Reduced Hospital Readmissions Through Personalized Care: Implementation of a Patient, Risk-Focused Hospital-Wide Discharge Care Center

Abstract

Hospital readmission remains a common and costly event. For patients, hospital readmission is often associated with adverse outcomes in the short and long term. Despite broad interest in reducing hospital readmissions and the implementation of a wide range of programs to combat unplanned preventable readmissions, progress has been challenging. Multipronged, complex interventions have demonstrated effectiveness in published studies, yet they typically target a particular patient population (e.g., older adults), condition (e.g., heart failure), or hospital unit and are seldom implemented at the whole-hospital level. There is a pressing need for implementation studies of hospital-wide, comprehensive discharge care interventions. This article describes the design and first 2 years of implementation of the Discharge Care Center (DCC), a hospital-wide intervention designed to reduce readmissions at Vanderbilt University Hospital. Structured around a nurse-led Triage Team and a multidisciplinary Care Coordination Team specialized in post-discharge response to patient needs, the DCC utilizes a combination of automated surveillance and risk-based support to provide comprehensive and personalized coverage to adult patients discharged from the hospital to home. In its first 2 years of implementation, the DCC has cared for patients after 80,247 hospital discharges, providing 57,352 clinically relevant interventions. Vanderbilt University Hospital has seen a highly promising impact on readmissions, from a baseline (July 2019–June 2021) monthly mean 30-day unplanned readmission rate of 10.6% to a new rate of 9.9%, sustained for 2 consecutive years (July 2021–October 2023). This article presents the key functions of the DCC, the forms of implementation adopted, and practical advice for other healthcare systems aiming to implement a successful discharge care intervention at scale.

Hospital readmission has increasingly received attention in healthcare research, policy, and practice, especially since the Centers for Medicare and Medicaid Services began issuing public reports of 30-day readmission rates 15 years ago.¹ Hospital readmissions are frequent and expensive,² and preventable unplanned readmissions may be indicative of poor quality of care.^3,4 Across the United States, in 2020, there were over 3.4 million adult readmissions within 30 days of hospital discharge, which corresponds to a readmission rate of 14%; these readmissions incurred an average cost of $17,700 each, amounting to a total cost of over $60 billion for US readmissions in one year alone.⁵ Notably, nearly one in five readmissions (18.7%) are considered potentially preventable.⁵

In addition to its high cost, readmission within 30 days of hospital discharge has been linked to both short- and long-term mortality.⁶ Given its importance to patients and health systems alike, it is unsurprising that 30-day readmission has become a key measure for assessing the quality of care provided by hospitals. Readmission performance has concrete financial and reputational repercussions for hospitals, with direct implications in terms of penalties, public reporting, and national rankings. Further, it is believed that systemic gaps in care play a contributing role in excessive readmissions, revealing an opportunity for improvement.^7,8

Throughout the United States, at levels ranging from federal strategies to individual hospital care plans, efforts have been implemented to reduce readmissions, with discharge care and coordination as an area of particular focus.⁹ Despite these efforts, the decline in readmissions has been slow among Medicare beneficiaries (from 18.3% in 2010 to 17.2% in 2015 to 17.0% in 2020), and the overall rate of 30-day adult readmission has shown even less movement (14.2% in 2010; 14.0% in both 2015 and 2020).^5,10,11

Efforts to reduce readmissions may be hindered by the complexity of the causes of readmissions, with an average of 8.7 factors contributing to each readmission.¹² Previous work has consistently indicated that no single intervention is effective in reducing readmissions; however, this work has suggested a benefit of certain hospital-initiated post-discharge interventions—especially comprehensive programs involving multiple components and contacts—on hospital readmission.^13–16 A review of 42 studies of discharge interventions showed that multifaceted interventions were 40% more effective in reducing readmissions compared with simple interventions.¹⁷ In another study, an average of 3.5 components were found in interventions bridging the hospital and post-discharge periods, and the number of components was the only significant predictor of an intervention’s impact on readmission rate.¹⁸ Thus, reducing readmissions will likely require a multipronged approach with strategies spanning the continuum of care in both the hospital setting and the ambulatory/post-acute care setting.

