Adherence to Recommended Practices for Provision of Long-Acting Reversible Contraception Among Providers in a Large U.S. Health Care System

Abstract

Background: There is little research examining adherence to practice guidelines for long-acting reversible contraception (LARC). We assessed same-day LARC provision and adherence to other guideline-recommended practices among providers in a large academic health care system.

Materials and Methods: We surveyed 363 providers who had billed using LARC-related codes within the prior 12 months. Primary outcomes were, for women requesting an intrauterine device (IUD) or implant, the typical number of visits for method provision and ability to add an insertion procedure to an annual examination. We used chi-square tests and multivariable logistic regression to identify characteristics associated with primary outcomes and described other practices and barriers to same-day LARC.

Results: Our response rate was 42% (153/363). A typical single visit for women requesting an IUD or implant was reported by 37% of IUD providers and 51% of implant providers, respectively, and was associated with obstetrician-gynecologist specialty versus family medicine (IUD: 44% vs. 12%, p = 0.001; implant: 57% vs. 26%, p = 0.002) and practicing in the county of the main academic medical center versus any other county (IUD: 48% vs. 20%, p = 0.001; implant: 65% vs. 27%, p < 0.001). High ability to add LARC insertion to an annual examination was reported by 48% of IUD providers and 51% of implant providers, with similar associations of specialty and county. Barriers to same-day LARC included scheduling constraints, insurance and billing concerns, and device stocking. Nearly all respondents provide LARC to nulliparous women and adolescents. Among IUD providers, 73% schedule routine follow-up after insertion.

Conclusions: Same-day LARC provision is low among providers in a large academic health care system. Provider-identified barriers suggest interventions to improve LARC access, including incentivizing device stocking and billing and insurance education.

Introduction

Use of long-acting reversible contraceptive methods, including intrauterine devices (IUDs) and contraceptive implants, has increased more than fourfold among U.S. women since the early 2000s,^1,2 likely contributing to recent declines in unintended pregnancy.³ Health care providers trained in long-acting reversible contraception (LARC) provision are pivotal to facilitating access to these methods, which appeal to many patients for their high efficacy and low maintenance.⁴ The American College of Obstetricians and Gynecologists (ACOG) and the Centers for Disease Control and Prevention (CDC) maintain practice guidelines for LARC provision, which synthesize existing evidence and provide recommendations to maximize access while ensuring safety.^5,6 The American Academy of Family Physicians and the American Academy of Pediatrics have also published practice guidelines for LARC provision based on these documents.^7,8 Recommended practices include same-day LARC provision, placement at any point in the menstrual cycle when pregnancy can be reasonably excluded, provision to nulliparous and adolescent women, and limiting unnecessary follow-up.

More than 90% of obstetrician-gynecologists (OB/GYNs) provide IUDs and more than half provide implants,^9,10 compared with LARC provision by <25% of family physicians and ∼10% of pediatricians.^11,12 Prior research suggests substantial variation in knowledge of LARC practice guidelines among OB/GYNs^13,14 and across medical specialties and provider types.^15–17 However, there is less research examining specific practices for LARC provision or adherence to practice guidelines, which may impact contraceptive access and may vary according to provider characteristics. We assessed specific practices related to LARC provision among health care providers who provide LARC within a single, large health care system in Western Pennsylvania. Specifically, we aimed to describe adherence to guideline-recommended LARC practices, evaluate characteristics associated with guideline adherence, and identify barriers to implementation of guideline recommended practices.

Materials and Methods

Study design and population

The University of Pittsburgh Institutional Review Board approved this study. We conducted a cross-sectional survey of all providers affiliated with the University of Pittsburgh Medical Center (UPMC) health care system who were associated with at least one LARC-related diagnosis, procedure, or device code in electronic medical records between October 1, 2017, and September 30, 2018. We excluded providers currently enrolled in training programs and assumed that lack of a functioning UPMC e-mail address indicated lack of current affiliation. Exclusions resulted in a recruitment pool of 363 providers.

Recruitment and data collection

We administered an online survey using Qualtrics (Provo, UT). Potentially eligible providers received an e-mail inviting them to participate in the “UPMC LARC Provider Survey” via a personalized link. We offered participants a $10 physical or electronic Starbucks gift card upon survey completion. We recruited participants from December 11, 2018, through January 19, 2019, with three reminder e-mails distributed at ∼1, 3, and 5 weeks after the initial e-mail. An additional e-mail was sent from an OB/GYN division chief encouraging colleagues to participate.

