key: cord-269099-q5nyzlhs authors: Lee, Jisun; Lee, Hyunsoo; McCuskey Shepley, Mardelle title: Exploring the spatial arrangement of patient rooms for minimum nurse travel in hospital nursing units in Korea date: 2020-07-23 journal: nan DOI: 10.1016/j.foar.2020.06.003 sha: doc_id: 269099 cord_uid: q5nyzlhs Abstract With increasing demands on medical care services, one of the trends is the mixed patient room arrangement of single/double-bed and multi-bed rooms in a nursing unit on the same floor. This influences nurse-to-patient assignment and often causes an unbalanced workload and longer travel distances for nurses. The objective of this study was to investigate how floor configuration and room density influence nurse travel in the hospital's medical surgical units in Korea. This study presented a novel approach to measure nurse travel distances in eight existing nursing units. The agent-based simulation was conducted to model nurses' walking trails, and the distance of one nurse travel to assigned patient rooms was measured for each nurse. With revisions in the spatial arrangement of patient rooms, locating multi-bed rooms near the nurse station, symmetric room layout centering the nurse station, and planning both single/double-bed and multi-bed rooms on one side of corridors, nurse travel distance decreased more than 15%. This study contributed to the knowledge of agent-based simulation as an evaluation framework for spatial analysis. Apart from application to Korea, these results are particularly of interest in countries where private patient rooms are not commonly economically feasible. Abstract With increasing demands on medical care services, one of the trends is the mixed patient room arrangement of single/double-bed and multi-bed rooms in a nursing unit on the same floor. This influences nurse-to-patient assignment and often causes an unbalanced workload and longer travel distances for nurses. The objective of this study was to investigate how floor configuration and room density influence nurse travel in the hospital's medical surgical units in Korea. This study presented a novel approach to measure nurse travel distances in eight existing nursing units. The agent-based simulation was conducted to model nurses' walking trails, and the distance of one nurse travel to assigned patient rooms was measured for each nurse. With revisions in the spatial arrangement of patient rooms, locating multibed rooms near the nurse station, symmetric room layout centering the nurse station, and planning both single/double-bed and multi-bed rooms on one side of corridors, nurse travel distance decreased more than 15%. This study contributed to the knowledge of agent-based simulation as an evaluation framework for spatial analysis. Apart from application to Korea, these results are particularly of interest in countries where private patient rooms are not commonly economically feasible. ยช 2020 Higher Education Press Limited Company. Publishing Services by Elsevier B.V. on behalf of KeAi. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). Nurses are one of the most valuable resources in the healthcare industry, and they are responsible for direct patient care. Nurses' roles and responsibilities are essential in meeting the high demands of patient care and advancing improved patient outcomes (Institute of Medicine, 2011). However, the healthcare industry is facing a growing shortage of nursing staff and a high staff turnover rate, while the high demand for quality care services persists (Health Resources and Services Administration, 2013) . The primary reason nurses leave their jobs is a highly demanding workload (Mazurenko et al., 2015; Tao et al., 2015) . Their work is physically and emotionally intense with heavy daily duties, high emotional exhaustion, and physical burnout, which results in nurse dissatisfaction with their jobs (Aiken et al., 2002; Faller et al., 2011) . Factors such as extended working hours, continuous working shifts, lack of rest breaks, and long walking distances contribute to nurse fatigue and stress Pati et al., 2008; Witkoski and Dickson, 2010) . Poor environments result in nurse stress and a reduced sense of wellbeing Ulrich et al., 2004 Ulrich et al., , 2008 Zimring et al., 2004) . In a nursing unit, long travel distance has been one of the most critical issues that impact nurse fatigue and stress (Chaudhury et al., 2009; Zborowsky et al., 2010) . Nurse travel is also directly related to the time spent on patient care. The literature indicates that nurses who spend more time walking spend less time at patients' bedsides Shepley and Davies, 2003; Trites et al., 1970; Ulrich et al., 2004) . Researchers have tried to evaluate how much walking distance can be shortened and how much time saved, which can potentially be directed to patient care (Gurascio-Howard and Malloch, 2007; Lu and Zimring, 2012; Pati et al., 2012 Pati et al., , 2015 . Improving the environment to support a well-balanced patient load and reduce walking distance is necessary for the nurses' well-being as well as secured time for patient care. This study examined the effects of spatial features on nurse walking distance (NWD), particularly focusing on the arrangement of patient rooms. The patient room locations need to facilitate effective access from the center of the ward, i.e., the nurse station, and other assigned patient rooms for minimum nurse travel (Nazarian et al., 2018) . Over the last two decades, trends in patient room types have moved towards the preference for single-occupancy patient rooms. Single-occupancy rooms have been considered to be more appropriate than multi-occupancy rooms because they enhance privacy, prevent infections, reduce noise, and incorporate space for families (Chaudhury et al., 2005 (Chaudhury et al., , 2006 . However, the majority of patient room types differ by country depending on the level of medical service demands and cultural