Most interventions implemented to date have been relatively small in scale, involving only a certain patient population, a particular condition, or a specific hospital unit, and some studies continue to test single-component (e.g., nutrition support) or single-instance interventions.^14,19,20 To inform practice and policy aiming to reduce hospital readmissions, there is a pressing need for implementation studies of comprehensive, hospital-wide discharge care.

In 2019, Vanderbilt University Hospital began working to establish a multidisciplinary Discharge Care Center (DCC) as a centralized, accessible resource to help recently discharged patients and their caregivers navigate the healthcare system after leaving the hospital. The DCC’s primary goal is to improve the quality and safety of transitions of care and to reduce preventable readmissions. In this article, we 1) describe lessons learned from our experience with implementing this hospital-wide initiative, which has successfully reduced readmissions at our hospital in its first 2 years of implementation, and 2) distill the program to its essential components to present broadly applicable, practical advice for adapting and launching a discharge care center in other healthcare facilities.

The Discharge Care Center at Vanderbilt University Hospital

Vanderbilt University Medical Center (VUMC) is a quaternary care academic medical center located in Middle Tennessee. Its main campus 726-bed adult hospital, Vanderbilt University Hospital, admits more than 50,000 patients and receives over 85,000 emergency room visits annually. In 2019, Vanderbilt University Hospital tasked a Discharge and Transitions Executive Committee with assessing the current state of discharge support and designing interventions to improve the quality and safety of discharge. The team paired registered nurse (RN) and physician leaders to oversee workstreams and conduct baseline exploration, propose and design interventions, plan and execute the implementation, and perform an evaluation. In August 2020, early pilot studies to establish the initial services of the DCC began with a small group of patients at Vanderbilt University Hospital. In July 2021, after incorporating an automated messaging component into the program, the DCC began scaling availability of its services covering the first 30 days post-discharge to patients across Vanderbilt University Hospital discharged to home settings. Full implementation was reached by October 2021, with the core elements of care coordination, automated text messages and phone calls, and pharmacy support in place. Continued improvements to the DCC’s program have been made over time, including adaptations in patient risk stratification, the creation of the DCC eConsult (an electronic consultation order that inpatient staff place to the DCC to communicate anticipated patient post-discharge needs), and regular “close-the-loop” feedback meetings with inpatient and ambulatory leadership regarding leading metric performance and opportunities to improve quality of care at discharge. These components are described below.

Operational Model

Targeting the patients who most need enhanced support is crucial to maximize impact and to appropriately distribute limited resources. Accordingly, the DCC’s core workflow is planned around stratifying risk among patients being discharged, along with consideration of their personal communication preferences. The DCC is fully integrated into Vanderbilt University Hospital’s standard discharge process, which already included elements of medication management and patient education. The DCC supports Vanderbilt University Hospital patients discharged to home or home with home health services. Both inpatient and observation patient status types and elective and nonelective admission types are included. DCC interventions are carried out as needed throughout the 30-day period following discharge. Certain units and programs in our hospital provide their own transitional care services; DCC coverage thus excludes obstetric patients, patients involved in clinical research, those enrolled in home hospice, patients with a ventricular assist device in place, and patients who have received a transplant. All other patients, regardless of payer status and regular source of care, are in scope.

The DCC model relies on elements of both patient-driven and clinical staff-driven communication (see Figure 1). During the discharge process, discharging nurses introduce eligible patients to the DCC and ask them to share a post-discharge phone number for automated communication over the next 30 days; patients may opt out at any time during this period. These patients are also provided with a number to call the DCC hotline, which is in operation 24 hours a day, every day of the year, with any questions or problems. Patients flagged as at high risk for readmission because of the presence of an eConsult Order or a high calculated readmission risk score also receive a manual phone call. To enable both proactive and responsive communication, the DCC staff is configured as two teams: 1) the Triage Team of RNs, which responds to patient-initiated communication, and 2) the Care Coordination Team of social workers, care coordinators, case managers, and pharmacists, which reaches out proactively to a targeted group.

