Hospital-to-Home Interventions, Use, and Satisfaction: A Meta-analysis

Abstract

Context: Hospital-to-home transitions are critical opportunities to promote patient safety and high-quality care. However, such transitions are often fraught with difficulties associated with increased health care use and poor patient satisfaction.

Objective: In this review, we determine which pediatric hospital discharge interventions affect subsequent health care use or parental satisfaction compared with usual care.

Data sources: We searched 7 bibliographic databases and 5 pediatric journals.

Study selection: Inclusion criteria were: (1) available in English, (2) focused on children <18 years of age, (3) pediatric data reported separately from adult data, (4) not focused on normal newborns or pregnancy, (5) discharge intervention implemented in the inpatient setting, and (6) outcomes of health care use or caregiver satisfaction. Reviews, case studies, and commentaries were excluded.

Data extraction: Two reviewers independently abstracted data using modified Cochrane data collection forms and assessed quality using modified Downs and Black checklists.

Results: Seventy one articles met inclusion criteria. Although most interventions improved satisfaction, interventions variably reduced use. Interventions focused on follow-up care, discharge planning, teach back-based parental education, and contingency planning were associated with reduced use across patient groups. Bundled care coordination and family engagement interventions were associated with lower use in patients with chronic illnesses and neonates.

Limitations: Variability limited findings and reduced generalizability.

Conclusions: In this review, we highlight the utility of a pediatric discharge bundle in reducing health care use. Coordinating follow-up, discharge planning, teach back-based parental education, and contingency planning are potential foci for future efforts to improve hospital-to-home transitions.

FIGURE 1

PRISMA diagram showing article inclusion and exclusion.

FIGURE 2

Risk ratios for readmission by patient population and intervention type and forest plot for readmission by intervention type and patient group: RCT versus non-RCT, risk ratios, 95% CIs, and weight by patient population and intervention type for studies in which readmission is addressed. Risk ratios are indicated by the forest plots above, with risk ratios <1 favoring the intervention. Effect sizes were only calculated for those studies that provided sufficient information to calculate both the risk ratio and 95% CI. Nkoy et al performed separate analyses in 2 different patient populations (from a tertiary-care academic children’s hospital and a community hospital), and their study is therefore included twice. Similarly, because Brittan et al compared 2 groups of patients receiving follow-up in 2 different periods with those who did not receive follow-up, their study is also included twice (A and B). Pooled effect sizes are shown for each patient group (including patients with chronic illnesses, the neonatal population, and general patients who were hospitalized) by intervention type if effect sizes could be calculated for a minimum of 2 studies per group. ^a Studies that did not significantly impact readmission by this analysis but were reported as statistically significant by other statistical analysis methods not conducive to this meta-analysis.