Factors Associated with Choice of Sterilization Among Women Veterans

Abstract

Background: We sought to compare associations of contraceptive preferences, beliefs, self-efficacy, and knowledge with use of sterilization versus other methods of contraception.

Materials and Methods: This is a secondary analysis of a telephone-based survey of a nationally representative sample of women Veterans not desiring future pregnancy. Contraceptive method used at last sex was categorized as female sterilization, long-acting reversible contraception (LARC), short-acting methods, or nonprescription methods/no method. Multinomial regression models were performed to compare the association between independent variables (contraceptive preferences, beliefs, self-efficacy, and knowledge) and use of sterilization versus other contraceptive methods.

Results: Six hundred twelve women Veterans aged 18–44 years who were sexually active with men, had no history of hysterectomy or infertility, did not desire future pregnancy, and were not using male sterilization as their method of contraception were surveyed. A total of 208 women Veterans reported using female sterilization (34.0%). While method effectiveness was rated as extremely important by the majority of participants, there was no association between perceiving method effectiveness as extremely important and method selected in adjusted multinomial models. Women Veterans were more likely to use sterilization compared to hormonal methods of contraception if they reported that lack of hormones was an extremely important contraceptive method characteristic (aRRR 3.69, 95% CI 1.94–7.03). Women Veterans who strongly agreed with the belief that birth control decisions are mainly a woman's responsibility were less likely to use sterilization compared to LARC (aRRR 0.54, 95% CI 0.29–0.98).

Conclusion: Associations between contraceptive preferences, beliefs, self-efficacy, and knowledge and use of sterilization in a population of women Veterans not desiring future pregnancy are complex, and decisions may not solely be driven by desire to select a highly effective method.

Introduction

Recent studies have focused on the impact of preferences, beliefs, self-efficacy (confidence in one's ability to complete a task), and knowledge on contraceptive decision-making.^1–6 This emerging body of literature highlights that contraceptive method selection is influenced by multiple factors and considerations, beyond the method's effectiveness in preventing pregnancy. For example, in the Contraceptive CHOICE Project, a survey of 2,590 women demonstrated that effectiveness, safety, side effects, affordability, whether a method is long-lasting, and whether a method requires frequent adherence all influenced decision-making to various extents.¹

Patient-centered contraceptive counseling thus requires proactive assessment of patient preferences for method characteristics, rather than assuming women make contraceptive decisions based solely on effectiveness, as is emphasized in a strict “tier-based” counseling approach that prioritizes discussion of the effectiveness tier of a method.

It remains unclear, however, how and to what degree varied priorities impact selection of female sterilization, which is a unique method of birth control in that it is permanent and requires a surgical procedure. One prior small study suggested that lack of hormones, avoidance of the need for continued decision-making around child-bearing, and lack of information regarding reversible methods play a role in choice of sterilization compared to long-acting reversible contraception (LARC).⁴ Cultural background and current life situation impacted attitudes toward sterilization in another study.⁶ Prior qualitative analyses demonstrate that some providers discourage patients from choosing sterilization in favor of less invasive, nonpermanent methods.^7,8

However, it remains unknown how contraceptive preferences, beliefs, self-efficacy, and knowledge differ in women among using sterilization compared to all other methods of contraception. Therefore, a better understanding of these factors is important to facilitate patient-centered counseling.

To address this knowledge gap, we used data from The Examining Contraceptive Use and Unmet Need (ECUUN) study, a telephone-based, cross sectional survey with a nationally representative sample of women Veterans who use the Veterans Affairs (VA) Health Care System for primary care. Within the VA, women Veterans have access to comprehensive health care, including access to low or no cost contraception.⁹ As many system and cost barriers are thus minimized for Veterans, the ECUUN study represents a unique opportunity to isolate patient-level factors impacting sterilization choice in the largest integrated national health system in the United States. Our goal was to conduct a secondary analysis of ECUUN data to compare associations of preferences, beliefs, self-efficacy, and knowledge with use of female sterilization versus other methods in women Veterans not desiring future pregnancy.