Figure 1. Trajectories for Personalized Care in the Discharge Care Center (DCC).

Source: VUMC Discharge Care Center

The DCC intervention begins with an automated text message or phone call to all in-scope patients. The initial automated text message is sent 1 hour after the discharge order is placed. For patients indicating a preference for a phone call, the first call is placed the following day. This automated texting/calling continues throughout the first 30 days post-discharge. Patients preferring text messages receive a message about medications on Day 1 following discharge, a message inquiring about new or worsening symptoms on Day 2, and a message verifying receipt of any clinician-ordered durable medical equipment or home health services on Day 3. To continue monitoring for new or worsening symptoms, the text messages then continue every 2–3 days for the first 15 days, followed by every 5 days, for a total of 12 messages across the 30 days. Patients indicating a preference for phone calls receive less frequent automated communication. Details of the DCC’s use of automated text messaging, including examples of the messages used, have been published elsewhere.²¹ If the patient responds to these texts or calls indicating that they have a need, two-way communication is initiated. Depending on the type and complexity of help needed, concerns are addressed by the RN Triage Team or by the Care Coordination Team.

Working in parallel with the automated communication process, the DCC identifies patients with complex needs or high readmission risk who are likely to need additional support to address a post-discharge care need. These patients are identified through a combination of a VUMC-developed validated readmission risk prediction score and empirical clinical observations. Using clinical and demographic variables, the VUMC Cornelius readmission risk score quantifies risk of hospital readmission within 30 days of discharge for each patient admitted to Vanderbilt University Hospital. Because previous work highlights the unique insights of nurses and clinicians treating a patient,^22,23 Care Coordination Team DCC interventions can also be triggered through DCC eConsult Orders. These orders can be placed by any member of the inpatient care team to flag a high-risk or complex situation warranting personalized follow-up across the care continuum.

Discharge Care Center Services

Prior research at VUMC²⁴ demonstrated a transitional care intervention to be effective in reducing readmissions among patients with low levels of health literacy, pointing to a need to screen for patient needs and to target services.²⁵ Further empirical work including VUMC in a series of multicenter studies also informed the DCC program’s specific intervention categories and modes,^26–32 as well as the personnel required to carry out and personalize these interventions.

The main interventions provided by the DCC are 1) sign and symptom triage for new, concerning, or worsening symptoms; 2) patient education on lifestyle; 3) medication support including medication list accuracy, ensuring medications were obtained, and patient education; 4) verification of receipt of home care services or medical equipment; 5) follow-up appointment scheduling; and 6) connection to resources to meet socioeconomic or psychosocial needs. Additional support initiatives are currently being planned. Selections from this menu of services are triggered by a combination of a response to patient-initiated queries (i.e., hotline calls and indications of a need through the text messaging system) and proactive efforts to provide support to patients identified as having complex post-discharge care needs or being at high risk for readmission.

Staffing

The staffing model for the DCC aligns with the services offered and includes RNs, social workers, case managers, care coordinators, and pharmacists whose time is fully dedicated to the DCC. This enables the DCC to operate as an independent care center. DCC staffing needs were established by analysis of the key data points of average daily discharge volume of the in-scope patient population, historical post-discharge call length, desired number of touchpoints with patients, and estimated number of interactions that will result from touchpoints, followed by determining the most qualified staffing roles needed to prevent readmissions. The DCC has continued to adjust the staffing model and resource distribution to best fit the patient population and engagement needs throughout the center’s evolution. Additionally, each DCC team member has access to the patient’s electronic medical record and can contact Vanderbilt primary care providers, specialists, or the inpatient discharging team if needed to address patient needs.

