BACKGROUND AND OBJECTIVES

Factorial design of a natural experiment was used to quantify the benefit of individual and combined bundle elements from a 4-element discharge transition bundle (checklist, teach-back, handoff to outpatient providers, and postdischarge phone call) on 30-day readmission rates (RRs).

METHODS

A 24 factorial design matrix of 4 bundle element combinations was developed by using patient data (N = 7725) collected from January 2014 to December 2017 from 4 hospitals. Patients were classified into 3 clinical risk groups (CRGs): no chronic disease (CRG1), single chronic condition (CRG2), and complex chronic condition (CRG3). Estimated main effects of each bundle element and their interactions were evaluated by using Study-It software. Because of variation in subgroup size, important effects from the factorial analysis were determined by using weighted effect estimates.

RESULTS

RR in CRG1 was 3.5% (n = 4003), 4.1% in CRG2 (n = 1936), and 17.6% in CRG3 (n = 1786). Across the 3 CRGs, the number of subjects in the factorial groupings ranged from 16 to 674. The single most effective element in reducing RR was the checklist in CRG1 and CRG2 (reducing RR by 1.3% and 3.0%) and teach-back in CRG3 (by 4.7%) The combination of teach-back plus a checklist had the greatest effect on reducing RR in CRG3 by 5.3%.

CONCLUSIONS

The effect of bundle elements varied across risk groups, indicating that transition needs may vary on the basis of population. The combined use of teach-back plus a checklist had the greatest impact on reducing RR for medically complex patients.

Pediatric care transitions from the hospital to the home are complex events, and as such, they require multicomponent interventions to effect change.13  Failed transitions are associated with medication errors, communication deficits between hospital-based and primary care providers (PCPs), and poor patient adherence to follow-up.413  Collectively, these deficiencies may lead to increased hospital reuse and health care cost.1416 

The multifaceted nature of the transition from the hospital to the home requires a novel approach that can move care coordination beyond an individual venue.1723  A subcommittee of the American Academy of Pediatrics Section on Hospital Medicine launched the Improving Pediatric Patient-Centered Care Transitions (IMPACT) quality improvement (QI) research collaborative to design and study a multicomponent patient-centered pediatric care transition (PACT) bundle to improve this process.2425  The IMPACT collaborative used the framework for design and evaluation of complex interventions in QI research.27,28  First, the PACT bundle was designed by using the best available evidence published in the literature to consist of 4 elements: a transition readiness checklist, predischarge teach-back education, timely and complete written handoff to the PCP, and a postdischarge phone call.2934  Secondly, the feasibility of PACT bundle implementation was tested by using QI methodology at 4 pilot sites.26 

To validate the PACT bundle impact before large-scale implementation, the next important step and the aim of our present study was to assess the effect of the bundle elements and their combination on 30-day all-cause, unplanned readmissions for patients of different levels of medical complexity (no chronic condition, single chronic condition, and multiple chronic or complex medical conditions).27,28 

To achieve that goal, we used a retrospective factorial design model to analyze the data we collected throughout our QI initiative.

Factorial study design allows investigators to simultaneously evaluate the effect of >1 independent variable on a single dependent variable. In addition, interactions between independent variables can be explored. Therefore, factorial design is ideally suited to study a 4-component discharge bundle.3538  This method has been used extensively in other industries and, to a limited extent, in medicine.3545  Contextual considerations may also be examined by applying the rubric to background variables.46,47 

Although the benefits of factorial design are especially important when assessing complex systems with limited time and resources, ideally, this type of design is best used for prospective, randomized, multifactorial studies.37,38,43,44  For the IMPACT research collaborative 4-element PACT bundle, a prospective study would have required 16 hospital sites, simultaneously testing an assigned combination of bundle elements. As a grassroots, unfunded collaborative, this approach was not feasible or ethical. At each site, components of the bundle were already being used to some degree, and enforcing discontinuation of favorable practices belied the intent of the project. Instead, we measured variables of interest without assigning or controlling which bundle element or combination of elements each patient received. Data collected in this manner form a natural experiment suitable for retrospective factorial analysis (RFA). Therefore, like other investigators,49  we chose this method as the best available study design to assess the effect of each bundle element and combinations.

As outlined in the IMPACT pilot report, subjects were included from various populations at the 4 hospitals on the basis of each site’s annual patient volume, resource availability, and organizational priorities.26  Subpopulations included technology-supported (patients with ventriculoperitoneal shunt, tracheostomy tube, surgically placed feeding tube, and/or indwelling central venous catheter at age 0–18 years) and non–technology-supported patients (infants, toddlers, asthma patients age 2–17 years old or all patients admitted to a specific hospital unit).26  Three sites (A, B, C) collected data from January 2014 through December 2017; 1 site (D) contributed data through December 2016.26 

Because of expected difference in readmission rates (RRs) and discharge preparation needs, patients were manually classified by using a previously published pediatric medical complexity algorithm into 3 clinical risk groups (CRGs): no chronic disease (CRG1; eg, patient with bronchiolitis), single chronic condition (CRG2; eg, patient with asthma), and multiple or complex chronic conditions, including technology-supported patients (CRG3).49 

Patient Demographics

We calculated descriptive statistics on patient characteristics (age, sex, preferred language, insurance type, CRG, and hospital site) using proportions for categorical variables and medians for continuous variables.

