Systematic Improvement in the Patient Transfer Process to a Tertiary Care Children’s Hospital

We conducted a quality improvement initiative to optimize the communication processes for accepting interfacility transfers to our 150 bed tertiary care children’s hospital within a large academic medical center. We serve as a statewide referral center with an average of 1800 pediatric patients transferred from referring facilities to our hospital annually. Roughly 40% of inpatient admissions originate directly from a referring hospital. As part of an initiative to better serve our referral base, our clinical outreach team made site visits and held focus groups at referring hospitals to identify opportunities for improvement and promote referrals to our system. Referring providers reported dissatisfaction with the process of transferring pediatric patients to our medical center. Specifically, referring providers desired to speak with peers (attending physicians) rather than residents or fellows and requested a streamlined process that did not involve talking to multiple physicians to find the “correct” provider or specialist to accept a patient. They reported anecdotal examples of transfer communication taking more than 1 hour to have a call connected and a patient ultimately accepted for transfer.

Of note, at the time of our initial intervention, our hospital contracted with an external transfer center (TC) for all incoming admission or transfer requests. Beginning in 2015, our health care system began work on transitioning to an internal TC, which completed in 2018. Because of this large-scale system transition, changes to transfer processes for individual service lines could not be made from July 2015 to March 2019.

After review of the feedback from the focus groups at referring hospitals, improving the transfer of pediatric patients was named a top priority by senior leadership in the children’s hospital strategic plan beginning in 2013. In response, a multidisciplinary team composed of pediatric hospitalists and children’s hospital senior leadership was assembled in early 2013 to begin process improvement for transfers from referring hospitals directly to our pediatric units. An internal focus group of primary stakeholders (pediatric hospitalists, pediatric residents, nursing supervisors, care management, senior hospital leadership) mapped the baseline process and identified areas needing revision and gaps in the baseline process. A key driver diagram was used to identify interventions for improvement cycles (Fig 1). Interventions during the initial period of improvement (Phase 1, July 2013 to June 2016) were focused only on pediatric medicine service transfers (both hospitalist and subspecialty-focused teams). Key issues addressed included delays in speaking with a pediatric provider with authority to accept patients, frustration with having to repeat the patient history to multiple providers within our hospital, and overall length of time required to determine acceptance or disposition status for the request. Additionally, the stakeholders met regularly during this phase to elicit feedback and audit data and process performance, which was an ongoing intervention throughout the project. The active improvement phase was roughly 2 years and was followed by a year of sustainment during the initial planning phases of the system-wide TC transition.⁴ 

In Phase 2 (April 2019 to February 2022), efforts focused on updating protocols with the internal TC and spread of successful interventions from Phase 1 to transfers for 2 additional service lines: pediatric critical care (2019 to 2021) and pediatric surgery (2020 to 2022), selected because of strategic growth priorities for these areas. Phase 2 allowed for improved data access and opportunities to collaborate and review data between children’s and the internal TC leadership at monthly meetings.

After review of key drivers and in response to changes in residency staffing, our initial intervention in Phase 1 was creation of a designated pediatric hospitalist position with primary responsibility for patient throughput, including transfer calls. This position replaced a resident role that triaged initial transfer requests and then discussed requests with various service-line attending physicians to determine patient acceptance. The position was initially for an additional pediatric hospitalist on weekday mornings to coordinate patient flow and transfers during rounds. The primary responsibility of the hospitalist holding the transfer pager was to respond to transfer and admission requests (both externally and from our pediatric emergency department and clinics). Other responsibilities included completing general pediatrics consults and facilitating morning transfers out of the pediatric ICU and admissions to the floor while the accepting teams were rounding with residents. Designated coverage increased to full-day weekdays by mid-2015. Outside of this time, the transfer pager was covered by an in-house pediatric hospitalist additionally fulfilling other clinical duties.