and financial structures. Multi-occupancy rooms are still in the majority in many countries, including the UK, Germany, and Korea (Choi et al., 2017; Maben, 2009; Scott et al., 2012; Wagenaar and Mens, 2018) . In Korea, with high medical demands and limited land property, most general hospitals consist of a mix of single, double, and multi-bed rooms in one nursing unit. Experts report staff challenges with single room-only hospitals (Stephenson, 2015) , and argue for a flexible mix of room types for varied preferences and clinical goals (Pennington and Isles, 2013; Trant, 2010) Notably, in an epidemic condition like COVID-19, flexible patient room occupancy levels are needed. Healthcare facilities with single rooms only did not meet the demand, and patient beds were placed in corridors. In the mixed arrangement, both single/ double-bed and multi-bed rooms are assigned to each nurse for balanced workload (Choi et al., 2014 (Choi et al., , 2017 Shin and Kang, 2016) . In the racetrack type units where single/double-bed and multi-bed rooms are located on the opposite sides of the floor (with the service core in the middle), nurses have to travel along both sides to care assigned patients. However, studies on the relationship between patient room types and nurse travel distance are lacking. Nurses' movement patterns relate to nurse experience, patient care mix reflecting patient condition and patient load, and most of all, the location of assigned patient rooms (Choudhary et al., 2010; Heo et al., 2009; Nanda et al., 2015) . The distance between assigned rooms and patient room to patient room circulation are major contributors to nurse walking (Pati et al., 2008) . In this paper, an 'assignment' refers to a set of rooms assigned to a nurse. Current healthcare models represent proximity between spaces, typically reflecting a simple path for an effective sequence of activities. This is adequate in cases where the patient and staff movement are explicitly determined, such as the outpatient unit, emergency department, and operation room. However, for predicting and simulating movement in inpatient units, the simple approach of point-to-point distance is limited, because nurses move continuously during a shift, rarely following the same pattern. In this study, a novel approach of agent-based modeling is applied as a process-oriented analysis methodology to reflect a sequence of nurse activities. This study evaluates the effect of different layouts of hospital nursing units on nurse movement patterns, and also contributes to the field of agent-based simulations. In this study, a simulation-specific approach is taken to provide strong evidence to healthcare providers regarding the location of patient rooms in a nursing unit. 2. Literature review 2.1. Spatial features, nurse-patient assignment, and nurse travel distance The spatial relation between the nurse station and patient rooms is one of the most critical elements influencing nurse travel (Cai and Zimring, 2012; Gurascio-Howard and Malloch, 2007) . The literature reports that the most traveled paths of nurses are between the nurse station and patient rooms (Acar and Butt, 2016; Nazarian et al., 2018) . In an observational study, the travel between the nurse station and patient rooms took 34.5% of total nurses' travel during a 12-h shift (Acar and Butt, 2016 ). An individual nurse walks an average of 2.4e3.4 miles per daytime shift Shepley and Davies, 2003) , and this distance can be as much as 6 miles (Pati et al., 2012) . described how nurse travel distance, time with patients, and workload varied between different unit layouts in 36 medical-surgical units: 19 racetrack type units, 12 double-loaded units, single corridor, and five radial type units. It was obvious that the travel distance and patient time among nurses varied more within the same unit than between units. They found no statistically significant correlation between unit type, travel distance, and patient time; however, the difference in nurse travel distance between spatially contiguous patient assignment and non-contiguous room assignment was significant . Nurse managers and directors even walk longer than the average nurses walk, and long walking distances are a challenge for care coordination and overall management (Nanda et al., 2015) . The variability of distance traveled between individual nurses on the same unit is often greater than the variance across different hospital units . One of the causes of the large variability in distances traveled among individual nurses on the same unit is the spatial properties of room assignments. The associated spatial properties affect nurse travel patterns and distance traveled. Nurses travel between assigned patient rooms, nurse stations, and supply rooms (Butt et al., 2004) . The nurse-to-patient assignment involves a time-consuming and complicated process with many considerations, such as patient acuity, patient preference, continuity of care, and nurse experience. The patient's location in the unit is one of the primary factors to consider for patient assignments and often more critical than nurse experience and expertise (van Oostveen et al., 2014) . Patient rooms that lie outside the main circulation of a nursing unit contribute to added walking distance to and from the patient rooms to nurses' areas (Pati et al., 2008) , and patient assignments in two corridors result in additional walking (Gurascio-Howard and Malloch, 2007) . The travel from the center of the ward, i.e., the nurse station, and other assigned patient rooms impacts walking distance (Nazarian et al., 2018) . In nursing units with a flexible mix of room types for varied preferences and clinical goals, the arrangement of room types is one of the causes for scattered patient assignments. Even though patient assignments are complicated, for balanced workload, assigning both single/ double-bed and multi-bed rooms is prevalent in Korea