Survey development

We developed a 25-question survey to assess adherence to LARC provision practices as recommended by the CDC 2016 U.S. Selected Practice Recommendations (SPR) for Contraceptive Use⁵ and ACOG Practice Bulletin on LARC.⁶ The survey was developed based on previously published literature^9,18,19 and with iterative feedback from content experts, including OB/GYNs and family planning specialists. We used scenario-based questions when possible to assess real-world practice patterns as opposed to provider knowledge about guideline-recommended practices. We pilot-tested the survey with 10 trainee physicians with experience providing LARC (3 OB/GYN residents, 2 OB/GYN family planning fellows, 2 internal medicine women's health fellows, and 3 adolescent medicine fellows). Pilot participants provided feedback on survey format, question wording, and response options.

Measures

We assessed average monthly frequency of IUD and implant provision over the past 12 months (none, <1/month, 1–5/month, 6–10/month, >10/month). Providers who reported no provision of a given method were not asked the remainder of questions pertaining to that method but were still eligible to respond to questions about the other method.

Our primary outcome was same-day LARC provision, assessed using two-survey questions. First, we asked providers, “In your estimation, how many office visits does your typical patient have for IUD placement (i.e., from deciding she wants an IUD to having the device placed)? Please consider your primary clinical practice location.” Response options were “1 visit,” “2 visits,” or “3 or more visits,” which was dichotomized to 1 (i.e., a single visit) versus 2 or more visits for analysis due to low frequency of 3 or more visits (n = 1 for each method). A second item assessed providers' ability to add an LARC insertion to an annual examination visit at their primary practice location, using the following scenario:

Layla is a 26-year-old, established patient of yours. During her scheduled annual exam, she tells you she has “done a lot of research” and wants an [IUD/implant] for contraception. She has not been sexually active for three weeks and her menstrual period ended 2 days ago. Layla is asking if you can insert the [IUD/implant] today. Based on your current clinical practice, how likely is it that you can insert the [IUD/implant] during today's visit?

Response options were “very unlikely,” “unlikely,” “equally likely and unlikely,” “likely,” and “very likely,” which was dichotomized for analysis as “likely/very likely” versus all other responses. Of note, this scenario describes a patient who is clearly not pregnant per the CDC guidelines⁵; immediate LARC insertion is therefore appropriate if desired. Regardless of add-on ability, participants then identified reasons why they might be unable to complete the LARC insertion described in the scenario, using a select-all-that-apply question including an open-ended “other” option, for which responses were recoded to discrete options if applicable.

Other outcomes are based on the CDC (SPR) and include provision of IUDs to nulliparous women (recommended); LARC provision to women younger than 18 years (recommended); parental consent requirements for women younger than 18 years (not required under Pennsylvania law²⁰); timing of IUD insertion within the menstrual cycle (recommendation: insert at any time if reasonably certain not pregnant), scheduling of routine follow-up after IUD insertion (not recommended), and use of misoprostol for IUD insertion (not recommended for routine use, may be helpful after prior failed insertion).⁵ Regarding use of misoprostol, participants were asked, “In the past 12 months, have you ever used misoprostol to assist with an IUD insertion?” Those who replied “yes” were asked to indicate how often they used misoprostol to assist with IUD insertion for the following types of patients: “Nulliparous women (1st insertion attempt),” “Parous women (1st insertion attempt),” and “Women who have had a failed IUD insertion.”

Provider and practice characteristics were assessed via self-report. Variables included provider type (physician, nurse practitioner, physician assistant, midwife); primary medical specialty (OB/GYN, family medicine, internal medicine, pediatrics); gender; years in practice; average number of patients seen per week; time spent practicing in a federally qualified health center (FQHC), Title X, or public health clinic (yes/no); primary practice location (Allegheny County, a metropolitan area containing the health system's main academic medical center, vs. all other counties); and primary practice in an academic setting, that is, involved in training of medical students, residents, or other health professions trainees (yes/no).

