ASSESSING THE COMBINED PREDICTIVE CAPABILITY OF SPATIAL AND ORGANIZATIONAL STRUCTURAL FACTORS ON PATIENT OUTCOMES IN INTENSIVE CARE UNITS: A SPACE SYNTAX EXPLORATION

Background: The spatial structure of healthcare environments plays a critical role in shaping caregiver behavior and influencing patient outcomes. Rooted in structuralist philosophy, Space Syntax proposes a methodology using spatial units and their connections to define the spatial structure of environments (called Integration of each space) and extends the idea to investigate its role in human interaction. In healthcare, research has shown that Space Syntax variables such as room accessibility and visibility can predict ICU outcomes, including patient mortality and length of stay. However, spatial structure alone cannot fully explain outcome variance in complex care environments. Organizational culture (i.e., shared values and beliefs shaping policies) also influences behavior. Since nurses’ time and movement in the unit are crucial indicators of organizational culture, this study integrates these dimensions into Space Syntax analysis to potentially better predict patient outcomes. Objective: The primary aim was to investigate how ICU architectural layout, combined with nurses’ time and movement, could predict patient outcomes more effectively than spatial structure alone. Methodology: This study developed three new spatio-cultural variables by reconstructing Space Syntax analysis (Time Weighted Integration Within: TWIW, Time Weighted Integration Between Model One: TWIB1, and Time Weighted Integration Between Model two:TWIB2) to include the proportion of nurses’ entries and the time spent in the patients’ room Space Syntax. Two units of the Medical Intensive Care Unit (MICU) and the Trauma Surgical Intensive Care Unit (TSICU) were chosen as the research setting to be studied over a month. Observational data on nurse movement and time were collected and paired with architectural plans to calculate the newly developed variables and run correlation analysis with patient outcome data, including length of stay, ventilator-free days, discharge outcomes, and mortality. Result: Correlation analyses revealed that the reconstructed variables (particularly TWIB2) demonstrated stronger predictive power than the original Hillier and Hanson Integration model. TWIB2, which could predict mortality in both MICU and TSICU, explained 40.3% of mortality variance in severely ill patients (APACHE II > 15) in MICU, outperforming Hillier and Hanson’s integration analysis (R2 = 20%). Meanwhile, TWIB1 and TWIW could explain the length of stay for each room more powerfully than Hillier and Hanson's Spatial analysis in MICU. These findings suggest that nurse behavior alters spatial depth significantly to predict outcomes, concluding that space is not only passively experienced but dynamically co-constructed through organizational processes. Implications: While this methodology needs to be re-studied mathematically and replicated to gain generalizability, this study could develop Space Syntax analysis for predicting patients’ outcomes in healthcare settings by reframing spatial structure with organizational culture indicators. It offers practical tools for evaluating ICU performance, aligning spatial design with real-time workflows, and guiding evidence-based design, particularly in times of rapid policy shifts such as during pandemics.