Changes in Rates of Inpatient Postpartum Long-Acting Reversible Contraception and Sterilization in the USA, 2012–2016

Abstract

Objectives:

To examine recent rates of long-acting and permanent methods (LAPM) of contraception use during delivery hospitalization and correlates of their use.

Methods:

A retrospective cohort study utilizing the 2012–2016 National Inpatient Sample of hospitalizations in the United States of America. The International Classification of Diseases, 9^th and 10^th Revision, Clinical Modification codes were used to identify deliveries, inpatient long-acting reversible contraception (IPP LARC), and postpartum tubal ligation (PPTL). We conducted univariable and multivariable logistic regression to examine associations between demographic, clinical, hospital and geographical characteristics with likelihood of LAPM including IPP LARC and PPTL.

Results:

Our sample included 3,642,328 unweighted deliveries. The rate of IPP LARC increased from 34.6 to 54.9 per 10,000 deliveries (58.7%), while the rate of PPTL utilization decreased from 719.5 to 671.8 per 10,000 deliveries (6.6%) over the study period. In multivariable analysis of LAPM utilization versus neither, cesarean delivery (aOR=7.25, 95%CI: 7.08–7.43) was associated with greater utilization. Native American (aOR=4.01, 95% CI:2.91–5.53) race was associated with increased use of IPP LARC compared to a non-long-acting method of contraception. Age between 18 to 29 years (aOR=6.21, 95%CI: 5.42–7.11) was associated with greater use of IPP LARC versus PPTL. Delivering in a rural hospital ((aOR=0.09, 95%CI: 0.06–0.12) and cesarean delivery (aOR=0.09, 95%CI: 0.06–0.12) were associated with greater use PPTL versus IPP LARC.

Conclusions:

The IPP LARC rate remains at less than 10% the PPTL rates in our study timeframe. The demonstrated variation in uptake of highly effective methods of contraception inpatient after delivery offer possible opportunities for better understanding and improvement in access.

Introduction

Initiation of postpartum contraception is critical in preventing unintended and short-interval pregnancies that are associated with maternal and neonatal morbidity and mortality (Hanley et al., 2017; Janša et al., 2018; Shree et al., 2018). Long-acting and permanent methods of contraception (LAPM) including long-acting reversible contraception (LARC) and female sterilization via tubal ligation are the most effective methods of postpartum contraception and result in significantly fewer short-interval pregnancies (Brunson et al., 2017; Centers for Disease Control and Prevention, n.d.; White, Teal, et al., 2015). However, variation exists in use of postpartum LARC and inpatient postpartum tubal ligation (PPTL) use by age, parity, medical comorbidity, race/ethnicity, hospital, and geographical area (Borrero et al., 2011; Kavanaugh et al., 2015; Khan et al., 2018; Moniz et al., 2017; Potter et al., 2013; Starr et al., 2015; White, Potter, et al., 2015; White & Potter, 2014). Additionally, significant barriers to LAPM access exist, including financial barriers and poor access to health care (Dehlendorf et al., 2010).

Globally, low-resource countries have a greater prevalence of use of LAPM methods in relation to overall contraceptive use compared to the United States where short-acting methods remain the most commonly used methods (Joshi et al., 2015). Worldwide, rates of LARC use are increasing while rates of reliance on permanent contraception are decreasing (Joshi et al., 2015; Nations Department of Economic et al., n.d.; Trends in Contraceptive Use Worldwide 2015 United Nations, n.d.). However, this trend may not be mirrored in the United States. Inpatient immediate postpartum (IPP) LARC use increased from 1.86 to 13.5 per 10,000 deliveries and PPTL remained stable from 711 to 683 per 10,000 deliveries from 2008 to 2013 in the United States (Moniz et al., 2017). Recent policy changes in the United States such as Medicaid LARC reimbursement guidelines and improved contraceptive coverage under the Affordable Care Act as well as new practice guidelines regarding inpatient provision of LARC may additionally impact this trend. Further, whether such increased availability of IPP LARC has changed desire for and provision of PPTL in current practice in the United States is unclear. While IPP LARC and PPTL are similar efficacious, it is unclear whether any further increases in rates of IPP LARC would be as a result of a similar decrease in rates of PPTL or represent increased access to LAPM to additional patients.

