Evaluating Cost-Effectiveness of Interventions that Affect Fertility and Childbearing: How Health Effects are Measured Matters

Abstract

Background

Current guidelines for economic evaluations of health interventions define relevant outcomes as those accruing to individuals receiving interventions. Little consensus exists on counting health impacts on current and future fertility and childbearing.

Objective

To characterize current practices for counting such health outcomes.

Design

We developed a framework characterizing health interventions with direct and/or indirect effects on fertility and childbearing and how such outcomes are reported. We identified interventions spanning the framework and performed a targeted literature review for economic evaluations of these interventions. For each article, we characterized how the potential health outcomes from each intervention were considered, focusing on QALYs associated with fertility and childbearing.

Results

We reviewed 108 studies, identifying seven themes: 1) Studies were heterogeneous in reporting outcomes. 2) Studies often selected outcomes for inclusion that tend to bias toward finding the intervention to be cost-effective. 3) Studies often avoided the challenges of assigning QALYs for pregnancy and fertility by instead considering cost per intermediate outcome. 4) Even for the same intervention, studies took heterogeneous approaches to outcome evaluation. 5) Studies employed multiple, competing rationales for whether and how to include fertility-related QALYs and whose QALYs to include. 6) Studies examining interventions with indirect effects on fertility typically ignored such QALYs. 7) Even recent studies had these shortcomings.

Limitations

The review was targeted rather than systematic.

Conclusions

Economic evaluations inconsistently consider QALYs from current and future fertility and childbearing in ways that frequently appear biased towards the interventions considered. As the Panel on Cost-Effectiveness in Health and Medicine updates its guidelines, making the practice of cost-effectiveness analysis more consistent is a priority. Our study contributes to harmonizing methods in this respect.

INTRODUCTION

Whose health benefits should be counted in economic evaluations (i.e., cost-effectiveness analyses (CEAs) or cost-benefit analyses (CBAs)) of interventions that directly or indirectly affect current and future fertility and childbearing? Consider the example of in vitro fertilization (IVF). If a woman receives IVF and gives birth to a healthy child, should QALYs gained by the mother and child be counted? Should fewer QALYs be assigned if the child born has cystic fibrosis (CF)? Are QALYs lost if, instead, the woman who has had IVF then has fetal genetic testing and chooses to abort the fetus with CF? Are QALYs lost or gained if a young woman diagnosed with cancer has a hysterectomy and becomes infertile? What if her partner is treated for testicular cancer and becomes infertile?

These questions raise thorny and contentious issues. Yet, analysts conducting economic evaluations must answer them explicitly or implicitly. For interventions that have direct or indirect effects on fertility and childbearing, choices about whose health benefits to count can strongly influence conclusions about which interventions to invest in. Furthermore, the particular nature of such decision problems – for example, infertility treatment of a couple in which the male partner’s preferences are frequently relevant – make consideration of these questions distinctly challenging. For such interventions, it is important to conceptualize both whose benefits to quantify and how to do so.

Current guidelines for conducting economic evaluations of health interventions define relevant health outcomes primarily as changes in the quantity and quality of life for individuals receiving interventions.^1–3 However, strikingly little consensus guidance exists on whether and how to count the health impacts of interventions that affect current and future fertility and childbearing, despite the increasing use of technologies that directly or indirectly have such impacts. Literatures as diverse as welfare economics, law, and ethics have grappled with these fundamental challenges. In each area, debate remains, suggesting that applied economic evaluations are unlikely to take a uniform, consistent approach.^4–14

Our paper is not intended to resolve these debates, but rather to probe important questions in the context of the current update¹⁵ to the guidelines contained in Cost-Effectiveness in Health and Medicine.² To inform this discussion, we developed a framework describing situations in which these considerations are relevant. Then we conducted a targeted literature review of economic evaluations of interventions where fertility and reproductive issues are at the forefront. Our goal was to identify the range of practices in counting health benefits and harms of interventions in such analyses. Our review is targeted rather than systematic, and is intended to highlight the state of research in this area and to identify inconsistencies that may bias results, highlighting the need for harmonized guidance.

METHODS

We developed a framework describing interventions that have direct and/or indirect effects on fertility and childbearing. We distinguished interventions by the situation (women with a current pregnancy, women with a future pregnancy, or men at any time), effect on fertility and childbearing (increases or decreases in fertility, pregnancies ending before birth, or birth defects/perinatally acquired disease/preterm birth and newborn complications), and whether effects are intended or unintended (i.e., side effects). For example, contraception has the intended effect of decreasing a woman’s fertility, whereas radiation treatments can have the unintended effect of decreasing a woman’s fertility. For some possible combinations, no interventions are relevant: for example, no interventions intend to increase birth defects.

We next identified a set of health interventions that span the framework. These include interventions that affect a woman’s current pregnancy (e.g., preeclampsia screening and treatment, prenatal genetic testing) and interventions that affect a woman’s fertility (e.g., IVF and assisted reproductive technologies, contraception). We also identified interventions that affect a man’s fertility (e.g., vasectomy, vasectomy reversal, and radiation treatments). The set of interventions is not meant to be exhaustive but rather to collectively cover all logically consistent combinations in the framework.

We then developed a framework to characterize how such studies quantify outcomes. For the child, mother, and father, we determined whether a study reported intended and unintended outcomes separately, combined as QALYs or DALYs, monetized for CBA, or not at all. For reported outcomes, we determined the time horizon over which the outcomes were measured (through gestation/childbirth, lifetime, or other).

We performed a targeted (rather than systematic) literature review, searching for articles presenting economic evaluations of the selected interventions. We searched in high-impact general medical journals and journals whose focus is on obstetrics and gynecology or other issues related to fertility and childbearing. For each article, we identified which health outcomes were considered, with particular focus on QALY gains and losses associated with current and future fertility and childbearing.

Finally, we performed quantitative analyses (descriptive statistics and assessment of statistical differences using Fischer exact tests and logistic regressions as appropriate) and qualitative analyses of the selected articles to identify key patterns, with specific focus on consistency across intervention types.

RESULTS

Table 1 shows our framework and intervention examples that fit within it. Table 2 presents our framework for how health outcomes in the studies are measured and reported. From our literature review, we selected 108 studies for analysis that covered interventions spanning the logically feasible combinations in our framework. Table 3 summarizes groups of articles by intervention type. Table 4 provides information about each study we reviewed. Below, we provide quantitative and qualitative characterizations of key themes identified in the review.

Table 1.

Framework for Interventions that Can Affect Female and Male Fertility and Childbearing, with Selected Examples

Situation	Direct Effects (Intended purpose)
	Fertility		Pregnancy Ends before Birth*		Birth Defects/ Perinatally Acquired Disease/Preterm Birth and Newborn Complications**
	Increase	Decrease	Decrease	Increase	Decrease	Increase
Woman’s current pregnancy	NA	NA	Preeclampsia treatment Diabetes treatment Screening and treatment of chlamydia and other STDs Treatment of hypothyroidism Smoking cessation	Therapeutic abortion Selective embryo reduction	Diabetes treatment Folate supplementation Fetal surgery Screening and treatment of chlamydia and other STDs Smoking cessation Antiretroviral therapy for HIV+ pregnant women Genetic testing***	NA
Woman’s fertility	IVF and ARTs Screening and treatment of chlamydia and other STDs Obesity reduction	Reversible contraception Sterilization	Obesity reduction Smoking cessation Increasing the inter-pregnancy interval Treatment of diabetes	NA	Pre-implantation genetic screening Pre-pregnancy genetic screening Smoking cessation Folate supplementation Treatment of diabetes Contraception for HIV+ pregnant women	NA
Man’s fertility	IVF and ARTs Screening and treatment of chlamydia and other STDs Male fertility treatments Vasectomy reversal	Condoms Vasectomy	NA	NA	Male genetic screening	NA

	Indirect Effects (Not the intended purpose)
	Fertility	Pregnancy Ends before Birth*			Birth Defects/ Perinatally Acquired Disease/Preterm Birth and Newborn Complications**
	Increase	Decrease	Decrease	Increase	Decrease	Increase
Woman’s pregnancy	NA	Caesarean delivery	NA*	Imaging Radiation treatments Fetal surgery Appendectomy Chemotherapy for cancer treatment Selective embryo reduction Genetic testing†	NA	Imaging Radiation treatments Chemotherapy for cancer treatment
Woman’s fertility	Female HPV vaccination Obesity reduction Drugs that may decrease the effectiveness of birth control pills	Obesity reduction Imaging Radiation treatments Hysterectomy Chemotherapy for cancer	Female HPV vaccination Obesity reduction Smoking cessation Screening and treatment of chlamydia and other STDs Treatment of hypothyroidism	Imaging Radiation treatments	Smoking cessation Treatment of STDs	Imaging Radiation treatments Cancer treatment
Man’s fertility	Male HPV vaccination	Imaging Chemotherapy for testicular cancer Radiation treatments	NA	NA	NA	Radiation treatments

NA = not applicable, ART = assisted reproductive technologies, HIV = human immunodeficiency virus, HPV = human papilloma virus, IVF = in vitro fertilization, STD = sexually transmitted disease

^*The category “Pregnancy Ends before Birth” includes both miscarriage of pregnancies whose desired outcome is birth and those whose desired outcome is termination of the embryo/fetus. The category assumes that health care is provided to women who know they are pregnant and considers the fetus that the woman is currently carrying.