Materials and Methods

Study design and population

The ECUUN study was a national telephone survey of a random sample of 2,302 women Veterans across all regions and Veterans Integrated Service Networks (VISNs) in the United States to assess contraceptive use, pregnancy history, and experiences with VA reproductive health care. Full survey methodology has been previously published.¹⁰ In brief, women Veterans between the ages of 18 and 44 years with at least one VA primary care visit in the past year were randomly sampled using VA administrative data; comparisons between survey participants and nonparticipants from the sampling frame indicated that the study sample is representative of the larger population of reproductive-aged female VA users. Interviews were conducted from April 2014 to January 2016 by trained interviewers using computer-assisted telephone interview technology. Interviews lasted an average of 45 minutes and participants received a $30 honorarium. The study was approved by both the VA Pittsburgh and University of Pittsburgh Institutional Review Boards.

For this analysis, the study population was limited to those women who were heterosexually active in the past year (n = 491 excluded), had no history of hysterectomy (n = 167 excluded) or infertility (n = 169 excluded), and reported that they desired no future pregnancies (n = 760 excluded). We also excluded women reporting use of male sterilization (n = 103) at last sex as our study attempts to understand correlates of woman-controlled contraceptive methods.

Measures

The outcome of interest was use of female sterilization at last sex compared to three other categories of contraception: LARC (implant or intrauterine device); short-acting hormonal methods (injectable, pill, patch, and vaginal ring), or nonprescription/none (emergency contraception, male condom, female condom, spermicide, diaphragm, cervical cap, sponge, rhythm, withdrawal, no method). If multiple methods were used, participants were categorized according to the most effective method of contraception.¹¹ Women Veterans using nonprescription methods and those using no method of contraception were combined as they were similar with respect to demographic characteristics as well as their preferences, beliefs, and knowledge. Moreover, the number of women using no method of contraception was small (n = 25).

The independent variables of interest were measures of contraceptive preferences, beliefs, self-efficacy, and knowledge. Preferences and belief items were adapted from the National Longitudinal Study of Adolescent Health; these measures have been used previously in adult as well as adolescent populations.^12–16 Questions regarding preferences asked women to rate the importance of the following factors in contraceptive decisions using 4-point Likert scale: method effectiveness, ease of use, lack of hormones, lack of interrupting sex, and prevention of sexually transmitted infections (STIs). Beliefs about contraception were assessed using 5-point Likert scale and asked women to indicate the extent to which they agreed with the following statements: “it is too much of a hassle to use contraception”; that “when it is your time to get pregnant, you will”; and “it is mainly a woman's responsibility to make decisions about birth control.”

Two items assessed contraceptive self-efficacy using 5-point Likert scale to ask participants how sure they were that they could (1) “plan ahead to have some form of birth control consistently available over the next year” and (2) “use a method of contraception as indicated over the course of the next year.” Self-efficacy questions were based on similar items used in other studies of contraception.^13,17 Variables for preferences, beliefs, and self-efficacy were dichotomized as “extremely important”/“strongly agree”/“very sure” versus all other responses, similar to other published studies assessing contraceptive preferences.^15,18 Questions were asked of participants one at a time over the phone so that they were unaware of the preference, belief, self-efficacy, or knowledge question being asked next, but they were also unable to change an answer as the survey continued.

Given the prevalence of misinformation regarding sterilization as well as lack of knowledge of the similar efficacy alternative of LARC methods,¹⁹ we assessed knowledge about sterilization and LARC with five true/false questions. Knowledge questions were developed by the PI (S.B.), reviewed by national family planning experts, and pilot-tested with women Veterans at the VA Pittsburgh Women's Health Clinic. Knowledge measures were dichotomized as correct versus incorrect/do not know.

Demographic and health variables from the survey were examined as covariates. Demographic variables included age, race/ethnicity, marital status, education level, annual household income, religion, deployment status, branch of military service, additional non-VA insurance, and geographic census region. Health variables included parity, tobacco use, history of medical comorbid conditions that could potentially impact prescription of estrogen-containing contraceptives (e.g., hypertension, migraines), history of mental health condition, and history of military sexual trauma. Further details on nature of questions asked regarding medical and mental health history have been previously published.¹⁰

Data analysis

We first summarized demographic and health variables, overall and by contraceptive method type at last sex (sterilization, LARC, short-acting hormonal method, or nonprescription/none) using ANOVA and χ² tests to test differences across groups. We then assessed the association between the independent variables (dichotomized measures of preferences, beliefs, self-efficacy, and knowledge) and contraceptive method type at last sex using χ² tests. Multinomial logistic regression models were used to compare associations between each independent variable and use of sterilization compared to other contraceptive method types, while adjusting for the demographic and health variables associated with contraceptive method type in bivariate analysis at the p < 0.15 level. Model estimates are presented comparing sterilization to each of the other methods. Adjusted relative risk ratios (aRRRs) and 95% confidence intervals (CIs) were reported. Analyses were conducted using SAS, version 9.4 (SAS Institute, Cary, NC) and STATA 14 (College Station, TX).

