Rodriguez, Maria Isabel MD; Jensen, Jeffrey T. MD, MPH; Darney, Philip D. MD, Msc; Little, Sarah E. MD; Caughey, Aaron B. MD, PhD
Unintended pregnancy has direct and indirect costs for families and communities and occurs disproportionately in poor women with limited resources and minimal access to health care.1 Cost–benefit analyses repeatedly have shown public savings from investment in family planning programs for low-income women.2–4 Medicaid is the largest provider of contraceptive services nationwide but leaves vulnerable populations at risk for unintended pregnancy because of restrictions that limit access.5–7 One such group is immigrant families who have been in the United States for fewer than 5 years, termed new immigrants. Documented new immigrants are eligible for Emergency Medicaid coverage only.5 The majority of new immigrants are Latino.5
With respect to reproductive health care, Emergency Medicaid covers the cost of a delivery but provides no federal funds for prenatal care or contraception. The postpartum interval offers a key opportunity to provide contraception for women who do not routinely access the health care system.8,9 Previous research has well established the cost-effectiveness of these methods, with each method estimated to save at least $5.00 in public funds for every dollar spent.1,8,10
Although the cost of uncompensated obstetric care is high, it is not known what role migration plays in offsetting the costs of future pregnancies for local payers.5,6 The probability of migration out of the area introduces conflicting interests for the various stakeholders. Whereas society always will bear the future costs of unintended pregnancy, the hospital or the state Medicaid program might not incur future costs if the woman leaves the area.
Given this background, we conducted a cost–benefit analysis to estimate the costs of offering postpartum contraception to new Latina immigrants in states with a high proportion of new immigrants. Designing cost-effective health policy that provides optimal care requires careful consideration of multiple constituents and their interests. To account for this, our model follows women for 5 years postpartum from a term delivery and considers the perspectives of the hospital, the state Medicaid program, and society. A migratory population offers unique challenges in designing policy; our model accounts for the probability of migration out of the area where the birth occurred.
MATERIALS AND METHODS
A decision-analytic model was developed with TreeAgePro 2004 software (TreeAge Software, Williamstown, MA). Such models consider theoretical cohorts of women who can be provided with different health care options. The sequence of the model began with a woman postpartum from an obstetric delivery (Fig. 1). We used a Markov, or recursive, model to analyze the probability of pregnancy over a 5-year time period, accounting for changes in contraceptive use and method failure.11 Such a model assumes that, in each cycle of the model, a woman can either get pregnant or not. If not, she can get pregnant in the next cycle. If she does get pregnant, she can experience a miscarriage, an elective termination of pregnancy, an ectopic pregnancy, or a viable pregnancy. The probabilities of these outcomes occurring varied based on the availability predicted use of postpartum contraception. To keep the model manageable, we assumed, conservatively, that such viable pregnancies would be uncomplicated. If the model found potential savings from providing contraception, such savings estimates likely would be at the low-end of the spectrum.
We created three different models, each with the same template but with different cost inputs, to model the perspective of the costs of future pregnancy-related care for Latina immigrants for the hospital, the state Medicaid program, and society over the subsequent 5 years (Fig. 1). Costs for the hospital are the difference between the costs of the procedure and the money received from the state. State Medicaid costs were calculated as the payments made by the state for the procedures, and the societal perspective includes the total cost for care. Women were monitored postpartum from a term delivery, and their probabilities of contraceptive use and subsequent pregnancy were modeled.
Our model compared the baseline policy of covering an obstetrical delivery only with a program offering postpartum contraception. We considered the following methods in our model: the intrauterine device (IUD), female surgical sterilization, depot medroxyprogesterone acetate (DMPA), oral contraceptives, and male condoms. Our model includes several key assumptions. We conservatively assumed that all deliveries occurred at term and resulted in healthy newborns. Likewise we assumed that parents with eligible newborns would take advantage of publicly funded health and social service programs. We assumed that if contraception were free, women would be 50% more likely to choose a covered method then one they would have to pay for out of pocket.12,13 We also assumed that fewer than 3% of women would be able to afford the costs out of pocket for an IUD or sterilization.6 Sensitivity analysis was performed around all inputs.
The primary outcome was future reproductive health care costs due to pregnancy in the next 5 years. The institutional review boards at Oregon Health & Science University and the University of California, San Francisco, reviewed and approved the research protocol.