where most hospitals consist of a mix of room types within the same unit (Choi et al., 2014 (Choi et al., , 2017 Shin and Kang, 2016) . Patient room assignments are a basis for nurse travel, and a nurse is required to establish movement pattern strategies according to the spatial relationship between patient rooms and nurses' areas (Heo et al., 2009) . Also, the number of visits to patients is correlated with spatial properties (Choudhary et al., 2010; Heo et al., 2009 ). Agent-based simulations have been widely applied to simulate pedestrian walking behaviors in architecture from a building to an urban scale. They are often employed to generate realistic and autonomous pedestrian walking behavior. Using simple rules to simulate the decisionmaking processes while interacting with environments, agent-based models perform high levels of correlation to human behavior. Various successful models, such as the cellular automata (Dijkstra and Timmermans, 2002) , social force (Helbing and Molnar, 1995) , boids-type behavior (Reynold, 1987) , shortest path (Hoogendoorn and Bovy, 2004) , and exosomatic visual architecture (Turner and Penn, 2002) have been described in the literature. Previous case studies find strong correlations between agent-based models with human pedestrian behavior in comparison to real-world observations and computing movement models (Batty, 2001; Helbing et al., 2002; Turner and Penn, 2002; Turner, 2007; Torrens, 2012; Yang et al., 2011; Vizzari et al., 2015) . Due to the similarity of agentbased models and human movement behavior, evacuation modeling and wayfinding are the most common consolidated areas in which agent-based models are employed (Najian and Dean, 2017; Orellana and Alsayed, 2013; Raubal, 2001a Raubal, , 2001b Vizzari et al., 2020) . Although many researchers have conducted agent-based models using aggregated data, some researchers focus on individual human movement level. Orellana and Alsayed (2013) carried out research covering on-site observations and virtual simulations of pedestrian walking and demonstrated a strong relation between observed and virtual pedestrian movement at an individual level. Traditionally, in nursing unit studies, walking distances have been estimated based on point-topoint linear proximity between key areas Shepley and Davies, 2003; Pati et al., 2012; Yi and Seo, 2012) . Among Korean inpatient unit case studies, Shin and Kang (2016) assessed nurse walking distances using traditional point-to-point linear measurement based on field interviews of nurses' patient room assignments. However, human walking behavior typically does not follow the centerline and makes a curve at corners to move along the shortest path to a destination point (Helbing et al., 2001) . Agent-based modeling can accommodate multiple path options of every nurse and, at the same time, reflect more realistic human behavior patterns. Most studies on nurse travel assessment are based on field observations (e.g., Choudhary et al., 2010; Seo et al., 2011; Shepley and Davies, 2003; Shin and Kang, 2016; Sturdavant, 1960; Yi and Seo, 2012) , and few employ simulation-based experiments (Nanda et al., 2015; Pati et al., 2015) . For visualizing and measuring walking distances within buildings using a computational method, Dym et al. (1988) suggested travel-distance algorithms that examine a hospital floor plan via defining the shortest path between each wall's midpoints. Lee et al. (2010) employed a metric graph structure to represent building circulation and calculated walking distances with the shortest path and interpreted the building shape with a buffered spaceboundary polygon reflecting half the width of a person's shoulder from a wall as minimum buffer distance. Nanda et al. (2015) combined field research and spatial parametric modeling tools in their study on assessing nurse walking distance in a medical-surgical unit. A Rhino/ The spatial arrangement of patient rooms for minimum nurse travel 3 + MODEL Grasshopper model was developed to assess nurse walking distance from every patient room to the support areas. Observation studies on nurse walking distance reveal the complexity of their travel. The variations between the sequence of activities, frequencies of visits, unexpected interruptions for minor tasks during patient visits are variable. However, the literature agrees that the most frequently visited areas are patient rooms (Acar and Butt, 2016; Nanda et al., 2015; Nazarian et al., 2018; Pati et al., 2012) and the proximity between the nurse station and the patient room is critical (Cai and Zimring, 2012; Gurascio-Howard and Malloch, 2007) . Also, studies predicting and computing nurses' movement suggest that nurse walking routes are more relevant when based on a sequence of activities rather than simple proximity measures (Choudhary et al., 2010; Heo et al., 2009; Nanda et al., 2015; Shin and Kang, 2016) . A novel spatial analysis tool was implemented to visualize nurse walking behavior and spatial data on the current environment of hospital inpatient units. Using a parametric model in Rhino/Grasshopper, an agent-based simulation was conducted to assess each nurse's walking distance. The agentbased simulation generated virtual nurses' walking trails, and the parametric algorithm measured the distance (NWD) of each walking trail. A visualized nurse walking trail reflected one patient visit "round" of each nurse, which was regularly taken for activities such as doctor visits, shift handover, or medication during a shift. Those were paths from the nurse station toward assigned patient rooms, stopping by a clean supply/medication room, and returning to the nurse station. The selected cases included typical medical-surgical inpatient units of eight general hospitals (Fig. 4) . This article represented partial findings of a doctoral dissertation (Lee, 2019) , which applied mixed methods of data collection: literature review, case