RFA

Because there were important differences in RR for the 3 CRG groups, data from the 3 groups were analyzed separately (Table 1). A 24 factorial design matrix was created for each group, with 2 indicating the number of levels and 4 expressing the number of factors.36  In our study, the term “factor” referred to “bundle element,” whereas the 2 levels indicated whether bundle elements were used or not used for each patient. Our main outcome was 30-day all-cause, unplanned RR. Study-It software (McGraw-Hill Professional, New York, NY) was used to analyze the effects of the 16 possible combinations of bundle elements for each CRG. This can be visualized by examining the design matrix depicted in Fig 1: the test groups represent the 16 possible combinations of bundle elements (rows), with the RR for each shown in the final column (far right). The estimated unweighted effects (bottom row) of each of the 4 bundle elements and bundle element interactions (columns) is calculated by adding or subtracting (according to sign) the RR of each test group and dividing by the number of positive test groups.36  The distribution of these effects was viewed graphically on a dot diagram to separate the important effects (outliers or values farthest from 0) from background noise (values clustered around 0).36  The factors and combinations yielding important effects were then analyzed by using response plots, which visually display the effect of bundle elements and interactions between elements, thereby illuminating important cause-and-effect relationships. Paired comparison chart analyses then followed to further evaluate the consistency of bundle elements on reducing RR. The effect of each bundle element on RR in relation to other bundle elements provides support for the conclusions drawn from the response plots.

FIGURE 1

Design matrix for CRG3 revealing test groups for each individual bundle element, 2-, 3-, and 4-bundle element interactions, and corresponding RRs. There were no important 3- or 4-way interactions. Note the effects of teach-back (T) and T plus a checklist (C) on RRs (last row). H, handoff; P, phone call.

FIGURE 1

Design matrix for CRG3 revealing test groups for each individual bundle element, 2-, 3-, and 4-bundle element interactions, and corresponding RRs. There were no important 3- or 4-way interactions. Note the effects of teach-back (T) and T plus a checklist (C) on RRs (last row). H, handoff; P, phone call.

Close modal
TABLE 1

Summary of Matrix Data for CRGs 1, 2, and 3 Revealing the Number of Patients Receiving the 16 Combinations of the IMPACT Bundle Elements and Corresponding Number of Readmitted Patients

Test GroupTotal
12345678910111213141516
CRG1                  
 No. patients in each test group 138 153 296 535 124 116 328 674 67 117 263 406 51 79 234 422 4003 
 No. readmitted 10 27 12 29 14 13 142 (3.5% RR) 
CRG2                  
 No. patients in each test group 70 109 1491 310 46 67 185 303 17 38 81 199 18 32 95 217 1936 
 No. readmitted 12 13 10 79 (4.1% RR) 
CRG3                  
 No. patients in each test group 84 36 249 186 46 43 213 222 53 18 249 101 19 16 119 132 1786 
 No. readmitted 14 58 35 28 30 11 57 14 20 18 314 (17.6% RR) 
Test GroupTotal
12345678910111213141516
CRG1                  
 No. patients in each test group 138 153 296 535 124 116 328 674 67 117 263 406 51 79 234 422 4003 
 No. readmitted 10 27 12 29 14 13 142 (3.5% RR) 
CRG2                  
 No. patients in each test group 70 109 1491 310 46 67 185 303 17 38 81 199 18 32 95 217 1936 
 No. readmitted 12 13 10 79 (4.1% RR) 
CRG3                  
 No. patients in each test group 84 36 249 186 46 43 213 222 53 18 249 101 19 16 119 132 1786 
 No. readmitted 14 58 35 28 30 11 57 14 20 18 314 (17.6% RR) 

RRs varied by CRGs (CRG1 = 3.5%, CRG2 = 4.1%, and CRG3 = 17.6%). For information regarding the bundle elements included in each test group, see Fig 1.

The average effects and graphical methods of factorial analysis in Study-It are unweighted, treating each subgroup equally regardless of the number of patients. Because the number of patients in each of the 16 test groups in each analysis varied (Table 1), calculation of the important effects weighted by subgroup size was required to obtain the most accurate estimate. Therefore, for important individual bundle elements and element interactions identified by the factorial analysis, the original data were subgrouped, and the weighted RRs were calculated by using the raw data.

Contextual considerations and background variables were evaluated in the manner described above. In each CRG, this analysis considered hospital site (A, B, C, or D), preferred language (English or non- English), and medical insurance provider (Medicaid versus non-Medicaid) as background variables.

To avoid the influence of a single readmission on the results, data were considered sufficient for further analysis if the test group had a minimum of 10 patients. When the subgroup size is >10, a single readmission will change the RR percentage by <10% minimizing the impact of highly variable estimates.

Ethical Considerations

Central institutional review board approval was obtained from the primary investigator’s institution. Site-specific institutional review board approval or exemption was obtained.

Characteristics of 7725 patients included in the analysis are presented by site in Table 2.

