BACKGROUND AND OBJECTIVE: Children with medical complexity have unique needs when facilitating transitions from hospital to home. Defining readiness for discharge is challenging, and preparation requires coordination of family, education, equipment, and medications. Our multidisciplinary team aimed to increase the percentage of medically complex hospital medicine patients discharged within 2 hours of meeting medical discharge goals from 50% to 80%.
METHODS: We used quality improvement methods to identify key drivers and inform interventions. Medical discharge goals were defined on admission for each patient. Interventions included implementation of a complex care inpatient team with electronic admission order set, weekly care coordination rounds, needs assessment tool, and medication pathway. The primary measure, percentage of patients discharged within 2 hours of meeting medical discharge goals, was followed on a run chart. The secondary measures, pre- and post-intervention length of stay and 30-day readmission rate, were compared by using Wilcoxon rank-sum and χ2 tests, respectively.
RESULTS: The percentage of medically complex patients discharged within 2 hours of meeting medical discharge goals improved from 50% to 88% over 17 months and sustained for 6 months. In preintervention–postintervention comparison, median length of stay did not change (3.1 days [interquartile range, 1.8–7.0] vs 2.9 days [interquartile range, 1.7–6.1]; P = .67) and 30-day readmission rate was not impacted (30.7% vs 26.4%; P = .51).
CONCLUSIONS: Efficient discharge for medically complex patients requires support of a multidisciplinary team to proactively address discharge needs, ensuring patients are ready for discharge when medical goals are met.
- CCC —
- complex chronic condition
- CCHMC —
- Cincinnati Children’s Hospital Medical Center
- EHR —
- electronic health record
- HM —
- hospital medicine
- IQR —
- interquartile range
- LOS —
- length of stay
- RN —
- registered nurse
Children with medical complexity have multisystem, chronic disease that can result in frequent hospitalizations.1,2 With multiple diagnoses, need for technology assistance, many subspecialty providers, and numerous medications, the discharge planning process for this population is different than for otherwise healthy children hospitalized with acute illnesses. Although it is essential to address discharge needs to ensure safe and effective transition from hospital to home, hospital physicians often prioritize treatment of acute medical problems over discharge planning.3,4 This poses challenges to providing timely, efficient, and safe hospital discharges, 3 care characteristics prioritized by the Institute of Medicine.5 Furthermore, discharge delays negatively impact patient flow and family experience.6
In previous work, we improved discharge efficiency in our general pediatric hospital medicine (HM) patients.7 Through standardization of discharge goals and implementation of high-reliability interventions focused on physician and nursing processes in the electronic health record (EHR), 80% of patients are now discharged from the hospital within 2 hours of meeting medical discharge goals.
However, the discharge process for medically complex patients remained inefficient; only 50% of patients on the HM service with neurologic impairment8 and/or technology dependence9,10 were discharged within 2 hours of meeting medical discharge goals. Preliminary work revealed that the medical team often overlooked the particular discharge needs of these medically complex patients and their families until after a child was medically ready for discharge. Discharge planning, including changes to home care orders with need for new equipment and teaching, multiple medication refills with need for previous authorization, and specialized transport home, was not approached in a standard manner nor addressed until the end of the stay. We hypothesized that interventions focused on optimization of a standardized discharge infrastructure for medically complex patients would improve discharge efficiency. By using improvement methods and reliability science, our multidisciplinary team aimed to increase the percentage of medically complex HM patients discharged within 2 hours of meeting medical discharge goals from 50% to 80% within 12 months.
Cincinnati Children’s Hospital Medical Center (CCHMC) is a 522-bed, free-standing children’s hospital. Children with medical complexity, defined as children with neurologic impairment and/or technology dependence for the purpose of this study, are admitted primarily to 2 general HM units staffed by pediatric registered nurses (RNs), with HM attending physicians that supervise a total of 5 teams of pediatric residents providing direct care. Neurologic impairment is defined as “functional and/or intellectual impairments that result from a variety of neurologic diseases” (eg, anoxic brain injury, lissencephaly).8 Patients with technology dependence “depend on medical technology to live or remain in their current state of health” (eg, tracheostomy, enteral feeding tube, cerebral spinal fluid shunt).9,10 The majority of these children (55%) receive outpatient care at CCHMC’s Complex Care Center, a medical home that provides primary care to 620 children with severe, chronic disease who receive care from ≥3 subspecialists.
Planning the Intervention
Previous process improvement on acute care patients7 included identification of medical goals for discharge and real-time documentation of when goals were met by bedside RNs via an EHR timestamp (Fig 1); the same process was applied to complex patients. We created a multidisciplinary group that included HM attending physicians, RNs, care managers, pharmacists, pediatric residents, social workers, and parents of children with medical complexity. The group defined the process of efficient discharge for children with medical complexity and identified key drivers (Fig 2). Interventions were designed to address top failure reasons for not leaving within 2 hours of meeting medical goals before the process was implemented, specifically transportation concerns, patient/parent factors, physician delay, and medication delay (Fig 3). Successful interventions were modified through sequential plan-do-study-act cycles based on the model for improvement11 before adopting into the process.
