Primary Care Utilization Within 1 Year After a Facilitated Postpartum-to-Primary Care Transition

Abstract

Objective:

To evaluate the effect of a behavioral science-informed intervention designed to facilitate postpartum-to-primary care transitions on primary care visits and screenings within 1 year postpartum for individuals with chronic conditions or pregnancy conditions with long-term health risks.

Methods:

This was a planned secondary analysis of a randomized controlled trial of a behavioral science-informed intervention designed to increase primary care practitioner (PCP) visits within 4 months postpartum versus routine care. The intervention included default PCP visit scheduling with nudge reminders and use of tailored language. The primary outcome for this secondary analysis was attending an “annual exam” or health care maintenance visit with a PCP within 1 year postpartum. Visits with a PCP for any reason and receipt of the following screenings or services by a PCP (e.g., weight, blood pressure, mood screening) were also compared. Outcomes were compared between groups with chi-square testing.

Results:

All 353 participants were followed through 1 year after their due date: 173 in the control group and 180 in the intervention group. More patients in the intervention group attended an annual exam with a PCP within 1 year versus the control group (59.0 vs 39.3%, p<0.001) and had a PCP visit for any reason (72.8 vs. 61.3%, p=0.02). A significantly higher rate of mental health disorder screening was observed in the intervention group (63.9 vs. 55.5%, p=0.046); significant differences in other screenings were not observed.

Conclusion:

This relatively simple and low-cost intervention designed to facilitate transition from postpartum to primary care within the first 4 months demonstrated benefits for PCP engagement within the first year postpartum.

Clinical Trial Registration:

ClinicalTrials.gov, NCT05543265.

Precis

An intervention designed to reduce patient administrative barriers and facilitate an individual’s transition from postpartum to primary care increased PCP visits within the first year postpartum.

Introduction

Postpartum to primary care transitions are important for individuals with ongoing care needs after pregnancy, but there are many barriers to effective transitions in practice.^1–7 Previously, we demonstrated that a behavioral science-informed intervention, which consisted of default (“opt-out”) scheduling of primary care appointments, along with patient-directed tailored messages and nudge reminders, increases the percentage of postpartum individuals attending a primary care visit within 4 months postpartum by 18.7% compared with those receiving standard care.⁸ It was also associated with increased receipt of recommended screenings and services by a primary care practitioner (PCP), such as blood pressure and mental health screenings.⁸

The evaluation of this intervention did not extend past the 4 months after delivery, and it is unknown if the increase in primary care visits among the intervention group persisted when examining visits within the first year postpartum. Thus, the primary objective for this study was to understand if the intervention resulted in more individuals attending a primary care visit overall or, alternatively, simply shifted the timing of the first PCP visit after delivery to within 4 months among those who would have already attended a PCP visit within the upcoming year. Secondarily, our objective was to determine if the intervention increased the receipt of routine screening and services by a PCP and if the intervention increased ongoing PCP engagement within the first year postpartum. We hypothesized that the intervention would increase PCP visits within the first year postpartum, increase receipt of PCP screenings and services, and increase ongoing PCP engagement.

Methods:

This was a planned secondary analysis of an individual-level randomized controlled trial (RCT) of a behavioral science-informed intervention to improve the postpartum-to-primary care transition versus routine care (NCT05543265) completed between November 2022 and October 2023.⁸ Individuals were eligible if they had a chronic or pregnancy-related comorbidity with known long-term health risks, which included pre-pregnancy body mass index (BMI) ≥30 kg/m², anxiety or depression, chronic or pregnancy-related hypertension, or pre-existing or gestational diabetes. Also, individuals had to have a PCP identified in the electronic health record (EHR) to be enrolled, as this study was focused on re-establishing primary care; the intervention was not designed to find and establish care with a new PCP, as discussed in detail in the discussion of primary trial results.⁸ Characteristics of patients enrolled in the primary trial, including race and ethnicity, are reported in the intervention and control groups to facilitate interpretation.