TABLE 2

Characteristics of Study Sample by Hospital Site

SiteNo. PatientsMedian Age,a yFemale Sex, n (%)Preferred Language English, n (%)Medicaid, n (%)Readmitted Patients (RR), n (%)CRG1, n (%)CRG2, n (%)CRG3, n (%)
A: children’s hospital within larger hospital 3514 1.9 1612 (45.9) 3282 (93.4) 1923 (54.7) 250 (7.1) 1990 (56.6) 831 (23.7) 693 (19.7) 
B: freestanding children’s hospital 1710 2.4 761 (44.5) 1677 (98.1) 1077 (63) 58 (3.4) 883 (51.6) 714 (41.8) 113 (6.6) 
C: children’s hospital within larger hospital 1694 0.7 727 (42.9) 1500 (88.5) 699 (41.3) 65 (3.8) 1111 (65.6) 352 (20.8) 231 (13.6) 
D: freestanding children’s hospital 807 3.6 327 (40.5) 686 (85) 792 (98.1) 162 (20.1) 19 (2.4) 39 (4.8) 749 (92.8) 
SiteNo. PatientsMedian Age,a yFemale Sex, n (%)Preferred Language English, n (%)Medicaid, n (%)Readmitted Patients (RR), n (%)CRG1, n (%)CRG2, n (%)CRG3, n (%)
A: children’s hospital within larger hospital 3514 1.9 1612 (45.9) 3282 (93.4) 1923 (54.7) 250 (7.1) 1990 (56.6) 831 (23.7) 693 (19.7) 
B: freestanding children’s hospital 1710 2.4 761 (44.5) 1677 (98.1) 1077 (63) 58 (3.4) 883 (51.6) 714 (41.8) 113 (6.6) 
C: children’s hospital within larger hospital 1694 0.7 727 (42.9) 1500 (88.5) 699 (41.3) 65 (3.8) 1111 (65.6) 352 (20.8) 231 (13.6) 
D: freestanding children’s hospital 807 3.6 327 (40.5) 686 (85) 792 (98.1) 162 (20.1) 19 (2.4) 39 (4.8) 749 (92.8) 
a

Median age on initial hospitalization.

We performed RFA for each CRG. We will first discuss the results of the CRG3 RFA in detail, followed by a summary of the effects of the PACT bundle elements on RR in CRG1 and CRG2.

Step 1: Design Matrix Analysis

In CRG3, the 24 factorial design matrix provided unweighted effects of bundle elements and the interactions between each element. Important effects of the single element, teach-back, and 2 combinations of bundle elements (teach-back plus a checklist; handoff plus a phone call) on pediatric readmissions were noted. No important effects were noted with 3- or 4-factor combinations.

Steps 2 and 3: Graphical Analysis of Estimated Effects

The dot diagram (Fig 2) provided visual representation of the estimated effects obtained from the 24 design matrix. This diagram indicated that teach-back was a crucial factor (RR reduction by 6.9%), as well as the combination (interaction) of teach-back plus a checklist (RR reduction by 4.8%). This analysis also indicated a potential effect of the handoff plus a phone call interaction on RR (reduction by 4.1%). The response plots (Fig 3) captured the strength and the cause-and-effect relationship between bundle elements and RR, including teach-back and the combination of teach-back plus a checklist on RR. The response plots also revealed a strong interaction between bundle elements handoff plus a phone call as evidenced by crossing lines in Fig 3, but no clear cause-and-effect relationship with RR was noted. Therefore, this interaction was not analyzed further.

FIGURE 2

Dot diagram revealing important effects on reducing RR (left to right): teach-back (T), interaction effects T plus a checklist (C), and handoff (H) plus a phone call (P). Cluster of bundle elements around 0 represents “noise.”

FIGURE 2

Dot diagram revealing important effects on reducing RR (left to right): teach-back (T), interaction effects T plus a checklist (C), and handoff (H) plus a phone call (P). Cluster of bundle elements around 0 represents “noise.”

Close modal
FIGURE 3

Response plots revealing the average estimated effects of single bundle element teach-back (A) and the interaction effects of teach-back plus a checklist (B) and phone call plus handoff (C) on reducing RRs. Solid versus dotted lines and + versus − signs indicate bundle element use. Teach-back plus a checklist interaction had the greatest effect on reducing RRs.

FIGURE 3

Response plots revealing the average estimated effects of single bundle element teach-back (A) and the interaction effects of teach-back plus a checklist (B) and phone call plus handoff (C) on reducing RRs. Solid versus dotted lines and + versus − signs indicate bundle element use. Teach-back plus a checklist interaction had the greatest effect on reducing RRs.

Close modal

Step 4: Paired Comparison

Paired comparison plots revealed a consistent effect of teach-back on reducing RR across almost all test groups (Fig 4A). There was only 1 paired comparison (teach-back plus handoff, plus a phone call, minus checklist) in which teach-back had no effect on RR. In contrast, other bundle elements (handoff, phone call, checklist) revealed variable effect on RR by increasing or decreasing RR (eg, handoff, Fig 4B).

FIGURE 4

A, Consistent teach-back effect: paired comparisons show improvement between teach-back and 7 of the 8 combinations of the other three factors. B, Inconsistent handoff effect: handoff combined with the other factors increases RR for five of the other factors combinations, while reducing RR in three of the comparisons.

FIGURE 4

A, Consistent teach-back effect: paired comparisons show improvement between teach-back and 7 of the 8 combinations of the other three factors. B, Inconsistent handoff effect: handoff combined with the other factors increases RR for five of the other factors combinations, while reducing RR in three of the comparisons.

Close modal

Step 5: Estimation of Important Bundle Effects Weighted by Subgroup Size

To develop accurate estimates of the important effects identified by factorial analysis (steps 1–4), we calculated the weighted estimate effects of teach-back and the interaction between teach-back plus a checklist using the raw data. The weighted estimated effect of teach-back was a reduction of RR by 4.7% (the presence of teach-back [RR 14.9%] – the absence of teach-back [RR 19.6%]). The combined effect of teach-back plus a checklist resulted in a reduction of RR by 5.3% (the presence of both teach-back and checklist [RR 13.1%] – the absence of both teach-back and checklist [RR 18.4%]).

In CRG1 and CRG2, the checklist was the only effective bundle element in reducing RR (weighted effects 1.3% and 3.0%, respectively). Although these effects are smaller than those seen in CRG3, their relative importance remains high because of the much smaller baseline RR seen in these groups. In CRG1, the paired comparison revealed that in almost all pairs, the checklist consistently reduced RR, unlike that shown by handoff, teach-back, and phone call. In fact, when assessed in this manner, teach-back and handoff consistently increased RR.