Thus, recent changes in rates of inpatient postpartum LAPM use and demographics of use due to such policy and practice changes in the United States are presently unknown. Furthermore, given disparities in maternal and child health outcomes based on clinical and demographic factors, ongoing variation in rates of IPP LAPM used by these factors is important to identify as a potential area for improvement (Howell, 2018). Postpartum contraceptive uptake, particularly of highly effective methods such as LAPM, has been identified as a lever to reduce disparities in perinatal health outcomes (Council on Patient Safety in Women’s Health Care. Alliance for Innovation on Maternal Health (AIM), 2015).

Therefore, the objective of this study was to use recent nationally representative data in the United States to estimate changes in LAPM rates during the postpartum hospitalization period and identify demographic, clinical, and geographical covariates that are associated with use. We hypothesized that inpatient PPTL rates would continue to decline and IPP LARC rates continue to increase as they have for the past twenty years (Chan & Westhoff, 2010; Kavanaugh & Jerman, 2018; Khan et al., 2018; Moniz et al., 2017). We also hypothesized that IPP LARC use would be more common in those women who were younger, Caucasian, delivering vaginally, and delivering in urban hospitals in the North and West.

Materials and Methods

This study was a retrospective cohort study to determine rates of utilization of LAPM after live term deliveries using data from the National Inpatient Sample (NIS) between the years 2012–2016. The NIS is the largest all-payer database and represents a 20% random sample of discharges from hospitals in the United States, calculated from data received. The NIS uses a self-weighting design to provide estimates for the majority of hospitalizations annually and represents over 95% of the U.S. population (Healthcare Cost and Utilization Project [HCUP], n.d.).

Each individual patient record contains up to fifteen procedures and up to thirty diagnostic codes based on the International Classification of Disease, 9^th Revision, Clinical Modification (ICD-9-CM) for years 2012 to the first three quarters of 2015, and ICD-10-CM for the last quarter of 2015 and through 2016. The NIS also contains census demographic data and data from the American Hospital Association Annual Hospital Survey which includes hospital location, size, and teaching status.

Billing codes used to identify the study cohort are included in Supplemental Table 1. Briefly, women who underwent either vaginal or cesarean delivery were identified. Patients who underwent hysterectomy were identified by the DRG code 359, and were excluded from analysis. Those women with an intrauterine fetal demise or preterm delivery were excluded given the potential impact to their contraceptive decision-making. Patient who underwent IUD insertion, contraceptive implant insertion, or postpartum tubal sterilization, including salpingectomy, were identified. Salpingectomy was included as the intent of the procedure is the same as tubal ligation, that of permanent sterilization. The identification of patients who underwent LARC and PPTL in the NIS dataset were similarly previously reported by Moniz et al (Moniz et al., 2017).

Additional variables included in the analysis were age; race; United States census division: New England, Mid-Atlantic, East and West North Central, South Atlantic, East and West South Central, Mountain and Pacific; hospital location: rural, urban-teaching, and urban non-teaching; hospital size: large, medium, and small; patient location relating to population densities: central and fringe metropolitan counties with >1,000,000, counties with metros of 250k–999k people, counties with metros of 50k to 249k people, and micropolitan counties; median household income divided into quartiles; pre-existing comorbidities; and length of hospitalization. Parity is not included in the NIS.

Comorbidities in NIS are identified using the Agency for Healthcare Research and Quality comorbidity software using Elixhauser comorbidity measures (Elixhauser et al., 1998). These measures identify co-existing conditions unrelated to and believed to exist prior to the principal reason for admission. These comorbidities include the following conditions: anemia, cancer, chronic lung disease, coagulation disorders, diabetes mellitus, heart disease, Human Immunodeficiency Virus (HIV) or Acquired Immune Deficiency Syndrome (AIDS), hypertension, hypothyroidism, liver disorders, obesity, neurologic disorders, renal failure, and rheumatoid arthritis or collagen vascular diseases. In our analysis we included individual pre-existing conditions if they occurred in >1% of the total population, otherwise patients with these conditions were excluded from analysis. Additionally, we included the following delivery related comorbidities in our analysis: postpartum hemorrhage, preeclampsia, eclampsia, and transfusion of blood products during hospitalization for delivery.

