Targeted Visibility Analysis in Buildings Correlating Targeted Visibility Analysis with Distribution of People and Their Interactions within an Intensive Care Unit

This is a study of the spatial properties of environments that habitual users of a setting tune their behavior towards. The particular properties under consideration are those that affect visual fields. Previous studies in exhibition and office settings suggest that routine space use is affected by the structure of visual fields. In this study we test the proposition that the impact of spatial organization becomes clearer when we draw a distinction between generic visibility patterns and targeted visibility patterns. In studying generic visibility patterns we are considering all parts of a setting that are visible from each occupiable location. In studying targeted visibility patterns we focus on a previously specified set of visual targets and ask how many become visible from each occupiable location. A script was developed in order to compute targeted visibility measures using Depthmap, which normally computes generic visibility values. The Neurological Intensive Care Unit (ICU) at a large hospital in Atlanta was chosen as a test setting. Behavioral data were collected by a group of 5 observers over a period of 2 weeks in the 20-bed unit. Each observer recorded the location and activity of all people he/ she saw on the observation sheets while walking through the setting according to a pre-defined path. A total of 46 observation rounds were completed. 946 counts of people were observed in the setting including 320 doctors and 626 nurses. The targeted visibility measure, which represents number of pre-selected foci that are visible (i.e. patient beds in this case), is more strongly correlated with the density of all staff members compared with standard generic visibility. It is also more strongly correlated with the density of nurses and the density of interacting nurses. Generic visibility is more strongly correlated with both interacting and not-interacting doctors. The finding is all the more consistent because the above mentioned pattern remains unaffected in an additional analysis where all staff members engaging work surfaces were excluded. The result indicates people with different roles (nurses vs. doctors) are tuned to different features of environment. Nurses, especially when interacting, tend to position themselves where they can have high visual access to multiple patients. The distribution of doctors can be explained by the