In CRG2, the checklist was the single most effective bundle element in reducing RR (by 3.0%). Paired comparisons confirmed that the checklist alone consistently reduced RR, whereas all other bundle elements revealed an inconsistent effect. Furthermore, in paired comparisons, teach-back combined with the checklist mitigated the effect of the checklist.

Factorial analysis was used to analyze the effects of bundle elements on all-cause 30-day readmissions across background variables in each CRG. However, because of below minimum sample size requirement in some background subgroups (<10 patients in each test group), we were only able to evaluate effects of background subgroups Medicaid (CRG1 and CRG3), non-Medicaid insurance in CRG1, and English speaking in all CRGs. The performance of bundle elements within these background variables was consistent with the performance within each CRG except non-Medicaid (private and self-pay) in CRG1, for which none of the elements revealed reduction in RR.

We were not able to perform direct comparisons among associated background subgroups because of insufficient data.

Data, figures, and graphs for all analyses described here are available on request (S.N.O, unpublished observations).

In this RFA, we assessed the effect of individual and combined pediatric discharge bundle elements on 30-day all-cause, unplanned readmissions. To our knowledge, this is the first report in which the effectiveness of pediatric discharge bundle elements on reducing RR is described. First, we confirmed previous reports15,16  that RR varied across CRG. In medically complex patients, the use of teach-back alone reduced RR; however, the combination of teach-back plus a transition readiness checklist had the greatest effect on reduction. The checklist alone was the only bundle element or interaction that reduced RR in CRG1 and CRG2.

Our findings in CRG3 support the body of knowledge regarding the efficacy of teach-back as an evidence-based patient-centered communication tool.29,30  Qualitative research reveals that caregivers desire more instruction about signs and symptoms to look for after hospitalization and more hands-on experience in caring for children with medical complexity (CRG3).5053  To prevent readmissions in CRG3, teach-back addressed complex patient needs, such as medication management, feeding regimens, and/or technology-related instruction.26 

By contrast, in patients with nonchronic conditions (CRG1), use of teach-back and handoff to outpatient providers seemed to increase RR. In our study, teach-back for children in CRG1 primarily consisted of educating caregivers on contingency plans and reinforcing the importance and timing of follow-up appointments, whereas teach-back for patients in CRG3 also included advanced education about their disease, multiple medications, home nursing, and necessary equipment.26  Although teach-back can be a highly effective communication tool, it is plausible that increased attention to concerns at discharge for patients with simple conditions may heighten anxiety in caregivers while not equipping them to distinguish between worrisome and nonworrisome issues. Other plausible explanations include confounding by indication (the initial condition was not fully resolved before discharge) and overdiagnosis during follow-up visits leading to return to the hospital.5457  Therefore, the content and use of teach-back may need to be tailored for each CRG. This is particularly important for patients with single chronic conditions (CRG2), as confirming disease-specific knowledge before discharge is essential for these patients. It is plausible that the use of teach-back in this population alerted families to the appropriate need for readmission. Further study is needed to optimize the use of teach-back in these populations, realizing that RR may not be the appropriate metric to evaluate its usefulness.

For all groups, handoff ensured that PCPs received timely information. However, despite the obvious importance of clear and complete communication between inpatient and outpatient providers, handoff did not decrease RR in any of our study groups. Furthermore, the inconsistent effect of handoff on RR was evident in our paired comparisons. A review of literature revealed mixed effects of handoff on RR, including increases, in other studies as well.54  Selection bias was proposed as one plausible explanation because of heightened concern by inpatient providers about patient status or access to care prompting increased efforts at PCP handoff.54  Together, these data suggest that RR may not be the best way to evaluate the effect of handoff on patient outcomes.

The checklist was the only single bundle element that reduced RR in CRG1 and CRG2. The study checklist was used as a tool to facilitate care coordination, predominantly addressing logistic and social barriers.58  The checklist included basic information about the use of an interpreter, designated PCP and follow-up appointment, medications, equipment, and home nursing as indicated.26  Its use also reflected the completion of care coordination tasks by medical providers, additionally considering caregiver perceptions of discharge readiness, preferred language for teaching, assurance of prescription availability, and identification of transportation, financial, or social barriers.26  These items are known risk factors for readmission, and as our data suggest, they may be more important in preventing readmission for less medically complex children than confirmation of discharge education. The smaller size of the checklist effects (compared with effects for CRG3) are due to the lower average RR for patients in CRG1 and CRG2 (3.5% and 4.1%, respectively). As a percentage of these averages, both checklist effects are large (37% of CRG1 and 87% of CRG2). Additionally, the effect of the checklist was consistent, confirming its likely clinical relevance.

Interestingly, completion of the checklist alone for complex patients had no significant impact on RR but did augment the decrease in RR when paired with teach-back. It appears that the value of completing coordination tasks by medical providers is only realized when caregiver education is adequately ensured. We hypothesize that to achieve optimal outcomes for medically complex patients, rigorous attention to care coordination and caregiver education should be addressed simultaneously.

Regarding the lack of effect of the postdischarge telephone call on reducing RR, we identified low telephone connection rates as one of the major limiting factors. Individual site connection rates varied primarily in relation to the availability of a dedicated caller. Local context, such as allocation of hospital resources and follow-up phone call program prioritization, was reflective of connectivity success. In addition, a follow-up phone call is likely a safety tool that can be used to catch and correct misunderstanding but not necessarily reduce readmissions.59  It is possible that the effects of the postdischarge phone call may be better captured by assessing a different metric.