Patients Grouped onto the Complex Care Inpatient Team
In July 2013, we grouped children with medical complexity into 1 HM team supervised by a subset of 15 HM attending physicians to provide specialized care, including proactive discharge planning, to this patient population. The patients are identified at the time of admission by the RNs who manage bed placement in our hospital using clinical information from the admitting provider. Before this work, these patients were scattered among all HM resident teams. Additional staff, including a dedicated pharmacist, dietician, care manager, and social worker, were hired through hospital investment in improving chronic care. We also partnered closely with the outpatient Complex Care Center team, with their attending physicians and care managers frequently joining us for patient rounds.
Complex Care–Specific Order Set
In September 2013, we tested a complex care–specific admission order set in our EHR. The order set included a medical discharge goal order (Fig 4), specific to the needs of complex patients (eg, baseline oxygen requirement for 12 hours, tolerating enteral feeds for 24 hours). This order was placed on admission, and the provider, with input from caregivers and other team members, chose medical discharge goals from this list or added other goals relevant to the patient’s diagnoses. It was then modified as the patient’s course evolved. It focused only on medically relevant items with the intent that other discharge tasks (eg, home care orders, medications) were completed in advance of the patient meeting medical discharge goals.
Weekly Multidisciplinary Care Coordination Rounds
In October 2013, the improvement team implemented weekly multidisciplinary care coordination rounds. All team members attended this meeting to discuss discharge goals and complete discharge-related tasks, including sending medications to the pharmacy and completing home care orders. Any clarifying questions, such as transportation needs, were then reviewed with families at the bedside. Additional interventions were needed to coordinate care for patients with shorter lengths of stay whose admission did not overlap this weekly meeting.
Needs Assessment Tool
In January 2014, a needs assessment tool was created to help structure care coordination rounds and ensure comprehensive discharge for patients with shorter hospitalizations. This checklist included 8 essential discharge tasks specific to patients with medical complexity (Fig 5). Although the items included in the needs assessment tool were distinct from medical discharge goals, these tasks ensured the logistics of discharge were addressed throughout the hospitalization. Initially a paper document, the needs assessment tool was later incorporated as a modifiable document in the EHR, allowing all members of the team to see the status of each task. The assessment was started on admission and reviewed regularly throughout the patient’s hospitalization, including weekly team review at care coordination rounds, to facilitate completion of all tasks (eg, new equipment, home nursing orders) 24 hours before the child was medically ready for discharge. Specific sections of the needs assessment tool were assigned to team members (eg, home care needs were the primary responsibility of the care manager) to improve reliable task completion. At time of discharge, any outstanding tasks were completed by the discharging resident or nurse practitioner.
Because discharge medication prescribing was a frequent cause for delay, a medication pathway was introduced in late January 2014 to identify barriers (eg, previous authorization) or changes in regimen (eg, new prescriptions) in advance of discharge. The team pharmacist oversaw medication reconciliation after admission and tracked medication changes through hospitalization. The pharmacist also led a weekly meeting separate from care coordination rounds in which all medications were reviewed. Additionally, our pharmacist worked with families to identify home medications requiring refill, encouraged the team to prescribe discharge medications early in the stay, and called pharmacies to ensure medications were available.
Planning Study of the Intervention
Baseline data before the advent of the new inpatient complex care team included medically complex HM patients, identified by their primary care relationship with the Complex Care Center, from July 2012 through June 2013.
Data describing our cohort were extracted from the EHR, including age, gender, primary insurer, reported race and ethnicity, complex chronic conditions (CCCs),12 technology dependence, and discharge diagnoses. CCCs were defined as “any medical condition that can be reasonably expected to last at least 12 months (unless death intervenes) and to involve either several different organ systems or one system severely enough to require specialty pediatric care and probably some period of hospitalization in a tertiary care center.”12 CCCs were grouped into 11 categories (eg, gastrointestinal, respiratory). Technology dependence (eg, tracheostomy) was defined using the “dependence upon medical technology” or “device” subcategory within relevant CCC categories. CCCs and technology dependence categories are not mutually exclusive (ie, a patient may have a diagnosis in >1 CCC or technology dependence category).
The primary outcome measure was defined as the percentage of medically complex patients, admitted to the 2 primary units for HM patients, who were discharged within 2 hours of meeting medical discharge goals. We focused on these 2 units becuase they already followed the discharge process based on medical goals from our previous work.7 Median length of stay (LOS) was a secondary outcome measure. To ensure that our work in expediting discharge did not negatively impact readmission, 30-day readmission rate was evaluated as a balancing measure.