studies, and simulations. In the agent-based simulation, virtual agents represent nurses. Nurses' walking trails were generated and measured in a Rhino modeling space of each unit. The Grasshopper plug-in program "PedSim" is used to model nurse walking behavior. PedSim has been used to simulate walking paths in healthcare facilities for evaluating accumulated walking trails of patients (Lee and Lee, 2020) . Fig. 1 demonstrates how the possible routes where agents can move within a given environment are analyzed in PedSim. In the simulation, the building components, such as walls and fixed furniture, are set as obstacles, i.e., a series of closed polygons that form a space boundary. The edges and vertices are automatically generated at a buffered distance from the building polygons, and the edges become the possible routes that agents can choose. The buffered distance is a half-width of a shoulder and reflects the distance humans offset from walls to avoid collisions while walking in indoor space. PedSim also employs multiple forces for agents' walking behavior. It is based on the social force model (Helbing and Molnar, 1995) with agents driven by multiple forces, namely a target force, person repulsion, and obstacle repulsion for their basic movement mechanism, and an anticipatory collision avoidance force to make agents aware of potential collisions and take actions earlier (Wang, 2019) . In the simulation, an agent chooses the shortest path among the possible routes within the given environment and moves to a target point, avoiding collisions with multiple forces (Wang, 2019) . In the agent-based modeling, nurse travel is set with a sequence of movements per patient assignment. The movement sequence is programmed by a set of parameters: visit locations, number of visits, and a flow of the assignment. The methodology is process oriented. When the movement path is visualized on the layout following the key sequence of movement, the distance is measured automatically by the parametric components. The agents also Fig. 1 Example of the environment setting. Agents move from starting to destination points, and edges and vertices of space boundary at a buffered distance from building polygons are possible routes that agents can choose (Wang, 2019) . have a vision, and they move to the target point in possible visible routes. This study simulates nurse travel based on the most obvious sequence of visits during a shift, which is a patient visit round. Each nurse movement path reflects a set of patient assignments. Four or five different patient rooms, which are assigned to one nurse, are visited in each round, and four or five sets of assignments cover all the unit's rooms. The simulation process (Fig. 2) for a patient visit round is as follows: (1) the agent finds the shortest path to the destination point, (2) it visits points of interest if it sees those and avoids obstacles and other agents on its way, and (3) goes to the destination point. In the simulation, an agent, which is a nurse, (1) travels from a starting point: the nurse station, (2) visits points of interest: the supply/ medication room and assigned patient rooms of each nurse, and (3) goes to a destination point: the nurse station, avoiding obstacles such as walls and furniture in the plans (Fig. 3) . The typical double-loaded corridor type nursing unit was the standard design for many years from the early design of hospital wards, because of the need for cross-ventilation and natural lighting. With an increase in the demands for hospitals, planners have tried various plan configurations to achieve efficient activity patterns. The more compact plans, such as radial (Valley Presbyterian Hospital in California and Brigham And Women's Hospital in Boston), Lshaped (Aspirus Wausau Hospital in Wisconsin), square (Providence Hospital in Alaska), and triangular-shaped (Jacobs Medical Center in San Diego) nursing units have been developed with groupings of concentric support areas in inpatient units (Kobus et al., 2008) . Valley Presbyterian Hospital in California (1956) developed the first compact radial type unit, and recently Brigham And Women's Hospital in Boston developed radial intensive care units. Recent trends have shown a shift toward the racetrack type. Racetrack type design maximizes the perimeter wall of the unit while providing moderate visibility and accessibility from the nurse working area to patient rooms. Another trend is toward pod configuration, which provides better patient monitoring opportunities (Cama, 2009; Hamilton and Shepley, 2010; Thompson et al., 2012) . However, in Korea, with its enormous scale and high land prices, high-rise hospitals with a racetrack or triangular shape units are the most common. This study included four rectangular and one triangular-shaped inpatient unit layouts of the most highly ranked hospitals among 42 tertiary care hospitals in Korea (Ministry of Health and Welfare, 2018) and three representative triangular layouts from secondary care The spatial arrangement of patient rooms for minimum nurse travel 5 + MODEL hospitals (Fig. 4) . Among the selected medical-surgical inpatient units of eight hospitals, four are in Seoul and four in other cities. The typical floor plans of the units follow the criteria: 40 to 50 beds per unit, 500 to 1000 inpatient beds in total, and being built after 2000. Each ward floor has two nursing units and two centralized nurse stations, two medication and clean supply rooms, one nourishment room, two soiled linen rooms, and two equipment rooms at the center of each unit. Each nursing unit had a double-corridor design with a triangular or rectangular shape, and mixed arrangement of patient room types: single, double, or multi-occupancy (Tables 1 and 2 ). The characteristics of the units were categorized by (1) unit shape -triangular or rectangular and (2) location of the nurse station -near the elevator core or center of the