It was important to assess the consistency of the effects of bundle elements across subgroups of patients defined by background variables. The size of these subgroups within each CRG was only large enough to reliably assess background variables of English as a primary language and insurance of Medicaid. Bundle elements teach-back and checklist performed consistently across background variables mirroring the effects of these elements in each CRG as a whole, except for non-Medicaid. This was likely due to the predominance of these populations in the respective analyses. To fully assess the impact of background variables (including non-Medicaid), a larger sample size is needed.

To expand on this research, future iterations of the teach-back bundle element may include enhanced teach-back via simulation-based education and equipment-specific contingency plans to improve parent activation.60  This approach can aid in developing long-term family self-sufficiency by allowing caregivers to learn and practice new skills in a safe environment.60  In future studies, researchers may include measuring parent activation as a more proximal outcome measure.61,62  Additionally, as suggested in adult literature, telehealth interventions in combination with chronic care nursing management may be used to support patients who are medically complex.63 

This was a retrospective multifactorial analysis, and the number of patients varied in the 16 factorial groups. We were not able to randomize the factor combinations to patients3538  because several were already in use at different sites before the outset of the initiative. Secondly, using a previously published pediatric medical complexity algorithm, we manually classified patients into 3 CRGs, which might be subject to human error.49  However, investigators were experienced in this classification and equivocal cases were resolved by consensus. Thirdly, like previous studies that were underpowered to detect a change in readmissions, low RR in CRG1 and CRG2 presented challenges to data interpretation.20  Lastly, small numbers in subgroups of patients defined by background variables prevented full analysis of contextual factors.

Factorial analysis of the IMPACT database allowed for determination of which bundle element and combination of bundle elements worked the best in reducing RR. RRs and the effects of bundle elements varied across CRGs, indicating unique patient needs in each CRG for safe transition from the hospital to the home. Although, children with medical complexity (CRG3) account for a small percentage of the pediatric population, they are particularly vulnerable during the discharge transition, resulting in high RRs, poor health outcomes, and death.60  Our research indicates that teach-back in combination with checklist completion before hospital discharge may be important tools in preventing hospital readmissions within 30 days.

To allow investigators to learn more and learn faster with fewer resources, we propose that a prospective factorial design is uniquely suited to study these suggested interventions for medically complex patients. Because of the retrospective nature of our analysis, further study is needed to confirm our findings by using a prospective factorial design applied to other groups of patients and in different clinical settings.3538 ,45,64 

IMPACT Study Group

Department of Pediatrics, Weill Cornell Medicine and New York-Presbyterian Komansky Children’s Hospital:

Elisa Hampton, MD; Felicia Alleyne, MSN, RN; Lisa Schmutter, MPA, BSN, RN, CPN; Jennifer DiPace, MD; Jennie Ono, MD; Brooke Spector, MD; Cori Green, MD; Thanakorn Jirasevijinda, MD; Rizwana Popatia, MD; Mackenzi Preston, MD; Jessie Lee, MS; Amy Whiffen, MPH, CPXP; Sarah Smith, PharmD, BCPS; Elena-Mendez Rico, PharmD; Brianne Genow, MS, RN; Sara Powers, LCSW; Susan Costomiris, LCSW; Drisdy Key, LCSW; Jennifer Giannini, MSN, RN, NP-C, CNS; and Family Advisory Council members Courtney Nataraj, Jennifer Small, Kimberly LaRose, and Mariela Guerra.

Department of Pediatrics, Tufts University School of Medicine and The Barbara Bush Children’s Hospital:

Nicole Manchester, MSN, RN, CNL; Agatha Bellevue, RN, BS; Melanie Lord, BSN, RN, CPN; Teresa Morgan, RN; Nancy Bouthot, RN; Anna Martens, BA; Danielle DiCesare, BS; Clare Ronan, BA; Abihijit Bhattacharya, MD, PhD; Lorraine McElwain, MD; Jennifer Jewell, MD, MS; Jennifer Hayman, MD; Shannon Bennett, MD; Logan Murray, MD; Noah Diminick, MD; and Jonathan Bausman, DO.

Department of Pediatrics, Drexel University College of Medicine and St Christopher’s Hospital for Children:

Monica Kondrad, RN, BSN; Sharon Cray, BBA; William Woodall, MPH; Kayla Burley, MPH; Anna Marie Carr, MD; E. Douglass Thompson Jr, MD, MMM; Renee Turchi, MD, MPH; Francis McNesby, MD; Benjamin Sanders, MD; Nicola Brodie, MD; Daria Ferro, MD; Hannah Neubauer, MD; Carlyn Todorow, MD; Alana Cordeiro, MPH; Nathanael Koelper, MPH; Andrew Arnao, MPH; Alexis Runowski, MPH; Nadia Darova, MPH; Megan Mansfield, MPH; Anis Ansari, MPH; Samantha Serrao, MPH; Melissa Klass, MPH; Kristy Lightner, MPH; Saloni Mansuri, MPH; and Pooja Shah, MPH.

Department of Pediatrics, The Medical College of Wisconsin and Children’s Hospital of Wisconsin:

Patricia Lye, MD, MS; Emily Densmore, MD; Brittany Player, DO; Danita Hahn, MD; Jennifer Hadjiev, MD; Sarah Vepraskas, MD; Sara Lauck, MD; Jacquelyn Kuzminski, MD; Erica Chou, MD; Cori Gibson, MSN, RN, CNL; Kimberly Zimmanck, MS, RN; Amanda Quesnell, RN; Lori Smrz, RN, AE-C; Debbie Bakalarski, RRT; Khris O’Brien, RRT, NPS; Carla Settimi, BS, RRT; Casandra ZumMallen, MD; Alexis R. Baker, BS; Hana Millen, BS; Lisa Tello, RN; Kelsey Porada, BA, CRC; Steven Finkenbinder, PharmD, AE-C; and Peter O’Day, MD.