Patients were stratified into IPP LARC, PPTL, and no long-term contraception groups. Pairwise analysis was performed using Student’s t-test. We then developed three multi-variable logistic regression models. The first model evaluated predictors of the likelihood of receiving LAPM versus non-LAPM, the second evaluated predictors of IPP LARC versus no long-term contraception and excluding patients who underwent PPTL, and the third evaluated predictors of the likelihood of undergoing IPP LARC compared to PPTL.

Statistical significance was set at p <0.05. All statistical analysis was performed using STATA version 14.1 (Stata Corp, College Station, TX). Studies using the NIS database are considered exempt by our institutional review board. All research was completed with the ethical standards of the 1964 Declaration of Helsinki and its later amendments. Informed consent from patients was not pertinent to our study methodology.

Results

During the study period there occurred a total of 3,642,328 unweighted deliveries representing an estimated 18,211,640 deliveries in the United States. The rate of IPP LARC use increased from 34.6 to 54.9 per 10,000 deliveries (58.7%) between 2012 and 2016; while the rate of PPTL utilization decreased from 719.5 to 671.8 per 10,000 deliveries (6.6%) during the same time period.

Demographic characteristics of the LARC and PPTL groups are presented in Table 1. Patients with IPP LARC versus PPTL tended to be younger, with women between 18 to 29 years of age more likely to undergo IPP LARC placement (66.6% vs 35.9%, p<0.001). Women utilizing IPP LARC were less likely to be Caucasian (26.2% vs 47.5%, p<0.001), more likely to be in the lowest median income quartile, (45.1% vs 33.4%, p<0.001), and less likely to undergo cesarean delivery (35.0% vs 77.8%, p<0.001) than those women using PPTL. IPP LARC and PPTL utilization was significantly more common in urban teaching hospitals and large hospitals (87.6% vs 51.1%, p<0.001) and (67.0% vs 58.0%, p<0.001), respectively.

Table 1:

Comparison of Patients Receiving Inpatient Immediate Postpartum Long Acting Reversible Contraception (IPP LARC) vs Postpartum Tubal Ligation (PPTL)

	IPP LARC (n=14,380)	PPTL (n=234,737)	p-value
Age, years
18–29	9,546(66.6)	84,271(35.9)	<0.001
30–34	3,078(21.4)	76,524(32.6)	<0.001
> 35	1,783(12.4)	73,942(31.5)	<0.001
Race
Caucasian	3,770(26.2)	111,500(47.5)	<0.001
African-American	5,029(35.0)	33,802(14.4)	<0.001
Hispanic	4,266(29.7)	58,684(25.0)	0.001
Asian/Pacific Islander	515(3.6)	7,981(3.4)	0.31
Native American	254(17.7)	1,878(0.8)	<0.001
Region
North	3,280(22.8)	30,516(13.0)	<0.001
Midwest	1,986(13.8)	44,131(18.8)	<0.001
South	5,320(37.0)	68,309(29.1)	<0.001
West	3,462(24.1)	49,765(21.0)	<0.001
Hospital Location
Rural	412(2.9)	32,863(14.0)	<0.001
Urban Non-Teaching	1,367(9.5)	81,923(34.9)	<0.001
Urban Teaching	12,603(87.6)	119,951(51.1)	<0.001
Hospital Size
Large	9,639(67.0)	136,147(58.0)	<0.001
Medium	3,992(27.8)	67,370(28.7)	0.03
Small	1,014(7.1)	31,361(13.3)	<0.001
Population Density Around Hospital
Central County	7,810(54.3)	66,900(28.5)	<0.001
Fringe County	2,115(14.7)	49,764(21.2)	<0.001
Large Metro County	2,628(18.3)	50,234(21.4)	<0.001
Small Metro County	624(4.3)	23,708(10.1)	<0.001
Micropolitan	584(4.1)	25,821(11.0)	<0.001
Median Household Income
1st Quartile	6,479(45.1)	78,402(33.4)	<0.001
2nd Quartile	3,224(22.4)	62,675(26.7)	<0.001
3rd Quartile	2,870(20.0)	52,816(22.5)	<0.001
4th Quartile	1,754(12.2)	37,558(16.0)	<0.001
Mode of Delivery
Cesarean	5,030(35.0)	182,625(77.8)	<0.001
Vaginal	9,347(65.0)	52,111(22.2)	<0.001
Length of Stay, days*	2(2–3)	3(2–3)	<0.001

Presented as n (%)

^*Median (interquartile range)

In general, patients undergoing IPP LARC placement had significantly higher rates of both pre-existing comorbidities and delivery-related complications compared to patients undergoing PPTL (Table 2). Most notably: a history of drug abuse (5.5% vs 2.3%, p<0.001), anemia (30.4% vs 22.3%, p<0.001), chronic lung disease (9.9% vs 6.5%, p<0.001), postpartum hemorrhage (7.0% vs 2.5%, p<0.001), and preeclampsia (16.2% vs 9.9%, p<0.001).

