Patient satisfaction is an increasingly important aspect of health care and is influenced by multiple individual and environmental factors. The concept of environmental turbulence (ET) in clinical work environments has been described as instability and random change in the internal environment. For this study, ET is considered as a construct comprising: (1) Average Daily Census (ADC); (2) Admissions, Discharges, and Transfers (ADT); (3) Case Mix Index (CMI); and (4) Length of Stay (LOS). Little is known about the relationships among ET variables such as ADC, ADT, CMI, and LOS, or the relationship between these ET variables and patient satisfaction. Moreover, the study of patient satisfaction has not traditionally been informed by the principles of Complexity Science.

Study aims were to: (a) characterize the relationships among ADC, ADT, CMI, and LOS; and (b) examine ET variables as predictors of individual-level and unit-level patient satisfaction. This secondary data analysis utilized existing de-identified data of patients who were admitted to 15 medical and surgical patient care units at New York-Presbyterian/Columbia University Medical Center for the 36-month period (January 1, 2007 through December 31, 2009). Patient satisfaction was operationalized as the Press-Ganey total score. Study methods included Principal Components Analysis, Linear Regression, and Multilevel Analysis to determine the percent of variance in individual-level and unit-level patient satisfaction explained by ADC, ADT, CMI, and LOS and two composite variables resulting from the Principal Components Analysis: Clinical Stress (ADT, CMI, and LOS) and Capacity Stress (ADC).

All ET variables were significantly associated with patient satisfaction scores (n=12,197). The Linear Regression model containing the four ET variables and selected demographic characteristics explained 3% of the variance in the individual-level patient satisfaction. Multilevel models for each ET variable accounted for a significant amount of between unit variance in unit-level patient satisfaction: CMI - 6.39%, LOS - 11.78%, ADT - 12.66%, and ADC - 10.70%. This was also true for Clinical Stress. However, Capacity Stress decreased the amount of between unit variance explained in unit-level patient satisfaction.

A comprehensive understanding of the relationship between ET and patient satisfaction can inform the design of patient satisfaction interventions. Complexity Science offers an innovative perspective to understanding such relationships.