We acknowledge the following individuals: Kelsey Porada and Jessie Lee for their analytic and data management support and Family Advisory Council members, Courtney Nataraj, Jennifer Small, and Kimberly LaRose, for partnering with us in promoting patient- and family-centered care.

FUNDING: No external funding.

Dr Osorio conceptualized and designed the study, performed data collection and classification, conducted complete data analysis, drafted the initial manuscript, and revised the manuscript; Dr Gage participated in conceptualization and design of the study, data collection and classification, and refinement of the data collection, conducted initial analysis, and reviewed and revised the manuscript; Dr Mallory participated in conceptualization and design of the study and data collection and classification, conducted initial analysis, and reviewed and revised the manuscript; Drs Satty and Soung participated in acquisition of data, refined data collection tools, performed manual data classification, conducted initial analysis, and critically reviewed the manuscript; Dr Abramson performed data collection, participated in the initial drafting of the manuscript, and critically reviewed the manuscript for the important intellectual content; Mr Provost participated in conceptualization and design of the study, performed data analysis, and critically reviewed and revised the manuscript; Dr Cooperberg led conceptualization and design of the study, participated in manual data collection and classification, conducted initial analysis, and critically reviewed and revised the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

     
  • CRG

    clinical risk group

  •  
  • IMPACT

    Improving Pediatric Patient-Centered Care Transitions

  •  
  • PACT

    pediatric patient-centered care transition

  •  
  • PCP

    primary care provider

  •  
  • QI

    quality improvement

  •  
  • RFA

    retrospective factorial analysis

  •  
  • RR

    readmission rate

1
Coleman
EA
,
Parry
C
,
Chalmers
S
,
Min
SJ
.
The care transitions intervention: results of a randomized controlled trial
.
Arch Intern Med
.
2006
;
166
(
17
):
1822
1828
2
Jack
BW
,
Chetty
VK
,
Anthony
D
, et al
.
A reengineered hospital discharge program to decrease rehospitalization: a randomized trial
.
Ann Intern Med
.
2009
;
150
(
3
):
178
187
3
Voss
R
,
Gardner
R
,
Baier
R
,
Butterfield
K
,
Lehrman
S
,
Gravenstein
S
.
The care transitions intervention: translating from efficacy to effectiveness
.
Arch Intern Med
.
2011
;
171
(
14
):
1232
1237
4
Moore
C
,
Wisnivesky
J
,
Williams
S
,
McGinn
T
.
Medical errors related to discontinuity of care from an inpatient to an outpatient setting
.
J Gen Intern Med
.
2003
;
18
(
8
):
646
651
5
Tsilimingras
D
,
Miller
MR
,
Brooks
RG
.
Postdischarge adverse events in children: a cause for concern
.
Jt Comm J Qual Patient Saf
.
2009
;
35
(
12
):
620
621
6
Forster
AJ
,
Clark
HD
,
Menard
A
, et al
.
Adverse events among medical patients after discharge from hospital
.
CMAJ
.
2004
;
170
(
3
):
345
349
7
Forster
AJ
,
Murff
HJ
,
Peterson
JF
,
Gandhi
TK
,
Bates
DW
.
The incidence and severity of adverse events affecting patients after discharge from the hospital
.
Ann Intern Med
.
2003
;
138
(
3
):
161
167
8
Kripalani
S
,
Price
M
,
Vigil
V
,
Epstein
KR
.
Frequency and predictors of prescription-related issues after hospital discharge
.
J Hosp Med
.
2008
;
3
(
1
):
12
19
9
Kripalani
S
,
LeFevre
F
,
Phillips
CO
,
Williams
MV
,
Basaviah
P
,
Baker
DW
.
Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care
.
JAMA
.
2007
;
297
(
8
):
831
841
10
Harlan
GA
,
Nkoy
FL
,
Srivastava
R
, et al
.
Improving transitions of care at hospital discharge--implications for pediatric hospitalists and primary care providers
.
J Healthc Qual
.
2010
;
32
(
5
):
51
60
11
Roy
CL
,
Poon
EG
,
Karson
AS
, et al
.
Patient safety concerns arising from test results that return after hospital discharge
.
Ann Intern Med
.
2005
;
143
(
2
):
121
128
12
Moore
C
,
McGinn
T
,
Halm
E
.
Tying up loose ends: discharging patients with unresolved medical issues
.
Arch Intern Med
.
2007
;
167
(
12
):
1305
1311
13
Ruth
JL
,
Geskey
JM
,
Shaffer
ML
,
Bramley
HP
,
Paul
IM
.
Evaluating communication between pediatric primary care physicians and hospitalists
.
Clin Pediatr (Phila)
.
2011
;
50
(
10
):
923
928
14
Berry
JG
,
Toomey
SL
,
Zaslavsky
AM
, et al
.
Pediatric readmission prevalence and variability across hospitals. [published correction appears in JAMA. 2013;309(10):986]
.
JAMA
.
2013
;
309
(
4
):
372
380
15
Berry
JG
,
Hall
DE
,
Kuo
DZ