The intervention included the following components: 1) default scheduling of an annual exam with the patient’s PCP within 4 months of a patient’s estimated due date (EDD) or 1 year after their last annual exam if they had a recent annual exam; 2) salient labeling of the PCP appointment as the “Pregnancy-to-Primary Care Transition Appointment;” 3) targeted language (e.g., “appointment has been reserved for you”); and 4) nudge reminders for the appointment delivered via SMS and electronic health record messaging 2 weeks postpartum, and at 4 weeks and 1 week before the appointment. Of note, the study only scheduled annual exam visits on a patient’s behalf (as opposed to a follow-up or “problem” visit) because these visits are recommended annually and offered without cost-sharing. In communications, patients were instructed to contact their PCP’s office or research staff to cancel or change the appointment if necessary. The control group received routine pregnancy and postpartum care. The complete details of the original trial and the intervention have been previously published.⁸

The primary outcome in this secondary analysis was attending an “annual exam” by a PCP within 1 year after a patient’s estimated due date. Practitioners classified as PCPs included physicians and advanced practice clinicians affiliated with the following medical specialties: internal medicine, family medicine, pediatrics and adolescent medicine, and gynecology. The clinical documentation from each PCP encounter was reviewed; visits were classified as an “annual exam” or “health care maintenance” visit if the terms appeared in the notes or encounter-specific affiliated diagnosis codes. We also examined PCP visits for any reason (i.e., not limited to annual exams).

We examined the time to the first PCP visit at two intervals: 0–4 months after the patient’s EDD (the primary trial’s follow-up period, when the majority of the intervention group participants had their appointment scheduled and received nudge reminders) and 5–12 months (the remaining months in the first year postpartum). As a measure of engagement with primary care, we compared PCP visit frequency between the two groups, categorized as no visits with PCP, 1 PCP visit, or ≥2 PCP visits. Finally, we examined the receipt of routine screenings and services by a PCP: weight measurement, blood pressure assessment, mental health disorder screening, and documentation of a discussion or plan for diabetes screening and contraception.

All outcomes were compared between the two groups (intervention versus usual care) using chi-squared tests. While this was a pre-planned secondary analysis, the primary trial was not specifically powered for the primary outcome of this secondary analysis. Sample size calculations for the primary trial were previously published.⁸ As a sensitivity analysis, we limited the cohort to individuals whose PCP was affiliated with the same health system as where they received their obstetric care. This smaller group represents a cohort in which the likelihood was low that a patient attended a PCP visit that was not ascertained in the EHR review (i.e., a PCP visit occurred at a clinic that is not integrated with the health system’s EHR and thus was not observed during EHR review).

STROBE reporting guidelines were followed. The Mass General Brigham Institutional Review Board reviewed and approved the original trial and its planned secondary analyses. Stata v16.1 was used for the statistical analyses. P-values <0.05 were considered statistically significant.

Results:

All 353 participants in the primary trial were followed through 1 year after their due date: 173 in the control group and 180 in the intervention group. A comparison of patient demographics at the time of enrollment between the two groups is included in Table 1. Individuals had a mean (standard deviation (SD)) age of 34.1 (4.9) years old, and the distribution of self-reported races was as follows: 7.4% Asian, 6.8% Black, 15.0% multiple races or “Other,” and 68.6% White; 2.3% declined to report their race. Overall, 75.8% had anxiety or depression, 15.9% had a chronic or pregnancy-related hypertensive disorder, 19.8% had pre-existing or gestational diabetes, and 40.4% had a pre-pregnancy BMI ≥30 kg/m²; some participants met more than one eligibility criterion. Medicaid was the primary payer for the delivery encounter for 21.9% of patients, and 93.2% attended a postpartum care visit between 4–8 weeks postpartum.

