Linking Inter-professional Newborn and Contraception Care (LINCC) trial: Protocol for a stepped wedge cluster randomized trial to link postpartum contraception care with routine Well-Baby Visits

Abstract

Background:

Pregnancies conceived within 18 months of a prior delivery (termed short inter-pregnancy interval [IPI]) place mothers and infants at high risk for poor health outcomes. Despite this, nearly one third of U.S. women experience a short IPI.

Objective:

To address the gap in the current model of postpartum (PP) contraception care by developing and implementing a novel approach to link (co-schedule) PP contraception care with newborn well-baby care to improve access to timely PP contraception.

Methods:

The LINCC Trial will take place in seven clinical locations across five community health centers within the U.S. PP patients (planned n = 3150) who are attending a Well-Baby Visit between 0 and 6 months will be enrolled. The LINCC Trial aims to leverage the Electronic Health Record to prompt providers to ask PP patients attending a Well-Baby Visit about their PP contraception needs and facilitate co-scheduling of PP contraception care with routine newborn care visits. The study includes a cluster randomized, cross-sectional stepped wedge design to roll out the intervention across the seven sites. The outcomes of the study include receipt of most or moderately effective methods of contraception by two and six months PP; and rate of short IPI pregnancies. Implementation outcomes will be assessed at baseline and 6 months after site enters intervention period.

Conclusions:

The LINCC Trial seeks to evaluate the effectiveness and feasibility of a linked care model in comparison to usual care.

1. Introduction

A short inter-pregnancy interval ([IPI] i.e., a pregnancy occurring within 18 months of a prior delivery) can result in poor maternal and fetal outcomes including maternal anemia [1] and preeclampsia [2]; preterm birth, low birth weight infant, and congenital anomalies [3-7]. Despite these negative health consequences, nearly one third of U.S. women experience a short IPI, highlighting a significant public health problem [8].

The majority (74.4%) of short IPIs are unintended [9]. While consistent contraception use is the most effective way to prevent unintended pregnancies [10], rates of postpartum (PP) contraception use in the U.S. remain low [11]. Typically, the current standard of care (i.e., addressing contraception needs PP) has been scheduled between 4 and 6 weeks after delivery, however there is very little documented evidence pertaining to improved patient outcomes at this timepoint [12-14]. Today, many women resume sexual intercourse before six-weeks PP, prior to initiation of contraception [15,16] and data suggests <60% of low-income women covered by Medicaid women attend the six-week PP visit [17-19]. In addition, few women who return for PP care receive contraception at this visit due to a provider barrier (i.e., inadequate time at the visit, lack of knowledge about medically appropriate PP contraception or inability to provide same-day insertion of long-acting methods) [11]. Professional societies such as the American College of Obstetricians and Gynecologists have since noted that the timing of the comprehensive PP visit (i.e. addressing contraception needs PP) should be “individualized and woman centered” and ideally contact should be made with a maternal health provider within the first 3 weeks PP [20].

Despite these recommendations for earlier contact with a provider, efforts to improve PP care to date have been unsuccessful. Specific to contraception, methods such as email and text reminders, or antenatal discussion of the importance of contraception have had limited impact on visit attendance or contraception uptake [21,22]. This has led to testing alternative strategies to improve uptake of PP contraception, such as providing contraception prior to hospital discharge after delivery [23,24]. Unfortunately, efforts to increase immediate PP contraception provision have also faced significant implementation challenges due to a lack of knowledge among providers of appropriate immediate PP contraception options, financial concerns regarding billing and reimbursement, and competing clinical and administrative priorities [25].

A highly promising opportunity to reach PP patients is through newborn and Well-Baby visits. Given the high rate of attendance at Well-Baby visits, routine infant health care visits represent the most regular contact new PP patients have with the health care system [26,27], making it an ideal venue for implementing timely PP contraception care. Indeed, addressing select maternal health needs during a pediatric visit is now standard practice. For instance, PP depression screening for mothers has been incorporated into routine pediatric care, which has been found to be both reliable and feasible [28] and provides an opportunity to integrate PP care into pediatric care.