^**The category “Birth Defects/Perinatally Acquired Disease/Preterm Birth and Newborn Complications” is intended to reflect outcomes that occur among infants alive at the time of birth including higher subsequent risks from prematurity. Additionally, a wide range of interventions delivered during labor and delivery (e.g., caesarean section, operative vaginal delivery, fetal heart monitoring, and elective induction of labor) are also part of this category as they are intended to prevent adverse outcomes resulting from the delivery process to the newborn and/or mother.

^†“Genetic testing” when conducted during pregnancy encompasses a complex set of activities that are frequently performed sequentially depending on the outcomes of those performed previously. Hence, the relevant outcomes from each step of the testing and consequent intervention decisions based on results may be difficult to specify for all activities here. For example, in addition to serum screening for chromosomal abnormalities, chorionic villus sampling and amniocentesis can be used to collect material from amniotic fluid for testing. These latter collection mechanisms present unintended miscarriage risks.

Table 2.

Framework for How Health Outcomes Are Measured and Reported

	Whose Outcomes?
	Mother’s		Child’s		Father’s
	Intended Outcome	Unintended Outcome	Intended Outcome	Unintended Outcome	Intended Outcome	Unintended Outcome
How Reported
Not reported
Reported separately
Combined as QALYs or DALYs
Monetized for cost-benefit analysis
	Time Horizon
	Mother’s Outcomes		Newborn/Child Outcomes		Father’s Outcomes
Time Horizon
Not reported
Gestation and childbirth
Lifetime
Other

Table 3.

Overview of Studies Selected for Review

Intervention	Summary of Reviewed Articles
Treatment of chlamydia among women and men	8 articles.54–61 Infertility and ectopic pregnancy included as a loss of maternal QALYs. QALY gains for infants born if they avoid health harms. Male QALYs gained due to chlamydia treatment; no male QALY losses for fetal loss or caring for sick child.
Antiretroviral therapy for HIV-infected pregnant women	6 articles.62–67 Infant QALYs gained (DALYs averted) due to avoiding HIV infection. No maternal QALYs gained due to child avoiding HIV infection.
Contraception for HIV-infected pregnant women	2 articles.75, 76 One study considers births averted; one considers HIV-infected births averted. Neither considers antiretroviral therapy to attenuate newborn HIV infection, nor benefits to mother of ongoing treatment nor of having a child without HIV.
Therapeutic abortion	3 articles.68–70 Studies evaluate different abortion methods so all strategies will equally result in pregnancy termination. Strategies may result in higher risks of future infertility but these future lost QALYs are not considered.
Contraception for women and men	4 articles.71–74 Cost per pregnancy averted or cost per woman’s QALY gained. No studies consider child QALY losses from intentionally prevented pregnancy nor maternal QALY losses for complications such as infertility due to ectopic pregnancy.
Preeclampsia treatment	3 articles.82–84 One study included offspring QALYs; one did not include these QALYs due to lack of data; one reported intermediate outcome only.
Treatment of diabetes in pregnant women	8 articles.16–23 Studies generally include offspring QALYs and/or maternal QALYs lost from loss of fetus/baby and maternal QALYs gained from improved maternal health.
Treatment of diabetes in women of childbearing age	3 articles.85–87 Studies consider net gain/loss of value. Benefits include lower costs for mother during pregnancy and for mother/child after successful birth and perhaps longer term better outcomes for mother/child. It is not always clear how lost pregnancies are to be valued in terms of lost maternal QALYs versus lost future QALYs for a fetus surviving to birth.
Smoking cessation for pregnant women	11 articles.43–53 Often cost-benefit analysis (costs of outcomes such as preterm newborns but not necessarily lost QALYs if they die) or intermediate outcomes considered. Much more focused on maternal QALYs than potential offspring QALYs.
Smoking cessation for women of childbearing age	1 article.88 Considers net gain/loss of value. Benefit is reduced short-term cost of low birth weight children.
Genetic testing of pregnant women and their fetuses	18 articles.24–41 Studies generally include the cost of having a child with disabilities. Some studies include maternal QALY gains/losses due to false positive screening result, miscarriage, pregnancy termination, and giving birth to an affected child. Some studies assume the possibility of a “replacement child” after pregnancy termination.
Preconceptional genetic testing	5 articles.41, 103–106 Cost per future affected birth avoided. One study also reports cost per carrier detected. No studies consider QALYs.
IVF and ARTs	10 articles.107–116 Cost per intermediate outcome (live birth or live healthy birth) but no loss of QALYs for fetal loss or children with birth complications (low birth weight). Sometimes maternal QALYs considered. One study that considers maternal QALYs presents philosophical reason for not including offspring QALYs.
Pre-implantation genetic screening	1 article.110 Cost per healthy infant. No loss for termination or problems after birth.
Pre-pregnancy genetic screening	5 articles.41, 103–106 Cost per case averted or per carrier detected. One study mentions that other outcomes not included because no consensus about how people behave with this information in terms of pregnancy/termination.
Folate supplementation for pregnant women	7 articles.117–123 Generally QALYs gained from children born healthy and avoiding neural tube defects because of the intervention. Studies do not all count fetuses lost before birth. No studies count maternal/paternal QALYs from having such a loss nor for having children with or without a neural tube defect.
Folate supplementation for women of childbearing age	1 article.88 Cost per neural tube defect averted. Short-term costs only.
Vasectomy	2 articles.74, 77 Costs per pregnancy averted or couple-years of protection. Studies do not consider potential demand for future vasectomy reversal/inability to reverse in all cases/loss of unwanted future infertility. Studies do not consider loss of child QALYs.
Vasectomy reversal	4 articles.78–81 Cost per live birth or cost per pregnancy.
Imaging in pregnant women	1 article on imaging for preoperative appendix removal.134 Maternal QALYs from direct health benefits. Newborn QALYs conditional on birth. No maternal QALY loss for leukemia, preterm birth or miscarriage caused by imaging.
Female HPV vaccination	3 articles.89–91 Female QALY gains including those from avoiding hysterectomy but no fertility-specific effect of this nor child QALYs.
Male HPV vaccination	2 articles.92, 93 Cost per male or female QALY gained. Female QALYs gained when males do not transmit HPV to them.
Hysterectomy	5 articles on hysterectomy for benign conditions.94–98 Women’s QALYs gained due to treatment of the benign conditions. No fertility issues considered even though some of the comparators are less invasive and more fertility sparing.
Chemotherapy for cancer in women of childbearing age	2 articles.99, 100 Studies consider woman’s QALYs gained due to treatment but do not consider fertility losses due to treatment.
Chemotherapy for testicular cancer	2 articles.101, 102 Studies consider gains in men’s health due to cancer treatment, but do not consider potential health losses due to future infertility.

ART = assisted reproductive technologies, IVF = in vitro fertilization, QALY = quality-adjusted life year

Table 4.