Table 1:

Characteristics of individuals in the control vs. intervention group at the time of enrollment

Characteristic	Control Group (n=173)	Intervention Group (n=180)
Patient Age at Estimated Due Date, Mean (SD), Years	34.0 (5.0)	34.2 (4.8)
Primary Site of Prenatal Care
Hospital-based Clinic	121 (69.9)	129 (71.7)
Community-based Clinic	52 (30.1)	51 (28.3)
PCP Visit within 3 Years Prior to Enrollment	121 (69.9)	111 (61.7)
PCP Affiliation
PCP within Health Care System	122 (70.5)	117 (65.0)
Health Condition
Anxiety or Depression	128 (74.0)	138 (76.7)
Chronic or Gestational Hypertensive Disorder	26 (15.0)	31 (17.2)
Chronic or Gestational Diabetes Mellitus	38 (22.0)	31 (17.2)
Obesity (Pre-pregnancy BMI >= 30 kg/m2)	75 (43.4)	69 (38.3)
Racea
Asian	13 (7.5)	11 (6.1)
Black	12 (6.9)	14 (7.8)
Multiple Races or “Other”b	28 (16.2)	25 (13.9)
White	115 (66.5)	127 (70.6)
Declined / Not reported	5 (2.9)	3 (1.7)
Ethnicitya
Hispanic	41 (23.7)	37 (20.6)
Non-Hispanic	127 (73.4)	139 (77.2)
Not reported	5 (2.9)	4 (2.2)
Preferred Languagea
English	161 (93.1)	167 (92.8)
Spanish	12 (6.9)	13 (7.2)
Marital Statusa
Married	125 (72.3)	137 (76.1)
Not married	48 (27.7)	43 (23.8)
Educationa
High School Graduate or Some High School	30 (17.3)	22 (12.2)
Some College	11 (6.4)	22 (12.2)
Bachelor’s or Associate Degree	69 (39.9)	71 (39.4)
Graduate School Degree	63 (36.4)	65 (36.1)
Individual Annual Earningsa
<$30,000	32 (18.5)	36 (20.0)
$30,000 – $74,999	37 (21.4)	55 (30.6)
≥$75,000	82 (47.4)	76 (42.2)
Not reported	22 (12.7)	13 (7.2)
Primary Payer for Delivery Hospitalization
Medicaid	40 (23.1)	35 (19.4)
Private / Other	130 (75.1)	138 (76.7)
Unknown	3 (1.7)	7 (3.9)
Mode of Delivery
Vaginal	107 (61.9)	117 (65.0)
Cesarean	66 (38.2)	63 (35.0)
Preterm Delivery	10 (5.8)	16 (8.9)
Attended Obstetric Postpartum Care Visit	163 (94.2)	166 (92.2)

^aSelf-reported.

^bPatients could select “Other” as a race option if they did not self-identify with the other options: American Indian or Alaskan Native, Asian, Black, Native Hawaiian or Other Pacific Islander, White.

More patients in the intervention group received an annual exam with a PCP within 1 year compared with the control group (59.0 vs 39.3%, p<0.001) (Table 2). Similarly, there was a significant increase in a primary care visit for any reason (72.8 vs. 61.3%, p=0.02). Findings were similar in the sensitivity analysis limiting the cohort to the group of patients with a PCP affiliated with the same health system (Table 2).

Table 2:

Comparison of Primary Care Practitioner visits between the control and intervention groups within 1 year postpartum.

Overall Cohort
Outcome	Control Group N=173	Intervention Group N=180	P value
Primary Outcome
Annual exam within 1 year	68 (39.3%)	107 (59.4%)	<0.001
Visit timing: 0–4 months	25 (14.5%)	64 (35.6%)	<0.001
Visit timing: 5–12 months	43 (24.9%)	43 (23.9%)	0.83
Secondary Outcomes
Visit for any reason	106 (61.3%)	131 (72.8%)	0.02
PCP Engagement
No visits	67 (38.7%)	49 (27.2%)	0.07
1 visit	53 (30.6%)	64 (35.6%)
2+ visits	53 (30.6%)	67 (37.2%)
Subgroup of Individuals with a PCP in the Health System
	N=122	N=117
Outcome	Control Group N=122	Intervention Group N=117	P value
Primary Outcome
Annual exam within 1 year	60 (49.2%)	85 (72.6%)	<0.001
Visit timing: 0–4 months	20 (16.4%)	52 (44.4%)	<0.001
Visit timing: 5–12 months	40 (32.8%)	33 (28.2%)	0.44
Secondary Outcomes
Visit for any reason	87 (71.3%)	101 (86.3%)	0.005
PCP Engagement
No visits	35 (28.7%)	16 (13.7%)	0.01
1 visit	45 (36.9%)	47 (40.2%)
2+ visits	42 (34.4%)	54 (46.2%)

PCP, primary care practitioner.

Figure 1 shows the timing of PCP visits for an annual exam (Panel A) and visits for any reason with a PCP (Panel B) in the overall cohort (solid lines) and in the subgroup of individuals whose PCP was in the same health system (dashed lines). PCP visit rates were higher for the intervention group than the control group throughout the year postpartum, but the gap grew fastest within the first 0–4 months after delivery and then remained steady in the 5–12 months after delivery. When stratified by follow-up period, the rates of the annual exam with a PCP were 35.6 vs. 14.5% from 0–4 months (p<0.001) and 23.9 vs. 24.9% (p=0.83) from 5–12 months postpartum between the intervention versus control group, respectively.

Figure 1:

Time to first primary care practitioner (PCP) visit within 12 months postpartum between intervention group and control group. Annual visits (A) and visits for any reason (B). The cumulative distribution of annual visits (A) and visits for any reason with a PCP (B) are shown among the control (orange lines) and intervention (blue lines) groups for the overall cohort (solid lines) and the subgroup of individuals with a PCP affiliated with the same larger health care system as their obstetric practice (dashed lines). EDD, estimated due date.

Figure 2A compares the receipt of common services and screenings with a PCP within 12 months postpartum. Those in the intervention group were more likely to receive mental health disorder screening than the control group (63.9 vs. 55.5%, p=0.046); other services were numerically higher in the intervention group though not statistically different. Figure 2B shows the same comparisons in the subgroup analysis. Among individuals with a PCP within the same health system, those in the intervention group were more likely to have a blood pressure screening (79.5 vs. 67.2%, p=0.03), weight assessment (78.6 vs. 65.6%, p=0.03), mental health disorder screening (50.8 vs. 65.0%, p=0.03), and have a documented plan about depression and or mental health in the EHR (63.2 vs. 47.5%, p=0.02) than those in the control group.

Figure 2:

Comparison of receipt of primary care services at a primary care practitioner (PCP) visit within 12 months postpartum between intervention group and control group participants. Overall cohort (A) and subgroup (B). The components of primary care screenings and assessment between the intervention and control groups of the overall cohort (A) and the subgroup of individuals with a PCP affiliated with same larger health care system as their obstetric practice (B). *P-value <.05.

PCP engagement was measured by comparing the number of PCP visits between the groups. The distribution of PCP visits between the two groups is shown in Table 2 for the overall cohort and for the sensitivity analysis. In the overall cohort, 67 (37.2%) in the intervention group vs. 53 (30.6%) individuals in the control group had more than one PCP visit within 12 months postpartum (p=0.07 for the comparison of the distribution of visits between the groups). In the sensitivity analysis including only participants with a PCP in the health system, 54/117 (46.2%) in the intervention group versus 42/122 (34.4%) in the control group had more than one PCP visit (p=0.01 for the comparison of the distribution of visits between the groups).

Discussion:

An intervention designed to facilitate the postpartum-to-primary care transition among individuals with chronic conditions resulted in higher rates of PCP annual visits and visits for any reason within the first year after delivery compared with routine postpartum care. This increase was largely driven by individuals in the intervention group re-connecting with a PCP within the first 4 months postpartum, which is likely attributed to the components and design of the bundled intervention, including default scheduling of the PCP visit within 4 months, tailored messaging, and nudge reminders.⁸ Patterns of primary care use between the groups were similar between 5–12 months postpartum.