The Linking Inter-professional Newborn and Contraception Care (LINCC) Trial creates a novel, patient-centered approach to improve access to PP contraception by co-scheduling PP care with newborn and well child visits using the electronic health record (EHR). The LINCC Trial intervention aims to leverage the EHR platform to prompt providers to ask mothers attending a Well-Baby Visit about their PP contraception needs and facilitate co-scheduling of PP contraception care with the next infant visit. To facilitate implementation and limit use of clinic resources, the LINCC Trial intervention will evaluate the effectiveness of the LINCC Trial by comparing pre- and post-intervention outcomes including clinic-level provision of most or moderately effective methods of contraception at two- and six-months PP and rates of short IPI’s. Intervention fidelity as well as facilitators, barriers, and costs of implementation will also be evaluated using an implementation-effectiveness hybrid model.

2. Methods

2.1. Overall approach

Our main hypothesis is that implementation of the LINCC Trial intervention will facilitate women’s access to care, specifically improving receipt of PP contraception and decreasing short IPI pregnancies, and that this approach (i.e. linked care model) could be applicable to addressing other maternal health needs during the PP period.

The LINCC Trial consists of three main aims:

Aim 1 - Conduct formative work (key informant interviews) to develop and refine delivery models that will integrate the LINCC Trial intervention into participating community health centers (CHCs).

Aim 2 - Implement a cluster-randomized, stepped wedge hybrid effectiveness-implementation trial in participating CHCs across the U.S.

Aim 3 - Assess the implementation of the intervention and report on key barriers and facilitators related to successful implementation.

The study will be conducted in three phases, corresponding with our aims: (1) intervention development and refinement through key informant interviews with CHC staff and patients (months 1–18); (2) Roll out of intervention to evaluate primary clinical outcomes (receipt of most or moderately effective methods of contraception by two months PP) and secondary clinical outcomes (receipt of most or moderately effective methods of contraception by six months PP; and rate of short IPI pregnancies (months 19–48); and (3) Assess intervention fidelity as well as facilitators and barriers through key informant interviews with CHC staff and patients (months 19–60).

2.2. Stakeholder collaboration

The LINCC study is a collaboration between the two academic medical centers and AllianceChicago, a national Health Center Controlled Network (HCCN). A HCCN is a group of safety net community health centers (CHCs) that share common infrastructure including Health Information Technology, clinical collaboration to improve quality of care, and resources to enhance access to care. AllianceChicago supports nearly 200 urban and rural sites in 16 states. The CHCs serve diverse, underserved, and high-risk populations including the uninsured, poor, minority, non-English speaking, and homeless. CHCs also serve as an ideal setting for a program intended to reach the women most in need of improved contraception services in the PP period. While the scope and delivery of family planning services across CHCs varies, virtually all CHCs have reported that they provide at least one contraceptive method (99.8%) at one or more clinical sites [29]. Working in partnership with a network with shared infrastructure and CHC staff members will help to ensure the LINCC Trial is designed and delivered in a way that is feasible to scale-up and implement more broadly across the CHC network, if the intervention proves to be effective.

Our partnership leverages complementary strengths and builds on previous, synergistic work conducted to address long-standing disparities in PP care. The CHCs partnered with the HCCN utilize a common EHR system with uniform data definitions and capture methods. In addition, a customized data infrastructure, comprised of data from participating health centers, provides for robust data analysis and reporting against national performance measures at the provider, health center, and population levels. In addition to serving as a tool for clinical decision support and quality measurement, the data warehouse provides the ideal infrastructure for supporting the LINCC Trial by ensuring feasibility of the hybrid-trial but also aiding in the translation and dissemination of the LINCC Trial into a sustainable care model within the CHC network. Additionally, the Multiple-PI model in this program brings together PIs from two academic medical centers with complementary expertise.

2.3. Study design

This study is a Type 1 hybrid clinical effectiveness-implementation trial 30 which aims to improve timely, high quality contraceptive care among patients through implementation of an intervention aimed at facilitating co-scheduling of PP contraception care at the Well-Baby Visit. The primary outcome of clinical effectiveness will be evaluated using data abstracted from EHR. Our secondary outcome of intervention implementation uses mixed methods to measure intervention uptake in real world clinical settings.