Data analysis

We generated descriptive statistics for all variables. We tested bivariate associations between provider and practice characteristics and same-day LARC provision and other outcomes using chi-square or Fisher's exact tests if expected counts were less than n = 5, with a significance level of p < 0.05. We used multivariable logistic regression to identify characteristics associated with same-day LARC outcomes (typical single visit and high add-on ability) while adjusting for variables associated with these outcomes in bivariate analyses at the p < 0.2 level. Due to prohibitively small sample sizes, we excluded internal medicine (n = 5) and pediatric providers (n = 4) from analyses examining medical specialty as a predictor. We used Stata 14.2 (College Station, TX) for all analyses.

Results

Study population

Of 363 providers contacted via e-mail, 167 (46%) consented to participation (Fig. 1). Ten were ineligible (6%) and 4 did not complete the survey, resulting in 153 completed surveys (response rate 42%). The majority of respondents were physicians (n = 103, 67%) and reported a primary medical specialty of OB/GYN (n = 104, 68%) (Table 1). All advanced practice providers (APPs) reported working in OB/GYN except for one family medicine physician assistant. Most providers place an average of one to five IUDs or implants per month (Fig. 2). Ten participants (7%) provided no IUDs and 23 (15%) provided no implants over the past 12 months.

FIG. 1.

Recruitment flow diagram. FM, family medicine; IM, internal medicine; IUD, intrauterine device.

Table 1.

Sample Characteristics of Long-Acting Reversible Contraception Providers

Characteristic	n (%)a
Provider type
Physician (MD or DO)	103 (67)
Advanced practice provider	50 (33)
Nurse practitioner	19 (12)
Certified nurse midwife	16 (10)
Physician assistant	15 (10)
Primary medical specialty
Obstetrics and gynecology	104 (68)
Family medicine	40 (26)
Internal medicine	5 (3)
Pediatrics	4 (3)
Gender
Female	122 (80)
Male	31 (20)
Years in practice
0–4	35 (23)
5–9	32 (21)
10–14	24 (16)
15–19	16 (10)
20 or more	46 (30)
Patients seen per week
1–10	12 (8)
11–50	50 (33)
51–100	78 (51)
>100	13 (9)
Any time practicing in FQHC, Title X, or public health clinic
Yes	24 (16)
No	129 (84)
Primary practice in academic setting
Yes	121 (79)
No/unsure	32 (21)
County of primary practiceb
Allegheny county	92 (61)
Any other county	58 (39)

^aN = 153. Percentages may not add to 100% due to rounding.

^bMissing data for county of primary practice (n = 3). Allegheny County is a metropolitan location containing the health system's main academic medical center and educational programs. Fourteen PA counties were represented overall: Allegheny (n = 61), Beaver (n = 2), Bedford (n = 3), Blair (n = 9), Butler (n = 4), Clarion (n = 2), Clinton (n = 2), Erie (n = 6), Fayette (n = 4), Lawrence (n = 4), Lycoming (n = 14), Mercer (n = 3), Venango (n = 3), and Westmoreland (n = 2). FQHC, federally qualified health center; LARC, long-acting reversible contraception.

FIG. 2.

Average frequency of LARC provision over the past 12 months. LARC, long-acting reversible contraception.

Same-day LARC provision

Typical number of visits to obtain LARC

When asked how many office visits a “typical” patient has for IUD or implant placement after deciding she wants the method, 37% (53/143) of IUD providers and 51% (66/130) of implant providers reported that a single visit was typical (Table 2). For both methods, a typical single visit was associated with OB/GYN specialty compared with family medicine (IUD: 44% vs. 12%, p = 0.001; implant: 57% vs. 26%, p = 0.002) and practicing in Allegheny County versus any other county (IUD: 48% vs. 20%, p = 0.001; implant: 65% vs. 27%, p < 0.001). For IUDs, a typical single visit was additionally associated with seeing fewer patients per week (p = 0.02). In multivariable models, OB/GYN specialty and practicing in Allegheny County remained associated with a typical single visit for both methods.

Table 2.