Results

Among the sample of 2,302 women Veterans who completed the interview, 612 women met the inclusion criteria and were included in the analytic sample. Of these, 208 women (34.0%) were using sterilization, 153 (25.0%) were using LARC, 116 (19.0%) were using hormonal methods, and 135 (22.1%) were using a nonprescription or no method of contraception at last sex (Table 1). Women using sterilization as their method of contraception were more likely to be older and parous and less likely to be single compared to those who were not using sterilization.

Table 1.

Patient Demographic and Health Factors, Overall and by Sterilization Use, n (%)

	Total	Sterilization	LARC	Hormonal methods	Nonprescription/Nonea
	N = 612 (100%)	n = 208 (34.0%)	n = 153 (25.0%)	n = 116 (19.0%)	n = 135 (22.1%)	p value
Demographic factors
Age, mean (SD)	36.1 (5.1)	37.3 (4.7)	35.2 (5.0)	34.4 (5.7)	36.8 4.8)	<0.0001
Age, %						<0.0001
20–29	10.8	4.3	13.1	20.7	9.6
30–34	26.0	25.0	28.8	31.9	19.3
35–39	32.8	31.3	36.6	23.3	39.3
40–44	30.4	39.4	21.6	24.1	31.9
Race/ethnicity, %						0.26
Non-Hispanic white	52.3	53.4	56.2	56.0	43.0
Non-Hispanic black	31.0	32.7	30.1	26.7	33.3
Hispanic	10.1	8.7	7.2	12.1	14.1
Other	6.5	5.3	6.5	5.2	9.6
Marital status, %						0.0007
Married/living with partner	58.3	63.9	66.0	55.7	43.0
Single	12.9	8.2	11.8	14.8	20.0
Divorced/separated/widowed	28.8	27.9	22.2	29.6	37.0
Education—college degree or above, %	53.1	50.5	52.9	53.4	57.0	0.71
Annual household income, %						0.34
<$20,000	16.3	16.9	12.6	17.2	18.8
$20,000–<$40,000	32.5	29.0	37.7	37.1	27.8
$40,000+	51.2	54.1	49.7	45.7	53.4
Religion, %						0.31
None	18.8	14.9	20.3	22.4	20.0
Christian	68.5	74.0	64.1	62.9	69.6
Other	12.7	11.1	15.7	14.7	10.4
Ever deployed, %	49.8	43.8	57.5	49.1	51.1	0.08
Military branch, %						0.37
Army	49.3	54.3	49.7	45.7	44.4
Navy	22.9	20.2	23.5	20.7	28.1
Marine corps/coast  guard/air force	27.8	25.5	26.8	33.6	27.4
Outside insurance, %	51.1	50.0	56.2	50.9	47.4	0.49
Geographic census regions, %						0.20
Northeast	6.4	6.3	9.2	6.0	3.7
Midwest	19.0	21.2	18.3	23.3	12.6
South	54.1	55.3	50.3	48.3	61.5
West	20.6	17.3	22.2	22.4	22.2
Health/medical factors
Parity, %						<0.0001
0	18.3	7.7	18.3	31.9	23.0
1	21.4	9.1	22.9	28.4	32.6
2	36.9	44.7	42.5	33.6	21.5
3+	23.4	38.5	16.3	6.0	23.0
Tobacco use, %	19.0	18.3	17.6	20.7	20.0	0.91
≥1 medical condition, %	60.0	62.0	62.7	55.2	57.8	0.52
≥1 mental health condition, %	70.1	71.2	71.9	69.8	66.7	0.77
History of military sexual trauma, %	51.5	50.5	52.9	57.8	45.9	0.29

Values in bold denote statistical significance.

^aNonprescription/none includes 110 (18.0%) women with nonprescribed method and 25 (4.1%) women with no method.

LARC, long-acting reversible contraception.