Once the responsibility for transfer requests was transitioned to a designated hospitalist role, our next intervention was to streamline the external TC protocols for pediatric patients by decreasing the time from call initiation to connection with an attending physician. At baseline, our hospital had 13 different protocols and processes for pediatric medical subspecialties and levels of care for the TC to follow to gain patient acceptance. To reduce the options for the TC for contacting an internal physician and standardize handling of transfers for pediatric medical patients, individual pediatric subspecialty protocols were eliminated and replaced by 1 general pediatric medicine protocol for all admission requests to nonsurgical services. This single protocol began with initially contacting the pediatric hospitalist transfer attending for all medical transfers, rather than attending physicians from the various subspecialty services. Initially, the hospitalist would request for a subspecialist to be paged into the call if necessary to determine acceptance to their primary service or for consultation. After the single protocol was created, our next intervention was to eliminate logistical TC intake tasks before contacting the pediatric hospitalist, including: standard questions that were not high-yield for pediatrics (ie, screening for myocardial infarction at the start of the call), obtaining a full demographics sheet, and verifying insurance. Some of these logistical tasks remained in the transfer process but were moved to occur after the initial provider discussion determined if the patient was appropriate for transfer. Although pediatric surgery and pediatric critical care were not a specific focus of Phase 1 interventions, all pediatric protocols underwent similar revisions of logistical task order and prioritization of provider connection.

Although the pediatric hospitalist could facilitate many transfers independently, we noted ongoing delays for subspecialty transfer requests because a second physician was required to respond to the transfer request before acceptance could be offered. In response, the pediatrician-in-chief of our children’s hospital granted the hospitalist role the broad ability to accept transfers on behalf of all nonsurgical pediatric acute care services without their initial approval (“mutual acceptance”). Under mutual acceptance, the hospitalist receiving pediatric medical calls focused on efficiently accepting the patient to our hospital if appropriate. We moved discussion of which service would ultimately admit the patient to after the acceptance and did not include the referring providers in these often-lengthy internal discussions. The hospitalist could request participation of a subspecialist in the call if needed for patient management or disposition questions.

We noted ongoing delays on some transfer requests because of page callback and received feedback from internal providers that it often took time to receive the page, find a phone, and then wait to be connected to the referring provider once they called the TC back. Therefore, we obtained a shared cell phone for direct access to the hospitalist managing transfers to attempt to eliminate delays related to the paging system and reconnection of providers after callback.

Given the success of this intervention in Phase 1, we subsequently implemented direct dial access for pediatric critical care and pediatric surgery transfers in Phase 2. For pediatric critical care, we updated system contact listings and protocols to include the direct dial phone carried by the critical care provider staffing the unit and eliminated the pager listing that previously was used. Beginning 1 year later, we added direct dial access for general pediatric surgery. At baseline, transfer calls were directed to the on-call pediatric surgeon who was frequently operating or occupied with other tasks. During Phase 2, independent of this project, pediatric surgery began assigning a surgeon each week to round on inpatients. Using this improvement, responsibility for transfer calls moved from the daily on-call to the rounding attending, who was given a shared direct dial phone for transfer calls.

As above, key stakeholders met throughout Phase 1 of the intervention to review or audit data and develop interventions. We restructured the stakeholder group to include internal TC leadership after the transition in Phase 2 and resumed monthly meetings for data review and identification of any issues requiring resolution because of variation from expected TC process.

After noting delays in some calls when the first-line provider was unable to answer, the TC implemented a rapid escalation process system-wide. Each service line had a designated second- and third-line provider to take a transfer call in the event the primary provider was unavailable to answer. For pediatric transfer requests, we set the timing to move through the escalation pathway to 5 minutes.

Because of the need for manual extraction of data, we reviewed a random sample of 10 calls per month for pediatric medical and critical care transfers, representing roughly 5% to 10% of all pediatric medical calls and 40% to 50% of critical care calls monthly. We reviewed all inpatient transfers for general pediatric surgery because of lower volume. In Phase 1, the external TC used a software algorithm for random call selection for review and sent transcripts for 10 randomly selected transfer calls to the improvement team for manual review and extraction. Because of this arduous process and delays in receipt, additional transcripts could not be requested to replace any that were incomplete or did not meet inclusion criteria. As such, some months had <10 pediatric medical calls included during Phase 1. In Phase 2, given the internal TC module integration with our electronic medical record, we generated a monthly TC report for each service line of interest. Calls were randomly ordered in the report, and the first 10 calls meeting inclusion criteria were reviewed for manual extraction. If a randomly selected call did not follow the process of interest (eg, an internal provider accepted a patient outside of the referral line or call was marked as pediatric hospitalist but was actually emergency department), it was excluded at that stage and in Phase 2, replaced with the next call.