Associations of Provider Characteristics with Same-Day Long-Acting Reversible Contraception: Typical Single Visit

Characteristic	Single visit is typical for LARC provision
	IUD (n = 143)			Implant (n = 130)
	%	pa	aOR (95% CI)b	%	pa	aOR (95% CI)b
Total	37	—	—	51	—	—
Provider type		0.25			0.29
Physician	40		—	54		—
Advanced practice provider	31		—	44		—
Primary medical specialtyc		0.001			0.002
Obstetrics and gynecology	44		7.94 (2.17–29.06)	57		4.77 (1.83–12.46)
Family medicine	12		Ref.	26		Ref.
Gender		0.66			0.74
Female	36		—	50		—
Male	41		—	54		—
Years in practice		0.80			0.61
0–4	39		—	45		—
5–9	33		—	59		—
10–14	43		—	59		—
15–19	25		—	38		—
20 or more	40		—	51		—
Patients seen per week		0.02			0.36
1–10	64		Ref.	73		—
11–50	49		1.13 (0.24–5.36)	54		—
51–100	27		0.47 (0.10–2.12)	45		—
>100	31		0.78 (0.11–5.48)	56		—
Any time practicing in FQHC, Title X, or public health clinic		0.38			0.75
Yes	45		—	48		—
No	36		—	51		—
Primary practice in academic setting		0.30			0.11
Yes	39		—	54		1.20 (0.44–3.23)
No/unsure	29		—	37		Ref.
County of primary practiced		0.001			<0.001
Allegheny	48		3.06 (1.29–7.28)	65		4.68 (1.96–11.17)
All others	20		Ref.	27		Ref.

Bold cells indicate statistically significant results.

^ap-Values are from chi-square tests or Fisher's exact tests if expected number <5.

^bLogistic regression model with outcome of single visit typical versus two or more visits for IUD or implant provision. n = 133 for IUD model and n = 119 for implant model due to missing data and exclusions. Models are adjusted for the variables shown based on associations with the outcome in bivariate analysis at the p < 0.2 level.

^cDue to prohibitively small sample sizes, participants reporting a medical specialty of internal medicine (n = 5) or pediatrics (n = 4) were excluded from analyses examining provider specialty as an independent variable.

^dMissing data for county (n = 3).

aOR, adjusted odds ratio; CI, confidence interval; IUD, intrauterine device.

Ability to add LARC to an annual examination

In the scenario-based question, 48% (68/143) of IUD providers and 51% (66/130) of implant providers reported that it was “likely” or “very likely” they could add an LARC insertion to the patient's annual examination (Table 3). High add-on ability was associated with OB/GYN specialty compared with family medicine (IUD: 56% vs. 18%, p < 0.001; implant: 59% vs. 20%, p < 0.001) and practicing in Allegheny County versus any other county (IUD: 55% vs. 35%, p = 0.02; implant: 61% vs. 33%, p = 0.002). For implants, high likelihood of add-on ability was additionally associated with practicing in an academic compared with a nonacademic setting (56% vs. 30%, p = 0.01). In multivariable models, only OB/GYN specialty remained associated with high likelihood of add-on ability for both methods. Reasons for potential inability to add LARC insertion to an annual examination are shown in Figure 3.

Table 3.

Associations of Provider Characteristics with Same-Day Long-Acting Reversible Contraception: Add-On Ability

Characteristic	Provider is likely/very likely able to add LARC to annual examination
	IUD (n = 143)			Implant (n = 130)
	%	pa	aOR (95% CI)b	%	pa	aOR (95% CI)b
Total	48	—	—	51	—	—
Provider type		0.34			0.23
Physician	45		—	47		—
Advanced practice provider	53		—	59		—
Primary medical specialtyc		<0.001			<0.001
Obstetrics and gynecology	56		7.89 (2.66–23.37)	59		7.27 (2.58–20.47)
Family medicine	18		Ref.	20		Ref.
Gender		0.72			0.93
Female	48		—	51		—
Male	44		—	50		—
Years in practice		0.09			0.18
0–4	58		Ref.	61		Ref.
5–9	63		1.37 (0.39–4.86)	62		0.91 (0.26–3.24)
10–14	43		0.39 (0.11–1.34)	47		0.41 (0.10–1.74)
15–19	38		0.79 (0.20–3.18)	31		0.58 (0.14–2.47)
20 or more	35		0.32 (0.10–0.99)	43		0.46 (0.14–1.53)
Patients seen per week		0.18			0.20
1–10	64		Ref.	73		Ref.
11–50	58		1.47 (0.30–7.21)	59		0.64 (0.11–3.86)
51–100	41		0.79 (0.17–3.69)	43		0.51 (0.09–2.81)
>100	38		1.17 (0.16–8.44)	44		0.53 (0.06–4.91)
Any time practicing in FQHC, Title X, or public health clinic		0.48			0.87
Yes	55		—	52		—
No	46		—	50		—
Primary practice in academic setting		0.05			0.01
Yes	52		2.13 (0.79–5.70)	56		2.41 (0.82–7.13)
No/unsure	32		Ref.	30		Ref.
County of primary practiced		0.02			0.002
Allegheny	55		1.77 (0.75–4.17)	61		2.19 (0.88–5.43)
All others	35		Ref.	33		Ref.