unit. A nursing unit refers to the number of beds the nurses in a group are accountable to take care of, and the units of general hospitals in Korea have 40e60 beds per nursing unit due to a large number of patients. The staff to patient ratio ranges from 1:7 to 1:12 for registered nurses (National Health Insurance Service, 2019). The nurse-patient assignments of selected units are based on Shin and Kang's (2016) research on the nurse-patient assignments. The nurse-room ratio of the eight units was set with respect to three criteria: (a) nurse-patient ratio was 10e12 patients per nurse, (b) both single/double-bed patient rooms and multibed patient rooms were assigned to one nurse for workload balance, and (c) all patients in a single room assigned to the same nurse (Table 3) . To compare the nurse travel distance between units, one patient visit "round" of each nurse was measured in the simulation. In the delivery of patient care and related activities, the locations nurses visit may vary based on urgent medication and patient requests. According to empirical studies on the traveled paths of nurses (Acar and Butt, 2016; Kim and Chai, 2018; Nanda et al., 2015) , the most frequently visited paths during both day and night shifts are between: (i) patient rooms and nurse station; (ii) nurse station and clean supply/medication room; and (iii) patient rooms and clean supply/medication room, in this order. In this study, the simulation setting was limited to one patient visit round of each nurse to the assigned patient room visits at each shift change, excluding irregular travel. The distance to the patient room doors was calculated, and the movement inside the patient room was excluded from the measured distance. The movement process for a round is shown in Fig. 5 : (1) Starting from the centralized nurse station, a nurse (an agent) visits a clean supply/medication room for preparation prior to an assigned patient room round. (2) The nurse (agent) moves to the assigned patient rooms. (3) After visiting the assigned patient rooms, the nurse returns to the nurse station. Nurse walking trails for one round of each nurse (N Z 33) were simulated, and the distance of each round path was measured ( Fig. 6 and Table 4 ). The shortest average NWD of each unit was measured in Unit 4, and the longest in Unit 7 Unit Nurse-patient assignment Total The spatial arrangement of patient rooms for minimum nurse travel 7 + MODEL (41.56 and 68.97). Units 1, 2, and 4 showed relatively short average NWD. The shortest NWD of one nurse was captured in N-1 of Unit 2, and the longest in N-1 of Units 8 (26.06 and 82.51). The variations of NWD among nurses were in a broader range within a layout than between other units. Considering nurses travel between the nurse station and assigned patient rooms approximately fourteen times during a shift (Kim and Chai, 2018) , the variations will be even more substantial. Unit 4 shows the most equivalent distances among nurses within a unit. Units 3 and 7 show lower SD than the average; however, Unit 7 had the longest average NWD, which indicated that nurses in Unit 7 walked more than nurses in other units. The NWD in Unit 6 varied in the largest differences among nurses, and the variations of NWD among nurses in Units 1, 2 were also larger than other units. Units 7 and 8 had long average NWD, and variations of NWD within the units were at the average level. Although the triangular units showed shorter NWD than the rectangular units, unit typology may not be the most fundamental reason for the short walking distance. The simulation results presented strong evidence that the average NWD was shorter in the units with a nurse station located at the center of the unit (Table 5) , and the locations of nurse stations were closer to the center of the units in the triangular units (Fig. 6) . Therefore, the distance to the patient rooms from the nurse station was shorter in the triangular units than the rectangular units. In rectangular units, the nurse station was closer to the unit entrance rather than the center of the units. With a categorized comparison of two groups of 'contiguous' and 'non-contiguous' patient room assignments, the average NWD was 46.34 and 59.98 m (Table 5) . This study defined the nurse-patient assignments between immediately adjacent patient rooms as "contiguous", and with nearby but not contiguously neighboring patient rooms as "non-contiguous". Obviously, the units with geographically contiguous patient room assignments had shorter NWD than the units with non-contiguous patient room assignments. Unit 3, a triangular unit with a non-contiguous patient room assignment, had a longer NWD than other triangular units. Unit 6 was the only rectangular unit with a contiguous patient assignment, and it had a shorter NWD than other rectangular units. Unit 6 had the lounges located in the middle of the patient room sections, which increased the distance between the nurse station and patient rooms. The simulation results revealed that these measures did not consequently result in longer NWD. The units of the long distances from nurse station to end of the unit (NS to EU) were Units 1, 6, and 8, and of the long perimeter (length of patient room walls on door side) were Units 3, 6, and 8, compared to the average (Fig. 7 and Table 4 ). However, Unit 1 showed a short NWD with a long NS to EU distance, while Unit 7 showed the longest NWD with one of the shortest NS to EU distance. The average ranges of NWD were captured in Units 3 and 6 despite the long perimeter length. Unit 6 showed an average range of NWD with a long NS to EU distance and perimeter length. One of the critical spatial attributes influencing nurse travel distance was the patient room arrangement. In the cases where single/double-bed patient rooms and multibed rooms were planned in the same unit, an equivalent number of single/double-bed and multi-bed patient rooms were to be assigned