The DCC staffing model centers on the Triage Team and the Care Coordination Team (Table 1). The Triage Team of RNs is responsible for incoming calls and messages. Tasks include initial sign and symptom triage, patient education, verification of receipt of ordered services or medication in response to a patient’s expression of a problem, and escalating problems that require additional clinical intervention to the Care Coordination Team. The multidisciplinary Care Coordination Team reaches out to patients identified as needing additional support, assigning a healthcare professional appropriate to the need. Social workers focus on socioeconomic and psychosocial needs, commonly providing interventions such as counseling and referral to resources for housing, food, transportation, and substance abuse treatment. Care coordinators and case managers handle issues that require longer follow-up such as follow-up appointments, durable medical equipment needs, or home health services. Pharmacists address medication management needs, including prescription coordination, clarification of medication discrepancies, counseling on side effects, assistance with medication cost barriers, and help with medication administration questions.

Table 1.

Discharge Care Center Staffing Model

	Staff	Responsibilities
Triage Team (responsive)	Registered nurses	• Answer unprompted incoming patient calls (address or refer to the Care Coordination Team) • Monitor incoming patient text messages/calls through the automated texting platform (address or refer to the Care Coordination Team)
Care Coordination Team (proactive)	Social workers, care coordinators, case managers, pharmacists (deployed per case as needed)	• Place manual phone calls to recently discharged patients who are identified as having complex care needs or being high risk (by risk score or referral) • Address concerns raised by patients referred for care from the Triage Team

Monitoring the Discharge Care Center’s Performance

Multiple types of data are utilized to monitor and evaluate the effectiveness of the DCC. In addition to monitoring readmission within 30 days of inpatient discharge as the key outcome of interest, we gather data on patients contacted by the DCC through VUMC’s electronic health record. These data include patients’ basic demographic characteristics, length of hospital stay, discharging unit, treatment team, diagnosis, Cornelius readmission risk score, and other basic clinical characteristics.

The DCC also directly collects data at multiple points throughout the 30-day post-discharge intervention process. First, data on the reach and response rates of the automated communications are provided by VUMC’s patient text and call outreach software partner, Artera. Second, the DCC gathers data from eConsult Orders, tracking the type of concern and requesting hospital unit for each order; the health literacy, readmission rates, and Cornelius risk scores of all patients with eConsult Orders; and the overall volume of orders over time. Third, when a patient has indicated that they need help (through the DCC hotline or response to a text/phone call), the interventions provided are recorded by the Triage Team or the Care Coordination Team. These interventions are classified by general intervention category (e.g., pharmacy intervention topic, prescription coordination, scheduling assistance, durable medical equipment issues, or home care services coordination). Fourth, throughout the 30-day post-discharge period, the DCC collects call center data regarding patients’ efforts to reach the hotline (e.g., call abandonment rate, speed to answer, and call volume), as well as data on the proactive outgoing calls placed by the Care Coordination Team (e.g., call volume and average length of call). Finally, information is collected on secondary outcomes, such as patient engagement and pharmacy interventions, as well as rich qualitative data on the interventions performed and their outcomes. Overall, the DCC-specific data collection allows for tracking of the frequency and type of transitional care problems identified and provides a record of responses. Regular DCC team touchpoints and meetings with health system leadership offer mechanisms for reviewing and interpreting process and outcome data to drive continuous improvement.

The Discharge Care Center’s First 2 Years of Operation

Description of Participants

This evaluation, which addresses the first 2 years of operation of the DCC following full implementation in October 2021, was reviewed by the Vanderbilt Institutional Review Board and determined to be quality improvement, not requiring informed consent (#241270). From October 2021 to October 2023, the DCC contacted over 100 newly discharged patients each day, supporting a panel of approximately 3,000 to 3,200 patients at any given time. Table 2 presents basic demographic and clinical characteristics for discharged patients receiving a message from the DCC during this period and those requiring an intervention. Overall, patients were contacted by the DCC following 80,247 hospital discharges, and 23,050 discharges (29%) required an intervention. Of note, a single patient could have more than one discharge during the observation period, and more than one intervention could be performed per discharge.