Bold cells indicate statistically significant results.

^ap-Values are from chi-square tests or Fisher's exact tests if expected number <5.

^bLogistic regression model with outcome of likely/very likely able to add (IUD or implant) to annual examination versus very unlikely, unlikely, or equally likely and unlikely. n = 133 for IUD model and n = 119 for implant model due to missing data and exclusions. Models are adjusted for the variables shown based on significant associations with the outcome in bivariate analysis at the p < 0.2 level.

^cDue to prohibitively small sample sizes, participants reporting a medical specialty of internal medicine (n = 5) or pediatrics (n = 4) were excluded from analyses examining provider specialty as an independent variable.

^dMissing data for county (n = 3).

FIG. 3.

Reasons for potential inability to add LARC insertion to an annual examination. Participants were asked to select all reasons they might be unable to add an IUD or implant insertion to the annual examination scenario. Data are percentage of respondents selecting each reason among providers of IUDs (n = 143) and implants (n = 130). Responses were not mutually exclusive with the exception of “No reason—I am confident I would always be able to insert.”

Adherence to other recommended practices for LARC provision

LARC for nulliparous women and women younger than 18 years

All participants reported providing IUDs to nulliparous women, except four OB/GYN providers who see only pregnant and postpartum women. Ninety-six percent (137/143) of IUD providers and 98% (128/130) of implant providers provide LARC to women younger than 18 years; those who did not reported seeing few or no nonadult patients (data not shown). Among participants who provide LARC to minors, 12% (17/137) reported requiring parental consent for IUDs and 15% (19/128) for implants.

IUD insertion timing within the menstrual cycle

Ninety percent (128/143) of IUD providers place IUDs at any time during the menstrual cycle; seven (5%) place IUDs during the first half of the cycle only and eight (6%) place IUDs during menses only. All eight providers who insert only during menses reported that they find it easier to insert IUDs during menses, and six reported that they require active menstruation to ensure the patient is not pregnant.

Use of misoprostol for IUD insertion

Among IUD providers, 35% (50/143) reported using misoprostol to assist with an IUD insertion in the prior 12 months. Use of misoprostol was more prevalent among APPs (47% vs. 29% physicians, p = 0.03), OB/GYN providers (47% vs. 6% family medicine, p < 0.001), and providers working in nonacademic settings (52% vs. 30% providers in academic settings, p = 0.028), and less prevalent among providers who practice in an FQHC, Title X, or public health clinic (9% vs. 40% among providers with no time in these settings, p = 0.006).

Among the 50 providers who used misoprostol, 62% (31/50) reported routine use (half of the time or more) for patients with a prior failed IUD insertion (appropriate usage per guidelines). However, 42% (21/50) reported routine use for first-attempt insertions in nulliparous women, and 6% (3/50) reported routine use for first-attempt insertions in parous women.

Routine follow-up after IUD insertion

Seventy-three percent of IUD providers (104/143) reported scheduling routine follow-up appointments after IUD insertion. OB/GYN providers were less likely to schedule follow-up compared with family medicine providers (67% vs. 88%, p = 0.02), as were providers practicing in Allegheny County (67% vs. 84% other counties, p = 0.03) or academic settings (68% vs. 90% in nonacademic settings, p = 0.01).

Discussion

In our survey of LARC providers within a single, large, academic health care system, fewer than half reported typical same-day LARC provision or high ability to add LARC to an annual examination, while adherence to other recommended practices was higher. Barriers to same-day provision were largely structural or logistical in nature, including scheduling constraints, insurance and billing concerns, and device stocking, suggesting substantial room for improving LARC access via practice- or system-level interventions.

