Validation of the PHASE Measure for Admissions to an ICU and After Interfacility Transfer

The home-to-hospital transition before hospitalization is complex, involving clinician handoffs, changes in medical therapies, and stress to patients and families.1,2 This stress is intensified in admission to the ICU, where families are engaged in high-stakes decision-making and complex communication is occurring between clinicians.3 Similarly, interfacility transfers (IFT), where a child presents to one institution but requires transfer to a separate hospital for definitive care, carries stress of transport, long commutes for families, and increased risk of medical miscommunication.4,5 Understanding families’ experience during the admission process in these settings is paramount for optimizing patient outcomes in high-risk situations.The Pediatric Hospital Admission Survey of Experience (PHASE) measure is a 4-domain, 14-item quality measure that evaluates quality of care delivered at the time of admission.6 Although the PHASE measure has established validity evidence for pediatric admissions to general medical units in community and children’s hospitals, its use has not been validated in other hospital settings. Our aim was to psychometrically validate the PHASE measure for patients admitted to an ICU (neonatal or pediatric) or through IFT.We conducted a prospective, cross-sectional study from July 2021 to March 2023. Patients were eligible if admitted to an ICU (via the emergency department or IFT) or to the general medical or surgical ward after an IFT. One caregiver of eligible patients completed the survey within 72 hours from the time of admission if they were present during the admission process. [Redacted] Institutional Review Board approved this study.Caregivers completed the PHASE measure (Supplement 1), an instrument that measures the admission experience across 4 domains: (1) patient and family engagement, (2) information sharing, (3) effectiveness of care delivery, and (4) timeliness of care. Caregivers also completed a 6-question demographic survey (Supplement 2), including self-reported race and ethnicity, as caregivers may experience racial and ethnic disparities during the admissions process.7 Data were collected electronically via REDCap.8We completed a Confirmatory Factor Analysis (CFA) with polychoric correlations to test the 4-domain model of the PHASE measure.9,10 We then evaluated how well the model fit the data, calculating the goodness of fit index (study criterion ≥ 0.90), root mean squared error of approximation (study criterion ≤ 0.90), and the comparative fit index (study criterion ≥ 0.90). Domain scores were calculated using a top-box approach, ranging from 0 to 1, with 1 representing the most-favorable answer to all domain questions. CFA was conducted using R (version 4.2.2).Caregivers of 211 eligible patients completed the survey, 162 from caregivers of patients admitted to an ICU and 110 from caregivers of patients admitted after an IFT, including 52 patients admitted to the ICU via an IFT (Table 1). Comparative fit index resulted in confirmation of the 14-item measure with 4 domains; domains demonstrated acceptable correlations ranging 0.75 to 0.93 (Table 2). Indices of fit demonstrated robust fit using the goodness of fit (0.998), comparative fit index (0.997), and root mean squared error of approximation (0.057). Mean domain scores ranged from 0.65 (95% confidence interval 0.59–0.71) for timeliness of care to 0.87 (95% confidence interval 0.83–0.91) in information sharing (Table 3).The PHASE measure demonstrates reliability and validity in the measurement of caregiver experience during the hospital admission process to ICU settings and among patients admitted through IFT, building on previous validation work in general medical units.6Strengths of this analysis include robust methodology in the initial PHASE validation and in this confirmatory work. Limitations include that study participants were obtained through convenience sampling at a single center, which cannot adequately represent all experiences. The 72-hour sampling frame may have introduced recall bias for those surveyed later during hospitalization, particularly for those with high-acuity ICU admissions. Our hospital is not a birthing hospital, so neonatal ICU admissions present through the emergency department or IFT. These parents’ experiences may differ from those whose infants are admitted directly after delivery.To be consistent with the original PHASE measure psychometric analyses and to facilitate ease of interpretation, we applied top-box analytics. However, future applications of the PHASE measure could evaluate the distribution of scores, which may provide more granular data. When evaluating groups at high risk for healthcare disparities, this may identify inequities experienced during the admission process and inform focused efforts to reduce disparities.