to each nurse for balanced workload distribution. In Units 5, 7, and 8, where single/double-bed and multi-bed rooms were located in separate two hallways, the nurses had to move up and down across the center support areas to visit patients in two corridors. These geographically non-contiguous patient assignments caused long travel distances. Among the rectangular units, Unit 6 presented an average level of NWD despite the lounges located in between patient rooms, the long perimeter of patient rooms, and long NS to EU distance. Unit 6 had the patient room arrangement, which supported a contiguous patient assignment plan, locating a mixed arrangement of patient rooms along the same corridor, as nurses get assigned for patient rooms of both single/ double-bed and multi-bed rooms. However, in Unit 3, even though both room types were located along the same corridor, they were placed in distant locations. Consequently, Unit 3 presented non-contiguous patient assignments with longer NWD than others among the triangular units. The patient room arrangement contributed to the geographical continuity of the nurse-patient assignment, and non-contiguous patient assignments often caused inefficient travel distances for nurses. Therefore, the spatial arrangement of patient room types was one of the critical features for nurse walking efficiency, as this was highly related to the patient assignment plan. The frequency of nurse visits to patient rooms from the nurse station was higher in double-bed rooms than singlebed rooms (Nazarian et al., 2018) , which means locating multi-bed patient rooms closer to the nurse station would reduce nurse walking. To explore the effects of patient room arrangement on NWD, the room arrangements of three rectangular units, Units 5, 7, and 8, were revised to support geographically contiguous nurse-patient room assignment through the following modifications: The spatial arrangement of patient rooms for minimum nurse travel 9 + MODEL 1) Symmetric room layout centering the nurse station, 2) Mixed planning of single/double-bed and multi-bed patient rooms on one side of the corridor, and 3) Multi-bed rooms as near as possible to the main nurse station to shorten nurse travel. In Unit 5, two multi-bed patient rooms on the upper corridor were moved to the lower corridor for nurses to care for patients in a single corridor. In Units 7 and 8, similar changes to the layouts were made. One patient visit round was simulated for each nurse, and the travel distance was measured (Tables 6 and 7) . A notable reduction in NWD was evident for all three units, with around 15% decrease in the revised layouts. Of the three units, Unit 7 had the largest difference. In Unit 7, the average NWD decreased from 68.97 to 55.95 m for one round, with an 18.88% reduction in distance. Also, SD decreased from 10.21 to 3.84, which is a noticeable improvement for equivalent NWD among nurses within Unit 7. In Units 5 and 8, the reduction was also evident: the averages of NWD decreased by 14.64% and 16.31%, respectively. However, SD slightly increased as the shortest NWD further decreased. Considering this was based on a single trip, the reduction of NWD would be more substantial with multiple trips in real situations. Based on the results, mixed planning of single/ double-bed rooms with multi-bed rooms on one side of the corridor with consideration of contiguous nurse-patient room assignment influenced shorter nurses' travel distance. . 7 Comparisons of NWD of each unit: average, longest, shortest NWD, and SD measures. The x-axis represents the unit number and y-axis the distance in meters. Note. (DA: the difference between the average NWD of the existing and revised layouts). The results of this study indicated that changes in spatial features, such as nurse station location, patient room arrangement, affect nurses' walking distances, and the size of the units, may not merely increase nurse travel with supportive spatial planning. Also, this study discussed the relationship between nurse-patient room assignment and nurse travel distance. Nurse travel is based on the patient assignment during a shift, and the assignment frequently involves geographically non-contiguous patient rooms (Pati et al., 2008) . Non-contiguous patient room assignments often increase walking for nurses . Although it is assumed that nurses are assigned to a set of contiguous patient rooms, actual patient assignments vary in regard to other factors, such as the acuity level of a patient, competency of a nurse, and a nurse-to-patient ratio (Pati et al., 2008) . This study emphasized that patient assignment is one of the critical factors to consider in the design stage for reducing nurses' walking, even though it is an operational issue and not a spatial issue. With an understanding that the actual patient assignment may vary and not result in contiguous rooms, depending on the balance between criteria, this study explored the optimal layout that healthcare designers could provide to support contiguous room assignments, initially. In a mix of single/double and multi-bed patient rooms in one nursing unit, this study demonstrated two spatial factors impacting minimum nurse travel distance: 1) locating multi-bed patient rooms near the nurse station and 2) planning both single/double-bed and multi-bed rooms on one side of corridors instead of separating single/doublebed and multi-bed patient rooms on two different corridors. The critical design factors in a nursing unit design are space layout that supports efficient nursing activities, reduced walking distances, organized supply areas, controlled noise level, and visibility with ease of supervision (Zborowsky et al., 2010) . This study has focused on the issue of walking distances, and the results demonstrated that planning multi-occupancy patient rooms near the nurse station were more effective in reducing nurse travel distance. This approach can