Carl Galloway, MD
Clinical Question: Can a collaborative quality improvement project improve the quality and efficiency of pediatric hospital discharges?
Background: Transitions of care, including at the time of hospital discharge, are a potential source of risk and can be associated with adverse events including medication errors and preventable readmissions. Some studies have shown that 10–20% of patients had an adverse event after discharge, and half of those were preventable; one adult study found nearly half of the discharged patients had at least one medication error.1,2 Although multiple projects to improve the discharge process have been published in adult literature, few have focused on the pediatric population. In this study, the Children’s Hospital Association (CHA) formed a pediatric quality improvement collaborative across multiple facilities to examine whether shared improvement strategies would affect failures of discharge-related care, parent-reported readiness for discharge, and readmission rates.
Study Design: Multicenter quality improvement collaborative.
Setting: 11 freestanding tertiary-care children’s hospitals in the United States.
Synopsis: Each of the 11 participating sites chose a specific target population, such as patients with sickle cell disease, asthma, or all discharged pediatric patients. Populations were selected at the discretion of the sites. A multidisciplinary expert advisory panel reviewed literature and developed a change package that included being proactive about discharge planning during hospitalization; improving throughput; arranging post-discharge treatment and support; and communicating post-discharge plan with patients, families, and providers. Each site selected elements of the change package to implement based on individual needs and preferences and incorporated via plan-do-study-act cycles during three action periods. Elements that were implemented by most or all sites included family education on diagnosis and discharge plans, use of discharge checklists, improvement of written discharge instructions, post-discharge follow-up phone calls to reinforce discharge instructions, and identifying and obtaining medications. Virtual learning conferences and monthly Web conferences were held for participants in the collaborative, and experienced improvement coaches guided teams through implementation.
The primary aim of the study was to reduce discharge-related care failures by 50% in 12 months. Failures were measured by phone calls to families two to seven days following discharge, and if any problem related to discharge occurred, the discharge was considered a failure (all-or-none measure). Components of this measure included understanding the diagnosis, receiving discharge instructions and education, complying with instructions, receiving necessary equipment, planning for follow-up pending tests, receiving help with appointments, and not requiring a related unplanned medical visit. Other measures evaluated in this study included patient/family readiness for discharge and unplanned readmission rates (72 hours and 30 days).
Overall, the rate of failures of discharge care was 34% at baseline, which decreased to 21% at the end of the collaborative, for a reduction of 40%. Some individual hospitals exceeded this mark as well. Among the hospitals reporting data on family readiness for discharge, there was a statistically significant improvement, with 85% of families at baseline rating readiness in the highest category and 91% in the last quarter of the study. There was no improvement in rates of unplanned readmission, with 72-hour readmission rates steady across the project (0.7% at onset, 1.1% at end of study; P = 0.29) and slight worsening of the 30-day rate (4.5% to 6.3%; P = 0.05).
Potential explanations for the findings related to readmission rates include seasonal variability in readmissions as well as high variability in patients included in the study. For example, one site focused on patients with sickle cell disease, another on patients with asthma, and others included all diagnoses. Overall, unplanned readmission rates were low (around 1% for 72-hour, 5% for 30-day), which is consistent with other pediatric studies.