Summary of All Selected Studies

				Outcomes Considered*	Time Horizon**
Author/Year	Intervention	Cost-Effectiveness Measure(s)	Comments	MI	MU	CI	CU	PI	PU	M	C	P
INTERVENTIONS TARGETED TO PREGNANT WOMEN
Direct Effects: Decrease Miscarriage
	Preeclampsia screening and treatment
Adam 200582	Treatment of severe preeclampsia	Cost per DALY averted	Looks at cost-effectiveness of treatment for severe preeclampsia in a number of African and Asian countries. Measures DALYs stemming from maternal morbidity and mortality and newborn mortality. Does not include stillbirths or newborn morbidity, due to lack of data.	Q	†	Q	†			L	L
Simon 200684	Preeclampsia treatment	Cost per eclampsia case prevented	Also measures maternal and newborn deaths, but does not include these outcomes in the cost-effectiveness ratio. Likewise mentions side effects but does not include them.	S		S				G	G
Shmueli 201283	Screening and subsequent treatment for preeclampsia	Cost per preeclampsia case prevented Cost per offspring QALY gained	Two analyses. First analysis measures cost per preeclampsia case prevented. Second analysis measures QALY gains among offspring.	S		Q				G	L
	Screening for and treatment of diabetes in pregnant women
Poncet 200221	Screening for gestational diabetes	Cost per negative health outcome averted	Four negative health outcomes considered: hypertensive disease (mothers) and, for newborns, macrosomia, prematurity and perinatal mortality.	S		S				G	G?
Nicholson 200519	Screening for gestational diabetes	Cost per maternal QALY gained Cost per infant QALY gained	Maternal outcomes include hypertensive disease, polyhydramnios, caesarian delivery and vaginal delivery with or without complications (including hysterectomy). Three maternal health states: perfect health, perfect health following hysterectomy, death. Neonatal outcomes include hypoglycemia, macrosomia, respiratory distress syndrome, shoulder dystocia. Three neonatal health states: mild/no morbidity, moderate morbidity, severe morbidity/infant death.	Q		Q				L	L
Round 201122	Screening for gestational diabetes	Cost per QALY gained	Outcomes include serious perinatal complications including neonatal death/stillbirth (resulting in a QALY loss). Maternal outcomes not explicitly modeled but authors assume disutility due to maternal complications and having a baby experience adverse effects.	Q		Q				L?	L
Werner 201223	Screening for gestational diabetes	Cost per QALY gained	Maternal outcomes include caesarian and vaginal delivery, avoidance of future diabetes. No maternal QALY losses associated with child born with complications. Neonatal outcomes include preterm birth, shoulder dystocia, brachial plexus injury, and stillbirth.	Q		Q				L	L
Marseille 201317	Screening for gestational diabetes	Cost per DALY averted	Maternal outcomes include prenatal adverse events (including maternal death) and avoidance of future diabetes. Neonatal outcomes include perinatal adverse events (including fetal/newborn death), shoulder dystocia, brachial plexus injury, and avoidance of future diabetes.	Q		Q				L	L
Lloyd 200916	Treatment of type 1 diabetes during pregnancy	Cost per live birth Cost per newborn QALY gained	Outcomes used to estimate costs include fetal death, antenatal admissions, normal delivery vs. complications of labor vs. caesarean delivery, NICU days, maternal hypoglycemia, neonatal hypoglycemia. QALYs estimated based only on birth defects avoided and paper states that NICE (National Institute for Health and Care Excellence) deemed these results irrelevant to the National Health Service.	S		Q				G	L
Moss 200718	Treatment of mild gestational diabetes	Cost per additional serious perinatal complication prevented Cost per perinatal death prevented Cost per newborn LY gained	Maternal outcomes were induction of labor and caesarean delivery. Neonatal outcomes were serious perinatal complications (death, shoulder dystocia, bone fracture, nerve palsy), admission to neonatal nursery; jaundice requiring phototherapy. Newborn LYs gained were estimated based on neonatal deaths avoided.	S		S				G	L?
Ohno 201120	Treatment of mild gestational diabetes	Cost per QALY gained	Maternal outcomes include preeclampsia, cesarean delivery, and death. Also maternal QALY losses from neonatal death. Neonatal outcomes used to calculate costs and QALYs include macrosomia, shoulder dystocia, brachial plexus injury, hypoglycemia, hyperbilirubinemia, NICU admissions, and death.	Q		Q				L	L
	Smoking cessation programs for pregnant women
Ershoff 198344	Smoking cessation program for pregnant women	Net gain/loss of value	Includes cost of smoking cessation program. Benefits are hospital cost savings from fewer low birth weight newborns.	$		$				G	G
Windsor 198853	Smoking cessation program for pregnant women	Cost per woman quitting smoking	Includes cost of smoking cessation program only. Benefit is number of women who quit smoking.	S						G
Ershoff 199045	Smoking cessation program for pregnant women	Net gain/loss of value	Includes cost of smoking cessation program. Benefits are hospital cost savings from fewer low birth weight newborns.	$		$				G	G
Marks 199047	Smoking cessation program for pregnant women	Net gain/loss of value Cost per low birth weight case averted Cost per perinatal death averted	Includes cost of smoking cessation program. Benefits are hospital cost savings from fewer low birth weight newborns. No value assigned to averted perinatal deaths. Time horizon for children is Other since the study examines benefits of reducing long-term care for infants with disabilities secondary to low birth weight.	$		$				G	O
Shipp 199251	Smoking cessation program for pregnant women	Net gain/loss of value	Includes cost of smoking cessation program. Benefits are hospital cost savings from fewer low birth weight newborns.	$		$				G	G
Windsor 199352	Smoking cessation program for pregnant women	Net gain/loss of value	Includes cost of smoking cessation program. Benefits are hospital cost savings from fewer low birth weight newborns. Time horizon for children is Other since the study examine costs which include costs of long-term care for infants with low birth weight.	$		$				G	O
Hueston 199446	Smoking cessation program for pregnant women	Net gain/loss of value	Includes cost of smoking cessation program. Benefits are hospital cost savings from fewer low birth weight newborns.	$		$				G	O
Pollack 200149	Smoking cessation program for pregnant women	Cost per case of sudden infant death syndrome (SIDS) averted	Includes cost of smoking cessation program. Benefit is sudden infant death syndrome (SIDS) cases averted – equivalent to lives saved.			S					L?
Dornelas 200643	Smoking cessation program for pregnant women	Cost per woman who quits smoking	Includes cost of smoking cessation program only. Benefit is number of women who quit smoking.	S						G
Parker 200748	Smoking cessation program for pregnant women	Cost per woman who quits smoking	Includes cost of smoking cessation program only. Benefit is number of women who quit smoking.	S						G
Ruger 200850	Smoking cessation program for pregnant women	Cost per maternal QALY gained	Costs include intervention cost, lifetime maternal health cost. Maternal QALYs gained: among women who quit smoking. Infant outcomes are mentioned in terms of birth weight and post-delivery status but are not considered because their differences were not statistically significant in studies of intervention effect.	Q						L
Direct Effects: Pregnancy Ends before Birth
	Therapeutic abortion
Cowett 200668	Therapeutic abortion	Cost per maternal QALY gained	Comparison of various procedures for elective abortion in the second trimester. Includes direct costs of abortion, hysterectomy, and complications. Maternal QALYs gained/lost due to complications from procedures, and death.	Q	Q					L
Hu 200970	Therapeutic abortion	Cost per maternal QALY gained	Comparison of various procedures for elective abortion in the first trimester. Includes costs from procedures and complications. Maternal QALYs gained/lost due to procedures, complications, and infertility.	Q	Q					L
Hu 201069	Therapeutic abortion	Cost per maternal LY gained	Comparison of various procedures for elective abortion in the first trimester. Includes costs from procedures and complications, including cost of future infertility. Maternal LYs gained/lost due to changes in maternal life expectancy.	S	S					L
Direct Effects: Decrease Birth Defects/Perinatally Acquired Disease/Preterm Birth and Newborn Complications
	Folate supplementation for pregnant women
Romano 1995122	Folic acid fortification of grain	Net gain/loss of value	Costs include cost of grain fortification. Savings due to averted lifetime care costs for children born with neural tube defects (NTDs) (spina bifida, anencephaly).			$					L
Llanos 2007120	Folic acid fortification of grain	Cost per newborn DALY averted	Costs include cost of grain fortification and lifetime cost of care for children with NTDs. DALYs averted when newborn NTDs are averted. Ignores stillbirths.			Q					L
Sayed 2008123	Folic acid fortification of grain	Net gain/loss of value	Costs include cost of grain fortification. Savings due to averted lifetime care costs for children born with NTDs (spina bifida, anencephaly, orofacial clefts).			$					L
Jentink 2008119	Folic acid fortification of grain	Cost per newborn QALY gained	Costs include cost of grain fortification and lifetime cost of care for children with NTDs. QALYs gained when newborn NTDs are averted.			Q					L
Bentley 2009117	Folic acid fortification of grain	Cost per population QALY gained	Considers population-wide health effects of grain fortification, including not only reductions in newborn NTDs but also reductions in colon cancer, myocardial infarction, and increases in B-12 masking in the general population. Costs include cost of grain fortification and lifetime cost of care for each of the four conditions. QALYs gained/lost due to any of the four conditions. Since fathers and mothers are included in the general population we count general population QALY effects as accruing to them as well.	Q	Q	Q		Q	Q	L	L	L
Dalziel 2010118	Folic acid fortification of grain or folate supplementation for pregnant women	Cost per newborn DALY averted	Costs include cost of fortification/supplementation and lifetime cost of care for children with NTDs. DALYs averted when newborn NTDs are averted. Includes DALYs associated with pregnancy termination and stillbirth.			Q					L
Rabovskaja 2013121	Folic acid fortification of grain	Cost per population QALY gained	Costs include costs of fortification, lifetime healthcare costs for children with NTDs, and healthcare costs for adult conditions caused by excess folic acid (vitamin B-12 deficiency, neuropathy). Includes QALYs gained among newborns when NTDs are avoided and QALYs lost among adults due to complications from excess folic acid.		Q	Q			Q	L	L	L
	Genetic testing of pregnant women
Ginsberg 199430	Screening for cystic fibrosis (CF)	Net gain/loss of value	Costs of screening, counseling, diagnostics, abortions, and screening of subsequent pregnancies. Also, if the couple decides to have more children, costs of lost/delayed years of life for elective abortions of false positives and spontaneous abortions of normal fetuses during screening; in effect, partial compensation with a replacement child. Benefits include averted lifetime healthcare for CF cases, plus averted caregiver costs.			$	$				L
Garber 199827	Screening for CF	Net gain/loss of value	Includes direct costs of testing, abortion or birth, health care and lifetime costs of CF care. Includes indirect costs of lifetime earnings of newborn with or without CF; value depends on whether a woman aborts and later has a “replacement child.”			$	$				L
Rowley 199838	Screening for CF	Cost per QALY gained	Includes costs of screening and diagnostics, procedure costs including termination and delivery costs, and lifetime costs of CF care. Includes costs and benefits for the case when a woman aborts and later has a “replacement child.” Benefits are lifetime QALYs saved for affected child, its mother and father; QALY analysis assumes “replacement child.”	Q		Q		Q		L	L	L
Nielsen 200234	Screening for CF	Net gain/loss of value	Includes direct costs of screening, diagnostics, and abortion. Benefits are lifetime healthcare costs of child with CF. Considers different scenarios for “replacement children.”			$	$				L
Doyle 200326	Screening for CF	Net gain/loss of value	Includes direct costs of screening, diagnostics, and abortion. Benefits are lifetime healthcare costs of child with CF.			$					L
