480 - Construct Validation of the PediHome Measure of Pediatric Home Healthcare Quality

Publication Number: 1460.480

S. Margaret Wright, Children's Mercy Hospitals and Clinics, Kansas City, MO, United States; Emily Cramer, Children's Mercy Kansas City, Kansas City, MO, United States; Joanna Thomson, Cincinnati Children's, Cincinnati, OH, United States; Amanda Warniment, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, United States; Erica Stearns, University of Illinois Chicago, Anna, IL, United States; Stefanie Ames, University of Utah School of Medicine, Salt Lake City, UT, United States; Caitlin Koob, Medical University of South Carolina, Charleston, SC, United States; Carolyn Foster, Ann & Robert H. Lurie Children's Hospital of Chicago, Chicago, IL, United States

Background: There is a critical gap in validated measures of access to high quality home healthcare (HHC) for children. PediHome is a content-valid family-reported measure of HHC quality. PediHome domains include overall, relational, and clinical quality for both "nursing" and "aide" (certified nursing assistant/home health aide) care.

Objective: To assess the reliability and construct validity of the PediHome measure in a multi-site cohort of children with ≥1 complex chronic condition (CCC).

Design/Methods: We analyzed PediHome measure responses within a cross-sectional survey distributed among the 9-site Family CIRCLE longitudinal cohort composed of family caregivers of children with ≥1 CCC. Construct validity was evaluated with confirmatory factor analysis (CFA), while concurrent validity was assessed by correlations with overall quality ratings. Reliability of scales was assessed using McDonald's omega calculated from CFA estimates.

Results: Of the 165 caregivers who received HHC services in the past year and completed the PediHome survey, 42% received nursing care alone, 26% received aide care alone, and 32% received both (Table 1). The relational quality scales for both nursing (Table 2) and aide (Table 3) conform to a single factor structure. The nursing relational quality scale model fit is acceptable (CFI=0.92, SRMR=0.056), with high reliability (ω=0.924), and strong concurrent validity correlation (0.59, p< 0.001. The aide relational quality model demonstrated good fit (CFI=0.97, SRMR=0.014), very high reliability (ω=0.985), and strong concurrent validity (r=0.67, p< 0.001). The clinical quality scale for nursing is a 2-factor structure with acceptable overall fit (CFI=0.90, SRMR=0.036). The two factors include a task scale and monitoring scale. Both subscales as well as the total scale score have high reliability (Task: ω=0.956, Monitoring: ω=0.960, Total Scale: ω=0.978) and are strongly correlated with overall quality ratings suggesting good concurrent validity (Task: r=0.511, p< 0.001; Monitoring: r=0.482, p< 0.001; Total Scale: r=0.502, p< 0.001). The aide clinical quality scale conforms to a single-factor structure with overall good fit (CFI = 1.00, SRMR = 0.000), high reliability (ω=0.917), and moderate concurrent validity (r=0.316, p=0.012).

Conclusion(s): The confirmatory factor analysis provides support for a single factor structure for both nursing and aide relational quality, a single factor structure for aide clinical quality, and a 2-factor structure for nursing clinical quality. Future research will use a second sample for further validation and invariance testing.

Table 1. Association Between Study Population Characteristics and receipt of Home Healthcare Services
Nursing: private duty nurse, registered nurse or licensed practical nurse
Aide: certified nursing assistant/home health aide

Table 2. Nursing Relational and Clinical Quality


Table 3. Certified Nursing Assistant/Home Health Aide Relational and Clinical Quality