Probability inputs were derived from literature review and from hospital and state Emergency Medicaid records. (Table 1). The distribution of pregnancy outcomes for Hispanic women (term delivery, ectopic pregnancy, spontaneous abortion, induced abortion) was obtained from the 2004 National Survey of Family Growth and is shown in Table 2.8,14
The probability of contraceptive uptake was collected for a cohort of recent Emergency Medicaid patients delivering at a university hospital in a new-growth state (Table 2). These data were used for our baseline estimates of the proportion of women who would accept long-acting contraception postpartum. Sensitivity analyses were performed around these inputs and all others.
A variable to reflect the probability of migration out of the state in which the case pregnancy occurred was included in the model. The baseline probability for immigrants to return to their country of origin was obtained from data from the U.S. Social Security Administration (summarized in Table 2).15 The proportion of reproductive-age women in the immigrant population was estimated from the literature for 10 states with a high proportion of immigrants (summarized in Table 3).16 The total number of recent immigrants, mean age, and the proportion that are female are known from the literature.16 We assumed conservatively that 10% of all women would be of reproductive age and at risk for pregnancy in our calculations.16
A study database of 5,412 women with Emergency Medicaid who had an obstetric diagnosis at a university hospital between July 2001 and December 2006 was created (Table 1). Records were identified based on insurance type and sorted by diagnosis related group (DRG) codes for pregnancy outcome type. These are the only obstetrical diagnoses for which services are covered for Emergency Medicaid patients. Hospital charges and state payments were identified for each procedure, and mean amounts by DRG were calculated. Hospital revenue is the difference between costs and payment collected from the state. All dollar amounts were converted to 2008 dollars using the Consumer Price Index.
All procedure costs were calculated from state and hospital Emergency Medicaid payments and costs. We considered only the direct costs associated with providing care at the time of admission. Hospital costs were calculated from the department's cost-to-charge ratio for each obstetric outcome and IUD. The hospital-cost input used in our model included revenue received from the state for future pregnancy and newborn care. State costs are the mean payments made to the hospital for each obstetrical DRG. Social costs are the true costs of the procedure and reflect indirect and direct costs associated with care. As a conservative estimate of social costs, we used mean hospital costs (Table 4). All procedures that result in a live birth include the cost of newborn care. Because these newborns are U.S. citizens and are eligible for social programs for needy families, the social costs of a live birth include an estimate of the costs of these programs from birth to age 5.2
Contraceptive costs for long-acting methods were derived from hospital billing records and the literature. The costs for an IUD are the average cost for the device and insertion from the Oregon Family Planning Expansion program. The costs for sterilization are considered as an obstetric procedure. Therefore, the costs for the hospital are the direct costs of providing the service minus the revenue received. State costs are payments made to the hospital. Social costs are estimated by hospital charges. The cost inputs for DMPA, contraceptive pills, and condoms are for a 1-year supply. As used in previous research, 80 acts of intercourse per year are assumed in calculating condom costs.8 The prices for condoms, DMPA, and oral contraceptives were obtained from the Medi-Span Master Drug Database.17
Sensitivity analysis is a statistical tool that allows us to estimate how a change in one of our model parameters would affect results. A threshold value marks the point to which change in a variable would alter the conclusion.18 We tested the robustness of our model by performing univariable sensitivity analysis on all inputs and with Monte Carlo simulations.
For the sensitivity analysis, we varied costs down to one half and up to two times baseline estimates. Three key inputs were identified: probability of immigration, probability of abortion, and probability of uptake of long-acting contraception. Bivariable sensitivity analyses were performed on these variables to determine whether changes in multiple variables would affect our conclusion. When varying each cost category, we maintained the baseline cost proportions. We performed a Monte Carlo simulation using 10,000 trials to test the model's robustness to simultaneous multivariable changes. A Monte Carlo simulation varies all probability estimates simultaneously by sampling distributions around the baseline estimate.
Provision of postpartum contraception, regardless of citizenship status, is always cost saving for society regardless of how inputs are varied. The same is not true for the hospital that provides the contraception but receives inadequate reimbursement. From the hospital's perspective, provision of family-planning services over a 5-year period would cost $632 compared with $265 to continue the current policy of covering only the obstetric delivery (a loss of $367). Conversely, from a societal perspective, offering the option of family planning coverage in our model leads to societal savings of $17,792 per woman in future obstetric and neonatal costs ($9,776 compared with $27,568). Providing the option of postpartum family-planning services also saved the state money: $108 per person ($929 to $821). We applied these potential savings to a range of states with both established and new populations of immigrants. Using our model, we projected the savings for 10 states with large immigrant populations; the range of savings extended from $259,000 in New Mexico to $12,700,800 in California (Table 3).