Norman 201235	Screening for CF	Net gain/loss of value	Includes direct costs of screening, diagnostics, abortion, and future IVF. Benefits are lifetime healthcare costs of child with CF.			$					L
Hagard 197631	Screening for Down syndrome	Net gain/loss of value	Includes direct costs of screening and diagnostics. Benefits are lifetime healthcare costs of a child with Down syndrome. Considers the possibility of a “replacement child.”			$					L
Vintzileos 200040	Screening for Down syndrome	Net gain/loss of value	Includes costs of screening, diagnostics, abortion. Benefits are lifetime costs of caring for a child with Down syndrome (direct costs of care plus indirect costs of lost wages).			$					L
Ball 200724	Screening for Down syndrome	Cost per maternal QALY gained	Maternal QALY gains/losses due to: false positive screening result (6 months), procedure-related miscarriage (2 years), pregnancy termination (2 years), Down syndrome birth (mother’s lifetime).	Q	Q					L
Gekas 200928	Screening for Down syndrome	Cost per Down syndrome case detected	Includes costs of screening, diagnosis, abortion. Reports outcomes including unnecessary terminations but does not include this in the cost-effectiveness measure. Considers fetal loss from amniocentesis.			S	S				G
Gekas 201129	Screening for Down syndrome	Cost per Down syndrome case detected	Includes costs of screening, diagnosis, abortion. Reports procedure-related miscarriages but does not include this in the cost-effectiveness measure.			S	S				G
O’Leary 201337	Screening for Down syndrome	Cost per Down syndrome case confirmed	Includes costs of screening and diagnosis only. Reports the number of procedure-related miscarriages but does not include this in the cost-effectiveness measure.			S	S				G
Ohno 201336	Screening for Down syndrome	Cost per maternal QALY gained	Maternal QALY gains/losses due to: procedure-related pregnancy loss (2 years), elective termination (2 years), spontaneous abortion (2 years), false-positive diagnosis followed by birth of normal child (1 year), Down syndrome live birth (mother’s lifetime), birth of normal child (mother’s lifetime).	Q	Q					L
Chamberlain 197825	Screening for neural tube defects	Cost per affected birth averted	Includes screening program cost. Reports affected births averted as well as fetal loss and infant morbidity due to amniocentesis. Does not make a final conclusion because of difficulty in quantifying benefits and including intangibles (e.g., “parental distress”).			S	S				G
Little 201032	Screening for spinal muscular atrophy (SMA)	Cost per maternal QALY gained	Maternal QALY gains/losses due to: procedure-related fetal loss (2 years), pregnancy termination (2 years), giving birth to a child with SMA (mild, severe; mother’s lifetime).	Q	Q					L
Wildhagen 199841	Screening pregnant women to find those who are Fragile X carriers	Cost per case of Fragile X syndrome averted	Includes cost of screening, diagnosis and abortion. Reports the number of births of normal children averted and number of fetal losses due to screening but does not include these outcomes in the cost-effectiveness measure.	S	S	S	S			O	G
Vintzileos 199939	Screening pregnant women for fetuses with Fragile X syndrome	Net gain/loss of value	Includes direct costs of screening and diagnostics. Benefits are lifetime direct and indirect costs incurred by a child with Fragile X syndrome. Reports the number of fetal losses due to screening but does not include this in the cost-benefit ratio.			$	S				L
Musci 200533	Screening for Fragile X syndrome	Net gain/loss of value Cost per maternal QALY gained	Maternal QALY gains/losses due to: false-positive screening test, procedure-related pregnancy loss, elective abortion, giving birth to a child with Fragile X syndrome. QALY time horizons not reported.	Q	Q					L?
	Antiretroviral therapy for HIV-Infected Pregnant Women
Marseille 199966	Antiretroviral therapy for HIV-infected pregnant women	Cost per newborn DALY averted	Includes program cost plus lifetime cost of treating an HIV-infected child. Newborn DALYs averted when mother-to-child HIV transmission is averted.			Q					L
Shah 201167	Antiretroviral therapy for HIV-infected pregnant women	Cost per newborn DALY averted	Includes program cost plus lifetime cost of treating an HIV-infected child. Newborn DALYs averted when mother-to-child HIV transmission is averted.			Q					L
Kuznik 201265	Antiretroviral therapy for HIV-infected pregnant women	Cost per newborn DALY averted	Includes program cost plus lifetime cost of treating an HIV-infected child. Newborn DALYs averted when mother-to-child HIV transmission is averted.			Q					L
Fasawe 201362	Antiretroviral therapy for HIV-infected pregnant women	Cost per newborn DALY averted Cost per maternal LY gained	Includes program cost plus lifetime cost of treating an HIV-infected child. Newborn DALYs averted when mother-to-child HIV transmission is averted. Maternal LYs gained when an HIV-infected woman remains on antiretroviral therapy after pregnancy.	S		Q				L	L
Gopalappa 201463	Antiretroviral therapy for HIV-infected pregnant women	Cost per newborn infection averted Cost per child and adult HIV transmission averted	Includes program cost plus lifetime cost of treating an HIV-infected child. Newborn infections averted when mother-to-child HIV transmission is averted. Adult and child HIV transmissions averted when HIV-infected women remain on antiretroviral therapy after pregnancy.	S		S		S		L	L	L
Ishikawa 201464	Antiretroviral therapy for HIV-infected pregnant women	Cost per newborn QALY gained Cost per child and adult QALY gained	Includes program cost plus lifetime cost of treating an HIV-infected child. Newborn QALYs gained when mother-to-child HIV transmission is averted. Additional (adult partner) QALYs gained when an HIV-infected woman remains on antiretroviral therapy after pregnancy.	Q		Q		Q		L	L	L
INTERVENTIONS TARGETED TO WOMEN OF CHILDBEARING AGE, AND MEN
Direct Effects: Increase Fertility
	IVF and Assisted Reproduction Technologies (ARTs)
Mersereau 2008110	Preimplantation genetic screening for aneuploidy	Cost per healthy infant	Includes costs of IVF, genetic screening, abortion, and miscarriage. Health benefit is birth of a healthy infant within four IVF cycles.	S		S				O	O
Mol 1997111	Two methods for treating ectopic pregnancy (conservative surgery alone, and salpingectomy followed by IVF and embryo transfer)	Cost per subsequent intrauterine pregnancy	Includes costs of surgery and salpingectomy, procedure complications, IVF, and repeat ectopic pregnancy. Health benefit is an intrauterine pregnancy (either spontaneous or through IVF) within three years.	S						O
Fiddelers 2009107	Seven IVF strategies	Cost per live birth	Includes costs of IVF, pregnancy and birth until age 6 weeks. Health benefit is a live birth within three VF cycles.			S					O
Finger 2010108	Two IVF strategies	Cost per live birth	Includes costs of IVF. Health benefit is a live birth within three IVF cycles.			S					O
Griffiths 2010109	Two IVF strategies	Cost per live birth	Includes costs of IVF, pregnancy, embryo reduction, abortion, ectopic pregnancy, stillbirth, live birth, neonatal care (as a function of newborn birth weight), and ovarian hyperstimulation syndrome. Health benefit is a live birth within one, two, or three IVF cycles. Also reports number of miscarriages, stillbirths and LBW infants.			S	S
Moolenaar 2011112	Three IVF strategies	Cost per live birth	Includes costs of IVF. Health benefit is a live birth within one year.			S					O
van Loendersloot 2011114	Three IVF strategies	Cost per live birth	Includes costs of IVF and miscarriage. Health benefit is live births within 5 years.			S					O
Scotland 2011113	Comparison of single vs. double embryo transfer	Cost per woman’s QALY gained	Includes costs of IVF, miscarriage, treatment of ovarian hyperstimulation, still birth, neonatal death, neonatal care, child health care up to age 18 (healthy or disabled). Women’s QALY gains/losses associated with the following states: infertile but wanting a child, undergoing IVF, discontinuation of IVF due to lack of success, successful healthy birth (singleton or twins), successful birth of a child with a disability (singleton or twins, with one or two of the children disabled), stillbirth, neonatal death.	Q	Q					O
Vitek 2013116	Two IVF strategies	Cost per live birth	Includes costs of IVF. Health benefit is a live birth in one or two IFV cycles.			S					O
Verhoeve 2013115	Tubal patency testing and treatment	Cost per live birth	Includes costs of tubal patency testing and treatment, and IVF. Health benefit is measured as live births within 3 years.			S					O
	Male fertility treatments for men who have had vasectomies
Heidenreich	Assisted reproductive	Cost per live birth	Includes costs of ART, surgery,		S	S	S		S		O
78 2000	technology (ART), surgical vasectomy reversal		procedural complications, delivery, and therapeutic abortion. Health benefit is measured as any live births that subsequently occurred (retrospective study). The study also reports the number of therapeutic abortions that occurred.
Meng 200581	ART, two methods of surgical vasectomy reversal	Cost per pregnancy	Includes ART and surgery costs. Health benefit is a successful pregnancy within four ART cycles.	S						O
Hsieh 200779	ART, surgical vasectomy reversal	Cost per pregnancy	Includes ART and surgery costs. Health benefit is a successful pregnancy within three years.	S						O
Lee 200880	ART, surgical vasectomy reversal	Cost per live delivery	Includes direct costs of ART and surgery and indirect costs of lost productivity, complications for men due to ART or surgical vasectomy reversal, complications for women undergoing IVF, and costs of multiple gestation pregnancies. Health benefit is measured as a live birth.			S					O
	Chlamydia screening and treatment
Hu 200458	Chlamydia screening of women	Cost per QALY gained among women Reductions in neonatal and male partner conditions and costs	Includes direct costs of screening program, and treatment of chlamydia and complications including ectopic pregnancy and tubal infertility. Calculates total discounted QALYs for a woman in her lifetime. QALY decrements associated with asymptomatic chlamydia, pelvic inflammatory disease (PID), chronic pelvic pain, ectopic pregnancy, and infertility. QALY decrement for infertility assumed to last until age 50. Other QALY decrements calculated over shorter time (e.g., chronic pelvic pain – 5 years; acute PID – 11 days). Study also reports cases of urethritis, epididymitis, neonatal conjunctivitis and neonatal pneumonia averted.	Q		S		S		L	G	L
de Vries 200655	Chlamydia screening of men and women, and their partners	Cost per major outcome averted among women, newborns and men	Includes direct costs of screening program, and direct and indirect costs for treatment of chlamydia and complications in women, newborns and men. Major outcomes in women: infertility (which may have an associated IVF cost), ectopic pregnancy, PID, and chronic pelvic pain. Major outcomes in newborns: neonatal conjunctivitis, neonatal pneumonia. Major outcomes in men: epididymitis.	S		S		S		O	G	O?
Adams 200754	Chlamydia screening of men and women	Cost per QALY gained Cost per major outcome averted in women, newborns and men	Includes direct costs of screening program and treatment of chlamydia and complications in women, newborns and men. Major outcomes in women: symptomatic PID, ectopic pregnancy, tubal factor infertility. Major outcomes in newborns: neonatal conjunctivitis, neonatal pneumonia. Major outcomes in men: epididymitis. QALY decrements associated with occurrence of major outcomes.	Q		Q		Q		O	G	O
Roberts 200760	Chlamydia screening of men and women	Cost per major outcome averted in women, newborns and men	Includes direct costs of screening program and treatment of chlamydia and complications in women, newborns and men. Major outcomes in women: PID, ectopic pregnancy, infertility. Major outcomes in newborns: “neonatal complications.” Major outcomes in men: epididymitis.	S		S		S		L	G	L
Deogan 201056	Chlamydia screening of men and women, and their partners	Cost per QALY gained among women, newborns and men	Includes direct costs of screening program and treatment of chlamydia and complications in women, newborns and men. Women’s QALYs: PID, infertility, chronic pelvic pain, ectopic pregnancy, cervicitis. Men’s QALYs: urethritis, epididymitis. Newborn QALYs: conjunctivitis, neonatal pneumonia.	Q		Q		Q		L	G	L
