6 ( ) 1.0 ( ) 0.6 ( ) a Adjusted for patient age, embryo availability, prior ART procedure, prior live birth, use of ICSI, use of assisted hatching, cryopreservation of nontransferred embryos, and patient diagnosis. Reynolds. Insurance and IVF multiple birth risk. Fertil Steril Reynolds et al. Insurance and IVF multiple birth risk Vol. 80, No. 1, July 2003

6 TABLE 4 Comparison of Massachusetts and noninsurance states on factors associated with multiple birth risk. Characteristic Massachusetts Noninsurance states P value Triplet or higher order live-birth deliveries (%) 3 embryos transferred embryos transferred embryos transferred NS Pregnancies resulting from transfer of 3 embryos that were triplet or higher order gestation (%) Outcomes of triplet or higher order gestations (%) Complete pregnancy loss Partial pregnancy loss (singleton or twin live birth) Triplet or higher order live birth Note: NS not significant. Reynolds. Insurance and IVF multiple birth risk. Fertil Steril DISCUSSION Overall, our findings suggest that mandated coverage for ART affects embryo transfer practices. Patients younger than 36 years of age undergoing IVF in two of three states with mandated insurance coverage for ART were found to have transfer of fewer embryos than did patients in states without mandated coverage. The negative finding for the third insurance state (Illinois) may indicate that the effect of insurance on reducing the number of embryos transferred is negated when the number of procedures covered is subject to limitations. The findings pertaining to the effect of insurance coverage on reducing risk of multiple births are less conclusive. The proportion of triplet or higher order live births was lower in all three insurance states than in the noninsurance states, but the difference was significant only for Massachusetts. Similarly, the adjusted odds ratios for triplet or higher order live birth were less than 1.0 for all three insurance states, but only Massachusetts reached significance. Of note, the odds ratio for the state that limits the number of procedures covered is closest to 1.0. In addition, the analyses for Rhode Island are limited with respect to statistical power. The results demonstrate that when insurance coverage is associated with reduced risk of multiple birth, several factors may be at play. In Massachusetts, for example, at least three factors seemed to affect this risk. First, patients in Massachusetts tended to have fewer embryos transferred than did women in noninsurance states. Second, when patients in Massachusetts did have higher numbers of embryos ( 3) transferred, they were less likely to become pregnant with triplet or higher order gestations. Finally, the triplet or higher order pregnancies of Massachusetts patients were less likely to result in higher order multiple births because of higher rates of complete and partial pregnancy loss. These differences in implantation rates and fetal losses may reflect important differences in patient selection between Massachusetts and noninsurance states with respect to which patients received higher-order embryo transfers. In Massachusetts, clinicians and patients may be more conservative in transferring higher-order embryos for patients deemed to have a good prognosis. If so, in Massachusetts, patients with a poor prognosis would be disproportionately represented in the higher-order embryo transfer group. We were unable to evaluate patient prognosis based on clinical or laboratory data, such as embryo morphology. Of note, the significantly reduced risk of multiple birth for Massachusetts patients was not accompanied by a large decrease in the chance of a live birth. Although the adjusted odds ratio for a live-birth delivery was reduced among all insurance states, the reduction was small for Massachusetts, as was the difference between overall live-birth rate in Massachusetts and that in the noninsurance states (35% vs. 40%). When considering the decreased chance for a live birth among patients in insurance states overall, one must take into account that these figures reflect per-procedure success rates, not cumulative live-birth rates across multiple procedures. We were unable to link multiple procedures performed on the same patient. In addition, we focused on ART procedures and associated risk for multiple births. We did not have data