, et al
.
Hospital utilization and characteristics of patients experiencing recurrent readmissions within children’s hospitals
.
JAMA
.
2011
;
305
(
7
):
682
690
16
Cohen
E
,
Berry
JG
,
Camacho
X
,
Anderson
G
,
Wodchis
W
,
Guttmann
A
.
Patterns and costs of health care use of children with medical complexity
.
Pediatrics
.
2012
;
130
(
6
).
17
Berry
JG
,
Blaine
K
,
Rogers
J
, et al
.
A framework of pediatric hospital discharge care informed by legislation, research, and practice
.
JAMA Pediatr
.
2014
;
168
(
10
):
955
962
,
quiz 965–966
18
Institute for Healthcare Improvement
.
State Actions on Avoidable Rehospitalizations (STAAR) Initiative
.
Cambridge, MA
:
Institute for Healthcare Improvement
;
2012
. Available at: wwwi.ihi.org/staar
19
Kripalani
S
,
Jackson
AT
,
Schnipper
JL
,
Coleman
EA
.
Promoting effective transitions of care at hospital discharge: a review of key issues for hospitalists
.
J Hosp Med
.
2007
;
2
(
5
):
314
323
20
Auger
KA
,
Simon
TD
,
Cooperberg
D
, et al
.
Summary of STARNet: Seamless Transitions and (Re)admissions Network
.
Pediatrics
.
2015
;
135
(
1
):
164
175
21
Wu
S
,
Tyler
A
,
Logsdon
T
, et al
.
A quality improvement collaborative to improve the discharge process for hospitalized children
.
Pediatrics
.
2016
;
138
(
2
):
e20143604
22
Society of Hospital Medicine
.
BOOSTing care transitions resource room
.
2008
.
23
Solution for Patient Safety
.
Do no harm
.
24
Shen
M
,
Hershey
D
,
Bergert
L
, et al
.
Pediatric hospitalists collaborate to improve timeliness of discharge communication
.
Hosp Pediatr
.
2013
;
3
(
3
):
258
265
25
Coghlin
DT
,
Leyenaar
JK
,
Shen
M
, et al
.
Pediatric discharge content: a multisite assessment of physician preferences and experiences
.
Hosp Pediatr
.
2014
;
4
(
1
):
9
15
26
Mallory
LA
,
Osorio
SN
,
Prato
BS
, et al;
IMPACT Pilot Study Group
.
Project IMPACT pilot report: feasibility of implementing a hospital-to-home transition bundle. [published correction appears in Pediatrics. 2017;139(6):e20170816]
.
Pediatrics
.
2017
;
139
(
3
):
e20154626
27
Campbell
M
,
Fitzpatrick
R
,
Haines
A
, et al
.
Framework for design and evaluation of complex interventions to improve health
.
BMJ
.
2000
;
321
(
7262
):
694
696
28
Margolis
P
,
Provost
LP
,
Schoettker
PJ
,
Britto
MT
.
Quality improvement, clinical research, and quality improvement research--opportunities for integration
.
Pediatr Clin North Am
.
2009
;
56
(
4
):
831
841
29
Schillinger
D
,
Piette
J
,
Grumbach
K
, et al
.
Closing the loop: physician communication with diabetic patients who have low health literacy
.
Arch Intern Med
.
2003
;
163
(
1
):
83
90
30
Kornburger
C
,
Gibson
C
,
Sadowski
S
,
Maletta
K
,
Klingbeil
C
.
Using “teach-back” to promote a safe transition from hospital to home: an evidence-based approach to improving the discharge process
.
J Pediatr Nurs
.
2013
;
28
(
3
):
282
291
31
Braun
E
,
Baidusi
A
,
Alroy
G
,
Azzam
ZS
.
Telephone follow-up improves patients satisfaction following hospital discharge
.
Eur J Intern Med
.
2009
;
20
(
2
):
221
225
32
Auger
KA
,
Kenyon
CC
,
Feudtner
C
,
Davis
MM
.
Pediatric hospital discharge interventions to reduce subsequent utilization: a systematic review
.
J Hosp Med
.
2014
;
9
(
4
):
251
260
33
Balaban
RB
,
Weissman
JS
,
Samuel
PA
,
Woolhandler
S
.
Redefining and redesigning hospital discharge to enhance patient care: a randomized controlled study
.
J Gen Intern Med
.
2008
;
23
(
8
):
1228
1233
34
Dudas
V
,
Bookwalter
T
,
Kerr
KM
,
Pantilat
SZ
.
The impact of follow-up telephone calls to patients after hospitalization
.
Am J Med
.
2001
;
111
(
9B
):
26S
30S
35
Fisher
RA
.
The Design of Experiments
. 1st ed.
Edinburgh, Scotland
:
Oliver and Boyd
;
1935
36
Moen
RM
,
Nolan
TW
,
Provost
LP
.
Quality Improvement through Planned Experimentation
. 2nd ed.
New York, NY
:
McGraw Hill
;
1998
37
Provost
LP
.
Analytical studies: a framework for quality improvement design and analysis
.
BMJ Qual Saf
.
2011
;
20
(
suppl 1
):
i92
i96
38
Apfel
CC
,
Korttila
K
,
Abdalla
M
, et al
.
An international multicenter protocol to assess the single and combined benefits of antiemetic interventions in a controlled clinical trial of a 2x2x2x2x2x2 factorial design (IMPACT)
.
Control Clin Trials
.
2003
;
24
(
6
):
736
751
39
Korttila
K
,
Apfel
CC
.
Factorial design provides evidence to guide practice of anaesthesia
.
Acta Anaesthesiol Scand
.
2005
;
49
(
7
):
927
929
40
Zurovac
J
,
Moreno
L
,
Crosson
J
,
Brown
R
,
Schmitz
R
.
Using multifactorial experiments for comparative effectiveness research in physician practices with electronic health record
.
EGEMS (Wash DC)