Table 2.

Demographic and Clinical Characteristics of Patients Contacted by the Discharge Care Center (DCC) from October 2021 to October 2023

	Discharges followed by automated DCC communication to patient	Discharges requiring a DCC intervention
Demographic Characteristics
Gender
Women	39,090 (48.7%)	11,251 (48.8%)
Men	41,152 (51.3%)	11,798 (51.2%)
Race/ethnicity
White	62,838 (78.3%)	17,668 (76.7%)
Black or African American	12,611 (15.7%)	3,907 (17.0%)
Hispanic/Latino	1,623 (2.0%)	523 (2.3%)
Asian	765 (1.0%)	211 (0.9%)
American Indian or Alaska Native	340 (0.4%)	112 (0.5%)
Middle Eastern or North African	234 (0.3%)	67 (0.3%)
Native Hawaiian or other Pacific Islander	46 (0.1%)	13 (0.1%)
Asian Indian	2 (0.0%)	2 (0.0%)
Other	420 (0.5%)	134 (0.6%)
Prefer not to answer	245 (0.3%)	96 (0.4%)
Unable to provide	1,123 (1.4%)	317 (1.4%)
Age group
<18 years	25 (0.0%)	3 (0.0%)
18–29 years	6,552 (8.2%)	1,567 (6.8%)
30–39 years	9,135 (11.4%)	2,357 (10.2%)
40–49 years	10,721 (13.4%)	2,989 (13.0%)
50–59 years	14,230 (17.7%)	4,188 (18.2%)
60–69 years	18,281 (22.8%)	5,644 (24.5%)
70–79 years	14,701 (18.3%)	4,446 (19.3%)
≥80 years	6,602 (8.2%)	1,856 (8.1%)
Clinical Characteristics
Top diagnosis-related groups
Observation/extended recoverya	24,154 (30.1%)	6,305 (27.4%)
Major small/large bowel procedures	1,651 (2.1%)	523 (2.3%)
Septicemia without mechanical ventilation >96 hours	1,483 (1.8%)	415 (1.8%)
Craniotomy/endovascular intracranial procedures	1,260 (1.6%)	399 (1.7%)
Operating room procedures for obesity	1,256 (1.6%)	221 (1.0%)
Percutaneous cardiovascular procedures without coronary artery stent/acute myocardial infarction	1,119 (1.4%)	297 (1.3%)
Heart failure/shock	1,052 (1.3%)	331 (1.4%)
Other kidney/urinary tract diagnoses	981 (1.2%)	264 (1.1%)
Total	80,247	23,050

^aNo diagnosis-related group is assigned to observation/extended recovery patients.

Table 3 presents hospital length of stay for all patients contacted by the DCC and for discharges requiring/not requiring an intervention. The median length of stay among patients contacted by the DCC was 2.5 days. Hospital stays were approximately half a day longer for discharges requiring a DCC intervention than for discharges not requiring an intervention (median of 2.8 vs. 2.3 days, respectively).

Table 3.

Hospital Length of Stay for Patients Contacted by the Discharge Care Center (DCC) from October 2021 to October 2023

	Discharges with automated DCC communication	Discharges requiring a DCC intervention	Discharges not requiring a DCC intervention
Mean	4.1	4.5	3.9
Median	2.5	2.8	2.3
25th Percentile	1.2	1.3	1.2
75th Percentile	4.9	5.4	4.7

Note: All numbers reflect the number of days a patient remained in the hospital following a single admission.

Frequency and Nature of Use

Of the 80,247 initial patient post-discharge contacts made by the DCC from October 2021 to October 2023, almost all (97%) remained in the program beyond the first automated message, and 73% continued to receive messages through Day 30. For most discharges (79%), patients responded to a text or phone prompt at least once during the program.