This study adds to the literature by examining specific LARC practices and adherence to practice guidelines among APPs and physician providers of different medical specialties. Importantly, our primary outcomes were framed around providing LARC to women who had already decided that they want an IUD or implant, which eliminates patient-specific reasons for delaying insertion and may have lessened provider perceived barriers related to time for contraceptive counseling and method selection. Given this standardization, our results represent providers' practical abilities to provide same-day LARC when this clearly aligns with a patient's goals.

Two-visit insertion protocols are an established barrier to LARC access,^21–23 yet same-day LARC provision remains uncommon. A recent national survey of OB/GYN physicians found that only 29% offer same-day IUD placement,¹⁰ compared with 44% of OB/GYN providers in our sample who reported a typical single visit for IUDs and 57% for implants. Although providers in our system may be more likely to provide same-day LARC than are providers nationwide, same-day access is still not the norm for the majority of patients in our health system. Family medicine providers were significantly less likely to offer same-day LARC, suggesting that OB/GYN providers may be more aware of practice guidelines, as noted in prior studies,^15,16 or that OB/GYN practices are more adept at offering this service, possibly as a result of greater demand for LARC overall. Same-day provision was also associated with practicing in proximity to our system's academic medical center, which may reflect efforts by departmental leadership to increase LARC access.

Approximately half of providers reported a high likelihood of being able to add an LARC insertion to an annual examination for a woman who clearly desires the method. However, only 12%–13% were confident that they would always be able to complete the insertion, with the rest citing potential barriers, which were largely logistical in nature, rather than stemming from gaps in provider knowledge or training, as identified in other studies.^13–17 First, many participants were concerned about inadequate appointment durations and clinic flow disruption with add-on LARC provision, and more than 20% reported insufficient time for education, counseling, or consent as a potential barrier despite a scenario in which the patient clearly desired the method. To address these concerns, some practices routinely ask reproductive-age patients if they are considering LARC when making an appointment, to schedule additional time for the appointment. While sufficient time to address patient questions and a thorough consent process must be maintained, distribution of online educational materials or contraceptive decision aids²⁴ to individuals considering LARC could help reduce in person time required for counseling. Maintaining premade insertion kits can also reduce the time associated with LARC procedures, and formal time studies may be helpful to evaluate the financial and workflow impacts of add-on procedures.²⁵

Nearly half of providers cited verifying insurance coverage as a barrier to same-day LARC. Some practices designate an administrative employee to verify coverage or obtain pre-authorization, if necessary, before or during appointments. Alternatively, signs in the waiting room or pre-appointment contacts could encourage interested patients to confirm their own coverage in advance. Health systems could help reduce the need for verification by maintaining a list of insurers with exceptions to the Affordable Care Act's contraceptive coverage mandate, which guarantees zero cost-sharing for most insured U.S. women.²⁶ Although both providers and patients would benefit from transparent and accessible information regarding insurance policies for contraception, funding and manpower to maintain such a list remains an unsolved problem.

Same-day LARC provision is only possible if devices are stocked on site. Approximately 20% of respondents reported the need to order the method as a barrier to adding LARC insertion to an annual examination for both IUDs and implants, suggesting that ∼80% of providers' practices routinely stock LARC methods. Practices should be encouraged to purchase devices directly and bill for LARC as a medical benefit rather than ordering devices as pharmacy benefits as they are requested. While practices with lower demand may worry about stocking devices they may not use, resources exist to help forecast method demand and to order devices at reduced cost.²⁵ Health systems could also incentivize device stocking to improve access, for example, by absorbing the cost of unused devices. However, while on-site stocking is clearly necessary for same-day provision, the high rate of on-site stocking identified in our and other studies¹⁹ suggests that stocking is not sufficient to optimize same-day LARC access and emphasizes the impact of other barriers such as clinic flow disruptions and billing and insurance issues.

Finally, more than 10% of providers reported needing a second visit to be able to bill for both an annual examination and an LARC procedure. However, it is often possible to bill for both an Evaluation and Management (E/M) service and a procedure simultaneously. For example, if a patient presents for an annual examination (i.e., receives other preventive health services) and has an LARC method placed during that visit, both an E/M and a procedure code can be reported as long as documentation supports both billing codes. Established coding guides for LARC provision should be distributed to providers and practice managers to help correct misconceptions around billing and reimbursement for LARC methods and reduce the perceived need for a second visit.^27,28

Adherence to recommendations related to LARC eligibility criteria was substantially greater than for same-day provision. Nearly all providers reported placing IUDs at any time during the menstrual cycle, with only 6% inserting during menstruation only. Encouragingly, nearly all respondents provide IUDs to nulliparous women and adolescents. However, as Pennsylvania law allows minors to consent for contraceptive services without parental notification or consent,²⁰ the parental consent requirements reported by 12%–15% of providers represent an unnecessary provider-created barrier to LARC access for adolescents. Education around this topic could help improve availability of the confidential contraceptive services that adolescents deserve.