Bivariate relationships between women Veterans' preferences, beliefs, self-efficacy, and knowledge and contraceptive method type used at last sex are depicted in Appendix Table A1. In terms of preferences, method effectiveness (91.0%) followed by ease of use (83.0%) were most commonly reported as extremely important method attributes among women Veterans overall and across each contraceptive method category. However, the proportion of women reporting lack of hormones as extremely important was higher among women using sterilization (40.2%) or nonprescription/no method (42.1%) compared to women using LARC (31.5%) or hormonal methods of contraception (14.8%). The proportion of women who had a strong preference regarding a contraceptive method that was effective in preventing STIs was highest among women using sterilization compared to other methods.

A minority of women Veterans strongly agreed with the belief statements that contraception is too much of a hassle; that when it is your time to get pregnant, it will happen; and that birth control decisions are mainly a woman's responsibility. Fewer women Veterans using sterilization strongly agreed that birth control decisions were mainly a woman's responsibility (13.0%) compared to women using other methods of contraception, although this relationship was not statistically significant. In terms of self-efficacy, most women Veterans using sterilization reported feeling very sure of being able to have some form of birth control consistently available (84.1%) or use a method as indicated (87.9%). However, compared to women using sterilization, a higher percentage of women using LARC or hormonal methods and a much lower percentage of women using nonprescription/no method were very sure on the measures of self-efficacy. Finally, with regard to knowledge, women using sterilization were less likely to answer correctly about the duration of action of intrauterine devices (42.3%), compared to women using LARC (90.8%). The item most incorrectly answered was about sterilization reversibility; approximately half of women Veterans incorrectly answered that sterilization was easily reversible, including 56.3% of those using sterilization.

Results from the adjusted multinomial models are shown in Table 2. Women who rated method effectiveness and ease of use as extremely important were not more likely to use sterilization compared to other methods, including nonprescription methods or no method. Women who had a strong preference for lack of hormones in a method were more likely to use sterilization versus hormonal methods (aRRR 3.13, 95% CI 1.68–5.85). Women who rated STI prevention as extremely important were more likely to use sterilization versus LARC (aRRR 2.15, 95% CI 1.33–3.46). In terms of beliefs, women who strongly agreed that birth control was mainly a woman's responsibility were less likely to use sterilization relative to LARC (aRRR 0.55, 95% CI 0.31–0.99). Women who were very sure that they could have some form of contraception consistently available or very sure that they could use a method as indicated were more likely to use sterilization relative to nonprescription/none as their method of contraception (aRRR 2.26, 95% CI 1.27–4.02 and aRRR 2.47, 95% CI 1.30–4.68, respectively). Finally, women Veterans with lower knowledge about the frequency of intrauterine device replacement were less likely to use sterilization than LARC methods (aRRR 0.07, 95% CI 0.04–0.13).

Table 2.

Adjusted Association of Contraceptive Method Use with Preferences, Beliefs, Self-Efficacy, and Knowledge Among Women Who Did Not Plan to Get Pregnant in the Future

	Sterilization vs. LARC	Sterilization vs. hormonal	Sterilization vs. nonprescription/none
	aRRR (95% CI)	aRRR (95% CI)	aRRR (95% CI)
Preference: extremely important
Method effectiveness	0.78 (0.36–1.71)	0.60 (0.23–1.57)	1.06 (0.49–2.28)
Easy to use	1.01 (0.56–1.81)	0.75 (0.38–1.50)	0.91 (0.49–1.69)
Does not contain hormones	1.30 (0.81–2.08)	3.13 (1.68–5.85)***	0.78 (0.48–1.27)
Does not interrupt sex	1.13 (0.72–1.78)	0.99 (0.59–1.65)	0.83 (0.51–1.35)
Effective in preventing STIs	2.15 (1.33–3.46)**	1.50 (0.87–2.58)	1.52 (0.90–2.56)
Belief: strongly agree
Contraception is too much of a hassle	2.33 (0.78–6.92)	1.55 (0.50–4.81)	1.26 (0.47–3.40)
When it is your time to get pregnant, it will happen	1.25 (0.67–2.35)	1.36 (0.63–2.95)	0.81 (0.44–1.50)
Birth control decisions are mainly a woman's responsibility	0.55 (0.31–0.99)*	0.63 (0.33–1.20)	0.71 (0.38–1.34)
Self-efficacy: very sure
You can have some form of birth control consistently available	0.61 (0.30–1.22)	0.55 (0.24–1.26)	2.26 (1.27–4.02)**
You can use a method of contraception as indicated	0.60 (0.27–1.35)	0.52 (0.19–1.40)	2.47 (1.30–4.68)**
Knowledge: percent correct
There are methods of birth control that are as effective as tubal sterilization, or having your tubes tied, but that can be stopped or removed if a woman decides that she wants to get pregnant.	0.74 (0.44–1.24)	1.03 (0.59–1.80)	1.18 (0.70–1.97)
Contraceptive implants that are placed in your arm, such as Implanon or Nexplanon, can prevent pregnancy for over 1 year.	0.96 (0.62–1.50)	0.73 (0.44–1.22)	0.85 (0.53–1.36)
Only women who have delivered a baby can use an intrauterine device, or IUD	0.96 (0.60–1.55)	1.07 (0.62–1.83)	0.73 (0.44–1.21)
After a woman gets a tubal sterilization, or has her tubes tied, her doctor can easily reverse the procedure if she wants to have children in the future.	1.27 (0.81–1.98)	1.43 (0.86–2.37)	1.10 (0.69–1.77)
Intrauterine devices, or IUDs, available in the United States must be replaced every year	0.07 (0.04–0.13)***	0.88 (0.53–1.47)	0.88 (0.55–1.42)

Values in bold denote statistical significance.

Multivariable multinomial models were separately conducted for each predictor (row) and each comparison (column) with different method serving as the reference category, adjusting for demographic and health variables that associated with contraceptive method at p < = 0.15 from Table 1 (including age, marital status, deployment, parity).

^*p < 0.05.

^**p < 0.01.

^***p < 0.001.

Discussion

In this study of over 600 women Veterans who were heterosexually active and not desiring future pregnancy, associations between contraceptive preferences, beliefs, self-efficacy, and knowledge and use of sterilization were complex. Despite not wanting future pregnancy, women Veterans who rated method effectiveness as highly important were not more likely to use sterilization compared to other methods of contraception, including nonprescription/no method. While contraceptive counseling based on effectiveness is emphasized in “tier-based” counseling recommendations and effectiveness was the most important method attribute for women Veterans in our sample, our study adds to evidence that the decision to use sterilization is not solely based on the procedure's effectiveness at preventing pregnancy.^1–6

Discerning specific method choice, then, may be more influenced by other contraceptive attributes. In fact, we found the preference for lack of hormones to be more strongly associated with sterilization use than any other preference or belief. The preference for lack of hormones has previously been shown to be a driving factor in choosing between sterilization and LARC, although larger studies of comprehensive contraceptive decision-making have also demonstrated that attributes such as safety, affordability, convenience, and permanence are important to patients.^1,2,4 Thus, contraceptive counseling should be individualized by exploring the specific goals and values for each patient through shared decision-making.²⁰

Achieving fully informed consent through shared decision-making also entails ensuring an accurate understanding of the various contraceptive methods available. In contrast to previous studies, our study population was generally knowledgeable about LARC compared to sterilization in terms of efficacy.^4,6 However, like other studies, misinformation about sterilization reversibility was quite prevalent.^19,21,22 Half of all participants surveyed incorrectly believed sterilization was easily reversible, highlighting a key gap in patient education. While sterilization reversal is a covered benefit under the VA system (while in most cases, in vitro fertilization is not covered), it is important to counsel patients that sterilization should only be performed if desiring permanent, irreversible contraception.

We did find that women Veterans who desired a method that was effective in preventing STIs were more likely to use sterilization compared to LARC. Given that this question was phrased as a general preference, we are unable to determine whether women erroneously believed sterilization would protect against STIs or if women who choose sterilization tend to be more likely to also prioritize STI prevention. We are also unable to determine the reason (i.e., STI prevention or increased contraceptive benefit) why women used condoms in addition to another method of contraception. The strong preference given to STI prevention, however, underscores the need for ongoing sexual health discussion even after sterilization.

Several limitations to this study are important to note. First, as ECUUN included only Veterans who were recently seen in VA primary care, all participants had at least a high school education and were engaged in health care. This may reflect as differences in preferences, beliefs, self-efficacy, and knowledge compared to the general United States population and impact the generalizability of our study results. Furthermore, there are a lack of published validated scales to comprehensively measure contraceptive preferences, beliefs, and self-efficacy. Thus, questions posed may not have fully explored the complex decision-making involved in contraceptive choice. We also were unable to assess the degree to which women Veterans prioritized preferences relative to each other. Finally, given the limitations of observational, cross sectional studies, we are unable to determine the directionality or temporality of the association between predictors and sterilization. For example, it is unclear whether women who had greater self-efficacy preferred sterilization or conversely, reported greater self-efficacy as a result of using sterilization for contraception. Thus, preferences, beliefs, self-efficacy, and knowledge at time of study interview may not necessarily reflect those at the time of the decision to undergo sterilization.