Table 2:

Rates of Pre-Existing Comorbidity and Delivery-Related Complications Between Patients Receiving Inpatient Immediate Postpartum Long Acting Reversible Contraception (IPP LARC) or Postpartum Tubal Ligation (PPTL)

	IPP LARC (n=14,380)	PPTL (n=234,737)	p-value
Pre-Existing
Alcohol Abuse/Dependence	199(1.4)	2,140(0.9)	<0.001
Drug Abuse/Dependence	791(5.5)	5,490(2.3)	<0.001
Depression	799(5.6)	9,683(4.1)	<0.001
Psychosis	369(2.6)	3,831(1.6)	<0.001
Anemia	4,381(30.4)	52,362(22.3)	<0.001
Coagulopathy	505(3.5)	5,826(2.5)	<0.001
Heart Failure	267(1.9)	3,643(1.6)	0.06
Hypothyroidism	558(3.9)	10,266(4.5)	0.06
Seizure Disorder	339(2.4)	4,204(1.8)	0.03
Obesity	1,631(11.3)	22,644(9.6)	<0.001
Chronic Lung Disease	1,436(9.9)	15,273(6.5)	<0.001
Chronic Hypertension	1,569(10.1)	21,837(9.3)	<0.001
Pre-gestational Diabetes	918(6.3)	13,694(5.8)	0.12
Delivery Related
Post-partum Hemorrhage	1,008(7.0)	5,773(2.5)	<0.001
Preeclampsia	2,324(16.2)	23,237(9.9)	<0.001
Eclampsia	19(0.1)	92(0.04)	0.10
Transfusion	396(2.8)	4,342(1.8)	0.002

Presented as n (%)

Table 3 shows the multivariable logistic model for use of either PPTL or LARC. The strongest predictor of LARC or PPTL utilization was undergoing a cesarean delivery (aOR=7.25, 95%CI: 7.08–7.43). The strongest predictor for not utilizing either type of contraception was age 18 to 29 years (aOR=0.32, 95%CI: 0.31–0.33) and eclampsia (aOR=0.43, 95% CI: 0.27–0.68).

Table 3:

Independent Predictors of Women Receiving Inpatient Immediate Postpartum Long Acting Reversible Contraception (IPP LARC) or Postpartum Tubal Ligation (PPTL) versus Neither

