Factors Associated With Family Experience in Pediatric Inpatient Care

Abstract

Background: Hospitals are rapidly increasing efforts to improve the pediatric inpatient experience. However, hospitals often do not know what to target for improvement. To determine what matters most to families, we assessed which aspects of experience have the strongest relationships with parents' willingness to recommend a hospital.

Methods: Cross-sectional study of 17 727 surveys completed from November 2012 to January 2014 by parents of children hospitalized at 69 hospitals in 34 states using the Child Hospital Consumer Assessment of Healthcare Providers and Systems Survey. Hierarchical logistic regressions predicted the "top box" for willingness to recommend from measures of specific care dimensions (nurse-parent communication, doctor-parent communication, communication about medicines, keeping parents informed about the child's care, privacy with providers, preparing to leave the hospital, mistakes and concerns, child comfort, cleanliness, and quietness), adjusting for parent-child characteristics. Relative importance was assessed by using partially standardized adjusted odds ratios (aORs).

Results: Child comfort (aOR 1.50; 95% confidence interval 1.41-1.60) and nurse-parent communication (aOR 1.50; 95% confidence interval 1.42-1.58) showed the strongest relationships with willingness to recommend, followed by preparing to leave the hospital, doctor-parent communication, and keeping parents informed. Privacy and quietness were not significantly associated with willingness to recommend in multivariate analysis.

Conclusions: Our study uncovered highly valued dimensions that are distinct to pediatric care. Hospitals that care for children should consider using dedicated pediatric instruments to measure and track their performance. Improvement efforts should focus on creating an age-appropriate environment, improving the effectiveness of provider interactions, and engaging parents to share their values and concerns.

FIGURE 1

Bivariate and multivariate drivers of willingness to recommend. Bivariate and multivariate aORs per 1-SD change in predictor were used as measures of relative importance. Bivariate odds ratios were estimated by using separate hierarchical logistic regression models that predicted top-box recommendation from top-box scores of each composite experience measure, adjusting for child and/or parent characteristics (child age and health status; parent age, education, language, and relationship to the child), with hospital random effects. Multivariate odds ratios were similarly estimated with all experience measures entered simultaneously. Coefficient estimates of each composite experience measure were rescaled by SD to standardize comparisons.

FIGURE 2

Components of the 5 most important predictors of willingness to recommend. See Supplemental Fig 4 for a full list of Child HCAHPS Survey measures.

FIGURE 3

Variation of the 5 most important experience measures across hospitals. Intensity of color corresponds to the rank of each experience measure in predicting the top box for willingness to recommend at the individual level, as calculated by using the CAHPS Analysis Program 4.1. All scores were adjusted for child age, parent-reported child health status, parent age, parent relationship to the child, parent education, and parent’s preferred language. A hospital’s performance differs significantly from the mean if the 2-sided P value of t tests comparing the adjusted hospital mean and overall mean is <0.05.