ABSTRACT This paper discusses the design of ventilation systems for negatively and positively pressurized patient isolation rooms. The paper focuses on how to quantify and achieve target levels of protection for either the patient (positively pressurized rooms) or health care workers and other hospital occupants (negatively pressurized rooms). Attention is paid to the influence of ceiling supply diffuser selection. Thermal comfort issues are also discussed, and an alternative to “age-of-air” techniques using age-of-contaminant calculations is recommended for use in patient isolation room design. Practical considerations are illustrated through the presentation of two case studies. The first case study of a TB isolation room includes a CFD model analysis of different air distribution systems including an assessment of ventilation effectiveness and patient thermal comfort. This work includes simulation of a cough from a patient toward a health care worker and throughout theisolation suite. The second case study of a positive pressure isolation room assesses the throw of supply air around a patient bed in terms of providing protection for the patient while maintaining comfortable conditions. INTRODUCTION Challenges to health services around the world from monkey pox, severe acute respiratory syndrome (SARS), and continued cases of tuberculosis have meant that hospitals have had to deal with, and prevent the spread of, contagious diseases. Additionally, health care facilities are providing for people afflicted with diseases that suppress a patient’s immune system, either through the treatments (e.g., cancer) or due to

the illness itself (HIV/AIDS). To provide adequate protection of patients and caregivers, special purpose patient isolation rooms are designed with ventilation systems that maintain a negative pressure to protect caregivers or positive pressure to protect patients. While positive or negative pressurization can be used as a containment strategy, it leaves the occupants within the room (caregivers and patients) with risks from each other. Proper ventilation design can help to reduce those risks by providing either deflection of contaminated air or at least efficient removal of contaminants. The momentum from the supply diffuser can be used beneficially in this regard or, as often happens without careful design, can aggravate the problem. The case studies that follow illustrate some of the effects of supplymomentum, buoyancy, and room dimensions on ventilation within the space. BACKGROUND The purpose of this section is to review the means by which airborne infections spread, using tuberculosis (TB) as an example, and then present methods by which patient room ventilation systems are assessed. In that review, comparisons between the different methods are provided. TB Isolation Rooms (Case 1) Infectious diseases can spread from one person to another by aerosol droplets. The spread of tuberculosis (TB), for example, occurs when an otherwise healthy individual inhales a sufficient number of tubercle bacilli that are expelled by a patient infected with pulmonary TB. An infected patient continuously expels these particles when coughing, sneezing, talking, or spitting (Gammaitoni and Nucci 1997). The infec-

Duncan Phillips is an associate and senior specialist, Ray Sinclair is a principal and project director, and Glenn Schuyler is a principal and vice-president of research at Rowan Williams Davies and Irwin, Inc., Guelph, Ontario, Canada.

IAQ 2004

1


tious particles are estimated to be on the order of 1 to 5 µm in size (CDC 1994) and normal room air currents can keep them airborne for long periods of time. These particles are easily spread about a room or building unless adequate protection and control are provided. A number of outbreaks of TB in hospitals in the 1980s and 1990s (Conroy et al. 1997) prompted the CDC to issue a number of reports regardingthe prevention of TB transmission in health care facilities (CDC 1994). Infected patients are isolated from other individuals in hospitals and placed in special isolation rooms. A series of administrative and engineering controls are implemented to reduce airborne transmission. The administrative controls as described by CDC (1994) and summarized by Conroy et al. (1997) include (1) an infection control program identifying individuals likely infected, (2) training, (3) medical surveillance of at-risk health care workers, and (4) respiratory protection for those in immediate contact with infected patients. The engineering controls recommended include: (1) the room be at least 0.001 in. w.c. negative pressure with respect to adjacent spaces; (2) airflow should be designed for and tested such that it travels from hallways or anterooms into the patient room; (3) the exhaust flow should exceed supply air by 10% or a minimum of 50 cfm within the patient isolation room; (4) the velocity under the door when it is closed should be a minimum of 100 fpm; and (5) the dilution ventilation rate should be at least 6 ACH and 12 ACH in newer facilities (Conroy et al. 1997). The basic design philosophy of TB patient isolation rooms is relatively straightforward: a high ventilation rate within a room is used to dilute and flush the aerosol contaminants. ASHRAE (2003) notes that “the preferred design approach [to airborne infectious isolation rooms] emphasizes air mixingeffectiveness and dilution ventilation without attempting to establish unidirectional airflow.” The objective here is to maximize the mixing rate. Efforts are made to prevent the airborne contaminants from escaping the room by ensuring a net flow into the room at all times and in some cases the presence of an anteroom serves as an airlock. In cases where it is not possible to provide adequate dilution, or where additional preventative measures are desired, HEPA filtering (Gammaitoni and Nucci 1997) and ultraviolet germicidal irradiation (UVGI) (Memarzadeh and Jiang 2000) can be used to remove or kill the viable TB bacilli. The UVGI assessment is reported in more detail in NIH (2000). The required performance of a patient isolation room ventilation system is related to providing adequate dilution to minimize the risk that a caregiver may be exposed to an infectious dose. Kowalski et al. (1999) provide references that suggest that the infective dose for M. tuberculosis is between 1 and 10 bacilli. Given that the number of aerosol particles in the 1 to 5 m aerosol particle range released by a sneeze is on the order of 100,000 and a cough on the order of 1,000 particles (Duguid [1945] as cited by Kowalski and Bahnfleth [1998]), the dilution rate required in order to reduce the
2