also impact patient visibility and noise level. More patients are easily visible and accessible from the nurse station when locating multi-bed patient rooms near the nurse station (Lee, 2019) . A high noise level is often reported around the nurse station, and patients who prefer single-occupancy rooms would benefit from being at a quieter location than around the nurse station. The nurse station is a central hub for nursing activities in a hospital unit and the primary work area of each unit. The location and distance of the nurse station to the patient room has been considered as a critical influence on nurse walking distance. Planning the nurse station closer to the center of the units is a convincing design strategy to keep nurse travel distance to a minimum, especially in a linear and less concentric rectangular unit. However, one of the reasons for positioning the nurse station close to the entrance would be the high need for visual control over the entrance. Even though it is out of the scope of this study, planning both the entrance and nurse station at the center could be a robust design strategy to keep nurse travel distance to a minimum and gain sufficient visibility to the entrance. Among the cases of this study, Unit 5 has the nurse station close to both entrance and center of the unit. However, patient visibility was low due to the orientation of the nurse station, which was facing the entrance rather than patient rooms. In Unit 5, despite the long perimeter length, the NWD was at the average level. It could be a good strategy to develop this unit into a hybrid nursing station model to improve patient visibility. It will also satisfy visual control to the entrance and patients, as well as achieving moderate NWD. Representing the process of how nurses' movements are made, this study attempted to contribute to the field of agent-based simulations. While classic population-level modeling has been limited in its ability to integrate individuals' decision-making, this study expanded the application areas of agent-based modeling to an individual level. Even though this study has limited the application strategy to minimum nurse travel, agents can be set as patients, physicians, staff, visitors, and caregivers within the facilities as an expanded approach. In this sense, an agent can also be set with any spatial factors in the simulation, such as noise level, visibility, and patient preferences. Planners can develop a simulation model with a mixed approach for optimized layout planning. Simulation of healthcare in the design process is useful because it allows designers and planners to analyze the performance of the facilities at an organizational level and also enables relative comparisons among design options. This study investigated how floor configuration and room density influenced nurse walking distance in the hospital's medical-surgical units with the intent of reducing staff fatigue and securing more time for nurses to spend on direct patient care. The effects of the unit typology, nurse station location, and spatial arrangement of patient rooms on nurse travel were evaluated. In units with a nurse station placed at the center of the unit, shorter nurse walking distances were measured. In triangular units, the nurse stations were placed closer to the center of the units, compared to the rectangular units. Also, a contiguous patient assignment contributed to shorter nurse travel. A well-planned patient room arrangement was one of the spatial features that influenced the contiguous patient assignment. With increasing demands on medical care services, nursing units in Korea have tended to consist of mixed patient room arrangements of single/double-bed and multi-bed patient rooms in the same unit. In units with a mixed arrangement, nurses were often assigned to patients in scattered locations, which resulted in more walking. Rearranging spatial layout of patient rooms such as, locating multi-bed patient rooms near the nurse station and mixed planning of single/double-bed rooms and multi-bed rooms on one side of the corridor, helped to reduce nurse walking. One of the limitations of this study was that the patient assignment model was limited to one scenario. The study did not address patient acuity or unit specialization (maternity, oncology, etc.). Despite other constraints of patient room assignments such as patient case mix and nurse experience, this study limited the model to a balanced distribution of patient rooms of single/double or multioccupancy, which was a major distribution criterion in selected cases. Other limitations were that this was a simulation and not a real observation study, and all nurse walking behaviors were not taken into account. Another limitation is that the shapes of units in this research have been limited to two types, racetrack or triangular types. This was appropriate for the geographic region in which this research took place. In Korea, these are the most common configurations for an inpatient unit design due to the need for large-scale hospitals and high land prices. Methodologically, this study presented a novel application of agent-based simulation to simulate and measure nurse walking models, enabling automatic distance measuring via setting the start and destinations. This study aimed to help healthcare providers to test their design options and make improvements with evidence in decisionmaking. For future studies, decentralized and hybrid nursing station models in long linear units will be further explored in relation to the spatial arrangement of patient rooms to facilitate efficient nurse walking and sufficient patient visibility. Achieving equivalent nurse walking distances for nurses remains a challenge for both designers and healthcare providers. This work was supported by the BK 21 Plus funded by the Ministry of Education of Korea. Modeling nurse-patient assignments considering patient acuity and