Sensitivity analysis showed our results to be robust to a wide range of inputs. When considering the social perspective, provision of contraceptive was cost-effective across all ranges of inputs. Of note, this is true regardless of the probability of immigration, uptake of long-acting contraception, and the cost for an IUD. Children born to immigrants are eligible for a full array of public services as U.S. citizens. Social programs offered vary by state. We included in our model an estimate of public-program costs to age 5 based on California data. Sensitivity analysis was done around this input to evaluate how social costs would vary in a state that offers fewer assistance programs or if families did not take advantage of them. Even when no public programs are offered, society still saves $9,600 per woman by offering contraception.
When considering the state's perspective, offering postpartum contraception to new immigrants saved the state money—$108 dollars per woman. We examined how these results would vary with changes in the probability of immigration and abortion and IUD cost. Family planning remains cost saving for states when the probability that a woman will remain in the area for 5 years is 49% or higher. A threshold value also was noted for the probability of induced abortion. If 37% of women elected to have abortions, the state would not save money on family planning care. One-way sensitivity analysis on all inputs and bivariable analysis on our key variables did not identify a situation in which it would be cost saving for a hospital to provide postpartum contraception.
Finally, Monte Carlo simulation was performed to test the robustness of the model to simultaneous changes in inputs. When considering the hospital's perspective, provision of postpartum contraception was not cost beneficial in 98.8% of trials when compared with the baseline policy of just covering the obstetric delivery. From the state's perspective, extending contraception to postpartum new immigrants was cost saving in 64.5% of the trials. Considering society as the payer, providing contraception regardless of immigration status was cost beneficial in 100% of trials.
Among a theoretical cohort of new immigrant women delivering at term, provision of postpartum contraception is cost saving when compared with the current policy of covering the delivery only when considering the perspectives of the state or society as the potential payer. It is not cost saving for a hospital to fund a program offering postpartum contraception; they lose less money by covering the costs of and billing for subsequent obstetric deliveries. Thus, the economic incentives to provide a cost-saving intervention from a societal perspective are misaligned.
Federal legislation effectively transfers financial responsibility for care for a large and rapidly growing patient population to local hospitals and state governments. New immigrants are restricted to acute, hospital-based care only. Hospitals caring for a high proportion of the medically indigent are penalized by legislation that mandates care without adequate compensation.19 Despite the known public savings associated with contraception, our model demonstrates that hospitals do not have financial incentives to offer family-planning services because doing so would lead to further financial losses for them.
Contraception is well established as a cost-effective use of public-health dollars.8,10,20,21 Our model supports this conclusion, showing that provision of postpartum contraception is actually cost saving across the range of all inputs from a societal perspective. Offering the option of contraception was also cost saving from the state perspective, regardless of the contraceptive methods selected by women.
When considering the state as the payer for contraception, significant savings would occur in a state in which 49% of the immigrant population would continue to reside for a 5-year period. Uncertainty around the probability of immigrants becoming permanent state residents likely affects individual state's incentives to provide preventive care. Demographics of the immigrant population in the United States have changed dramatically over the past 10 years.16 There has been a significant increase in the number of immigrants and a change in the states in which they reside.5,16 Although U.S. Census data show that only 38% of immigrants leave the United States within 10 years, it is not known what interstate migratory patterns are like for new immigrants.15
Our study is an economic model and, as such, has inherent limitations, including the inability to account for every possible factor. Economic modeling also is limited by the imprecision of health care–cost estimates. Furthermore, the model incorporates Medicaid payments and costs in only one state. Sensitivity analyses around these inputs did not significantly change our results; thus, we believe they are generalizable to other states. We analyzed only the cost savings of contraception provision and not the additional benefits that family planning provides, such as prevention of unintended pregnancy.22 Further, we conservatively assumed normal, term pregnancies. Obviously some of these future pregnancies would lead to preterm births, which are likely to be far more expensive to all three payer perspectives.23
Data on interstate immigration and pregnancy intention among new immigrants is particularly limited. Our baseline assessments for immigration out of the area of the case birth and probability of abortion are based on U.S. Census data and the National Survey of Family Growth.15,16 Economic incentives, such as free contraception, have been shown to increase uptake, but limited research has been done. This limits our ability to predict the effect that free postpartum contraception would have on uptake.12,13 However, sensitivity analysis showed that, at all probabilities of contraceptive uptake, our results would remain robust.
Family planning is integral to the equality of women and the elimination of disparities in society.24–27 Improper financial incentives at the federal level determine policy that restricts access to health care and wastes scarce public resources. Further research into interstate migratory patterns and probabilities should be conducted to assess the economic value of a federal mandate for preventive coverage of new immigrants.