We will use a prospective, stepped wedge cluster randomized design [30-33] to evaluate clinical outcomes of the intervention across seven clinical sites within five participating CHCs throughout the U.S. In a stepped wedge cluster randomized design, the randomization occurs at the clinic (CHC) level. Participating sites (CHCs) are observed under the control (non-intervention) condition and the intervention condition sequentially. Sites are randomized with respect to the timing at which they ‘step’ or cross over from the control condition to the intervention.

The stepped wedge design is a practical design for a complex intervention because each cluster provides data under both the control and intervention condition and therefore may serve as its own control, improving the precision of the study. Additionally, no individual patient consent is required for this system-level health service delivery intervention, and therefore outcomes may be obtained through routinely collected de-identified data from the common EHR platform. The pragmatic stepped wedge study design (see Fig. 1 below) involves a three-month control period for all sites, followed by a sequential roll-out of the intervention across the seven participating sites at three-month intervals or “steps” over a 19-month period. Baseline control data will be collected from each site for a minimum of 3 months and maximum of 21 months prior to intervention roll-out, depending on their assigned crossover timing. After collecting baseline data, there will be a 3 month washout period to allow the study team to conduct one month of training before implementing the intervention at each site and to ensure that all patients have the opportunity to be exposed to the intervention for two full months postpartum before primary outcome ascertainment. Each site will be observed for a minimum of three months after the washout period, based on anticipated clinic volume of eligible patients at each clinic per quarter. The entire trial will take place over a 27-month period. (See Tables 1 and 2.)

Fig. 1.

Trial Design (Aim 2).

Table 1.

RE-AIM Evaluation of Implementation of Linked Care.

RE-AIM Construct	Data Source	Outcome
Reach	EHR	Percentage of pediatric visits at 0–6 months of age where EHR modifications were used
Effectiveness	EHR	Percentage of eligible women receiving a most or moderately effective form of contraception at 2 and 6 months PP
Adoption	EHR	Percentage of eligible women who make a co-scheduled appointment
Implementation	EHR field created for study	Work flow EHR changes, tailored to each site to offer co-scheduled appointments Acceptability and feasibility of the intervention as reported by PP women and clinic staff Barriers or facilitators of implementation that impact fidelity to the intervention as reported by CHC staff and PP women (i.e., length of visits, no shows)
	Key informant interviews
Maintenance	EHR	Reach and Implementation outcomes measured at years 4 and 5 of grant

Adapted from Glasgow et al. 2007

Table 2.

Data sources and their corresponding role in study analyses.

Source	Study Domains	Data Variables	Data collection
EHRs	Socio-demographic and clinical covariates	Race/ethnicity, age, SES	Automated extraction of discrete EHR fields
	Effectiveness (primary and secondary outcomes)	Primary: Receipt of contraception by 2-months PP; Secondary: Receipt of contraception by 6-months; rate of short IPIs	Automated extraction of discrete EHR fields
	Scheduling data points	Number of co-scheduled appointments	Automated extraction of discrete EHR fields
CHE platform (Patient Survey)	Implementation	Acceptance or declination of co-scheduled appointments; Reason for declination	Automated extraction of discrete CHE fields
Key informant interviews (Aim 1)	Implementation	Acceptability, Adoption	Transcripts from interviews with PP patients, CHC staff, champions
Key informant interview (Aim 3)	Implementation	Acceptability, Adoption, Fidelity and Maintenance	Transcripts from interviews with PP patients, CHC staff, champions

2.4. Randomization

As the LINCC Trial includes changes to healthcare delivery, the intervention itself is diffuse and patient randomization is not feasible. Therefore, randomization will occur at the location of care level. The cluster, or unit of analysis, is the location of care affiliated with each participating CHC, which is critical for a system-level intervention. In a stepped wedge study, the design is extended so that every cluster provides before and after observations and every cluster switches from the control (usual care) to the intervention condition (linked care) at regular intervals (“steps”), at different points in time. The order in which a cluster switches from control to intervention is randomized. A random number generator will be used to assign sites to the step in which they receive the intervention (see Fig. 1 for timeline).

