In February 2021, clinical leaders (MAA, JMK, BYK, and AC) met with scheduling staff and insurance authorization representatives to understand concerns regarding inpatient-to-outpatient care transitions. These insights were combined with informally collected anecdotal experiences of dermatologists and residents in the preintervention process map (Figure 1). In brief, the preintervention workflow had the following steps:

Key pain points included (1) multiple and inconsistent communication channels (eg, EHR-based messages, emails, and phone calls), (2) unclear roles and responsibilities, (3) burdensome workload for dermatology and scheduling teams, and (4) intensive manual tracking of the process to schedule follow-ups. These pain points may be exacerbated in the consultative context; the inpatient dermatology team consults for numerous services, such as medicine, surgery, intensive care, and obstetrics and gynecology, resulting in many workflows to navigate. Furthermore, skin issues will often not completely resolve by discharge, despite the dermatology consultative team following up with patients closely throughout their hospitalization. Thus, patients may benefit from outpatient skin-directed follow-up; enabling this care faces substantial barriers. In particular, patients’ dermatologic diagnoses may often be unrelated to their primary reason for hospitalization, which may result in a deprioritization (by both patients and providers) of skin issues after discharge.

Phase 1 also included formative qualitative interviews conducted to understand patients’ and caregivers’ experiences with inpatient-to-outpatient dermatology care transition and elucidate their needs. A total of 14 patients and 1 caregiver were interviewed; methods and patient characteristics can be found in Multimedia Appendix 1.

Patients were satisfied with discharge plans and communication with their dermatologists; however, issues persisted around 3 themes: communication and expectation setting during discharge planning; dermatology team support and postdischarge teledermatology; and care coordination and prioritization for medically complex patients. Multimedia Appendix 2 contains supporting quotations. Notably, patients who experienced serious, nondermatological medical events had limited recall of interactions with dermatology. Consequently, for some patients, the necessity of dermatology follow-up was unclear, but caregivers could provide indispensable support. Prioritizing other health issues and coordinating dermatology care with other teams also impacted follow-up success. Nevertheless, patients who followed up with dermatology were generally satisfied with their follow-up coordination experience, even if delayed. They appreciated the dermatology team’s accessibility during transition, facilitated for some patients by teledermatology and direct provider messaging. Despite not all patients following up via video and some still preferring in-person visits, most agreed that video visits are convenient, easy, and provide better access. In fact, almost all patients expressed interest in future video visits, although concerns remained around photo or video quality.

The developed intervention included a SmartPhrase (also known as a dot phrase), a flexible Epic EHR feature that creates templates (eg, fillable statement or table) that can be integrated into patient notes, referrals, or discharge instructions by typing a period and short phrase [18]. On the basis of the patient interviews and informal data collection with inpatient dermatologists, dermatology residents, NPCs, front office schedulers, nurses, and insurance representatives, a team drafted a SmartPhrase form that addressed the main pain points. In particular, the SmartPhrase addressed schedulers’ request for more information, including recommended follow-up timeline, preferred patient contact information, and preferred provider for scheduling and requests for overbooking. Including the reason for follow-up and follow-up timeline was particularly important for confirming patient’s understanding of the importance of follow-up. Patients’ perspectives prompted the inclusion of a field to indicate whether a video visit would be an acceptable option. To address the patients’ perspective that their dermatologic condition was not always their priority, the SmartPhrase was completed with the patient at the bedside to explain the importance of follow-up to the patient and caregiver. The team shared the SmartPhrase with residents and incorporated their feedback before integrating it into Epic.

The SmartPhrase (Table 1) prompted dermatologists to obtain and document pertinent patient information necessary for scheduling follow-up care in the new workflow:

The intervention, the SmartPhrase and associated workflow, was activated on March 22, 2021. Dermatologists (n=5), rotating residents (2 per month), and 15 scheduling staff received group verbal training with written documentation describing the workflow and SmartPhrase use from an improvement leader. A total of 13 scheduling staff, including front office scheduling staff and NPCs and their 2 managers, were also trained.

Mixed methods were used to evaluate the impact and sustainability of the intervention, the SmartPhrase and associated workflow. Specifically, qualitative interview data, scheduling data, and EHR messaging data were triangulated and consolidated and interpreted in parallel.

The semistructured interviews explored interviewees’ perceptions of the early impact of the intervention on follow-up timeliness, workflow and workload, its potential sustainability, and persisting challenges for phase 3 (see phase 2 for methods and analysis).