Conclusions

In summary, this study provides insights into associations between contraceptive preferences, beliefs, self-efficacy, and knowledge and sterilization use in a national sample of women Veterans not desiring future pregnancy within the VA system. Our findings suggest that sterilization decision-making may be influenced by multiple factors beyond consideration about efficacy of a method. Using shared decision-making during contraceptive counseling promotes the elicitation, discussion, and weighing of these varied factors to facilitate decisions that are value-concordant for each individual person.²⁰ Further efforts are needed to ensure that women Veterans possess accurate and comprehensive knowledge of available contraceptive methods and to better understand how to effectively assess and attend to the various factors patients consider in the process of contraceptive decision-making.

Appendix

Appendix Table A1.

Contraceptive Method Use by Preferences, Beliefs, Self-Efficacy, and Knowledge Among Women Who Did Not Plan to Get Pregnant in the Future

		Contraceptive method type
	Total	Sterilization	LARC	Hormonal methods	Nonprescription/Nonea	p value
	N = 612 (100%)	n = 208 (34.0%)	n = 153 (25.0%)	n = 116 (19.0%)	n = 135 (22.1%)
Preference: extremely important
Method effectiveness	556 (91.0)	88.9	92.1	94.0	90.4	0.46
Easy to use	507 (83.0)	82.6	82.4	84.5	83.0	0.97
Does not contain hormones	202 (33.6)	40.2	31.5	14.8	42.1	<0.0001
Does not interrupt sex	378 (61.8)	62.5	58.8	61.2	64.4	0.79
Effective in preventing STIs	415 (67.8)	74.5	58.2	66.4	69.6	0.011
Belief: strongly agree
Contraception is too much of a hassle	31 (5.1)	6.7	3.3	4.3	5.2	0.50
When it is your time to get pregnant, it will happen	90 (14.7)	17.8	12.4	9.5	17.0	0.15
Birth control decisions are mainly a woman's responsibility	109 (17.8)	13.0	21.6	20.7	18.5	0.14
Self-efficacy: very sure
You can have some form of birth control consistently available	517 (84.5)	84.1	90.8	92.2	71.1	<0.0001
You can use a method of contraception as indicated	539 (88.2)	87.9	93.5	94.8	77.0	<0.0001
Knowledge: percent correct
There are methods of birth control that are as effective as tubal sterilization, or having your tubes tied, but that can be stopped or removed if a woman decides that she wants to get pregnant.	449 (73.4)	74.5	78.4	69.8	68.9	0.23
Contraceptive implants that are placed in your arm, such as Implanon or Nexplanon, can prevent pregnancy for over 1year.	363 (59.3)	56.7	56.9	64.7	61.5	0.46
Only women who have delivered a baby can use an intrauterine device, or IUD	419 (68.5)	64.4	69.3	69.8	72.6	0.43
After a woman gets a tubal sterilization, or has her tubes tied, her doctor can easily reverse the procedure if she wants to have children in the future.	338 (55.2)	56.3	54.2	53.4	56.3	0.95
Intrauterine devices, or IUDs, available in the United States must be replaced every year	340 (55.6)	42.3	90.8	44.8	45.2	<0.0001

Values in bold denote statistical significance.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

Dr. Arora is funded by the Clinical and Translational Science Collaborative of Cleveland, KL2TR0002547 from the National Center for Advancing Translational Sciences (NCATS) component of the National Institutes of Health and NIH roadmap for Medical Research.

Colleen Judge is supported by the National Center For Advancing Translational Sciences of the National Institutes of Health under Award Number TL1TR001858 (PI: Wishwa Kapoor).

Dr. Callegari is supported by a VA Career Development Award, CDA 14–412.

This study was supported by the VA Health Services Research and Development Service (HSR&D) Merit Review Award, IIR 12–124 (PI: Dr. Sonya Borrero).

This article is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The findings and conclusions in this report are those of the authors and do not represent the views of the Department of Veterans Affairs or the United States government.