Turner 201161	Chlamydia screening of men and women, and their partners	Cost per infection diagnosed	Considers costs of screening and partner notification. Outcome is chlamydia infections diagnosed.	S				S		O		O
Huang 201159	Chlamydia screening of women	Cost per case of PID averted	Includes costs of chlamydia treatment and complications, including PID and its sequelae (ectopic pregnancy, chronic pelvic pain, infertility, use of fertility services). Outcome is cases of PID averted.	S						L
Gillespie 201257	Chlamydia screening of men and women, and their partners	Cost per major outcome averted in women, newborns and men Cost per QALY gained	Includes direct costs of screening program and treatment of chlamydia and complications in women, newborns and men. Major outcomes in women: infertility, ectopic pregnancy, PID. Major outcomes in men: epididymitis. Major outcomes in newborns: neonatal conjunctivitis and pneumonia. QALY losses due to all of these outcomes.	Q		Q		Q		L	G	L
Direct Effects: Decrease Fertility
	Contraception (including emergency contraception)
Chiou 200372	Nine contraception methods for women	Cost per % reduction in pregnancy rate	Includes cost of contraception methods and their side effects, plus cost of contraception failures including ectopic pregnancy, spontaneous and induced abortion, and delivery. Effectiveness is measured as reduction in pregnancy rate. Analysis also includes side effects such as venous thromboembolism, urinary tract infections, and postoperative complications.	S	S					O
Sonnenberg 200473	Thirteen contraception methods for women	Cost per woman’s QALY gained	Includes cost of contraception methods and their complications (PID, infections, menstrual disorders), costs of contraception failure (ectopic pregnancy, spontaneous or elective abortion, delivery of normal or premature newborn) plus costs of longer term health effects caused by contraception (increases in cardiovascular risk and several types of cancer). Short-term and long-term QALYs variously associated with these contraception method complications, contraception failure, and longer term health effects.	Q	Q					L
Trussell 200974	Sixteen contraception methods, for men and women§	Cost per % reduction in pregnancy rate	Includes cost of contraception methods and their side effects, plus cost of contraception failures including ectopic pregnancy, spontaneous and induced abortion, and delivery. Effectiveness is measured as reduction in pregnancy rate.	S	S					O
Bayer 201371	Emergency contraception	Cost per woman’s QALY gained	Includes costs of emergency contraception and its failure (ectopic pregnancy, spontaneous or elective abortion, vaginal or caesarian delivery). QALY decrements associated with emergency contraception failure and outcomes of unintended pregnancy (ectopic pregnancy, spontaneous or elective abortion, vaginal or caesarian delivery).	Q						L
	Vasectomy		Vasectomy
Seamans 200777	Four vasectomy methods	Cost per couple year of protection	Includes cost of vasectomy methods. Measure of effectiveness is couple year of protection. Raises complication rates of vasectomy in low volume, lower resourced settings but does not include in calculations/outcomes reported.	S				S		O		O
Trussell 200974	Vasectomy	Cost per % reduction in pregnancy rate	Includes costs of vasectomy and postoperative complications. Effectiveness is measured as reduction in pregnancy rate.	S				S	S	O		O
Direct Effect: Decrease Future Chance Pregnancy Ends before Birth
	Preconception diabetes management for women
Scheffler 199287	Preconception care for women with diabetes	Net gain/loss of value	Includes cost of preconception diabetes care, as well as diabetes care during pregnancy. Benefits are hospital cost savings for women and their newborns when a pregnancy does occur (takes into account newborn birth weights).	$		$				G	G
Elixhauser 199385	Preconception care for women with diabetes	Net gain/loss of value	Includes cost of diabetes care program. Benefits are hospital cost savings for women and their newborns when a pregnancy does occur. This includes cost savings for birth defects/newborn complications avoided (hydrocephalus, aorta problems, respiratory distress, etc.) and for potential maternal problems avoided.	$		$				L	L
Herman 199986	Preconception care for women with diabetes	Net gain/loss of value	Includes cost of preconception diabetes care. Benefits are inpatient and outpatient care costs averted for mothers and their infants when maternal and newborn complications are avoided.	$		$				G	G
	Preconception smoking cessation
de Weerd 200488	Counseling on smoking cessation for women desiring to become pregnant	Net gain/loss of value	Includes costs of the counseling intervention. Benefits are hospital cost savings for low birth weight infants in first year of life.			$					O?
Decrease Future Birth Defects/Perinatally Acquired Diseases/Preterm Birth and Newborn Complications
	Preconception folate supplementation
de Weerd 200488	Counseling on folate supplementation for women desiring to become pregnant	Net gain/loss of value	Includes costs of the counseling intervention. Benefits are hospital cost savings for first year of life for avoided births of infants with neural tube defects.			$					O?
	Contraception for HIV-infected women
Reynolds 200675	Family planning programs for HIV-infected women	Cost per HIV-infected birth averted	Includes family planning program cost and, for pregnancies that do occur, cost of a program for prevention of mother-to-child transmission. Health effects are measured as HIV-infected births avoided.			S					O
Shade 201376	Family planning programs for HIV-infected women	Cost per pregnancy averted	Includes cost of the family planning programs. Effectiveness is measured as pregnancies averted. Does not consider HIV vertical transmission.	S						O
	Preimplantation genetic screening
Mersereau 2008110	Preimplantation genetic screening for aneuploidy	Cost per healthy infant	Includes costs of IVF, genetic screening, and abortion. Health benefit is birth of a healthy infant within four IVF cycles.	S		S				O	O
	Preconceptional genetic testing (for women and/or men)
Wildhagen 199841	Preconceptional screening of women to find those who are Fragile X carriers Screening of (nonpregnant) female students to find those who are Fragile X carriers	Cost per case of Fragile X syndrome averted Net gain/loss of value	Includes costs of screening and diagnosis. Benefits are savings in lifetime cost of care for newborns with Fragile X syndrome. Reports the number of births of normal children averted but does not include this in the cost-effectiveness measure.	$		$	S			O	L
Wildhagen 1998106	Preconceptional screening of men and women to find CF carriers Screening of (nonpregnant) students to find those who are CF carriers	Cost per case of CF averted Net gain/loss of value	Includes costs of screening and diagnosis. Benefits are savings in lifetime cost of care for newborns with CF.			$					L
Verheij 1999104	Preconceptional screening of couples to find CF carriers	Net gain/loss of value	Includes costs of CF screening, counseling, prenatal diagnosis, abortion. Benefits are lifetime health care savings associated with CF births avoided.			$	S				L
van der Riet 1997103	Preconceptional DNA testing of women for four monogenic diseases	Net gain/loss of value	Includes costs of DNA tests, abortions, and lifetime care costs for a child with CF, muscular dystrophy, myotonic dystrophy, or fragile X syndrome.	S		$	$	S		O?	L	O?
Warren 2005105	Screening of (nonpregnant) high school students for Tay Sachs and CF	Cost per carrier detected Cost per potential future affected birth avoided	Includes costs of screening, counseling, and diagnosis. Effectiveness is measured as number of CF and Tay Sachs carriers detected, and future affected births avoided.	S		S		S		O	O	O
INTERVENTIONS WITH INDIRECT EFFECTS ON CURRENT AND FUTURE FERTILITY AND CHILDBEARING
Kim 200890	HPV vaccination of girls and women	Cost per female QALY gained	QALYs gained by avoidance of cervical and other HPV-related cancers in women, genital warts, and respiratory papillomatosis.	Q						L
Vanni 201291	HPV vaccination of preadolescent females	Cost per female QALY gained	QALYs gained by avoidance of cervical cancer, genital warts.	Q						L
Ginsberg 201389	HPV vaccination of girls	Cost per female DALY averted	DALYs associated with cervical cancer.	Q						L
	Male HPV vaccination
Kim 200792	Addition of boys to HPV vaccination programs	Cost per female LY gained	LYs gained by avoidance of cervical cancer due to diminished HPV transmission from vaccinated males.	S						L
Kim 200993	Addition of boys to HPV vaccination programs	Cost per female and male QALY gained	QALYs gained by avoidance of cervical cancer, penile cancer and several other cancers, genital warts, and respiratory papillomatosis.	Q				Q		L		L
	Hysterectomy for benign uterine conditions
Hirst 200896	Hysterectomy for uterine fibroids	Cost per woman’s QALY gained	The women chose hysterectomy instead of a procedure (uterine artery embolization) that would preserve their fertility, so they forewent the ability to have future children (but only 70% of these women said they did not want children). Women’s QALYs associated with fibroid symptoms, procedure complications (including death) and quality of life after hysterectomy, measured up to age of menopause.	Q	Q					O
Sculpher 199897	Hysterectomy for treatment of menorrhagia	Cost per woman’s QALY gained	Costs measured over a 2-year time horizon. Women’s QALYs associated with menorrhagia symptoms, treatment recovery and quality of life after hysterectomy, measured over a 2-year time horizon. Hysterectomy complications were also included.	Q	Q					O?
Beinfeld 200494	Hysterectomy for treatment of uterine fibroids, in women with no desire for future pregnancy	Cost per woman’s QALY gained	Women’s QALYs associated with menorrhagia symptoms and quality of life after hysterectomy, measured up to age of menopause. Hysterectomy complications were also included.	Q	Q					O
Taipale 200998	Hysterectomy for patients with benign uterine disorders	Cost per woman’s QALY gained	Women’s QALYs associated with menorrhagia symptoms and after hysterectomy, measured up to age of menopause. Quality-of-life assessment considers procedure complications.	Q	Q					O
Bhattacharya 201195	Hysterectomy for treatment of menorrhagia	Cost per woman’s QALY gained	The women chose hysterectomy instead of other procedures that would preserve their fertility, so they forewent the ability to have future children. Women’s QALYs associated with menorrhagia symptoms, procedure complications (including death) and quality of life after hysterectomy, measured up to age of menopause.	Q	Q					O
	Breast cancer chemotherapy in premenopausal women
Smith 1993100	Chemotherapy and tamoxifen for premenopausal women with early breast cancer	Cost per woman’s LY gained	Benefits are LYs gained due to treatment.	S						L
Cheng 201299	Chemotherapy and goserelin for premenopausal women with early breast cancer	Cost per woman’s QALY gained	Benefits are QALYs gained due to treatment. Quality of life measured via patient questionnaires. Study notes that fertility is important in this population but infertility risks not specifically cited as disutility (but could have been included in patients’ quality-of-life assessments.)	Q	†					L
	Treatment of testicular cancer
Shibley 1990102	Chemotherapy for treatment of advanced testicular cancer	Net gain/loss of value	Costs include costs of treatment over one year. Benefit measured as discounted future earnings associated with lives saved (calculated based on distribution of men’s age with testicular cancer, and associated total future earnings by age).					$				L
Baniel 1995101	Two methods for treatment of Stage II testicular cancer	Cost per men’s QALY gained	Article does not report health states and associated QALY values that were used. Authors compared two treatments: they gained similar numbers of QALYs, but one cost less. Authors mention that the less costly intervention also did not diminish men’s fertility, though they did not quantify this benefit.					Q				L
	Diagnostic imaging in pregnant women
Kastenber 2013134	Imaging for appendicitis in pregnant women	Cost per QALY gained (women’s and child’s)	Maternal QALYs gained/lost through untreated and treated appendicitis. Child QALYs lost through premature birth, development of leukemia or gained through increasing change of appropriate treatment for mothers with appendicitis and not for those without.	Q	Q	Q	Q			L	L

ART = assisted reproductive technologies, CF = cystic fibrosis, DALY = disability-adjusted life year, HIV = human immunodeficiency virus, HPV = human papilloma virus, IVF = in vitro fertilization, LY = life year, NICU = neonatal intensive care unit, NTD = neural tube defect, PID = pelvic inflammatory disease, QALY = quality-adjusted life year, SMA = spinal muscular atrophy

^*Outcomes considered: MI = mother’s intended outcomes, MU = mother’s unintended outcomes, CI = newborn/child’s intended outcomes, CU = newborn/child’s unintended outcomes, PI = paternal intended outcomes, PU = paternal unintended outcomes. Table entries: “ ” = not reported, S = outcomes reported separately, Q = combined as QALYs or DALYs, $ = monetized for cost-benefit analysis

^**Time horizon over which outcomes are measured: M = maternal outcomes, C = newborn/child outcomes, P = paternal outcomes. Table entries: “ ” = not reported, G = gestation and birth, L = lifetime, O = other. The use of “?” after an entry in the Time Horizon columns denotes that for at least one outcome, the time horizon is ambiguous (e.g., the outcome is perinatal mortality which implies a lost life but the magnitude of the loss is not quantified in life years, QALYs, or DALYs, nor are later lifetime events considered; or the outcome is maternal complications and QALYs are used but it is unclear whether any lifetime effects of a maternal complication are/should be propagated over the woman’s lifetime)

^†Outcome that is listed in the intended column may also reflect the net effect of unintended outcomes given that the net effect is modeled.

^§This study also considered vasectomy; that analysis is described below with other articles evaluating the cost-effectiveness of vasectomy.

Theme 1: Studies were heterogeneous in how they report outcomes

Reviewed studies most frequently included intended intervention outcomes for mothers (66%) and children (67%) and less frequently for partners/fathers (16%). Studies included unintended outcomes much less frequently: mothers (20%); children (16%); partners/fathers (4%). Most studies (62%) reported intended outcomes for only one recipient (mother, child, or partner/father); 27% reported for two recipients and 11% reported for three. For unintended outcomes, 68% reported no outcomes, 26% reported for one recipient, and 6% reported for two or three.