Manual review of TC transcripts was performed by 2 team members. Data extracted included time stamp data from the call (including discrete times documented in the transcript for call initiation, time of provider connection and time of patient acceptance), admitting service, and number of internal physicians on the call before acceptance. The number of internal physicians on the call before patient acceptance was measured for pediatric medical transfers, because this was a specific target of mutual acceptance and directly impacts time to acceptance. On initial data review for pediatric surgery and pediatric critical care, calls were largely managed by a single provider, given the nature of the well-defined populations for the service lines, so we did not monitor this measure for improvement for those services. For improvement purposes, any variations from the standard process found when reviewing transcripts were flagged and reviewed in the monthly improvement meetings.

As a balancing measure, rapid response team activations for all transferred patients within 24 hours of admission was monitored during Phase 1 with the implementation of the hospitalist transfer role. Data analyzed included the date of the rapid response occurring within 24 hours of any interfacility transfer admitted to the inpatient floor. This measure was selected to ensure the process maintained safety despite increased speed of decisions, less subspecialty involvement in transfer calls, and more direct admissions to the unit bypassing the emergency department. This measure was primarily monitored in Phase 1 given that Phase 2 did not involve changes to mutual acceptance.

Data for measures were monitored on SPC charts (X-bar S for pediatric medicine and pediatric critical care time to provider and time to acceptance in aggregates of 10, X-mR chart with monthly averages for time data pediatric surgery given variable call number per month, c-chart for average number of providers per month, and t-chart for rapid response data) that were maintained on a dashboard for team member and stakeholder review. We used standard rules for Shewart chart interpretation.⁵ 

Phase 1 included data from 335 pediatric medicine transfer calls. Notably, during Phase 1 there was an overall increase in floor transfers each year to a total increase of roughly 200 transfers annually being admitted directly to the floor. Time between transfer call initiations to connection to an internal physician was reduced from an average of 11 to 5 minutes (Fig 2). Time from the initiation of the transfer call to acceptance to pediatric medical services was reduced from a baseline average of 23 to 11 minutes (Supplemental Fig 5). Special cause variation was noted on both SPC charts temporally coinciding with the implementation of mutual acceptance by the pediatric hospitalist for subspecialty service and use of the direct dial phone, respectively. The number of internal physicians on each call before acceptance decreased from a mean of 2.1 to 1.3 physicians per call.

Average time from call initiation to connection with an internal physician was reduced from 11 to 4 minutes for pediatric critical care transfers (Fig 3), and 16 to 5 minutes for pediatric surgery transfers (Fig 4). Special cause variation was seen in pediatric critical care transfers correlating with the bundled implementation of direct dial for an accepting provider and beginning monthly stakeholder data review. Additional special cause variation was detected after ongoing improvement after resumption of normal TC staffing after notable variability at the onset of the COVID-19 pandemic. Time to provider connection for pediatric surgery transfers demonstrated special cause variation coinciding with the change in process to a direct dial phone being carried by the rounding surgeon for transfer requests. Both processes reached goal times in Fall 2020 and demonstrated sustained improvement without further interventions.

Average time from call initiation to acceptance for admission was reduced from 20 to 11 minutes for critical care (Supplemental Fig 6) and 20 to 9 minutes for pediatric surgery (Supplemental Fig 7). Special cause variation was noted to directly correlate with decreases seen in time to provider connection.

Days between rapid response team activations within 24 hours of admission for transferred patients decreased from an average of 5.6 to 4.8 days during Phase 1. Special cause was detected 2 weeks before the implementation of mutual acceptance.

We streamlined the communication processes for transferring pediatric patients from referring hospitals into our tertiary care children’s hospital by prioritizing provider-to-provider connection and early acceptance of patients. Although this process largely occurs external to the hospital, it is vital to patient flow and referring provider satisfaction, which has also been addressed by other hospitals.^6,7  Our establishment of a specific hospitalist role to broadly manage transfer calls and use of direct dial access to attending providers across pediatric service lines demonstrates an efficient and safe approach for transfer of pediatric patients to inpatient units without using the emergency department.