Figure 2 provides an overview of DCC care. Overall, 29% of eligible discharges required an intervention. This percentage increased from 26% at the start of the program to 32% at the end of the evaluation period. Patients were more likely to receive an intervention if they were classified as high risk by Cornelius score (N = 10,966 patients, 38% intervention rate) or if an eConsult Order was placed by their care team (N = 671, 79% intervention rate).

Figure 2. Summary of Activities of the Discharge Care Center (DCC), October 2021 to October 2023.

Source: VUMC Discharge Care Center

In total, as shown in Figure 3, of the 57,352 interventions (across 23,050 discharges), approximately 75% were provided by the Triage Team and 25% by the Care Coordination Team. The largest categories of needs addressed by the Triage Team were new or worsening symptoms (38%) and patient education (29%). For the Care Coordination Team, almost half (47%) of needs addressed fell into the “other” category, reflecting the highly varied types of personalized care required; medication management and education was the next most common category, accounting for 42% of interventions provided by the Care Coordination Team.

Figure 3. Discharge Care Center Interventions Performed by Team and Need Type, October 2021 to October 2023.

Source: VUMC Discharge Care Center

Change in 30-day All-cause Readmission

In the first 2 years of full implementation of the DCC, all-cause 30-day unplanned readmission within the VUMC system after a Vanderbilt University Hospital index discharge dropped substantially. During the baseline period (July 2019–June 2021), the monthly mean readmission rate was 10.6%, versus a rate of 9.9% during the 2-year period after rollout (July 2021–October 2023).

Soon after the July start of the scaled rollout of DCC services for patients discharged to home, we saw 8 consecutive months of decreased readmission rates from our previous 2-year average value, or centerline. According to the rules of statistical process control, 8 or more consecutive points below the centerline reflects a shift or sustained change in process. The downward shift in centerline (July 2019–June 2021 vs. July 2021–October 2023) reflected a 6.6% relative reduction in our hospital readmission rate. Figure 4 demonstrates maintenance of this shift over the 2-year period. Although we cannot definitively draw a causal relationship between our intervention and this change, we are aware of no other changes during this period that could account for this meaningful and sustained reduction in the rate of unplanned readmissions.

Figure 4. Trends in Unplanned 30-day Readmission at Vanderbilt University Hospital before and during the Implementation of the Discharge Care Center Interventions, July 2021 to October 2023.

Note: SPC = statistical process control; VDAT = Vanderbilt Discharge and Transitions

Source: VUMC Discharge Care Center

Key Lessons Learned and Implications for Implementation in Other Contexts

There are three key messages to share with other health systems seeking to implement a discharge care intervention to reduce hospital readmission. First, incorporating an element of automation in patient post-discharge communication enables the monitoring and reach of a large population of patients with limited staffing levels. Automated communication with patients was one of the earliest enhancements to the DCC and has remained a core element of the program. The automated communication enabled DCC staff to focus on patients requiring assistance and allowed the DCC to take the pilot concept to scale hospital-wide without increasing overall personnel during this initial rollout. Patient engagement through automated communication can be facilitated by determining the method of contact using the communication preference documented in each patient’s medical record. It is essential that this automated contact component be followed up with direct communication by clinical staff with the appropriate expertise when a patient indicates that they need assistance through engagement with the automated software.

Second, this work has highlighted the importance of ongoing real-time data review and process improvement. During daily work, the DCC identifies areas where patient care can continue to be improved. For example, the DCC discovered that discharge summaries were less consistently completed for patients boarded in the emergency department than for patients in other inpatient units. In partnership with the emergency department, we conducted a root cause analysis and found that the necessary electronic health record prompts for discharge summary completion were not being displayed as intended in emergency department areas; this problem was swiftly corrected. To enable quick resolution of identified issues, it is essential to have a mechanism in place to share such quality improvement insight with hospital leadership who can directly influence policy and procedures across both inpatient and outpatient settings, ensuring rapid adaptation and hospital-wide innovations with relevance to frontline providers and patients. The DCC team also established a regular Quality Discharge Trends Meeting with provider, nursing, and operations leaders to discuss quality discharge trends. Sharing quantitative and qualitative findings and providing feedback to leaders facilitate continuous improvement “upstream” so that data-driven redesigns can be made earlier in the process. These regular meetings also serve to further strengthen the DCC program, providing an ideal venue for interpreting process and outcome data in real time.