The time horizon for measuring benefits also differed. If reporting on maternal outcomes, 47% of studies used a lifetime horizon, 22% considered gestation and the perinatal period, and 31% used another horizon (e.g., through menopause). These percentages were roughly similar for child outcomes. Though reported less frequently, 63% of studies reporting partner/father outcomes used a lifetime time horizon, with the remainder using another horizon (e.g., period of childbearing for the couple).

Theme 2: Studies select outcomes for inclusion that tend to bias towards finding the intended purpose of the interventions to be cost-effective

We found evidence of systematic bias in the inclusion of outcomes for economic evaluations. Specifically, if the goal of an intervention was to increase fertility or prevent harms to fetuses, studies tended to count newborn QALYs as gains and fetal deaths as losses. If the goal was to decrease fertility or prevent the birth of babies with harmful genetic conditions, QALYs of current and future newborns were typically not counted.

Studies evaluating interventions intended to decrease fertility or childbearing in general or selectively (e.g., vasectomy, abortion, prenatal screening) were less likely to report on intended effects of the intervention on the fetus or child (p<0.018) than those that evaluated interventions intended to increase fertility or childbearing (e.g., folate supplementation, smoking cessation for pregnant women).

We expand on this point qualitatively by comparing studies of screening for and treatment of diabetes in pregnant women^16–23 with those on prenatal genetic testing for pregnant women^24–41.

Better treatment of diabetes and better glucose control during pregnancy has as one of its intended effects the prevention of miscarriage. The 8 studies we reviewed included both maternal and child QALYs and most deemed fetal loss to be a harm, quantified either as a loss of maternal QALYs or as a loss of QALYs of a future child or both.

In contrast, the most frequent intended purpose of prenatal genetic screening is to give prospective parents the opportunity to abort a fetus with a harmful genetic condition. The 18 studies we reviewed tended either to include only maternal QALYs or else to leave the outcome as cost per case of a given genetic condition averted.

Including only maternal QALYs accentuates the estimated cost-effectiveness of prenatal genetic screening; most studies included no loss of QALYs for fetuses with the genetic condition who might have become babies with the condition, thus making such screening appear more cost-effective than it would appear if these QALY losses were included. In their 2013 study, Ohno et al. provided the explicit rationale for this exclusion: “Only maternal QALYs were used because including neonatal QALYs would bias against such diagnostic tools since a possible outcome after diagnosis included termination.”³⁶ Other authors made similar statements in their studies of prenatal genetic screening.^{32, 35} Yet a 2011 study by Ohno et al. on diabetes treatment during pregnancy counted fetal loss as a loss of QALYs to the mother as well as a loss of QALYs for a future child.²⁰

Theme 3: Studies often avoid the challenges of assigning QALYs for pregnancy and fertility-related outcomes by instead reporting costs per intermediate outcome(s)

Many studies avoided expressing outcomes in terms of QALYs and instead reported only intermediate outcomes such as cost per successful fertilization or cost per negative health outcome averted. Among studies that reported an outcome for a given recipient, those reporting intended outcomes utilized QALYs or net benefit as measures of benefit only 59% of the time for mothers, 65% of the time for children, and 50% of the time for partners/fathers. Studies were generally self-consistent with the measure of benefit they used for multiple recipients: 86% of studies reporting outcomes for mothers and children used the same measures of benefit (p<0.0001 with Fisher’s exact test). We found similar within-study consistency for reporting outcomes of mothers and partners/fathers (94%; p=0.001) and for children and partners/fathers (92%; p=0.002).

One rationale for reporting intermediate, separate outcomes may be that, because issues regarding QALY gains and losses due to current and future fertility are thorny, analysts wish to allow policymakers and other readers to draw their own conclusions from the disaggregated set of outcomes resulting from each intervention. However, such an approach makes it difficult to combine parental, fetal, and neonatal outcomes into a single QALY measure, thus hindering comparison of health outcomes across interventions.^{2, 42} Whether aggregating the outcomes to QALYs or keeping them disaggregated, it is important to reiterate (per Theme 2) that studies often failed to report all relevant outcomes.

Theme 4: Even for the same intervention, studies take heterogeneous approaches to outcome inclusion

Studies of the same intervention did not all examine the same outcomes in the same ways. For intervention categories (e.g., prenatal genetic screening) that had at least 6 studies reporting outcomes, we found no category in which all studies reported maternal outcomes or child outcomes with the same measure of benefit (e.g., QALYs). For example, the 11 studies of smoking cessation programs for pregnant women that we reviewed^43–53 variously considered as health outcomes the number of women who quit smoking, the number of low birth weight cases averted, the number of perinatal deaths averted, the number of cases of sudden infant death syndrome averted, and maternal QALYs gained due to smoking cessation. The 8 articles of chlamydia screening and treatment^54–61 (among mainly non-pregnant individuals) variously considered as outcomes the number of chlamydia infections diagnosed, cases of pelvic inflammatory disease averted, QALYs gained among women due to improved health, QALYs gained among women due to improved fertility, QALYs gained among men due to improved health, and newborn QALYs gained due to improved health. Additionally, studies in the same category often measured outcomes over different time horizons. Such variability in outcome reporting makes it difficult to compare the results of different studies of the same intervention.

Theme 5: There exist multiple, competing rationales for whether and how to include fertility-related QALYs and whether these should be for the mother only or for both parents

It is fairly non-controversial that, once babies are born, streams of QALYs attach to them. It is less clear whether and how to count the loss of a current pregnancy or the development of infertility. One approach is to consider the risk of a fetus not coming to term and being born; and, conditional upon the child being born, to consider the probability distribution of the child’s potential health states. One could multiply these probabilities with future streams of QALYs over the child’s life expectancy, conditional upon the child being born (assigning a value of 0 for loss prior to birth) to compute QALYs from lost pregnancies and infertility. This approach assigns QALY losses to interventions that temporarily or permanently reduce fertility or that end pregnancies. One can also consider infertility from the parental perspective – the diminution in quality of life experienced by those desiring to become mothers and fathers when they are unable to do so. These two approaches are not mutually exclusive, though one or the other is used in most of the studies we reviewed.

We compared articles dealing with the treatment of STDs for pregnant women (8 articles on chlamydia^54–61 and 6 articles on HIV and vertical transmission^62–67) to articles dealing with termination or prevention of pregnancies (3 articles on abortion^68–70, 4 articles on contraception in the general population^71–74, and 2 articles on contraception for HIV-infected women^{75, 76}). Studies on STD treatment typically considered increased maternal QALYs due to intended health gains and sometimes also lost maternal QALYs from fetal loss and paternal QALYs gained from improved paternal health (for chlamydia treatment). Additionally, the studies assigned lifetime QALYs gained if a child is born healthy as opposed to HIV-infected or chlamydia-affected. Differential rates of miscarriage were not considered in these studies. In contrast, the studies on the termination or prevention of pregnancies limited themselves to maternal QALYs gained or to intermediate outcomes (e.g., pregnancies prevented). They generally considered complications from procedures as health harms but did not consider maternal QALYs lost due to long-term infertility resulting from ectopic pregnancy or other complications of short-term interventions to reduce fertility. They also did not consider paternal QALYs lost from infertility of a spouse/partner. Most importantly, the studies of STD treatment assigned QALY gains to healthy babies, whereas the studies of pregnancy prevention or termination did not assign a QALY loss to pregnancies averted or terminated.

We examined 2 studies of vasectomy^{74, 77} and 4 studies of vasectomy reversal.^78–81 The reviewed vasectomy studies focused on the intermediate outcomes of pregnancies averted and couple-years of protection. These studies essentially treated vasectomy as a final and irreversible procedure, and did not consider the possible future desire of a man to have a vasectomy reversal. Since successful vasectomy reversal may not be possible, this failure to consider such an outcome could lead to an overly optimistic assessment of the value of vasectomy. The studies on vasectomy reversal used intermediate outcomes of live births or pregnancies. The use of pregnancy as an outcome is notable because not all pregnancies result in live births.

As in Theme 1, a mother’s QALYs were more likely to be counted than a father’s. This makes more sense in some circumstances than in others. For example, women seeking to have children from anonymous sperm donors may achieve QALY gains whereas their donors do not. If a procedure is maternally focused, then by standard practice in CEA, the woman is the patient and hence her QALYs are central. Yet, for many situations, QALY losses attributable to not being able to be a parent or to the loss of an expected child are arguably experienced by both expectant parents. However, including QALY gains/losses for both parents would bias against mothers seeking to have children without a spouse or partner, as the number of QALYs gained would be fewer without the paternal QALYs.