We examined cohort demographic and clinical characteristics using descriptive statistics. A run chart was used for analysis of our primary outcome measure. Established rules identified special cause variation for run charts; specifically, 8 consecutive points above or below the centerline, which would occur <0.4% of the time by chance, led to a midline shift.13–17 For analysis of pre- and postintervention outcomes of LOS and 30-day readmission rate, we excluded patients admitted during the intervention period (September 22, 2014–March 23, 2015). Pre- and postintervention median LOS were compared by using Wilcoxon rank-sum test. Pre- and postintervention 30-day readmission rates were compared by using χ2 test.
Of the 385 encounters during the study period (July 2012–May 2015), there were 227 unique patients; 13 patients were admitted in both preintervention and postintervention timeframes. The 227 patients were 54% male with a median age of 5.3 years (interquartile range [IQR] 2.2–15.6). The majority were white (66.1%) and non-Hispanic (92.9%) with public primary insurance (71.4%). Nearly three-quarters of children had diagnoses in ≥4 CCC categories, with the most common being neuromuscular (75.8%), gastrointestinal (73.1%), and congenital (65.6%). Nearly 80% of children were technology dependent, most commonly in the gastrointestinal category (70.9%). There were no significant differences in demographics or clinical characteristics of admitted patients pre- versus post-intervention. The most common discharge diagnoses in both pre-and postintervention periods were: (1) pneumonia (30% vs 22%), (2) bronchiolitis (13% vs 13%) and (3) vomiting and/or diarrhea (12% vs 8%). Approximately 4% of total HM discharges were attributable to the group of medically complex patients included in the study, which accounted for ∼17% of our bed days.
The percentage of medically complex patients discharged within 2 hours on our 2 study units increased from 50% to 80% within 7 months (Fig 6). Our initial shift to goal occurred after the institution of the needs assessment tool and medication pathway.
Although we initially reached our goal in October 2014, we experienced a downward shift of our outcome measure, with the median percentage of eligible patients discharged within 2 hours of meeting medical discharge goals decreasing to 63%. This shift coincided with a rapid increase in our overall hospital census starting in August 2014. The increased census on the units of interest may have led bedside providers, such as RNs covering other patients with competing care demands, to stray from proactive discharge planning for our complex patients. With interventions aimed at increasing process reliability, including more directed role assignment to team members so that each provider was aware of his/her specific task responsibilities, we were able to increase our median back above goal, even with continued high census. This improvement has sustained at goal for 6 months.
Median LOS, our secondary measure, did not significantly change between pre- and postintervention (3.1 days [IQR, 1.8–7.0] vs 2.9 days [IQR, 1.7–6.1]; P = .67). In addition, our balancing measure, 30-day readmission rate, was not negatively impacted pre- and postintervention (30.7% vs 26.4%; P = .51).
Through interventions focused on proactive discharge planning for medically complex patients, we were successful in increasing the percentage of patients discharged within 2 hours of meeting medical goals from 50% to 88%. Our most impactful interventions included standardizing discharge planning processes and identifying discharge barriers earlier. Although patients left soon after meeting discharge goals, the decrease in LOS was not significant.
An overall improvement in the efficiency of our process is valuable even without LOS decline. By anticipating discharge needs early, we were better able to predict timing of discharge, which facilitates anticipating bed capacity on our units. Additionally, our providers noted a perceived workload decompression, because the tasks were no longer left for completion on day of discharge. Our process also allowed families to clearly delineate their home needs so that details were planned well in advance of medical readiness. With our detailed process and dedicated team members, we also believe we decreased the likelihood for errors in the postdischarge timeframe, such as inaccurate prescriptions or home nursing orders. Readmission rates were also not affected in our study, suggesting that tracking medical goals is a reasonable method to determine when patients are ready for discharge, and that our process change did not lead to patients being discharged too early.
In our previous efforts to improve discharge efficiency,7 we focused on acute care patients admitted with general medical diagnoses. Although medically complex patients were included in those efforts, we struggled to discharge this subset of patients in a timely fashion, due to previous interventions not being designed for coordination of extensive outpatient needs. By first standardizing the way we define medical discharge goals in this patient population and making this order readily available in the EHR order sets, admitting providers were better able to apply the previous process of defining discharge goals on admission without interfering with their workflow. This early intervention facilitated later changes aimed at standardizing discharge processes.
As experts in the care of their children, it was essential that family members be engaged in our improvement processes. Medical goals were discussed and modified with family input, and discharge needs were identified through interactions with our multidisciplinary team members, including our dietician, social worker, pharmacist, and care manager. Consideration for family schedules, home equipment and medication refills needed, and transportation availability allowed us to reach the common goal of readiness for discharge when medically appropriate.