We selected time from referral call to connection to a provider as 1 of our primary process measures, because referring providers in our focus groups felt this was a high priority. Both financially and in practice, hospital leadership prioritized the needs of our referring providers by funding and authorizing a hospitalist role with mutual acceptance capabilities who eventually would be able to accept patients to a variety of specialty services. Our approach expands upon that reported by Blarcom and colleagues in their 2014 study detailing a similar direct admission system which demonstrated that decreasing time spent waiting to speak to the admitting physician improved referring physician satisfaction.⁸  Prioritizing efficient acceptance is particularly important in a health care system in which transferred patients frequently originate outside of the immediate region. Actual arrival of the patient, which was not the focus of this study, often takes hours because of securing transportation and travel time, but these pieces are dependent on the initial acceptance. Allowing initial mutual acceptance by the hospitalist and moving the discussion of which specific service accepts to later in the process allows patient transport to start as quickly as possible. Prompt indication of acceptance also secures the transfer to our hospital and reassures referring providers that they do not need to contact other hospitals.

This model removed residents from decision-making in the transfer process, which might impact their educational experiences. Our referral base indicated a preference to communicate with an attending physician, and our hospital also eventually strongly recommended attending physician involvement with transfers. Having consistent hospitalist providers receiving transfer calls over time has increased familiarity with triage and managing patient flow, likely leading to intangible improvements in the system and ongoing improvements in efficiency handling transfer requests overtime. We currently offer senior residents the opportunity to participate in an elective rotation with the pediatric hospitalist role responsible for transfers if desired.

A dedicated physician focused on TC and other patient flow needs has associated cost. Implementation of a similar process elsewhere could be limited by the willingness of the hospital to support a specific physician role on weekdays to manage the large subset of pediatric medicine transfers as we implemented in our system. However, there are also likely associated financial benefits to this system, including increased efficiency of physicians who were previously disrupted by TC calls. The other roles of this hospitalist including transferring patients out of the intensive care units, and our own emergency department also can help improve flow and bed use while teams are rounding. Additionally, although we did not model the long-term impact of this cost investment, improving transfer to our hospital might bring patients into our system who otherwise would have received care at other systems.

Although transfer of children to hospitals with a higher level of care is often unavoidable, and there are benefits to direct floor admission for patients, there are also associated risks.^9–11  We attempted to monitor for increased risk by increased speed of acceptance by monitoring rapid response calls within the first 24 hours after transfer. Although the time spent on transfer discussions decreased, we also noted a slight decrease in the number of days between rapid responses in the day after transfer, whereas the time spent on the transfer discussions decreased. This finding is difficult to interpret, because it may be simply because the volume of transfers increased, given that we noted special cause variation before the implementation of mutual acceptance. It is also possible that other factors impacted this that were outside of our study, such as capacity in various levels of care in the hospital, skill of transport team in stabilization, or seasonal proportion of various illnesses with differing likelihood of needing escalation of care.

In Phase 2, we expanded the intervention to critical care and pediatric surgery services. These service lines never added a separately staffed physician role to manage transfers, although both developed a system to be available by direct dial phone. We found the transfer of responsibility to a surgical attending with rounding responsibilities and elimination of pager call-back was particularly effective for pediatric surgeons. The escalation process when the primary person responsible for transfers was not available allowed a back-up process if a surgeon or intensivist was in a procedure when the initial call came.

Our study outlines a distinct operational structure that we believe can be replicated at other hospitals and clinical services to prioritize the needs of patients and referring providers. This model was successfully applied to pediatric critical care and pediatric surgery transfers and has subsequently been adopted by adult hospitalist and emergency medicine services at our hospital after follow-up site visitations to referring providers noted communication and ease of transfers of pediatric patients as a strength of our health care system.

This is a single-center study at a tertiary care children’s hospital, which may limit generalizability, although the spread of the model of direct call access to a physician with mutual accepting authority internally across different service lines demonstrates its flexibility. Institutional cultural details may have impacted our improvements, such as our wide and varied referral base and the presence of other nearby competing children’s hospitals to motivate improving the experience for referring providers. We improved the process of communication for accepting transfers, but reducing the time to actual patient arrival was not in scope for this project as it relies on many variable components, including bed and transport availability. Notably, we have a significant time gap in data collection from 2016 to 2019 when we moved to an internal TC and were not adequately able to measure system performance. Similar system limitations are common in quality improvement work, and the ability to apply the previous improvements to a completely new system, if anything, solidified our belief that our improvements are generalizable.

Prioritization of process improvement for transfers calls from referring facilities to a tertiary care children’s hospital was effective in decreasing time to provider connection, which ultimately decreased the time to accepting patients for transfer. Key interventions included creation of a pediatric hospitalist role specifically for managing transfer requests and flow within the hospital and implementation of direct dial access to accepting providers to streamline the process.