The impact of the intervention on the timeliness of scheduling, completion of follow-up, and messaging workload was assessed by comparing two periods: (1) preimplementation (January 1 to February 28, 2021) and (2) postimplementation (April 1 to May 31, 2021). March 2021 was excluded, as the SmartPhrase was enabled on March 22, 2021. Data were extracted for all patients who received an inpatient dermatology consultation, were discharged from the hospital (ie, inpatient, observation, and emergency department [ED] encounters) within 1 of the 2 evaluation periods, and were expected to need an outpatient follow-up dermatology visit, that is, hospitalization had current procedural terminology codes indicating potential need for follow-up care (Multimedia Appendix 3). Follow-up visits scheduled and completed within 90 days of discharge were included; those scheduled or completed more than 90 days postdischarge were unlikely to be related to the hospitalization. Outcomes included (1) proportion of patients completing a follow-up visit within 90 days postdischarge, (2) proportion of patients completing a follow-up visit within 14 days postdischarge (postdischarge goal of department), and (3) days from inpatient discharge to completed follow-up. Descriptive statistics are reported. P values were calculated using chi-square tests for categorical outcomes and 2-tailed t tests for continuous outcomes.

EHR-based messaging volume data, specifically in-basket messaging in Epic [20], was used as a proxy for communication workload, as it was a commonly used and measurable. Sent messages were extracted for 5 inpatient dermatologists and 8 dermatology residents (2 per month) involved in inpatient care during the 2 periods. Of 13 scheduling staff, 12 schedulers sent messages during the preimplementation period and 11 schedulers during the postimplementation period. Messages related to scheduling patients who received an inpatient consultation and completed an outpatient follow-up visit within 90 days of discharge were identified using a keyword search (Multimedia Appendix 4). A total of 2 outcomes are reported for the two periods: (1) number of follow-up patients associated with staff messages and (2) number of in-basket messages sent.

Residents felt responsible for care transitions, but they were unsure whether the admitting (ie, nondermatology) or consulting dermatology team was responsible for initiating a referral for outpatient follow-up (Table 2). Consequently, several residents found it easier to submit referrals themselves with a separate message to schedulers and consulting dermatologist. When urgent, some residents also called patients directly or sent additional staff messages to accelerate the process. However, residents rotate monthly, creating opportunities for inconsistencies in workflow and to lose patients during transitions. Residents also did not have role in outpatient dermatology during their 1-month inpatient dermatology rotation. In contrast, other academic dermatology programs have created discharge clinics where residents can provide follow-up care to achieve care continuity [21].

Scheduling relied heavily on back-and-forth messaging among interviewees through various channels, including referrals, in-basket messaging, and email (Table 2). Communication channels used depended on who initiated the referral, whether the patient was considered a new or return patient to the outpatient dermatology clinic, and whether the request was marked with “ASAP”; referrals not marked with “ASAP” were deprioritized with patient outreach occurring within 1 to 2 weeks, delaying care. Within this complex process (Figure 2), referrals were occasionally sent to the wrong staff members, and routinely lacked sufficient information were sent to schedule patients requiring further messaging among the team.

Even with timely and adequate information, schedulers struggled to reach and schedule patients, as patients were unaware of the need or reason for follow-up care (Table 2). Clinicians recognized this may be especially challenging for complex patients juggling many medical issues. Engaging patients and caregivers in shared decision-making predischarge and gauging their interest or ability to attend a follow-up appointment was considered necessary to accommodate patients, improve response to schedulers’ phone calls, and decrease the number of patients who decline or miss follow-up. This may be particularly challenging in the context of consultative dermatology. Although the inpatient dermatology team closely followed most patients during their hospitalization, there was high variability exposure to each patient and their caregivers. There may be variability in the prioritization of dermatological conditions and follow-up care depending on other health conditions and their admitting care team.

Scheduling and dermatology team lacked closed-loop communication; the dermatology team rarely knew whether patients were scheduled for follow-up, leading to persistent worry about losing patients (Table 2). Primarily residents, but also dermatologists, manually kept lists of patients discharged to monitor scheduling activities and follow-up status. This required repeatedly checking the EHR and messaging other scheduling and dermatology team members. Discharge delays further disrupted scheduling of follow-up, but residents only knew of these delays through this tracking and they “probably micromanage[d] it” (dermatologist).

The lack of appointment availability within the desired time frame also contributed to additional messaging (Table 3). When suitable timeslots were not available, extra messages were sent between scheduling staff and clinicians to find additional availability (Figure 2). All interviews considered this process burdensome. Timeliness of follow-up care was considered more important than ensuring continuity of care, but some schedulers and residents believed scheduling with the consulting dermatologist is easier, as the dermatologist could suggest specific timeslots or allow overbooking. The limited appointment availability is particularly challenging for high-volume specialties, including dermatology, that receive referrals from a variety of sources with various urgency.

According to the dermatology team, “...a patient can get scheduled for a visit regardless of their insurance status” (dermatologist), but insurance authorization could disrupt and delay follow-up plans (Table 3). Thus, obtaining approval before discharge could facilitate appropriate follow-up care plan.