Measuring Variation in Staff Preferences for Respiratory Care Practices Within and Between Nicus

Background: Rates of bronchopulmonary dysplasia (BPD) have proven resistant to many quality improvement (QI) efforts, and wide variation exists in care practices and prevention strategies among neonatal intensive care units (NICUs). The psychology of change is an under-measured domain of improvement science, and may represent an opportunity to strengthen QI efforts. We sought to examine psychology by developing a tool to measure provider practice preferences for elements of NICU respiratory care.

Methods: The Neonatology Survey of Interdisciplinary Groups in Healthcare Tool (NSIGHT) is a vignette-based survey method to measure practice preferences on aspects of neonatal care. For each vignette, neonatologists, advanced practice providers (including neonatal nurse practitioners, physician assistants and neonatal hospitalists), registered nurses, and respiratory therapists are asked to indicate preference between two practice options, A or B. Eleven NICUs conducted the NSIGHT survey between 08/2019-04/2021. Two parameters were measured for each vignette, for each NICU: (1) net preference for practice option A or B and (2) agreement among staff, regardless of net preference. Staff responses for each vignette were assigned a numeric value from 1 (strong preference for option A) to 5 (strong preference for option B). Net preference for each NICU for each vignette was determined based on mean value among staff. Agreement was measured by standard deviation (SD) as an indication of distribution of spread of preferences, with SDs in the highest tercile considered low agreement, SDs in the middle tercile considered medium agreement, and SDs in the lowest tercile considered high agreement. The 14 vignettes are available at https://hms.az1.qualtrics.com/jfe/form/SV_0Io6Hp9QgjyGH7U and are included in abbreviated form in the figures.

Results: Figure 1 indicates net preference for all vignettes at the 11 NICUs. Vignettes 1, 3, and 6 most often yielded the same net practice preference between NICUs. In contrast, vignettes 8, 11, 12, and 13 most often yielded different net preferences between NICUs. Figure 2 indicates agreement level on practice for each vignette, within each of the 11 NICUs. NICUs varied widely in number of vignettes with high within-NICU agreement (median 5, range 2 – 8). Vignettes 4, 6, and 10 had high within-NICU agreement at most NICUs. Vignettes 1, 3, 7, and 14 had low within-NICU agreement at most NICUs.

Conclusion: Use of the NSIGHT survey method in 11 NICUs revealed wide variation in net preferences and level of agreement on respiratory care practices both within and between units. Practices without net preferences and those with low agreement within a NICU may represent opportunities for improvement; practices with differences in preferences and agreement between NICUs may represent opportunities for collaboration. Targeting unit culture and the psychology of change through practice preference surveys may inform and direct respiratory care quality improvement efforts.