.
2013
;
1
(
3
):
1037
41
McAlister
FA
,
Straus
SE
,
Sackett
DL
,
Altman
DG
.
Analysis and reporting of factorial trials: a systematic review
.
JAMA
.
2003
;
289
(
19
):
2545
2553
42
Reed
JL
,
Huppert
JS
,
Taylor
RG
, et al
.
Improving sexually transmitted infection results notification via mobile phone technology
.
J Adolesc Health
.
2014
;
55
(
5
):
690
697
43
Huppert
JS
,
Reed
JL
,
Munafo
JK
, et al
.
Improving notification of sexually transmitted infections: a quality improvement project and planned experiment
.
Pediatrics
.
2012
;
130
(
2
).
44
Piazza
AJ
,
Brozanski
B
,
Provost
L
, et al
.
SLUG bug: quality improvement with orchestrated testing leads to NICU CLABSI reduction
.
Pediatrics
.
2016
;
137
(
1
):
e20143642
45
Neuhauser
D
,
Provost
SM
,
Provost
LP
.
It is time to reconsider factorial designs: how Bradford Hill and R. A. Fisher shaped the standard of clinical evidence
.
Qual Manag Health Care
.
2020
;
29
(
2
):
109
122
46
Kleinman
LC
,
Dougherty
D
.
Assessing quality improvement in health care: theory for practice
.
Pediatrics
.
2013
;
131
(
suppl 1
):
S110
S119
47
McDonald
KM
.
Considering context in quality improvement interventions and implementation: concepts, frameworks, and application
.
Acad Pediatr
.
2013
;
13
(
6 suppl
):
S45
S53
48
Anderson
JB
,
Iyer
SB
,
Schidlow
DN
, et al;
National Pediatric Cardiology Quality Improvement Collaborative
.
Variation in growth of infants with a single ventricle
.
J Pediatr
.
2012
;
161
(
1
):
16
21.e1
,
quiz 21.e2–21.e3
49
Simon
TD
,
Cawthon
ML
,
Stanford
S
, et al;
Center of Excellence on Quality of Care Measures for Children with Complex Needs (COE4CCN) Medical Complexity Working Group
.
Pediatric medical complexity algorithm: a new method to stratify children by medical complexity
.
Pediatrics
.
2014
;
133
(
6
).
50
Solan
LG
,
Beck
AF
,
Brunswick
SA
, et al;
H2O Study Group
.
The family perspective on hospital to home transitions: a qualitative study
.
Pediatrics
.
2015
;
136
(
6
).
51
Leyenaar
JK
,
O’Brien
E
,
Leslie
LK
, et al
.
Families’ priorities regarding hospital-to-home transitions for children with medical complexity
.
Pediatrics
.
2017
;
139
(
1
):
e20161581
52
Desai
AD
,
Durkin
LK
,
Jacob-Files
EA
,
Mangione-Smith
R
.
Caregiver perceptions of hospital to home transitions according to medical complexity: a qualitative study
.
Acad Pediatr
.
2016
;
16
(
2
):
136
144
53
Nelson
BB
,
Coller
RJ
,
Saenz
AA
, et al
.
How avoidable are hospitalizations for children with medical complexity? Understanding parent perspectives
.
Acad Pediatr
.
2016
;
16
(
6
):
579
586
54
Coller
RJ
,
Klitzner
TS
,
Saenz
AA
, et al
.
Discharge handoff communication and pediatric readmissions
.
J Hosp Med
.
2017
;
12
(
1
):
29
35
55
Coon
ER
,
Quinonez
RA
,
Moyer
VA
,
Schroeder
AR
.
Overdiagnosis: how our compulsion for diagnosis may be harming children
.
Pediatrics
.
2014
;
134
(
5
):
1013
1023
56
Auger
KA
,
Simmons
JM
,
Tubbs-Cooley
HL
, et al;
H2O Trial study group
.
Postdischarge nurse home visits and reuse: the Hospital to Home Outcomes (H2O) trial. [published correction appears in Pediatrics. 2019;143(4):e20190092]
.
Pediatrics
.
2018
;
142
(
1
):
e20173919
57
Schroeder
AR
,
Destino
LA
,
Brooks
R
,
Wang
CJ
,
Coon
ER
.
Outcomes of follow-up visits after bronchiolitis hospitalizations
.
JAMA Pediatr
.
2018
;
172
(
3
):
296
297
58
Lax
Y
,
Martinez
M
,
Brown
NM
.
Social determinants of health and hospital readmission
.
Pediatrics
.
2017
;
140
(
5
):
e20171427
59
Bahr
SJ
,
Solverson
S
,
Schlidt
A
,
Hack
D
,
Smith
JL
,
Ryan
P
.
Integrated literature review of postdischarge telephone calls
.
West J Nurs Res
.
2014
;
36
(
1
):
84
104
60
Thrasher
J
,
Baker
J
,
Ventre
KM
, et al
.
Hospital to home: a quality improvement initiative to implement high-fidelity simulation training for caregivers of children requiring long-term mechanical ventilation
.
J Pediatr Nurs
.
2018
;
38
:
114
121
61
Ancker
JS
,
Osorio
SN
,
Cheriff
A
,
Cole
CL
,
Silver
M
,
Kaushal
R
.
Patient activation and use of an electronic patient portal
.
Inform Health Soc Care
2015
;
40
(
3
):
254
266
62
Barnert
ES
,
Coller
RJ
,
Nelson
BB
, et al
.
A healthy life for a child with medical complexity: 10 domains for conceptualizing health
.
Pediatrics
.
2018
;
142
(
3
):
e20180779
63
Coller
RJ
,
Nelson
BB
,
Sklansky
DJ
, et al
.
Preventing hospitalization in children with medical complexity: a systematic
.
Rev Pediatr
.
2014
;
134
:
1
18
64
Pallotto
EK
,
Chuo
J
,
Piazza
AJ
, et al
.
Orchestrated testing
.
Am J Med Qual
.
2017
;
32
(
1
):
87
92

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.