	Unadjusted OR (95%CI)	Adjusted OR (95%CI)
Age, years
18–29	0.42(0.40–0.43)	0.32 (0.31–0.33)
30–34	0.66(0.62–0.69)	0.61(0.59–0.63)
> 35*	--	--
Race
Caucasian*	--	--
African-American	1.17(1.15–1.19)	1.21(1.16–1.24)
Hispanic	1.45(1.44–1.47)	1.55(1.51–1.59)
Asian/Pacific Islander	0.61(0.59–0.63)	0.70(0.66-.074)
Native American	1.17(1.11–1.25)	1.32(1.17–1.51)
Region
North	0.80(0.79–0.81)	0.77(0.74–0.79)
MidWest	0.81(0.80–0.83)	0.81(0.78–0.84)
West	0.84(0.83–0.85)	0.77(0.75–0.79)
South*	--	--
Hospital Location
Rural	1.36(1.34–1.39)	1.52(1.45–1.59)
Urban Non-Teaching	0.93(0.92–0.94)	1.02 (1.01–1.05)
Urban Teaching*	--	--
Hospital Size
Large*	--	--
Medium	0.96(0.95–0.97)	0.95(0.93–0.97)
Small	0.96(0.95–0.98)	1.02 (0.99–1.05)
Population Density Around Hospital
Central County*		--
Fringe County	1.18(1.11–1.25)	1.08(1.04–1.11)
Large Metro County	1.05(1.04–1.06)	1.22(1.19–1.26)
Small Metro County	1.14(1.12–1.16)	1.36(1.31–1.42)
Micropolitan	1.32(1.29–1.34)	1.17(1.11–1.22)
Median Household Income
1st Quartile	1.37(1.36–1.39)	1.83(1.77–1.90)
2nd Quartile	1.12(1.10–1.13)	1.64(1.59–1.70)
3rd Quartile	0.87(0.87–0.89)	1.34(1.29–1.38)
4th Quartile *	--	--
Cesarean	7.31(7.22–7.40)	7.25(7.08–7.43)
Length of Stay per Day	1.04(1.03–1.05)	0.99 (0.99–1.00)
Pre-Existing Comorbidity
Alcohol Abuse/Dependence	1.01 (0.95–1.06)	0.98(0.93–1.06)
Drug Abuse /Dependence	1.16(1.13–1.21)	0.98(0.92–1.04)
Depression	1.22(1.19–1.26)	1.03(0.98–1.09)
Psychosis	1.16(1.11–1.20)	1.02(0.96–1.08)
Anemia	1.22(1.21–1.25)	1.01(0.96–1.05)
Coagulopathy	1.14(1.10–1.18)	0.99(0.93–1.06)
Heart Failure	0.96(0.92–1.01)	0.98(0.93–1.03)
Hypothyroidism	1.09(1.06–1.12)	1.00(0.96–1.05)
Seizure Disorder	1.06(1.02–1.10)	0.99(0.94–1.03)
Obesity	1.58(1.56–1.61)	1.01(0.97–1.06)
Chronic Lung Disease	1.13(1.11–1.15)	1.03(0.99–1.06)
Chronic Hypertension	1.15(1.13–1.17)	0.98(0.96–1.02)
Pre-gestational Diabetes	1.28(1.25–1.31)	1.02(0.96–1.09)
Delivery-Related Complications
Post-partum Hemorrhage	0.81(0.78–0.84)	1.05(0.98–1.12)
Preeclampsia	1.14(1.13–1.17)	0.82(0.79–0.85)
Eclampsia	0.50(0.39–0.64)	0.43(0.27–0.68)
Transfusion	1.61(1.56–1.68)	1.04(0.95–1.12)

CI = Confidence Interval

OR = Odds Ratio

^*= referent group

When evaluating the independent predictors of using LARC compared to no contraceptives and excluding women who underwent PPTL, the strongest predictors of LARC utilization were women of minority race or ethnicity: Native American (aOR=4.01, 95%CI: 2.91–5.53), African-American (aOR=2.53, 95%CI:2.28–2.80), and Hispanic (aOR=2.01, 95%CI: 1.82–2.23) (Table 4). Additionally, delivering in rural hospitals was the strongest predictor of not utilizing LARC (aOR=0.14, 95%CI: 0.10–0.20).

Table 4:

Independent Predictors of the Likelihood of Receiving Inpatient Immediate Postpartum Long Acting Reversible Contraception (IPP LARC) compared to non-LARC methods (excluding sterilization)