on non-art infertility treatments, such as controlled ovarian hyperstimulation and intrauterine insemination, and thus could not evaluate whether insurance coverage for ART affects the use of these procedures, which also carry high risk for multiple births. Because insurance companies may require some women to undergo non-art infertility treatments before attempting ART procedures, one might argue that insured women would move more slowly to ART and only after non-art infertility treatments had failed. Alternatively, one could speculate that women residing in noninsurance states who had to pay out of pocket for ART procedures would be more motivated to attempt less expensive FERTILITY & STERILITY 21

7 non-art infertility treatments first and only move to ART as a last resort. We were unable to evaluate either of these hypotheses because we did not have specific data on the history of non-art infertility treatment attempts. The strengths of our study include the representatives of the data and the large sample size, which allowed sufficient power to allow for multiple stratification. However, the data were observational, and patients with multiple IVF procedures in 1998 are represented multiple times in the data set. Because the unit of analysis is the procedure itself and procedures on the same woman cannot be linked, analyses of embryo transfer practices are necessarily limited. It does not, however, limit analyses of multiple or live birth outcomes, since it would be highly unlikely for a woman to have had more than one livebirth delivery after IVF procedures in a 12-month period. Furthermore, although we were familiar with general state laws regarding insurance and could identify in-state residents who would likely be covered by them if insured, we did not have individual insurance data on patients. Some patients in noninsurance states have insurance policies that cover infertility services. Finally, because we used surveillance data, we did not have detailed information on demographic and clinical factors at the patient level; thus, there may be unmeasured differences between patients in insurance and noninsurance states. Proponents of the argument that insurance coverage will reduce multiple birth rates have identified two potential mechanisms of impact: decreased pressure on patients to transfer more embryos because they would be allowed multiple attempts, and a watchdog effect exerted by the insurance industry that would increase pressure on providers to decrease risk of multiple gestation. We note, however, that the addition of insurance coverage in a state might have two effects that would go in different directions, albeit the first would be stronger. Insurance would provide the resources to allow more patients to make multiple attempts, but it would also create a new group of patients who previously would not have attempted ART due to prohibitive costs but could now afford a single procedure only (e.g., due to copayments and uncovered expenses). Insurance coverage for these patients would not alleviate pressure to transfer more embryos. In addition, some insurance policies limit the number of procedures covered. Patients covered by such policies would presumably experience the same pressure to avoid a failed cycle that insurance is purported to alleviate. We are unaware of any insurance policy that limits the number of embryos transferred in a given procedure. Debate on the merits of mandating insurance coverage will probably continue. Our results support the argument that mandated insurance coverage can affect practice patterns by resulting in the transfer of fewer embryos. Because a pattern of fewer embryos transferred and reduced multiple birth risk was seen in only one of the three insurance states, however, we cannot say that insurance coverage per se is associated with decreased multiple birth risk. It is noteworthy that this state, Massachusetts, has the most comprehensive insurance coverage of ART, does not limit the number of procedures covered, and has had this insurance policy for 2 to 4 years longer than the other two insurance states investigated. The lack of uniformity of results among the insurance states indicates a need to further elucidate the nature of the effect of insurance coverage on embryo transfer practices and how it relates to birth outcomes. Acknowledgments: The data used for this study were collected using the Society for Assisted Reproductive Technology (SART) ART reporting system. This data system was developed by SART in Since 1995, data from the SART system have been used by the Centers for Disease Control and Prevention to calculate pregnancy success rates for ART clinics operating in the U.S. This system is jointly supported by SART; the American Society for Reproductive Medicine; the Centers for Disease Control and Prevention; and RESOLVE, the National Infertility Association. The authors thank SART, American Society for Reproductive Medicine, and RESOLVE, without whose contributions this work would not have been possible. References 1. Family Building Act of 2001, H.R. 389, 107th Cong., 1st Sess. (2001). 2. Equity in Fertility Coverage Act of 2001, H.R. 568, 107th Cong., 1st Sess. (2001). 3. H.R. 1246, 107th Cong., 1st Sess. (2001). 4. Fair Access to Infertility Treatment and Hope Act of 2000 (FAITH), S. 874, 107th Cong., 1st Sess. (2001). 5. State infertility insurance laws. American Society for Reproductive Medicine. Available at: Accessed July 2, Hughes EG, Giacomini M. Funding in vitro fertilization treatment for persistent subfertility: the pain and the politics. Fertil Steril 2001;76: Blackwell R, Team WM. Hidden costs of infertility treatment in employee health benefits plans. Am J Obstet Gynecol 2000;182: Kyle BL. In vitro fertilization: a right or a privilege? J La State Med Soc 2000;152: Neumann PJ. Should health insurance cover IVF? Issues and options. Health Polit Policy Law 1997;22: In vitro fertilization: insurance and consumer protection. Harvard Law Rev 1996;109: Seibel MM, Zilberstein M, Kearnan M. In-vitro fertilisation and health care coverage. Lancet 1995;345: Callens S. IVF in the USA, who pays for it? Med Law 1992;11: Retsinas J. The mandation of insurance coverage for in vitro fertilization. Health Soc Policy 1991;3: Martin JA, Park MM. Trend in twin and triplet births: Natl Vital Stat Rep 1999;47: Contribution of assisted reproductive technology and ovulation-inducing drugs to triplet and higher-order multiple births United States, MMWR Morb Mortal Wkly Rep 2000;49: Kiely JL. What is the population-based risk of preterm birth among twins and other multiples? Clin Obstet Gynecol. 1998;41: Guyer B, Martin JA, MacDorman MF, Andersen RN, Strabobino DM. Annual summary of vital statistics Pediatrics 1997;100: Pharoah PO, Cooke T. Cerebral palsy and multiple births. Arch Dis Child Fetal Neonatal Ed 1996;75:F Gardner MO, Goldenberg RL, Cliver SP, Tucker JM, Nelson KG, Copper RL. The origin and outcome of preterm twin pregnancies. Obstet Gynecol 1995;85: Spellacy WN, Handler A, Ferre CD. A case-control study of 1,253 twin pregnancies from a perinatal data base. Obstet Gynecol 1990;75: Kinzler WL, Ananth CV, Vintzileos AM. Medical and economic effects of twin gestations. J Soc Gynecol Investig 2000;7: Reynolds et al. Insurance and IVF multiple birth risk Vol. 80, No. 1, July 2003

8 22. Senat MV, Ancel PY, Bouvier-Colle MH, Breart G. How does multiple pregnancy affect maternal mortality and morbidity? Clin Obstet Gynecol 1998;41: Multiple gestation pregnancy. The ESHRE Capri Workshop Group. Hum Reprod 2000;15: Goldfarb JM, Austin C, Lisbona H, Peskin B, Clapp M. Cost-effectiveness of in vitro fertilization. Obstet Gynecol 1996;87: Elster N. Less is more: the risks of multiple births. Institute for Science, Law, and Technology Working Group on Reproductive Technology. Fertil Steril 2000;74: Angard NT. Seeking coverage for infertility. Insurers should offer reasonable services to help couples achieve pregnancy. AWHONN Lifelines 2000;4: Faber K. IVF in the US: multiple gestation, economic competition, and the necessity of excess. Hum Reprod 1997;12: Frankfurter D, Barren CB, Alper MM, Berger MJ, Oskowitz SP, Penzias AS. Insurance mandates for IVF coverage effectively lower multiple births per embryo transfer [abstract]. Fertil Steril 1998;70 (Suppl):S Jain T, Harlow BL, Hornstein MD. Insurance coverage and outcomes of in vitro fertilization. N Engl J Med 2002;347: Fertility Clinic Success Rate and Certification Act of 1992 (FCSRCA), Pub. L. No , October 24, Schieve LA, Peterson HB, Meikle SF, Jeng G, Danel I, Burnett N, et al. Live-birth rates and multiple-birth risk using in vitro fertilization. JAMA 1999;282: FERTILITY & STERILITY 23