More than 70% of providers reported scheduling routine follow-up visits after IUD insertion. Follow-up visits are recommended by IUD manufacturers and typically include a “string check” examination in addition to assessment of satisfaction and side effects. However, additional health system interactions following contraceptive provision have not been shown to improve adverse event detection or method continuation,^29,30 and practice guidelines specifically advise against routine follow-up, instead recommending assessment as necessary or during “other routine visits.”⁵ Unnecessary return visits create additional time burdens for patients, often require copayments, are poorly attended,³¹ and may contribute to long appointment wait times. Education of providers and practice managers regarding appropriate follow-up may help reduce this barrier.

CDC guidelines recommend against routine misoprostol use for IUD insertions,⁵ citing high-quality evidence that it does not improve insertion ease or success but may increase pain and side effects^32,33 in addition to increasing total insertion time and patient inconvenience. Misoprostol may be helpful for women with a previously failed IUD insertion,^34,35 a scenario cited by the majority of providers who use misoprostol. Encouragingly, nearly two-thirds of IUD providers in our study reported no use of misoprostol for IUD insertion in the past year. Nevertheless, 42% of those who did use misoprostol (representing 15% of all IUD providers) reported routine use of misoprostol for first-attempt insertions in nulliparous women, suggesting that inappropriate misoprostol use is fairly common in our system, consistent with a prior study.³⁶ Education should be targeted to providers or practices routinely using misoprostol, perhaps in conjunction with additional insertion training to reduce the perceived need for a cervical ripening agent.

Strengths of this study include our ability to compare LARC practices across provider and practice characteristics within a single, large health care system to better understand and improve access locally. While our findings are not immediately generalizable to other health care systems, our sample comprises both physicians and APPs working in a variety of medical specialties and practice settings. Our results may therefore be informative for efforts to improve LARC access in other systems.

There are also several limitations. Our study may be limited by selection bias, as providers who elected to complete the survey may be more engaged in LARC provision, and therefore more likely to adhere to practice guidelines (e.g., more likely to work in a practice that routinely stocks LARC). This could lead to overestimation of guideline adherence and may account for the higher rates of same-day provision observed in our study compared with national estimates.^9,10 Additionally, desirability bias may have influenced participants' reporting of their practice patterns to align with knowledge of established guidelines, as opposed to how they actually practice. As such, our results may represent overestimates of the true availability of same-day LARC placement within our health care system; indeed, prior studies in California family planning clinics suggest that provider's reports of LARC availability and same-day access may be inflated compared with patient-facing information.^19,37 In our sample, the majority of participants worked in OB/GYN specialties, with more limited representation from primary care providers, who may face unique barriers to LARC provision; however, this is likely reflective of the distribution of LARC providers within our system. Next, although practice-level policies and logistics likely influence provider practices, we are unable to account for clustering of providers by practice location. Importantly, our study does not examine provider practices related to LARC removal, an overlooked component of access, which is vital to protect patients' bodily and reproductive autonomy. Finally, while our survey-based study allowed for evaluation of statistical differences in LARC practices across provider characteristics, the quantitative nature of our study precludes more nuanced insight into reasons for the observed patterns. A concurrent qualitative investigation of provider-perceived facilitators and barriers to LARC provision is ongoing.³⁸

In conclusion, adherence to practice guideline recommendations for LARC provision is mixed among providers in a single, large health care system, with substantial room for improvement in same-day LARC provision. The barriers to implementation of guideline-recommended practices identified in this study may be useful to inform interventions to improve access to LARC.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work and Colleen Judge-Golden were supported by the National Center for Advancing Translational Sciences of the National Institutes of Health (TL1TR001858, principal investigator: Wishwa Kapoor). The findings and conclusions are those of the authors. The funding source played no role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication.