Leveraging Primary Care Resources Through Quality Improvement to Increase Follow-up After Mental Health Emergency Department Visits

Abstract

Introduction:

Timely follow-up after mental behavioral health (MBH) crises can improve health outcomes; moreover, it is a national quality metric. We aimed to improve the completion of 7-day follow-up after emergency department (ED) discharge for youth with an MBH concern from 28.7% of discharges to 50.0% within 12 months.

Methods:

We used quality improvement methods from November 1, 2024, to October 31, 2025, across 3 primary care sites and 2 ED sites. Baseline data, literature reviews, and lived experience identified key drivers and failure modes. We developed 2 interventions: (1) direct phone outreach after discharge and (2) increased access in pediatric primary care (PPC). The primary outcome measure was the percentage of ED MBH discharges with a completed follow-up appointment within 7 days of discharge. Process measures included the percentage of patients who received a postdischarge outreach call and who followed up within PPC. The balancing measure was return ED MBH visits within 30 days.

Results:

The mean follow-up percentage increased from 28.7% (n = 258) to 40.3% (n = 334). Follow-up for Medicaid-insured patients increased from 24% to 35% and remained stable for commercially insured patients (42%), decreasing our equity gap by 11%. Outreach calls occurred following 52.4% of discharges. Follow-up visits within PPC increased from 3.5% of discharges to 10.7% of discharges. There was no change in return ED visits within 30 days of ED discharge.

Conclusions:

Increasing access to pediatric primary care and implementing timely postdischarge follow-up calls can improve follow-up within a large healthcare system for patients discharged from the ED for MBH concerns.

INTRODUCTION

Rising rates of mental and behavioral health (MBH) conditions among youth have led to increases in emergency care nationally.^1,2 Between 2010 and 2016, youth MBH emergency department (ED) visits increased by 45% per year, over triple the rate of increase for non-MBH pediatric ED visits in the same timeframe.³ In total, EDs across the United States see 5,000–7,000 adolescents per week for MBH concerns.⁴

In 2024, the National Committee for Quality Assurance (NCQA) developed a quality metric recommending that patients with MBH concerns follow-up within 7 days of ED discharge.⁵ Follow-up visits enable providers to identify changing or emerging issues while promptly addressing any treatment barriers.⁶ Prior ED improvement projects have identified outpatient follow-up as an important component of MBH crisis care.⁷

Varied efforts have been shown to improve follow-up completion. Professional outreach postdischarge, addressing social drivers of health, and improving access have facilitated successful follow-up.^8,9 Stigma, provider shortages, and poor discharge communication are common barriers.^10,11

Two pediatric subpopulations are less likely to receive follow-up care: patients enrolled in Medicaid^11–13 and children without an established MBH provider.^11,14,15 As our primary care clinics see a high percentage of youth with Medicaid insurance, many of whom experience challenges accessing MBH treatment outside of primary care, this NCQA metric was particularly pertinent to our population. We theorized that formalizing postdischarge outreach and improving access within primary care would increase follow-up completion.

Our specific aim was to improve the percentage of patients discharged from a pediatric ED with an MBH diagnosis who receive follow-up within 7 days from 28.7% to 50% within 12 months. We also aimed to reduce the follow-up gap between Medicaid-insured and commercially insured patients by 10% within 12 months.

METHODS

Setting and Context

Interventions took place at 3 pediatric primary care centers (PPCs) and 2 EDs at a freestanding pediatric academic medical center in the Midwestern United States. The 3 PPCs serve approximately 30,000 total patients (70% Black, 8% Hispanic, and 82% Medicaid-enrolled or uninsured) 0–18 years of age. PPC patients average 22 MBH ED visits per week, including those hospitalized and those discharged from the ED.

Two PPCs are urban, 1 suburban. All 3 have full-time social workers (SWs) who provide crisis stabilization and address social drivers of health. All 3 also have integrated behavioral health psychologists delivering prevention and intervention services during physician visits. One PPC has a co-located psychiatrist (available 1.5 d/wk) who supports the team through consultative services. The other 2 have a co-located psychiatric nurse practitioner (available 1 d/wk per center) who provides long-term care for PPC patients with complex MBH conditions.

The EDs have an embedded MBH team available 24 hours/d with psychiatric SWs and an on-call psychiatrist. The ED team partners with an internal psychiatric bridge clinic that offers crisis assessment, stabilization, and intervention services, including short-term psychotherapy soon after ED discharge. After providing short-term psychotherapy, bridge clinicians can refer patients to PPC or to higher levels of care based on their clinical judgment. On average, PPC patients no-show 33% of their bridge appointments compared with 12% for the bridge clinic overall, indicating unique access challenges for our patients.

To arrange follow-up, the ED MBH team often attempts to contact a patient’s existing MBH provider. Based on a manual audit of SW safety plans, ~43% of PPC patients report having an existing provider. However, the status and nature of these relationships (ie, active versus inactive versus establishing care) were unclear. For patients without an established MBH provider, the ED MBH team attempts to coordinate an appointment with either a bridge clinic or primary care. However, no standardized processes existed to involve primary care teams for follow-up care.

We designed and tested interventions for patients 6–18 years of age to address this gap using quality improvement (QI) methodology and plan-do-study-act (PDSA) cycles guided by the Model for Improvement.¹⁶ Baseline data collection began January 1, 2024. Interventions began November 1, 2024, and lasted until October 31, 2025.

Intervention Team

In 2024, our state formed a learning network focused on improving 7-day follow-up for MBH ED discharges. In response, we formed a local QI team with three parent representatives (each had a child with a history of MBH disorders), two primary care physicians, one primary care SW, ED MBH team leadership, and a data analyst. Two team members (L.B.K. and J.M.M.) are clinical co-leads in the statewide learning network’s behavioral health domain. Our site/hospital system was the first site to test interventions around the shared outcome, with plans to report learnings back to the collaborative.

We developed a current-state process map that included failure modes and team members’ experiences, completed a literature review, and reviewed lessons from the baseline data.^16,17 Based on these findings, we developed a key driver diagram with 6 key drivers (Fig. 1). Process mapping with caregivers identified varied preferences regarding the site of follow-up—some preferred seeing their “trusted” pediatrician, whereas others preferred seeing a “specialist.” Despite some preferring pediatricians for follow-up, baseline data showed that our patients rarely followed up within PPC.

Fig. 1.

Key driver diagram.

Interventions

Table 1 lists all interventions and PDSA cycles. Our two primary interventions were postdischarge outreach and increasing access in PPC.

Table 1.

Interventions and PDSA Cycles

Intervention	Subintervention	Cycle 1	Cycle 2	Cycle 3
Postdischarge outreach	SW* calls patients and families after discharge	Plan and do: SW from one clinic calls their clinic patients after discharge.	Plan and do: implement one page handout “why follow-up matters”; spread to all three clinics; do not call patient if they have a care manager assigned.	__
		Study: parents and SW unclear about the purpose of f/u which hinders buy-in; parents getting too many f/u calls from different people.	Study: parents appreciated the reasons behind the call; more patients answered the phone; too burdensome on SW staff to track patients if they do not answer.
		Act: adapt.	Act: adopt current tools; adopt plan that SW only calls once.
	Automated discharge list	Plan and do: add SWs to daily ED discharge list.	Plan and Do: create custom automated daily email based off operational definition that only includes MBH discharges and includes discharges over the last 3 d.	__
		Study: hard to decipher MBH patients; some patients missing from list due to discharge timing; forgot to call patients on high-volume days	Study: able to properly identify discharges.
		Act: adapt.	Act: adopt.
	E-visits	Plan and do: build an asynchronous visit to recommend f/u and check in on patients.	__	__
		Study: unable to include safety assessment in asynchronous visit.
		Act: abandon
Increasing access in the primary care medical home	Provider training	Plan and do: partnered with MBH ED team to educate pediatricians around goals of an f/u visit	Plan and do: educate all pediatricians in our practice; provide referral resource guide; build note template to guide visits
		Study: pediatricians wanted resource list for referrals if needed; no telehealth visits were scheduled within 1 month of intervention	Study: pediatricians comfortable with ED f/u visits in office if SW or psychology present; SW team began scheduling in-person visits
		Act: adapt.	Act: adapt/adopt; limit f/u visits to clinic hours when SW or psychology present
	Integrated psychiatry	Plan and do: create process for SWs to refer patients to our integrated psychiatrist	__	__
		Study: patients with higher complexity were being scheduled here rather than in PPC office
		Act: adapt/adopt: integrated psychiatry follow-up patients require 60-min appointment.
	Schedule before discharge	Plan and do: train ED team to offer PPC appointments before patient leaves the ED.	Plan and do: PPC 24-h overnight triage service can schedule patients for MBH follow-up.	Plan and do: display flyer with PPC practice names and 24/7 phone number
		Study: difficult for ED team to access primary care calendars; many patients come overnight and PPC office is closed.	Study: 24/7 service used successfully; ED team struggled to remember which PPC clinics used this service	Study: small increase in ED team using triage service to schedule MBH f/u visit
		Act: adapt.	Act: adapt.	Act: adopt.
Document follow-up provider	—	Plan and do: educate ED staff on where to enter identified f/u provider into safety plan.
		Study: >70% of safety plans included an f/u provider.
		Act: adopt.
Preoccupation with failure	Data review	Plan and do: principal investigator reviewed all patient charts after discharge to monitor SW calls and reasons for not calling	Plan and do: Monthly meeting with SW team to review Pareto Chart of failures
		Study: SW’s were accepting of method and did not feel micromanaged as a team	Study: One SW was only calling suicidal patients; SW’s found feedback helpful and felt motivated to increase call percentage
		Act: adapt. Create Pareto Chart (updated monthly) to identify reasons for not calling	Act: adopt monthly check-ins

f/u, follow-up.

Postdischarge Outreach

PPC SWs called patients within 48 hours of ED discharge using an automated daily email of all PPC patients discharged from the ED in the prior 24 hours. SWs prioritized those without an existing MBH provider on the ED safety plan. During the call, they could facilitate appointments with PPC pediatricians or a co-located psychiatrist.

The SW team and the principal investigator (a primary care physician) met monthly to obtain feedback and review PDSA data. A Pareto chart (Supplemental Digital Content 2, https://links.lww.com/PQ9/A835) tracked reasons for not calling discharged patients. High-risk patients in our system are often enrolled in care management. Practice-based care managers already have a process for calling empaneled patients after ED discharge about follow-up needs. The SWs did not call these patients to reduce redundancy.

We adopted interventions to address common failure points. Outreach calls dipped in March 2025 due to failures in the discharge email. SWs were confused about the diagnoses in the email, forgot to call during long weekends, or inadvertently missed MBH patients on high-volume days. Some discharges were omitted from the email due to technical issues with discharge timing and report generation timing. We implemented a custom daily discharge email on April 21, 2025, that included only MBH diagnoses and reported discharges from the last 72 hours (rather than 24 h).

Increasing Access in Primary Care

The second intervention aimed to improve follow-up access within PPC. Partnering with the MBH ED team, we delivered a 30-minute training to PPC telehealth pediatricians on follow-up visit goals, safety monitoring, and safety planning. Pediatricians were empowered to provide long-term care when appropriate and were given resource lists of community referral options (Supplemental Digital Content 1, https://links.lww.com/PQ9/A834). Our co-located psychiatrist also agreed to see MBH crisis follow-ups, a new service, as she previously focused on consultative visits.

Feedback and data informed intervention adaptations. After the initial 30-minute training, no telehealth visits occurred. In response, we expanded the training to all PPC pediatricians to support in-person follow-up visits. Because pediatricians expressed concerns about support during these visits, we required in-person follow-up visits only when SWs and/or integrated psychologists were readily available to assist.

Measures

The primary outcome measure was the percentage of patients 6–18 years of age discharged from the ED with an MBH diagnosis who completed a follow-up visit within 1–7 days of discharge. Our measure mirrored the statewide operational definition to facilitate the dissemination of local findings to the state level.

Primary Measure Denominator

The denominator was the number of patients in the PPC registry who were discharged from the ED with an MBH diagnosis. In line with the statewide measure, we excluded patients who returned to the ED for MBH concerns within 7 days from the denominator, as there was not a full opportunity to achieve the metric; this occurred, on average, 2.5 times per month (range 0–8). If a patient returned to the ED within 7 days and was discharged home again, we included the subsequent visit in the denominator and restarted the follow-up period from the most recent discharge date.

Primary Measure Numerator

The numerator counted patients who completed a follow-up visit within 1–7 days of ED discharge. Our measure (and the statewide measure) differs from the NCQA measure by excluding visits on the same day as the ED visit (day 0). Our statewide learning collaborative found that most day-0 visits involved patients who had an MBH outpatient visit before being referred to the ED that day. We excluded day-0 visits to avoid overestimating follow-up percentages.

We counted billable, completed visits documented in the hospital’s electronic health record (EHR, “Epic”; Epic Systems, Verona, Wis.) with an MBH diagnosis code at follow-up. Follow-up visits could be with an MBH specialist (developmental pediatrician, psychiatrist, social worker, or psychologist outside primary care), a pediatrician, or a co-located PPC psychiatrist. We could not capture follow-up visits with community providers outside the EHR. Our equity measure segmented our primary outcome data into Medicaid-insured and commercially insured patients.

Due to concurrent QI efforts to increase well-child visits in our PPCs, some MBH follow-up visits were converted into well-child visits. To account for this scenario, we counted visits with primary wellness diagnostic codes that also had secondary MBH diagnostic codes. This scenario occurred in 1 baseline visit and 7 intervention period visits.

Process and Balancing Measures

Process measures included the percentage of ED MBH discharges contacted by our SW team, the percentage reached by phone, and the percentage who followed up within PPC.

Team members expressed concern that ED recidivism would increase if follow-up visits within PPC did not meet a patient’s needs. Thus, our primary balancing measure was the percentage of patients with another MBH ED visit within 8–30 days of the index ED discharge. We also worried that an increase in PPC follow-up visits would reduce follow-up visits outside primary care, reflecting a shift in follow-up location rather than an overall increase in follow-up. Thus, another balancing measure was the percentage of ED discharges with a completed follow-up visit outside PPC (ie, psychiatry, psychology, and developmental pediatrics).

Analysis

Statistical process control charts (p charts) with 1-month intervals tracked the primary outcome measure, process measures, and the primary balancing measure using a mean as the centerline. We tracked our equity measure with a 2-line run chart, which uses a median as the centerline. Eight consecutive data points above or below the centerline indicated that change resulted from special cause variation (SCV), not common cause.¹⁷

We compared odds ratios (ORs) between baseline and intervention periods to strengthen our results. For our postdischarge outreach, we compared the likelihood of successful follow-up based on whether or not a patient was reached by phone. ORs were also calculated to compare follow-up locations between baseline and intervention periods to (1) monitor PPC access and (2) evaluate our balancing measure around follow-up external to PPC.

The institutional review board approved this QI study.

RESULTS

Baseline data collection occurred from January 1, 2024, to October 31, 2024 (n = 258 ED MBH discharges). The intervention period was November 1, 2024, to October 31, 2025 (n = 334 ED MBH discharges). During this period, the mean follow-up increased from a baseline of 28.7% to 40.3% after 8 consecutive months above the mean from January 2025 to August 2025 (Fig. 2). Improvement coincided with the introduction and spread of the interventions.

Fig. 2.

Primary outcome measure: Percentage of patients discharged from the emergency department with a mental behavioral health diagnosis who had a documented follow-up visit within 7 days of discharge (statistical process chart/P-chart).

On our 2-line run chart, Medicaid-insured patients had a 24% median baseline follow-up, compared with 42% for commercially insured patients. Following interventions, follow-up for Medicaid-insured patients increased to 35% due to SCV. Commercially insured patients had no SCV. Thus, the difference between median follow-up for Medicaid-insured patients and median follow-up for commercially insured patients (“equity gap”) decreased from 18% to 7% (Fig. 3).

Fig. 3.

Equity measure: patients with completed follow-up visit within 7 days of emergency department discharge, by insurance type (run chart). ^aCommercially insured patients had a median follow-up of 42% with no special cause variation during the study time frame. Due to low n’s and high variability (range 0%–100% follow-up) for this cohort, individual monthly data points for commercially insured patients are not shown on this figure.

SWs called, on average, 52.4% of ED MBH discharges (Fig. 4A). Two calls required immediate safety planning. The most common reasons for not calling were (1) a patient received a direct call from their practice-based care manager (n = 47, 32.6% of noncalls) and (2) a patient already completed follow-up (ie, on day 1 postdischarge) (n = 25, 17.4%) (Supplemental Digital Content 2, https://links.lww.com/PQ9/A835). Less common failure modes included SWs forgetting to call, errors in the daily email (see “Methods” section), or SW assessments that PPC was not the appropriate outreach (eg, a patient with severe autism followed by specialists).

Fig. 4.

Process measures. A, Percentage of discharged patients who received social work outreach. B, Percentage of discharged patients who followed up within the primary care setting (statistical process charts/P charts).

Odds of successful follow-up were significantly higher when SWs called and reached a family than when they did not (OR = 2.15, 95% confidence interval [CI] = 1.15–4.03). When comparing patients reached to those not called, follow-up did not differ significantly (Table 2).

Table 2.

Process Measures

Process Measure	Comparison Variable (Patients with Successful 7-d Follow-up [n, %])	Intervention (Patients with Successful 7-d Follow-up [n, %])	Odds Ratio (95% CI)
Social work outreach calls	Patients called but not reached [23/79, 29.1%]	Patients reached by phone [46/98, 46.9%]	2.15 (1.15, 4.03)
	Patients not called [61/157, 38.9%]	Patients reached by phone [46/98, 46.9%]	1.39 (0.84, 2.32)
Follow-up location	Primary care clinic (baseline period) [9/258, 3.5%]	Primary care clinic (intervention period) [28/334, 8.4%]	2.53 (1.17, 5.46)
	Outside primary care (baseline period) [65/258, 25.2%]	Outside primary care (intervention period) [102/334, 30.5%]	1.31 (0.91, 1.88)

Bolded numbers are statistically significant.

Follow-up within PPC increased from 3.5% of ED MBH discharges to 10.7% of discharges (Fig. 4B). SCV was met on September 30, 2025. Odds of follow-up within PPC were 2.53 times higher during the intervention period (95% CI 1.17–5.46) (Table 2). Show rates within PPC for MBH ED follow-up (72.5%; 37/51) were similar to show rates broadly within PPC (72.4%). Most (23/28 = 82.1%) PPC follow-ups during the intervention period occurred with a pediatrician, 7.1% (2/28) with the co-located psychiatrist, and 10.7% (3/28) with the co-located psychiatric nurse practitioner. Our balancing measure showed no significant change in follow-up in other MBH settings such as psychiatry, psychology, and the ED bridge clinic (OR = 1.31, 95% CI = 0.91–1.88) (Table 2).

There was no SCV in the percentage of discharged patients who returned to the ED for MBH needs within 8–30 days (Supplemental Digital Content 3, https://links.lww.com/PQ9/A836).

DISCUSSION

Summary

This study describes a QI project designed to improve pediatric ED discharge follow-up for patients with mental and behavioral health needs. We identified that patients without established MBH care required improved access, clear discharge communication, timely follow-up scheduling, and greater emphasis on the value of outpatient care after discharge.^10,11,14,15 To address these needs, we implemented provider training, postdischarge SW outreach, a co-located psychiatry partnership, and shared accountability across hospital divisions.

Interpretation

The core interventions included direct outreach calls from SWs to families after discharge and follow-up opportunities in PPC. Follow-ups were more likely when PPC SWs successfully reached families after discharge, increasing confidence that our intervention contributed to the overall improvement. However, follow-up did not differ significantly among families not called. This is likely because many of those families received outreach calls from existing care managers, or follow-up was already completed before SW could call.