2.5. Study Setting

We have partnered with seven CHCs located in Hawaii, Illinois, and New York. Six of the seven CHCs share a common EHR platform (AthenaPractice) and one uses AthenaOne; both platforms are offered by the HCCN. We estimate accrual of 1134 study subjects.

2.6. Eligibility criteria

2.6.1. Site-level eligibility and Site engagement and recruitment

Engagement efforts focused on study sites which met the eligibility criteria: 1) both pediatric and contraceptive care on-site; and 2) volume of at least 35 deliveries per quarter. In addition to eligibility criteria, the study team gauged interest in participation and assessed sites’ readiness for implementation and ability to adapt a scheduling workflow needed for the intervention. Sites with a centralized scheduling system declined to participate due to the inability to specifically adapt the workflow. Each site consented to participate in the study before randomization.

2.6.2. Target population

All PP women attending a Well-Baby Visit for an infant aged 0–6 months at a participating site during the study period will be eligible for inclusion.

2.7. Enrollment

As this is a health system level intervention, all eligible patients (noted above) are exposed to the intervention on a rolling basis. Each time a PP woman attends a Well-Baby Visit for an infant aged 0–6 months during the intervention period, providers will be prompted to ask about their PP contraception needs and facilitate co-scheduling of PP contraception care with routine newborn care visits (if needed), regardless of whether it’s been discussed at a previous visit. We anticipate each site will have tailored workflows that specify how each clinic team identifies PP women eligible for LINCC. The LINCC Trial intervention will be implemented as a system-level intervention at each participating site.

2.7.1. Intervention Development (Aim 1)

To identify implementation factors for the trial, key informant interviews will be conducted with staff and patients of participating CHCs (Aim 1). Staff interviews will identify barriers and facilitators to implementation of the intervention and allow the research team to better understand the workflow of each clinic. Interviews with PP patients (<9 months PP) will identify patient’s preferences and needs regarding PP contraception.

2.7.2. The intervention (Aim 2)

The intervention components are as follows (Appendix 1):

1. EHR Platform (Provider-side modifications):

Well Child Care (WCC) Form:

Using the standard age-specific WCC forms for an infant ≤6 months, a new field will be created to prompt providers to ask if PP patient is using contraception (yes/no field with option for free text) and offer patients a contraception visit. This field will also contain helper text that provides a brief description that reminds providers to offer all moms a contraception visit, regardless of current contraception use. All moms are offered a contraception visit, regardless of current use, in the event that they would like to discuss alternate, or more preferred contraception methods.

Assessment & Plan Form:

This standard form includes a drop-down to specify timing and reason for next visit. At the end of each infant visit, when selecting a reason for the next visit, study-specific drop-down options will be created to capture if a co-scheduled appointment is desired (or declined) by the patient. These fields are included in the after-visit summary (AVS), which is routed to scheduling staff at the completion of a visit.

a. Quick Text:

A study-specific dot phrase (“.LINK” or “.LINCC”) will be available to indicate to the scheduling staff that a co-scheduled appointment has been requested for mom and baby. This quick text can be used in a free text field in the Assessment and Plan form.

2. Scheduler Workflow and Platform (scheduler-side modifications):

The sites use a separate, linked scheduling platform, Community Health Enhancements (CHE). Modifications were built into the scheduling software as a back-up system to ensure patients were offered a follow-up contraception visit.

a. Patient Survey:

At the time of infant checkout, any staff member responsible forchecking out infant will be trained to document PP patient’s interest in scheduling a contraception visit. These questions are to be asked regardless of if the provider has recorded acceptance or declination of a co-scheduled visit within the AVS.

b. Acceptance of co-schedule appointment:

If ‘yes’ or ‘no preference’ is selected, staff will proceed with usual processes to schedule next appointment for PP patient and infant on same day when possible.

c. Declination of co-scheduled appointment:

If ‘no’ is selected, staff will proceed will capture reason for decline and select one of the following drop-down options: 1).