	Unadjusted OR (95%CI)	Adjusted OR (95%CI)
Age, years
18–29	1.51(1.43–1.58)	1.72(1.52–1.96)
30–34	0.70(0.67–0.75)	1.21(1.05–1.40)
> 35*	--	--
Race
Caucasian*	--	--
African-American	3.38(3.23–3.56)	2.53(2.28–2.80)
Hispanic	1.64(1.55–1.72)	2.01(1.82–2.23)
Asian/Pacific Islander	0.64(0.56–0.72)	0.92 (0.74–1.13)
Native American	2.21(1.85–2.66)	4.01(2.91–5.53)
Region
North	1.62(1.53–1.70)	1.80(1.63–2.00)
MidWest	0.51(0.47–0.54)	0.75(0.65–0.87)
West	0.90(0.85–0.95)	1.03 (0.92–1.16)
South*	--	--
Hospital Location
Rural	0.24(0.21–0.27)	0.14(0.10–0.20)
Urban Non-Teaching	0.19(0.18–0.21)	0.32(0.28–0.37)
Urban Teaching*	--	--
Hospital Size
Large*	--	--
Medium	0.80(0.76–0.85)	0.52(0.47–0.57)
Small	0.43(0.39–0.48)	0.38(0.33–0.44)
Population Density Around Hospital
Central County*		--
Fringe County	0.55(0.52–0.59)	0.66(0.59–0.74)
Large Metro County	0.87(0.82–0.92)	0.85(0.77–0.94)
Small Metro County	0.57(0.52–0.63)	0.60(0.49–0.73)
Micropolitan	0.48(0.43–0.54)	0.93(0.72–1.19)
Median Household Income
1st Quartile	2.03(1.94–2.13)	1.46(1.29–1.65)
2nd Quartile	0.89(0.85–0.95)	1.17 (1.02–1.33)
3rd Quartile	0.72(0.67–0.76)	1.17 (1.02–1.33)
4th Quartile *	--	--
Cesarean	1.23(1.17–1.29)	1.20(1.10–1.30)
Length of Stay per Day	1.06(1.05–1.07)	1.05(1.04–1.06)
Pre-Existing Comorbidity
Alcohol Abuse/Dependence	1.49(1.23–1.82)	0.87(0.69–1.11)
Drug Abuse /Dependence	2.67(2.40–2.94)	1.05(0.85–1.30)
Depression	1.64(1.48–1.81)	0.89(0.77–1.06)
Psychosis	1.81(1.57–2.09)	1.00(0.84–1.29)
Anemia	1.74(1.63–1.87)	1.09(0.96–1.24)
Coagulopathy	1.59(1.41–1.81)	1.01(0.82–1.27)
Heart Failure	1.13(0.96–1.34)	0.82(0.68–0.99)
Hypothyroidism	0.96(0.85–1.08)	0.94(0.80–1.09)
Seizure Disorder	1.40(1.21–1.63)	0.91(0.75–1.10)
Obesity	1.87(1.74–2.01)	1.09(0.95–1.27)
Chronic Lung Disease	1.78(1.64–1.91)	1.09(0.96–1.23)
Chronic Hypertension	1.36(1.27–1.47)	1.10(0.99–1.22)
Pre-gestational Diabetes	1.38(1.25–1.54)	0.85(0.67–1.08)
Delivery-Related Complications
Post-partum Hemorrhage	2.23(2.04–2.44)	1.81(1.54–2.11)
Preclampsia	1.92(1.81–2.04)	1.35(1.19–1.50)
Eclampsia	1.54(0.83–2.87	0.98(0.32–3.08)
Transfusion	2.32(2.02–2.67)	1.12(0.87–1.44)

Patients receiving postpartum tubal ligation were excluded from this analysis

CI = Confidence Interval

OR = Odds Ratio

^*=referent group

Finally, the strongest predictors of LARC utilization relative to PPTL were age between 18 to 29 years (aOR=6.21, 95%CI: 5.42–7.11), Native American race (aOR=2.90, 95%CI: 1.94–4.33), African American race (aOR=2.00; 95%CI: 1.77–2.26), and delivering in the Northern region of the United States (aOR=3.19, 95%CI: 2.83–3.61) (Table 5). Delivering in a rural (aOR=0.09, 95CI: 0.06–0.12) or urban non-teaching hospital (aOR=0.36, 95%CI: 0.31–0.41) and cesarean delivery (aOR=0.14, 95%CI: 0.13–0.16) were strongly associated with PPTL rather than LARC.

Table 5 -.

Independent Predictors of the Likelihood of Receiving Inpatient Immediate Postpartum Long Acting Reversible Contraception (IPP LARC) Compared to Postpartum Tubal Ligation (PPTL)