Other studies have examined primary care utilization after pregnancy and interventions to improve care transitions. Multiple studies have examined the role of postpartum care navigators for individuals with risk factors or with specific conditions.^9–13 These programs have largely been effective at assisting patients in receiving recommended care after their delivery, though they can be resource-intensive and thus may have limited scalability to a larger population. More similar to our study, Cameron et al. evaluated the effect of a referral scheduling pathway for patients with diabetes or hypertensive disorders and found an increase of 48.1% in attending a PCP visit in the 1^st year postpartum; this intervention utilized a team member to contact and schedule patients once referred by an obstetrician but did not provide reminders.¹⁴ The findings from Cameron et al. and our study highlight that the removal of administrative burden on patients may have a positive effect on primary care engagement while recognizing there are still barriers unaddressed amongst the highest risk groups.¹⁴ Compared with health care navigators, a less resource-intensive intervention such as the bundle tested in this study may be a scalable solution to increase primary care engagement in the postpartum period.

Unlike in the primary analysis, which focused on the first four months postpartum, only the rate of mental health disorder screening by a PCP was significantly different between the intervention and control group over the full postpartum year, with other PCP screenings numerically, but not significantly, higher in the intervention arm. However, in the sensitivity analysis restricting to individuals with a PCP within the study health system—where outcome ascertainment is strongest—we observed significantly higher rates of blood pressure, weight, and mental health screenings and documentation of a plan for mental health concerns or depression in the intervention group. This finding highlights that individuals generally receive these screenings as part of PCP visits for any reason, not just annual exams. This study was not designed or powered to measure the effect on longer-term health outcomes or the effect of attending an annual exam versus a visit for another reason (e.g., urgent visit for respiratory illness, etc.). This study also highlights that a significant share of individuals did not receive an annual visit within the 12 months after their pregnancy, highlighting the need for ongoing work and strategies to reduce barriers, catalyze increased health activation, and encourage ongoing care for chronic health conditions after pregnancy.

While this was a planned secondary analysis of a randomized controlled trial that enrolled individuals at a large hospital and its outlying affiliated community clinics, the fixed sample size may limit the detection of true differences between the groups. The study did not require individuals to have a PCP within the same health system, which ultimately may have limited ascertainment of outcomes for individuals who did have a PCP visit outside the health system; thus, it’s likely the sensitivity analysis limited to individuals who have an affiliated PCP more accurately represent the rates of PCP visits.

Also, several factors limit the generalizability of these findings. First, we excluded individuals without a PCP listed in the EHR because the administrative burdens associated with establishing care with a new PCP differ from reconnecting or transitioning back to an established PCP. Before the primary trial, our initial review demonstrated that >90% of patients had a PCP listed in the EHR, thus only a minority was excluded from enrollment due to this stipulation. Of note, we did not require patients to have ever seen the listed PCP to be eligible, and some patients were unaware they had an “assigned PCP.” Second, the majority of OB clinics default schedule a postpartum visit around 6 weeks. Thus, our population had high rates of obstetric postpartum visit attendance (>90%); however, similar rates are reported in the PRAMS survey on postpartum visit attendance.¹⁵ Last, the study recruited from multiple sites, including community health centers, and provided all study materials in English and Spanish. However, due to the catchment area and size of the main obstetric clinic, there was a relatively high percentage of individuals who self-reported their race as White, were privately insured, and had a college degree. We acknowledge that the study was underpowered to examine subgroup-specific treatment effects; however, results from the primary trial show similar effect sizes in individuals who speak Spanish as a primary language, use Medicaid for insurance, and who report lower annual incomes.⁸ Further work via a larger trial is underway to examine the intervention effects among at-risk populations, particularly those likely experiencing higher administrative burdens.

A relatively low-resource intervention designed to reduce patient administrative barriers and facilitate an individual’s transition from postpartum to primary care increased PCP visits within the first year postpartum. The intervention also showed signs of increasing ongoing engagement with primary care, as assessed by multiple visits with a PCP. Ongoing follow up will be required to determine if this type of facilitated transition improves health outcomes after pregnancy and sparks long-term engagement with primary care in individuals with conditions that are known to affect their lifelong health.