PPC follow-up increased from 3.5% to 10.7% after interventions. Two months, June 2025 and July 2025, contributed substantially to this increase. This may be secondary to less access to school-based services in the summer. Because follow-up was 0% during the same months in 2024, we believe this increase reflects a genuine change in the system.

The final follow-up percentage within our healthcare system was 40.3% (35% for Medicaid-insured patients). Although below our goal of 50%, it was higher than what Hoffmann and colleagues¹⁴ found (31.2%) in their study of 28,551 pediatric patients with Medicaid insurance and similar to the 40.2% seen in an adult study.¹⁸ Considering that our metric excluded day-zero outpatient visits and our data excluded external follow-up visits, our percentage would likely be even higher using the NCQA measure and with full access to external data.

PPC can be a valuable resource for addressing youth mental and emotional needs, even after crisis.^13,19 Prior work found >11% of ED discharges for suicide attempt and depression followed up within primary care.¹³ This effort, though, depends on adequate pediatrician support and strong partnerships with other levels of care.^19,20 Prior QI efforts in pediatric primary care have improved screening and management of MBH through standardized workflows, provider feedback, and education.²¹

The medical home also provides families and patients with a trusted setting to receive care. Anecdotally, our SWs noted that parents were thankful to hear a familiar voice after their child’s stressful MBH crisis. Further feedback from patients and families, including those who did not follow-up, will support improvements in follow-up, outreach, and how best to meet their needs.

Our balancing measure of ED reutilization remained at 10%, reassuring us that PPC visits were not harmful. Future interventions to reduce ED recidivism will need to address the complex social and demographic factors that lead to ED MBH visits, including varying support systems, MBH complexity, and care access.^20,22–25 Despite this process being a national quality measure, our literature review found limited evidence that improving 7-day follow-up leads to reduced ED utilization in pediatric populations. One study found fewer return ED visits within 5 days, but more thereafter.¹⁴ Other studies focused on adult populations or hospitalization discharges as opposed to ED discharges.^6,18

Another concern with conflating ED recidivism with outpatient follow-up is severity bias—patients with more severe MBH problems are more likely to visit the ED and more likely to have follow-up care. Rather than measuring return ED visits, researchers should measure the impact of timely follow-up on meaningful, distal health outcomes, such as long-term symptom reduction or care connection.

Limitations

Our study data were limited to one healthcare system’s EHR. However, the denominator is highly reliable because our system is the primary source of emergency MBH care for our patients. Still, we experienced a challenge common in the United States: no feasible way to ascertain MBH follow-up visits completed outside our system.²⁶ This likely underestimates the follow-up percentage at the population level. However, because PPC SWs prioritized patients without existing MBH providers and we observed no reduction in internal specialty follow-up visits (Table 2), we are confident that the increase in follow-up reflects a true system-wide gain. Recently, promising efforts have improved data sharing and patient attribution, enhancing QI work and supporting communication within a patient’s care network.²⁶

Another limitation is generalizability. Our PPCs are fortunate to have embedded SWs, integrated psychologists, and co-located psychiatry team members. However, although SWs provide enhanced expertise, other staff members could likely perform these calls and triage to SW as needed. Few outreach calls required immediate safety planning. Similarly, our co-located psychiatrists provided <20% of follow-up care in PPC. Therefore, the project’s strength can be interpreted broadly as enhanced outreach and improved access to primary care.

CONCLUSIONS

This project provides evidence that MBH ED follow-up can occur promptly in pediatric primary care without increasing ED return visits.

ACKNOWLEDGMENTS

The authors gratefully acknowledge QI statistical consultation from Elizabeth Goetz, MSN, RN. We gratefully acknowledge statistical assistance from Melinda C. MacDougall, MS.

Supplementary Material

pqs-11-e944-s001.pdf ^{(55.8KB, pdf)}

pqs-11-e944-s002.pdf ^{(536.9KB, pdf)}

pqs-11-e944-s003.pdf ^{(66KB, pdf)}