Care coordination rounds were instrumental in achieving reliable completion of tasks before discharge. By meeting as a team at a designated time outside of rounds, we confirmed that medical goals were updated and social barriers identified. Key to our success, the needs assessment tool allowed tasks to be outlined and tracked over time, facilitating efficient discussion. This is similar to adult studies that found success by incorporating needs assessments into their discharge planning bundles.18–22 In adults on a general medical service,21 in adults with heart failure,22 and in elderly patients, internal medicine teams demonstrated that by assessing patients’ needs, they were able to target interventions to individual patients19,20 Our study used a similar approach to identify individual patient needs in our pediatric population and target interventions (eg, assist with transportation arrangement) to facilitate a smooth transition to home. One area included in the needs assessment tool that often required extensive coordination was discharge medication preparedness. Medication errors can lead to confusion at home, adverse drug reactions, and increased reutilization,23 so attention to performing comprehensive reconciliation before discharge was essential. Our pharmacist-led medication pathway ensured communication among prescribers, families, and pharmacies. The input of a pharmacist in predischarge medication reconciliation is well-described in adult hospitals as a way to improve accuracy of medication lists.20,23–28 Our study adds to this literature, because our pharmacist-led medication pathway was critical to our process.
We limited this improvement initiative to patients on our 2 main HM units because those units used the medically ready discharge process from our previous work.7 This led to a relatively low number of patients included in this study, which may have led to an increase in variability, especially early in data collection. We noted even after an increase in our biweekly numbers, however, that the centerline of 50% remained consistent and thus feel this is reflective of the true baseline.
Our study population was limited in that it did not include patients with traditionally longer LOS, such as those with ventilator dependence, because they are admitted to other units. By this, we may have selected for a population of medically complex patients with shorter LOS, influencing our ability to detect significant changes in the secondary measure of LOS. Importantly, LOS did not increase during this project, nor was there an increase in readmissions, suggesting that patient care and discharge using this new process did not contribute to increased return for admission because of an expedited discharge. We also did not include other medically complex patient populations in our scope; by first applying the process to our HM patients, we now have experience to support buy-in from other specialty providers. We will continue to follow our secondary measures as we spread this process to other services at our hospital, which will allow us to measure our impact on a larger scale.
The creation of a multidisciplinary team with a variety of expertise influenced our ability to improve rapidly, which potentially limits the generalizability of our study. However, many of our key interventions, such as meetings to facilitate care coordination and a tool to track discharge task completion, could be easily implemented in environments where such a team is not available, and the failures we addressed in our process are likely common to many settings.
Finally, the target of our improvement, the discharge process, is limited in that it is inherently people dependent. Although we used the EHR to standardize as much as possible, our frontline providers must be engaged for it to be successful.
The discharge needs of medically complex patients require the support of a multidisciplinary team. By defining medical goals and discharge needs early, tracking tasks over time, and designating roles to team members, we ensured that discharge tasks were complete when patients were medically ready for discharge.
We appreciate the dedication of our team, including: Suzan DeCicca, MSW, LSW; Stacey Litman-Padnos, MSW, LSW; Julie Ostrye, PharmD; Becky Brehob-Bucker, RD; Derek Fletcher, MD; David Hall, MD; Michelle Cobble, BSN, RN; Matthew Carroll, MD; Steven Smith, MD; Emily Goodwin, MD; Meghan Hofto, MD; Hilary Flint, DO; Marshall Ashby; Shelly Miller; Margaret DeOliveira; CCHMC inpatient and Complex Care Center care managers; CCHMC pediatric residents; and HM advanced practice nurses and attending physicians.
- Accepted April 19, 2016.
- Address correspondence to: Angela M. Statile, MD, MEd, Division of Hospital Medicine, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave, ML 3024, Cincinnati, OH 45229. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated that they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated that they have no potential conflicts of interest to disclose.
- Cohen E,
- Kuo DZ,
- Agrawal R, et al
- Simon TD,
- Berry J,
- Feudtner C, et al
- Institute of Medicine
- Shepperd S,
- Lannin NA,
- Clemson LM,
- McCluskey A,
- Cameron ID,
- Barras SL
- White CM,
- Statile AM,
- White DL, et al
- Langley GJ,
- Moen R,
- Nolan KM,
- Nolan TW,
- Norman CL,
- Provost LP
- Feudtner C,
- Hays RM,
- Haynes G,
- Geyer JR,
- Neff JM,
- Koepsell TD
- Provost LP,
- Murray SK
- Benneyan JC
- Desai AD,
- Popalisky J,
- Simon TD,
- Mangione-Smith RM
- Murphy EM,
- Oxencis CJ,
- Klauck JA,
- Meyer DA,
- Zimmerman JM
- Keeys C,
- Kalejaiye B,
- Skinner M, et al.
- Sarangarm P,
- London MS,
- Snowden SS, et al.
- Copyright © 2016 by the American Academy of Pediatrics