Mom receives care elsewhere; 2) Mom already has visit scheduled; 3) Co-scheduling is not convenient; 4) Mom already has birth control; or 5) No reason provided.

2.8. Site implementation

Depending on each site’s unique workflow and interaction with the EHR/scheduling platforms, certain components will be tailored to best fit the staff’s workflow at each site. At each site, a clinical site champion will be identified. The clinical champion is a member of staff who will oversee the study activities. The research team will engage in ongoing communication with the clinical champion via monthly calls to troubleshoot any issues related to participation. We will offer virtual training sessions two months prior to the intervention start date for each site to train clinic staff on the intervention (i.e., co-scheduled care) and how their workflow will be modified during the implementation of this intervention. We will use evidence-based implementation strategies [33] to promote intervention adoption into existing workflows including: stakeholder engagement, development of a site-specific training with tip sheets and videos, iterative training and technical assistance, and continuous monitoring and feedback.

2.8.1. Intervention Assessment (Aim 3)

To assess implementation of linked PP contraception and newborn care, we will conduct key informant interviews with clinic staff and patients at approximately the midpoint of each site’s intervention period (Aim 3). These interviews will help us to identify key barriers and facilitators related to successful implementation of the intervention as well as implementation outcomes such as acceptability, feasibility, and fidelity over time.

2.9. Study outcomes

2.9.1. Outcome measures

Our study is powered on the primary clinical effectiveness outcome, receipt of most or moderately effective methods of contraception by two months PP, as defined in the Contraceptive Care Measure for Postpartum Most or Moderately Effective Contraceptive Methods developed by the Office of Population Affairs of the U.S. Department of Health and Human Services [35]. Contraception receipt will be ascertained using the following data elements from the EHR database: visit billing codes for ICD-10 diagnosis codes, Current Procedural Terminology (CPT) procedure codes, and Healthcare Common Procedure Coding System (HCPCS) codes; National Drug Codes (NDC) codes for prescriptions in the medication data; and/or an open-ended Current Contraception field from the Contraception History EHR form. In addition, we will study receipt of most or moderately effective methods of contraception by six months PP. We will study the rate of short IPI pregnancies by using EHR data to estimate the clinic-level rate of incoming prenatal patients presenting with a short IPI during the baseline period (control condition) and 18 months after the introduction of the intervention in each CHC.

Implementation outcomes include acceptability, local adoption/penetration, fidelity to the intervention protocol, and sustainability of the intervention across each site [36]. Acceptability will examine perceptions of the intervention components among site champions, patients, and other members of site teams. Adoption/penetration will measure the reach and integration of LINCC and its practices throughout each participating site. Fidelity is being conceptualized as staff adherence to the standard operating procedures. Sustainability will examine the extent to which LINCC is maintained within the CHC’s ongoing operations during the later months of the intervention period.

Mixed methods will be used to elaborate on quantitative RE-AIM findings (Reach, Effectiveness, Adoption and Measures) [37]with themes from qualitative data [38] as outlined in Table 1.

2.10. Data sources

LINCC study uses multiple data sources as outlined in Table 2. The EHR will serve as the data source for all socio-demographic and clinical covariates, as well as the primary and secondary clinical effectiveness outcomes. The HCCN’s information technology team will build custom reports in their data warehouse for the investigators to access and validate data in real-time.

The main sources of data for our implementation outcomes include: the EHR, the scheduling platform (CHE) and qualitative data collected through observations shared by site champions, and interviews (Aim 3). The EHR platforms will provide quantitative measures of fidelity to the intervention components, as documented by the providers and schedulers in their day-to-day activities. Observations of staff by site champions will further assess fidelity, as well as local adoption/penetration. Key informant interviews with patients and clinic staff at the mid-point of the intervention period will provide data to further examine acceptability, adoption, and sustainability (Aim 3).

2.11. Statistical analysis of effectiveness outcomes

2.11.1. Primary Outcome: Receipt of all methods of contraception by 2-months PP

We will compare the distribution of patient characteristics between the unexposed observations from the control period and the exposed observations during the intervention period, which are analogous to distributions across arms in a parallel cluster randomized trial. Then, we will plot rates of our outcomes over each three-month time period, comparing how rates differ from the control period to the intervention period, overall and within each clinic, to examine the effects of the intervention and time on the outcomes. To examine treatment effects for our primary and secondary outcomes related to receipt of contraception, we will use a generalized linear mixed model, adjusting for the confounding effect of time with the random effect for CHC site and fixed effect for each time interval [30].

2.12. Qualitative analyses

Semi-structured interview guides will be developed around CFIR and RE-AIM constructs to elicit information related to barriers and facilitators encountered during the implementation of the linked care model. Interviews will be conducted in English or Spanish depending on preference of participant and will be audio-recorded and transcribed. Qualitative data will be coded independently by two coders using Dedoose Software version 7.0.23 (Los Angeles, CA) [36] and then discussed to resolve coding discrepancies. Thematic analysis will focus on the perceptions of acceptability, adoption, and maintenance related to implementing the LINCC Trial intervention. After coding is finalized and consensus has been reached, codes will be reviewed to generate emerging themes. Emerging themes will then be categorized by CFIR construct and/or implementation outcome: acceptability, feasibility, and fidelity. Similarities and differences in the themes will be compared across sites. This will allow us to identify the extent to which LINCC is seen as acceptable or having substantial reach within the health system among the participating sites.

2.13. Sample size and power

We expect approximately 1134 total patients over the full intervention period (27 months). Sample size calculations for cross-sectional stepped wedge designs consider the intra-cluster correlation (ICC) like other cluster randomized designs, as well as the confounding effect of time [31]. An available web-based application for stepped wedge designs [39,40] was used to verify that a minimum sample size of 18 women every 3 months per clinic provides adequate power (>80%) to detect a change from a baseline of 25% of women using any contraception by 2 months PP to 40% assuming an ICC of 0.10 [34,41]. This minimum sample size will be part of the selection criteria in determining eligibility of CHCs in Aim 1 [17,31].

3. Discussion

There is an urgent need for novel interventions to improve timely access to contraception after pregnancy and reduce short IPIs. The LINCC Trial includes a multi-component patient-centered intervention designed to overcome barriers and improve access to PP care, specifically for vulnerable populations. Results from the LINCC Trial can be used to change how postpartum contraception needs are identified. Safe interpregnancy interval spacing impacts health outcomes for both mothers and infants. Routine pediatric care visits offer regular and frequent opportunities to ask mothers about contraception needs. Our approach is EMR-based and could be used broadly in pediatric visits nationally. The LINCC trial includes the voices of women and clinicians, both of which are critical when developing a clinical practice change.

3.1. Limitations

As with all trials, the LINCC trial may be subject to limitations. First, due to the stress put on CHCs by the COVID-19 pandemic, we were delayed in starting the trial and were only able to recruit seven of the ten sites we had initially planned for. Also related to the fallout from the pandemic, the competing priorities at CHCs may be exacerbated and affect staff members’ ability to prioritize the implementation of the LINCC intervention when their site is scheduled to cross over. In addition, due to the sequential nature of the stepped wedge design, the intervention effect is subject to confounding by underlying secular trends in the outcome [42]. This can be statistically addressed by controlling for time in the mixed regression model, which is part of our analytic plan. However, the functional form of the time trend must be appropriately specified and the assumption must be made that the time trend is the same across clusters, which may be violated given the relatively small number of sites in the trial. Finally, we may encounter data quality issues due to our reliance on an EHR, which is intended for clinical use rather than research, for our quantitative measures.

4. Conclusions

The LINCC Trial creates a novel, patient-centered approach to improve access to PP contraception care and in turn, reduce the poor health sequela associated with short IPI pregnancies. The goal of the LINCC Trial is to build and study a comprehensive fully integrated model of care for PP patients and children. This intervention is foundational research to study how best to integrate a new model of care while addressing what is often an immediate care need for PP patients – family planning.

Data availability

No data was used for the research described in the article.