Brief report: Fertility rates in women with intellectual and developmental disabilities in Wisconsin Medicaid

Abstract

Background

Women with intellectual and developmental disabilities (IDD) face stigma and inequity surrounding opportunity and care during pregnancy. Little work has quantified fertility rates among women with IDD which prevents proper allocation of care.

Objective

Our objective was to cross-sectionally describe fertility patterns among women with and without intellectual and developmental disabilities (IDD) in 10-years of Medicaid-linked birth records.

Study design

Our sample was Medicaid-enrolled women with live births in Wisconsin from 2007–2016. We identified IDD through prepregnancy Medicaid claims. We calculated general fertility-, age-specific-, and the total fertility-rates and 95% confidence intervals (95% CI) for women with and without IDD and generated estimates by year and IDD-type.

Results

General fertility rate in women with IDD was 62.1 births per 1000 (95% CI 59.2, 64.9 per 1000 women) and 77.1 per 1,000 for women without IDD (95% CI: 76.8, 77.4 per 1000 women). General fertility rate ratio was 0.81 (95% CI: 0.7, 0.9). Total fertility was 1.80 births per woman with and 2.05 births per woman without IDD (rate ratio: 0.89 95% CI: 0.5, 1.5). Peak fertility occurred later for autistic women (30–34 years), compared to women with other IDD (20–24 years).

Conclusion

In Wisconsin Medicaid, general fertility rate of women with IDD was lower than women without IDD: the difference was attenuated when accounting for differing age distributions. Results highlight the disparities women with IDD face and the importance of allocating pregnancy care within Medicaid

INTRODUCTION

It is difficult to study fertility in women with intellectual and developmental disabilities (IDD) because of underrepresentation in surveys and poor collection of disability statistics in census data [1]. Fertility, or the average number of live births to women in a specified time frame [2], is affected by physiological and behavioral factors [3]. For many people with IDD, there are no biological reasons for lowered fertility, but rather social and behavioral contexts that limit desires and opportunities for pregnancy [4–7].

Medicaid data is an efficient tool to conduct meaningful epidemiological studies on fertility in a population at high risk for poor pregnancy outcomes. Medicaid is a federal- and state-level social safety net program that provides health insurance for low-income and disabled people in the United States [8]. Women of reproductive age are eligible for Medicaid if they meet state-determined disability criteria or have low income. In many states, eligibility thresholds are more generous when a person is pregnant, making pregnancy care a priority within the Medicaid system [9]. Using a Medicaid sample, we can compare women with and without IDD in a similar socio-economic position (as women with IDD are disproportionately low income) and capture the extent to which pregnancy-related services are needed in a population at high-risk of poor health outcomes [10, 11].

Our objective was to use 10 years of birth-record linked Wisconsin Medicaid data to estimate general fertility rate (GFR), age specific fertility rate (ASFR), and total fertility rate (TFR) values between Medicaid enrolled women with IDD and compare to Medicaid enrolled women¹ without IDD. By using a Medicaid sample, we focus on a group with lower income and higher propensity to disability compared to the general population. We investigated differences by year in ASFR and evaluated differences by three prevalent IDD: intellectual disability, cerebral palsy, and autism spectrum disorder (ASD).

MATERIALS AND METHODS

Sample derivation

We used data from the Big Data for Little Kids (BD4LK) a longitudinal cohort of all Wisconsin birth records for live, in-state, resident deliveries during 2007–2016. Birth records are linked to multiple administrative data sources, including Wisconsin Medicaid claims and encounters (hereafter ‘claims’). Based on the mother’s information, birth records were deterministically matched to up to 12 months of maternal Medicaid claims prior to childbirth. We used pre-childbirth claims to calculate whether women were continuously enrolled in Medicaid prior to pregnancy or initiated enrollment in Medicaid while pregnant.

IDD case definition

We screened one year of prepregnancy Medicaid claims through delivery for International Classification of Disease 9 or 10 codes (ICD) for IDD (Supplement 1). Claims included inpatient services, outpatient services, hospitalization, and other services. IDD can be reliably identified in Medicaid data [10, 12] with high positive predictive value [13]. We had five categories of IDD: intellectual disability, genetic and chromosomal abnormalities, cerebral palsy, ASD, and ‘other’ (e.g., fetal alcohol syndrome; full list in supplement 1). We relied on hospitalization claims for some women in 2007–2008 deliveries because claims data were bounded on January 1, 2007, and for some 2013–2014 deliveries because a lack of some pre-childbirth claims we conducted a sensitivity analysis excluding births in those years.