Theme 6: QALYs related to pregnancy and fertility are more likely to be ignored by analyses that consider interventions whose impact on these outcomes is indirect or unintended

Economic evaluations of interventions where the main focus was maternal health or where the pregnancy and fertility effects were indirect (e.g., female HPV vaccination, treatment of testicular cancer) were less likely to report outcomes affecting a fetus or child (p<0.01).

We examined a variety of interventions intended to address health conditions in women and men that could impact pregnancy or fertility outcomes. We first examined studies where improved pregnancy or fertility outcomes were an intended effect beyond the main effect of improving a woman’s health; they included treatment of preeclampsia and diabetes and smoking cessation in pregnant women (3 articles^82–84, 8 articles^16–23, and 11 articles^43–53, respectively) and treatment of diabetes and smoking cessation in women intending to become pregnant (3 articles^85–87 and 1 article⁸⁸, respectively). For the studies targeting pregnant women, maternal QALYs from intended health effects were often considered, while QALYs resulting from having live and healthy offspring were sometimes but not always included. Newborn QALYs were most likely to be included in studies evaluating diabetes control and smoking cessation among pregnant women, less likely in studies of diabetes control or smoking cessation for women intending future pregnancies, and least likely for studies of preeclampsia screening and treatment (where the immediate health benefits to the mother were likely greatest).

For economic evaluations of interventions with only indirect effects on pregnancy or fertility, the inclusion of pregnancy and fertility outcomes was even less likely. We considered studies of female and male HPV vaccination (3 articles^89–91 and 2 articles^{92, 93}, respectively), hysterectomy to address fibroids or other conditions (5 articles^94–98), and chemotherapy for cancer in women of childbearing age (2 articles^{99, 100}) or for men with testicular cancer (2 articles^{101, 102}). The studies of female HPV vaccination considered women’s QALYs and included the direct health benefits of avoiding hysterectomies but not the health benefits of preserved fertility (to women or to resulting QALYs from offspring). The studies of male HPV vaccination considered male QALYs resulting from direct health benefits as well as the prevention of transmission to women whose health gains, also measured in QALYs, were considered. These again included avoided hysterectomies but did not include health benefits from preservation of a woman’s fertility. When hysterectomy for conditions like fibroids was considered, women’s QALY gains from direct health effects were included but fertility effects were not considered, despite the fact that fertility-sparing alternatives were sometimes comparators. One study avoided the issue by focusing exclusively on women with no desire for future children.⁹⁴ The studies of chemotherapy that could damage a current fetus or alter fertility considered only the direct health benefits to men and women.

We compared studies considering genetic testing during pregnancy (18 articles^24–41) and prior to pregnancy (5 articles^{41, 103–106}) along with studies on IVF and other assisted reproductive technology (10 articles^107–116), genetic screening of fertilized embryos prior to IVF (1 article¹¹⁰) and folate supplementation during or prior to pregnancy (7 articles^117–123 and 1 article⁸⁸, respectively). For studies of genetic testing during pregnancy, the cost of having a child with a disabling condition if testing is not done or fails was typically included. Studies often used intermediate outcomes including unnecessary termination of a healthy fetus or else focused exclusively on maternal QALYs. In some studies the concept of a “replacement child” was used to justify not counting lost QALYs of having a child with disabilities relative to having no child at all; this assumes that the couple can conceive a child again relatively quickly if they choose to terminate the current pregnancy due to a suspected genetic disability. Likewise, studies of genetic screening prior to pregnancy focused on intermediate outcomes such as affected births averted or carriers detected. In contrast, studies of IVF or other reproductive technologies focused on intermediate outcomes such as live births (or live healthy births) and generally did not consider the greater risks of low birth weight babies or other conditions resulting from carrying multiple fetuses to term. Somewhat inconsistently, the study examining the value of pre-implantation genetic screening in IVF focused on the outcome of healthy infants. The studies of folate supplementation likewise focused on the intermediate outcome of healthy children born or on QALYs gained. An approach that assigns QALYs based on the health status of children born ranks healthy children higher than disabled children, and disabled children higher than not having or losing a child. Using an intermediate outcome of healthy births essentially ranks births of children in ill health as equal to having no child at all even if the intent of the expectant parents is never to abort.

Theme 7: The above six themes appear in studies from all years of our review, including those recently published

Our literature review included studies published over the past four decades (1974 to 2014). If we limit our analysis to studies published in the past decade (68% of the reviewed studies), we continue to observe the six themes identified above. Thus, the shortcomings of early studies have largely not been remedied by more recent publications.

DISCUSSION

Economic evaluations of interventions that may impact current and future fertility and childbearing consider their outcomes in an inconsistent manner that frequently appears biased towards the interventions considered. Indirect effects on fertility are often ignored. Even for the same intervention, studies take heterogeneous approaches to outcome inclusion, and even recent studies often appear inconsistent and biased. While there have been major trends in standardization for economic evaluation of health interventions generally, standardization is lacking in this area.

Although we performed a targeted rather than systematic review, we looked across a broad spectrum of interventions and economic evaluations of those interventions. The patterns we identified appear repeatedly in the 108 studies we reviewed.

Our review contributes to the prior literature in several ways. We found few prior reviews of interventions with indirect effects on pregnancy and fertility other than Pynna et al.¹²⁴ and none that considered this issue across multiple conditions or interventions. In editorials and reviews of specific conditions or interventions related to pregnancy and fertility, other authors have examined the issue of which outcomes to count and how to value them, and have noted some of the themes we highlight. Mooney and Lange¹²⁵ described a variety of methods for valuing losses of fetuses in current pregnancies in the context of prenatal genetic testing. Ganiats¹²⁶ considered this issue in the context of abortion more generally, noting the inconsistency of counting streams of future costs for children born with disabilities but not QALYs lost or gained from fetuses aborted or for children with disabilities who might have been born in the current pregnancy (due to “false positives”) or in the future. Petrou¹²⁷ and Trussell¹²⁸ echoed earlier work noting that, when considering pregnancies prevented or ended, if streams of averted costs are included then so too should be streams of lost benefits. Garceau et al.¹²⁹ reviewed economic evaluations of assisted reproductive technologies and noted the frequent use of short-term perspectives and intermediate outcomes and the failure to consider long-term health risks of multiple births, especially from preterm delivery, for mother and children. Yi et al.¹³⁰ noted the heterogeneity and selectivity of intermediate outcomes included in economic evaluations of folate supplementation. Pynna et al.¹²⁴ observed that economic evaluations of hysterectomy for benign conditions also are selective about which outcomes they include. Regarding changes in QALYs or other measures of benefits for parents resulting from outcomes for fetuses prior to birth, Simon et al.¹³¹ noted that economic evaluations are conducted from the time point when the intervention is delivered – which is often during pregnancy itself – and hence it is inconsistent not to value miscarriage and stillbirths as losses both from the mother’s perspective and as future losses. Caughey et al.¹³² noted that screening and other interventions offered for fetuses can have psychological benefits to prospective parents by alleviating anxiety and providing greater opportunities to prepare for the child’s birth. Ungar considered a number of related issues and challenges in the context of child health.¹³³

We reviewed many example interventions that collectively span our framework. However, there are numerous additional relevant interventions. One area of future work is to examine interventions conducted during labor and delivery that are intended to improve health outcomes for mother or child. In this area, it may be fruitful to examine whether CEAs capture other features that are relevant to patients’ decisions such as preferences for process aspects (e.g., hospital vs. birthing center; obstetrician vs. midwife), independent of their effects on health outcomes. Another area of future work is to examine in much greater detail CEAs that consider the wide variety of prenatal genetic screening strategies, as some strategies only involve serum screening whereas others involve invasive sample collection that can cause fetal loss.

Our proposed frameworks allow flexibility in characterizing studies while maintaining parsimony. Our framework does not directly reflect the distinction between pregnancies that are desired versus not desired. Such disaggregation may explain differential reporting of outcomes between interventions such as IVF and those such as abortion if only desired pregnancies are intended to be counted. However, this distinction raises other challenges such as whether one individual’s feelings about health outcomes of another (albeit future) individual should alter the normative way QALYs are counted for that individual. This is a clear illustration of the thorny issues facing guideline setting groups in this area.

The selection of health outcomes to include in an economic evaluation of a health intervention that can affect fertility and childbearing can significantly influence the estimated cost-effectiveness of the intervention. Our analysis highlights the need for guidelines that can standardize the inclusion and measurement of the health outcomes of such interventions – the type of guidance that could be included in the update¹⁵ of Cost-Effectiveness in Health and Medicine.² Our work fits within past and ongoing discussions of where to draw the line for which effects and costs to include in economic evaluations of health interventions^4–8 as it has relevance, albeit relatively weak, for all interventions that impact survival of individuals who could otherwise have children now or in the future. In the absence of such guidelines, we recommend that analysts use the framework we have developed here to report in detail all direct and indirect outcomes considered, as well as those not considered, and perform sensitivity analysis with different health outcomes included. Such transparency in outcomes reporting is important in lieu of standardized inclusion and measurement criteria for health outcomes of such interventions.