Third, the experience of launching and operating the DCC underscores the important notion in healthcare quality and safety that “communication is key” and emphasizes the importance of dynamics across the broader care team that spans the continuum. Our initial work showed that a readmission risk stratification score alone was insufficient for identifying patients in need of enhanced discharge care, highlighting the need to have a mechanism in place that allows the patient’s care team to communicate with the DCC. For example, we observed many instances where patients classified as low risk by the risk stratification score required DCC interventions because of needs related to social determinants of health, such as not having access to reliable transportation. While we are working on incorporating social determinants of health into our modeling, these cases were the impetus for the development of the DCC eConsult Order process, whereby members of the patient’s hospital care team can easily communicate their concerns directly to the DCC team if they identify something that may threaten the patient’s ability to successfully manage their care after discharge. Reducing the barriers to this kind of communication between the hospital care team and the DCC teams has been one of the most important post-implementation enhancements in the DCC program. Likewise, to provide personalized, patient-focused care, communication channels are crucial in electronic medical record environments to enable the effective transfer of information across care settings. This communication ensures consistency of information across sources, avoiding patient confusion and showing that highest-quality patient care must involve the entire multidisciplinary care team across a patient’s journey. The DCC’s engagement with other parts of the care team enhances personalized patient care during the vulnerable post-discharge period.

Discussion and Conclusions

The DCC was designed to be a personalized care delivery center during the vulnerable time of patient transition out of the hospital. It is intended to fill gaps in care, identify and address new problems quickly, and improve patient outcomes. Although this article only presented the data on reducing readmissions, our process data on the DCC interventions suggest innumerable benefits to patients’ understanding of their conditions and medications, adherence, follow-through with appointments, and recovery.

When seeking to implement a successful DCC model in different organizational contexts, it is important to distinguish between the essential functions of the program, which should be maintained, and the form of program delivery, which may vary according to the staffing model and other health system characteristics. The DCC’s main functions are to 1) stratify patient risk to allow for targeted, personalized care; 2) screen for post-discharge needs among all patients, regardless of risk level and payer source; and 3) provide inclusive coordination of care and monitoring for patients who need enhanced care. The DCC applies a predictive risk model to efficiently identify patients in need and then responds swiftly through the appropriate application of triage and/or care coordination interventions. Patients classified as at high risk for readmission or with complex care needs should be channeled through the proactive arm of the program to ensure enhanced support, whereas low-risk patients can be monitored and assisted through the hospital-wide responsive arm unless a need arises. Including both proactive and responsive components is critical in this intervention, as is incorporating automation to feasibly monitor a large patient population. At minimum, RN staff members will be required to carry out this intervention, and additional specialized clinical staff can add value to the program. Other hospitals might choose different methods of risk stratification, automated outreach, and staffing models to achieve these goals.

The VUMC Discharge Care Center’s success is partially attributable to its multimodal nature, multidisciplinary composition, and dedicated staffing structure. Because patients and their needs vary, the DCC offers multimodal avenues for enhanced post-discharge care and provides accessible and responsive care 24/7, including overnight and on weekends. The DCC’s multidisciplinary composition means that expertise on varied patient concerns in the first 30 days after hospital discharge is available within the DCC team. The DCC’s dedicated staffing structure allows members of both the Triage Team and the Care Coordination Team to focus their work exclusively on improving patient care in the post-discharge period. Taken together, these features combine to produce an independently running care center in continuous operation, true to our goals of making healthcare personal and improving patient outcomes after hospitalization.