Denominators to calculate fertility rates

We used Medicaid enrollment data from the Wisconsin Department of Health as the denominator for fertility calculations. We received the total number of Medicaid enrolled women overall and by age stratum (15–19 years, 20–24 years, 25–29 years, 30–34 years, 35–39 years, 40–44 years) for each year during 2007–2016 for all women, for women with IDD, and by IDD subtype (identified using the same codes as in BD4LK).

Calculating fertility rate

We restricted our sample to women 15–44 each year because of low numbers of pregnancies in the IDD group for ages <15 and >44. We calculated three measures of fertility: GFR (total number of births / total number of women), ASFR (fertility rate within age stratum), and TFR (average number of children a woman would have over the 30-year reproductive period, calculated by summing ASFRs weighted by number of years in each age stratum). GFR, ASFR, and TFR are technically rates, but since our unit of time was years, they are equivalent to risks. We calculated ASFR and TFR ratios comparing women with IDD to women without IDD over the 10-year period with corresponding 95% confidence intervals (CI) for risk ratios. We calculated ASFR in each year for the IDD group and used linear regression to assess whether there is a linear association between year and the fertility rate for the group with and without IDD, as determined by t-test and corresponding confidence interval. Because our data were bounded by 2007, we restricted our trend analysis to 2009–2016 and ran sensitivity analysis excluding 2013. We examined GFR, ASFR, and TFR within intellectual disability, ASD, and CP subgroups. We conducted analysis using R 4.0.3 and the University of Wisconsin Madison institutional review board approved this study.

RESULTS

Within this Medicaid cohort, there were 1,751 live births to 1,032 unique women aged 15–44 years with IDD and 272,839 live births to 176,659 unique mothers without IDD (Table 1). For women with IDD, 52.6% were consistently enrolled in Medicaid for 12 months prepregnancy and 30.8% were not enrolled in Medicaid until pregnancy. Similarly, 47.9% of women without IDD were consistently enrolled in Medicaid in the year prepregnancy and 35.5% were not enrolled until pregnancy. GFR was slightly lower for women with compared to without IDD. ASFR were highest among women aged 20–24 years and lowest among women aged 40–44 years, regardless of IDD status. Compared to women without IDD, fertility rates were lower among women with IDD in the 15–19 years-old age stratum and similar in other age strata. TFR did not statistically differ between women with and without IDD. We observed minimal change in GFR, TFR, and ASFR in our sensitivity analyses (<10% change; Supplement 2).

Table 1.

Fertility rates for women with and without intellectual and developmental disabilities enrolled in Wisconsin Medicaid, 2007–2016

	Mothers with Intellectual and developmental disabilities N=1032				Mothers without intellectual and developmental disabilities N=176659
	Number of live births	Total enrolled	Fertility rate / 1000 women with IDD	95% CI	Number of live births	Total enrolled	Fertility rate / 1000 women without IDD	95% CI	Fertility rate ratio	95 % CI
General Fertility Rate
	1751	28203	62.1	59.4, 65.0	272839	3541450	77.0	76.8, 77.4	0.81	0.7, 0.9
Age Specific Fertility Rate
15–19	233	7188	32.4	28.3, 36.5	33239	694899	47.8	47.3, 48.3	0.68	0.6, 0.8
20–24	606	5572	108.8	100.6,116.9	94282	773099	122.0	121.2, 122.7	0.89	0.8, 1.0
25–29	486	4756	102.2	93.6,110.8	81042	706030	114.8	114.0, 115.5	0.89	0.8, 1.0
30–34	264	3741	70.6	62.4, 78.8	43932	563581	78.0	77.3, 78.7	0.91	0.8, 1.0
35–39	123	3434	35.8	29.6, 42.0	16847	442977	38.0	37.5, 38.6	0.94	0.8, 1.1
40–44	39	3512	11.1	7.6, 14.6	3497	360864	9.7	9.4, 10.0	1.14	0.8, 1.6
Total Fertility rate (average births per woman)			1.80	1.7, 1.9			2.05	1.9, 2.2	0.89	0.5, 1.5

IDD: Intellectual and developmental disability

CI: Confidence interval

GFR decreased overtime with a peak of 89.5 births per 1000 women with IDD in 2008 and a low of 36.8 births per 1000 women with IDD in 2016. AASFR decreased among women with IDD aged 15–19 years (8.8 fewer births per 1,000 women with IDD per year, 95% CI: −12.4, −5.3 births per 1,000 women with IDD), and 20–24 years (15.8 less births per 1,000 women with IDD per year, 95% CI: −22.9, −8.8 births per 1,000 women with IDD) (Figure 1). There were decreases in ASFR among women without IDD in the 15–19- and 20–24-years age strata (4.6 fewer births /1,000 and 2.7 fewer births /1,000 women respectively). The TFR in the IDD group was 2.38 children per woman in 2009 and fell to 1.13 children per woman in 2016. TFR in the group without IDD was 2.22 children per woman in 2009 and fell to 1.91 children per woman by 2016.

Figure 1.

Age specific fertility rates and corresponding 95% confidence intervals for women with intellectual and developmental disabilities enrolled in Wisconsin Medicaid by year, 2009–2016

Each box represents one year with the X axis representing age categories and the Y axis being births per 1000 women enrolled in Medicaid.

GFRs were 48.5 births per 1,000 women with intellectual disability (95% CI: 44.6, 52.5 births per 1,000 women with intellectual disability), 24.6 births per 1,000 autistic women (95% CI: 20.8, 28.4 births per 1,000 autistic women), and 29.2 births per 1,000 women with cerebral palsy (95% CI: 25.8, 32.6 births per 1,000 women with cerebral palsy). Younger autistic women and women with cerebral palsy had lower ASFRs from ages 15–29 compared to women without IDD or women with intellectual disability (Figure 2). Among autistic women, the 30–34 years-old stratum had the greatest ASFR, whereas ASFR was greatest in the 20–24-year-old stratum for all other IDD types. TFR was 1.5 children per woman with intellectual disability, 0.8 children per autistic woman, and 0.7 children for women with cerebral palsy.

Figure 2.

Age specific fertility rates and corresponding 95% confidence intervals for women enrolled in Wisconsin Medicaid from 2007–2016, by intellectual and developmental disability type

Each box represents a non-exclusive IDD type with the X axis representing age categories and the Y axis being births per 1000 women enrolled in Medicaid

IDD: intellectual and developmental disability

ID: Intellectual disability

CP: Cerebral palsy

Not enough births for the autism or cerebral palsy group to estimate fertility in the 40–44 age group.

DISCUSSION

In the Wisconsin Medicaid system, the GFR of women with IDD was 19% less than the fertility rate of women without IDD. However, the TFR- which accounts for differing age distributions-did not differ statistically, suggests fewer differences in fertility rates than reported in other studies. Women with IDD had a lower fertility rate compared to women without IDD in the youngest age stratum with no difference for the older age strata. Fertility rates were heterogenous across IDD, with reduced fertility rates and later peak fertility for autistic women and women with cerebral palsy.

The fertility rates reported in this study among women with IDD are higher than Brown et. al.’s estimates using an Ontario health care system [14]. The fertility rates we found in women without IDD was also considerably greater than the rates found by Brown et al. Our increased rates compared to the Ontario study may reflect differences in how the cohorts were created or, differing reproductive patterns for women with IDD in Wisconsin compared to Ontario. Medicaid serves low income and disabled people, with a focus on pregnant individuals; Medicaid does not cover the full range of potential incomes. In contrast the Ontario system does not determine eligibility based on income and pregnancy, rather, they have a universal health care system that covers all women with additional social supports for those in need. Therefore, our sample, especially the non-IDD comparison group, was of lower income by definition of entry into Medicaid as compared to the full Ontario sample. Additionally, Brown et al., had a slightly different definition of IDD, having not included cerebral palsy; however, if we excluded cerebral palsy from our sample the fertility estimates would be larger. We only examined claims for IDD in the year prior to pregnancy, while Brown et al. looked back further at all available claims. Patterns by age also differed; the youngest age stratum with IDD in the Ontario sample had the greatest relative risk. ASFR differences may be attributable to differing age distributions, or health system policy. Brown et al., also had a larger sample (N=8,919) which improved their precision and ability to find smaller effects. Our findings are important to understanding fertility within Medicaid but are not generalizable outside of a low-income population.

We observed declining fertility rates in women with and without IDD over our study period, driven largely by reduced fertility among women aged 20–24 years. With these data, the relative contributions of healthcare policy, biology, and social factors that reduce fertility among women with IDD in Medicaid cannot be discerned. An analysis of fertility rates calculated using data from all Wisconsin birth records (with and without Medicaid) demonstrated a decrease from 77.2 births per 1,000 women in 2009 to 59.7 births per 1,000 women in 2016 for women 20–24 [15] which suggests some of the reduction in fertility rates over time in the Medicaid-covered population is representative of the larger base population. Over this time period, Wisconsin Medicaid shifted from ICD-9 to ICD-10 [16, 17] in documenting claims, a change which may have affected diagnostic practice for IDD [18]. Because of these changes, there are temporal patterns in who is identified with which type of IDD [19, 20], which, if differential by pregnancy, could contribute to the declining fertility rate we saw. Further, both state and national health care policy impact fertility within the Medicaid system, For example, the Affordable Care Act allowed children to stay on their parent’s private insurance plans until age 26 starting in 2010 and offered more affordable insurance plans for uninsured consumers starting in 2014 [21]; these policy changes could have impacted observed fertility rates.

Our finding of no difference between women with and without IDD in fertility rate in the older age strata is in contrast with past research and other populations and may be a result of the socioeconomic differences between Wisconsin’s Medicaid population and populations in other studies. Low income women, which includes those enrolled in Medicaid, are more likely to have children at younger ages as compared to higher income women [22]. Therefore, our low-income Medicaid sample may miss higher income women who have increased fertility at older ages. It will be important to replicate this work in US women with private insurance.

We saw heterogenous patterns in fertility by IDD type. Women with different types of IDD likely differ in whether they qualify for Medicaid, affecting inclusion in our data [23]. Cerebral palsy had lower ASFR in all age strata compared to women without IDD. It is possible that women with cerebral palsy in Medicaid are more severely impacted physically by cerebral palsy compared to women with cerebral palsy not on Medicaid [24], and women with severe physical disability are less likely to become pregnant compared to women with less severe disability [25]. Our finding that autistic women had peak fertility later compared to other IDD types could be a result of social impairment inherent to autism and associated delays in forming intimate relationships [26] that lead to pregnancy. It may also be an artifact of enrollment patterns, with autistic women entering Medicaid at later ages [23].

Many parents of children with IDD, disability support staff, and healthcare professionals report reservations about their children or clients having children [27, 28], and these concerns may subtly or explicitly prevent women with IDD from having children. In addition, women with IDD may be more likely to have an unintended pregnancy [29], furthering the need for appropriate sexual education and reproductive health care. People with IDD have the same sexual and reproductive rights as peers, yet receive less sexual education and have fewer early prenatal care compared to peers [30]. Because of these social factors, sexual and reproductive health care for women with IDD should be informed by the history of sterilization and stigma that hamper rights of women with IDD. We note that many other marginalized and low-income populations have experienced similar inequities surrounding reproductive rights and sexual education [31] and it may be the case that targeting interventions broadly to low-income people may be the most useful approach to improving outcomes for pregnant people with IDD.

We were limited by a lack of demographic data in the Medicaid enrollment denominator, preventing us from assessing difference by race, urbanicity, and other factors. We did not have data on miscarriage and termination, which prevent us from measuring fecundity. Our results generalize only to women who receive care via the Wisconsin Medicaid system and may not be generalizable to women with IDD not enrolled in Medicaid. Wisconsin Medicaid has different income requirements and serves a population with an income distribution that may not be comparable to other states or regions [32]Women who entered Medicaid at time of pregnancy had less opportunity to receive IDD claims and there maybe under-identified IDD in claims that lead to misclassification.

CONCLUSION

In Wisconsin Medicaid, many women with IDD are enrolled during pregnancy, yet these women have 0.81 times the GFR of women without IDD. Fertility rates were similar after adjustment for differing age distributions between women with and without IDD. Lower general fertility in women with IDD may be due to social and biological factors, including age-related selection into the Medicaid system. Medicaid is a vital health insurer for women with IDD and increased pregnancy-specific care may be warranted.

Highlights:

• In the Wisconsin Medicaid system, the General Fertility Rate of women with intellectual and developmental disabilities was 19% less than the fertility rate of women without intellectual and developmental disabilities

• The Total Fertility Rate- which accounts for differing age distributions- did not differ, suggesting fewer differences in fertility rates than reported in other studies.

• Fertility rates were heterogenous across intellectual and developmental disabilities, with reduced fertility rates and later peak fertility for autistic women and women with cerebral palsy