States should institute health policy that does not restrict access to family-planning services for Medicaid participants by citizenship status. Such a program has the potential to save millions of dollars in state and social expenditures in states with a sizable proportion of new immigrants (Table 3). Our data support evidence-based policy that rationally allocates public funds and affords all individuals the opportunity to plan their families.
1. Foster DG, Rostovtseva DP, Brindis CD, Biggs MA, Hulett D, Darney PD. Cost savings from the provision of specific methods of contraception in a publicly funded program. Am J Public Health 2009;99:446–51.
2. Amaral G, Foster DG, Biggs MA, Jasik CB, Judd S, Brindis CD. Public savings from the prevention of unintended pregnancy: a cost analysis of family planning services in California. Health Serv Res 2007;42:1960–80.
3. Gold RB. Breaking new ground: ingenuity and innovation in Medicaid family planning expansions. New York (NY): Guttmacher Institute; 2008.
4. Frost JJ. Estimating the impact of expanding Medicaid eligibility for family planning services. New York (NY): Guttmacher Institute; 2006.
5. DuBard CA, Massing MW. Trends in emergency Medicaid expenditures for recent and undocumented immigrants. JAMA 2007;297:1085–92.
6. Rodriguez MI, Edelman A, Wallace N, Jensen JT. Denying postpartum sterilization to women with Emergency Medicaid does not reduce hospital charges. Contraception 2008;78:232–6.
7. Okie S. Immigrants and health care—at the intersection of two broken systems. N Engl J Med 2007 9;357:525–9.
8. Trussell J, Lalla AM, Doan QV, Reyes E, Pinto L, Gricar J. Cost effectiveness of contraceptives in the United States. Contraception 2009;79:5–14.
9. Darney P, Speroff L. A clinical guide for contraception. 4th ed. Philadelphia (PA): Lippincott Williams & Wilkins; 2005.
11. Dunson DB. Bayesian methods for latent trait modeling of longitudinal data. Stat Methods Med Res 2007;16:399–415.
12. Hohmann H, Reid L, Creinin MD. Women's motivation to participate in contraceptive efficacy trials. Contraception 2009;80:270–5.
13. Lagarde M, Haines A, Palmer N. Conditional cash transfers for improving uptake of health interventions in low- and middle-income countries: a systematic review. JAMA 2007;298:1900–10.
14. Hatcher RA TJ, Nelson AL. Contraceptive technology. 18th ed. New York (NY): Ardent Media; 2004.
15. Duleep H, Dowhan D. Adding immigrants to microsimulation models. Washington, DC: U.S Social Security Administration Office of Retirement and Disability Policy; 2008.
16. Fortuny K CR, Passell J. The characteristics of unauthorized immigrants in California, Los Angeles County, and the United States. Washington DC: The Urban Institute; 2007.
18. Lilford RJ, Pauker SG, Braunholtz DA, Chard J. Decision analysis and the implementation of research findings. BMJ 1998;317:405–9.
19. Jones JW, McCullough LB, Richman BW. My brother's keeper: uncompensated care for illegal immigrants. J Vasc Surg 2006;44:679–82.
20. Foster DG, Biggs MA, Amaral G, Brindis C, Navarro S, Bradsberry M, et al. Estimates of pregnancies averted through California's family planning waiver program in 2002. Perspect Sex Reprod Health 2006;38:126–31.
21. Trussell J, Leveque JA, Koenig JD, London R, Borden S, Henneberry J, et al. The economic value of contraception: a comparison of 15 methods. Am J Public Health 1995;85:494–503.
22. Schwarz EB, Smith R, Steinauer J, Reeves MF, Caughey AB. Measuring the effects of unintended pregnancy on women's quality of life. Contraception 2008;78:204–10.
23. Phibbs CS, Schmitt SK. Estimates of the cost and length of stay changes that can be attributed to one-week increases in gestational age for premature infants. Early Hum Dev 2006;82:85–95.
24. Wulf D, Donovan P. Women and children benefit when childbearing is planned. New York (NY): Guttmacher Institute; 2002.
25. Ku L. Reducing disparities in health coverage for legal immigrant children and pregnant women. Washington, DC: Center on Budget and Policy Priorities; 2007.
26. Klerman LV, Johnson KA, Chang CH, Wright-Slaughter P, Goodman DC. Accessibility of family planning services: impact of structural and organizational factors. Matern Child Health J 2007;11:19–26.
27. Kost K, Landry DJ, Darroch JE. The effects of pregnancy planning status on birth outcomes and infant care. Fam Plann Perspect 1998;30:223–30.
Figure. No caption available.