	Unadjusted OR (95%CI)	Adjusted OR (95%CI)
Age, years
18–29	3.61(3.44–3.79)	6.21(5.42–7.11)
30–34	0.54(0.51–0.57)	1.90(1.63–2.20)
> 35*	--	--
Race
Caucasian*	--	--
African-American	3.09(2.94–3.25)	2.00(1.77–2.26)
Hispanic	1.14(1.08–1.19)	0.94(0.84–1.06)
Asian/Pacific Islander	1.06(0.93–1.19)	1.50(1.17–1.91)
Native American	1.94(1.61–2.34)	2.90(1.94–4.33)
Region
North	2.13(2.01–2.25)	3.19(2.83–3.61)
MidWest	0.61(0.57–0.65)	0.97(0.84–1.12)
West	1.06(1.00–1.12)	1.69(1.48–1.94)
South*		--
Hospital Location
Rural	0.17(0.15–0.20)	0.09(0.06–0.12)
Urban Non-Teaching	0.20(0.19–0.28)	0.36(0.31–0.41)
Urban Teaching*	--	--
Hospital Size
Large*		--
Medium	0.82(0.78–0.86)	0.61(0.54–0.67)
Small	0.44(0.40–0.49)	0.38(0.32–0.44)
Population Density Around Hospital
Central County*		--
Fringe County	0.65(0.61–0.69)	0.55(0.48–0.62)
Large Metro County	0.83(0.78–0.88)	0.58(0.51–0.66)
Small Metro County	0.49(0.45–0.55)	0.35(0.28–0.44)
Micropolitan	0.36(0.32–0.40)	0.62(0.47–0.79)
Median Household Income
1st Quartile	1.48(1.42–1.56)	0.67(0.58–0.78)
2nd Quartile	0.80(0.76–0.85)	0.64(.54–0.74)
3rd Quartile	0.82(0.78–0.87)	0.81(0.70–0.95)
4th Quartile *	--	--
Cesarean	0.16(0.15–0.17)	0.14(0.13–0.15)
Length of Stay Per Day	1.05(1.04–1.06)	1.07(1.06–1.08)
Pre-Existing Comorbidity
Alcohol Abuse/Dependence	1.51(1.23–1.86)	0.89(0.67–1.17)
Drug Abuse /Dependence	2.43(2.19–2.71)	1.03(0.80–1.31)
Depression	1.36(1.23–1.51)	0.84(0.69–1.01)
Psychosis	1.61 (1.38–1.87)	1.09(0.84–1.41)
Anemia	1.45 (1.36–1.55)	1.12(0.96–1.31)
Coagulopathy	1.43(1.26–1.62)	1.02(0.79–1.32)
Heart Failure	1.18(0.99–1.41)	0.83(0.67–1.03)
Hypothyroidism	0.87(0.77–0.98)	0.98(0.81–1.18)
Seizure Disorder	1.35 (1.15–1.57)	0.88(0.70–1.09)
Obesity	1.19(1.10–1.28)	1.14(0.97–1.35)
Chronic Lung Disease	1.61(1.49–1.75)	1.06(0.91–1.22)
Chronic Hypertension	1.20 (1.11–1.30)	1.15(1.02–1.30)
Pre-gestational Diabetes	1.08(0.98–1.21)	0.80(0.61–1.05)
Delivery-Related Complications
Post-partum Hemorrhage	3.01(2.73–3.31)	1.88(1.54–2.29)
Preclampsia	1.73(1.62–1.84)	1.71(1.49–1.96)
Eclampsia	3.46(1.76–6.80)	2.79(0.64–12.23)
Transfusion	1.47(1.27–1.70)	0.88(0.65–1.19)

CI = Confidence Interval OR = Odds Ratio

^*=referent group

Discussion

Our findings demonstrate that IPP LARC rates have dramatically increased (58.7%) from 2012 to 2016 whereas inpatient PPTL rates have slightly declined (6.6%) in the same time period. This mirrors global trends in reliance on LARC compared to permanent contraception (Joshi et al., 2015). While the IPP LARC rate remains at less than 10% the PPTL rates during this time period, this rate represents an increase from the less than 2% rates demonstrated in the 2008–2013 time period (Moniz et al., 2017). Our data demonstrate variations in LAPM use by age, race, mode of delivery, preexisting comorbidity, delivery-related complications, income, hospital, and geographical area. Women receiving IPP LARC tended to have higher rates of pre-existing and delivery-related complications than those women receiving PPTL. Younger age, as demonstrated previously in the literature, and preeclampsia were associated with decreased likelihood of use of IPP LAPM versus non-LAPM of contraception and an increased likelihood of LARC versus both non-LARC (excluding PPTL) and PPTL (Moniz et al., 2017; Starr et al., 2015).