travel distance metrics Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction Exploring isovist fields: space and shape in architectural and urban morphology An evaluation of physiological work demands on registered nurses over a 12-hour shift Out of Sight, Out of Reach: correlating spatial metrics of nurse station typology with nurses' communication and co-awareness in an intensive care unit Evidence-Based Healthcare Design Advantages and disadvantages of single-versus multiple-occupancy rooms in acute care environments: a review and analysis of the literature Nurses' perception of single-occupancy versus multi-occupancy rooms in acute care environments: an exploratory comparative assessment The effect of environmental design on reducing nursing errors and increasing efficiency in acute care settings: a review and analysis of the literature A Study on the Architectural Planning Guidelines of the Wards in Central Public Hospitals of the Communities A Study on the facility guidelines of the ward in hospitals A predictive model for computing the influence of space layouts on nurses' movement in hospital units Towards a multi-agent model for visualizing simulated user behavior to support the assessment of design performance A knowledge-based system for automated architectural code checking Work-related burnout, job satisfaction, intent to leave, and nurse-assessed quality of care among travel nurses Centralized and decentralized nurse station design: an examination of caregiver communication, work activities, and technology. Health Environ National Center for Health Workforce Analysis. Projecting the Supply and Demand for Primary Care Practitioners through 2020. U.S. Department of Health and Human Services Self-organizing pedestrian movement Simulation of pedestrian crowds in normal and evacuation situations Healthcare Leadership: Maximizing the Impact of Nursing Care Quality: a Closer Look at the Hospital Work Environment and the Nurse's Impact on Patient Care Quality. The Center for Health Design A 36-hospital time and motion study: how do medical-surgical nurses spend their time? A modeling approach for estimating the impact of spatial configuration on nurses' movement Pedestrian route-choice and activity scheduling theory and models A study on the change of physical environment in Seoul Medical Center by providing comprehensive nursing service Building Type Basics for Healthcare Facilities Visibility and Circulation Efficiency in Hospital Nursing Units: Visibility and Accessibility Analysis Using Agent-Based Simulation (Doctoral dissertation) Employing visibility and agent-based accessibility analysis to enhance social interactions in older adult care facilities Computing walking distances within buildings using the universal circulation network Can intensive care staff see their patients? An improved visibility analysis methodology Splendid isolation? The pros and cons of single rooms for the NHS Analyzing US nurse turnover: are nurses leaving their jobs or the profession itself Simulation of human wayfinding uncertainties: operationalizing a wandering disutility function Policy Direction of Comprehensive Nursing Care. Ministry of Health and Welfare Design lessons from the analysis of nurse journeys in a hospital ward On Spatial Wayfinding: agent and human navigation patterns in virtual and real worlds Inpatient unit flexibility: design characteristics of a successful flexible unit Estimating design impact on waste reduction: examining decentralized nursing An empirical examination of the impacts of decentralized nursing unit design Should hospitals provide all patients with single rooms? Human wayfinding in unfamiliar buildings: a simulation with a cognizing agent Ontology and epistemology for agent-based wayfinding simulation Flocks, herds and schools: a distributed behavioral model Finance. The profitability of private patients Impact of hospital unit design for patient-centered care on nurses' behavior Nursing unit configuration and its relationship to noise and nurse walking behavior: an AIDS/HIV unit case study A study improving a nurse's walking path and staff visibility in general hospital wards Single rooms in hospitals "increase anxiety" among nursing staff Comparisons of Intensive Nursing Service in a Circular and a Rectangular Unit: Rochester Methodist Hospital Examining perception of job satisfaction and intention to leave among ICU nurses in China Guidelines for intensive care unit design Moving agent pedestrians through space and time Ward design must not be restricted to single rooms Influence of nursing unit design on the activities and subjective feelings of nursing personnel From axial to road-centre lines: a new representation for space syntax and a new model of route choice for transport network analysis Encoding natural movement as an agent-based system: an investigation into human pedestrian behaviour in the built environment The Role of the Physical Environment in the Hospital of the 21st Century: a Once-In-a-Lifetime Opportunity. The Center for Health Design A review of the research literature on evidence-based healthcare design Developing and testing a computerized decision support system for nurse-to-patient assignment: a multimethod study An agentbased pedestrian and group dynamics model applied to experimental and real-world scenarios Hospitals: a Design Manual Hospital staff nurses' work hours, meal periods, and rest breaks: a review from an occupational health nurse perspective A spatial agent-based model for the simulation of adults' daily walking within a city The effect of hospital unit layout on nurse walking behavior Centralized vs. decentralized nurse stations: effects on nurses' functional use of space and work environment The Role of the Physical Environment in the Hospital of the 21st Century: A Once-In-a-Lifetime Opportunity. The Center for Health Design The authors have no conflicts of interest to report.