As many of these clinical and demographic factors are also associated with disparities in perinatal health outcomes, improved access to IPP LAPM may be a potential lever to reduce disparities by optimizing the interpregnancy interval as a potentially modifiable risk factor for adverse perinatal health outcomes (Bryant et al., 2019; Howell, 2018; Teal, 2014). For example, in this study, women with preeclampsia or eclampsia were less likely to receive IPP LAPM, despite the recommendation for ideal birth spacing in this population to improve maternal and neonatal health outcomes (Bryant et al., 2019). Increased awareness of the interplay between disparities in perinatal health outcomes and IPP LAPM provision thus offers a potential avenue for improved patient counseling and potential uptake. Both in the United States and globally, clinical and demographic factors such as age, race/ethnicity, and social determinants of health are linked to unmet need for contraception and thus present an opportunity for improvement to reduce disparities in perinatal health outcomes (Arora et al., 2020; Nations Department of Economic et al., n.d.; Ngendahimana et al., 2021; Verbus et al., 2019; Wulifan et al., 2016)

Cesarean delivery was most strongly associated with use of LAPM versus non-LAPM of contraception, largely related to provision of tubal ligation. Surgical access to the fallopian tubes during time of cesarean section may have contributed to this trend. However, disparities in tubal ligation in women delivering by cesarean have been noted due to age, adequacy of prenatal care, marital status, and insurance status, and therefore, clinicians should be mindful of the potential barriers to PPTL including the required Medicaid sterilization consent policy, paternalistic counseling based on an over-emphasis of the risk of regret, and implicit bias in terms of characteristics of women who are done childbearing (Arora, Castleberry, et al., 2018; Block-Abraham et al., 2015; Morris et al., 2019).

Previously, African American women were reported as more likely to undergo PPTL than white women. However, our data and other more contemporary studies demonstrate a shift in the racial/ethnic demographics of women desiring sterilization to the white population (Arora, Wilkinson, et al., 2018; Borrero et al., 2007; Moniz et al., 2017; White & Potter, 2014). In our study, African American and women of other minority racial/ethnic groups were more likely to use LAPM versus non-LAPM methods, largely due to the provision of LARC. Additionally, delivery in an urban/teaching setting was associated with decreased use of LAPM versus non-LAPM methods, though an increased use IPP LARC. Thus, whether the racial/ethnic trends demonstrated are due to racial differences in contraceptive preferences or covariation of race with delivery in a teaching hospital with increased comfort and availability of IPP LARC is unclear (Cole et al., 2019; Dehlendorf et al., 2014; Jatlaoui et al., 2018; Wilkinson et al., 2019). As women of color are at increased risk of short interpregnancy intervals, better understanding IPP LAPM desire and uptake in this population is imperative to reduce disparities in perinatal health outcomes (Bryant et al., 2019).

Strengths of this study include the use of the National Inpatient Sample, which allows for the analysis of a large cohort of women receiving inpatient postpartum contraception. Given that these women represent a cross-section of the United States, the external generalizability of our findings are limited to this population. Compared to prior studies, we also used additional codes for delivery and excluded clinical situations such as fetal demise or preterm birth that may change contraceptive (especially tubal ligation) counseling and utilization (Moniz et al., 2017). Limitations include the fact that the NIS relies on ICD codes. This methodology could lead to erroneous inclusion or exclusion of clinical information such as parity that has relevance in contraceptive decision-making. Furthermore, given variation in timeline of states implementing Medicaid expansion and authorizing reimbursement of IPP LARC, we were unable to study associations of insurance status with LAPM use.

Despite these limitations, our study provides contemporary national estimates of postpartum LARC and tubal ligation in the United States. While changes in uptake of LARC versus tubal ligation continue, the reasons (i.e. patient choice, physician counseling, and/or access) for this trend remain unclear. Variation in uptake of highly effective methods of contraception inpatient after delivery offer possible opportunities for further study of disparities and reduction of barriers to access and counseling in order to reduce unintended short-interval pregnancies and ultimately, improve perinatal health outcomes.

Figure 1 –

Changes in Inpatient Immediate Postpartum Long-Acting Reversible Contraception Utilization from 2012 to 2016 by Geographic Region

Figure 2 –

Changes in Postpartum Tubal Ligation Utilization from 2012 to 2016 by Geographic Region

Significance:

Over the past twenty years, rates of inpatient postpartum tubal ligation (PPTL) have steadily decreased while rates of inpatient immediate postpartum long-acting reversible contraception (IPP LARC) have increased, though recent changes and demographics in use are unknown. We report a 58.7% increase in utilization of inpatient postpartum long-acting reversible contraception and 6.6% decrease in inpatient postpartum sterilization rates from 2012–2016. A shift in the racial/ethnic demographics of LARC versus PPTL use was noted compared to prior time periods. As access to postpartum contraception is an important lever to decrease disparities in perinatal health outcomes, improved and equitable access to IPP LAPM is necessary.