concentration of particles below 10 is at least 100 and more likely 10,000 for a single breath. Multiple coughs or sneezes and a prolonged exposure to the contaminated air means thatthe dilution rate must be significantly higher. Unfortunately, the use of a high ventilation rate in a room does not guarantee that all regions of the room are ventilated at a high rate. Stagnant zones, isolated corners, and shortcircuiting can lead to a reduction in ventilation efficiency and, as a result, lower ventilation rates in large sections of a ventilated room. Some close proximity CFD modeling work has been conducted in the past. Bjørn and Nielsen (1998) report CFD simulations of the transport of gases between two people breathing and how they compare to physical experiments using breathing manikins. Bjørn and Nielsen (1998) report that while it was difficult to match the CFD simulation to the data, the simulations did permit them to assess the sensitivity of the results to various room configurations. They determined that the convective heat outputs of the individuals, crosssectional exhalation area (size of mouth opening), and pulmonary ventilation rate were all important factors in the level of contamination from one person to another. The distances between the manikins was comparable to that of a health care worker tending to a patient. Their results also indicated that the simulated exposure is not very sensitive to variations of exhalation temperature at small mutual distances but is more so as the distance increases. In this paper, a study is presented of ventilation conditions in a TB isolation room. The purpose of the CFD modelingcarried out was to assess the efficacy of the ventilation in a typical room planned for construction at an existing health care facility. The goal of the assessment was to determine the level of protection afforded health care workers by the ventilation system under varying operating conditions (heating and cooling modes) and system configurations and supply diffuser types. The design objectives were to achieve a local ventilation rate of 12 or greater in all parts of the room with a volumetric supply air change rate of 15 ACH or less to the room. Given the mixing type of ventilation system implemented, this requires that the stagnant zones be eliminated. It may be necessary to have a dilution rate within the room of at least 100,000:1 in order to reduce the number of aerosol particles at the health care worker’s face from multiple coughs to which they may be exposed over a period of time. This dilution rate between the patient’s mouth and the health care worker was known to be unachievable. That said, it was of interest to know what levels of dilution are likely and whether the different ventilation configurations play a role in this close proximity dilution. Of equal importance was the concentration of particles in the vicinity of the door. If the door region were contaminated with particles, some may escape into the hallway when the door is opened, since no anteroom was planned for the isolation rooms in this facility.
IAQ 2004


Positive PressurePatient Isolation Room (Case 2) A second type of patient room is one in which the care is provided to patients who are immuno-suppressed. In these rooms, the ventilation system of the room and connected spaces are configured to prevent air from entering the room through any means except the room ventilation system. The ventilation system in the room is typically configured with HEPA filters for the incoming air and the room is positively pressurized with respect to the adjacent spaces. It is common for these facilities to have significant ventilation rates (e.g., 15 ACH) and thus the risk of drafts is high. Complaints of drafts are common in facilities of this type, and experience suggests that improper sizing and placement of ceiling diffusers can lead to draft. These patients are often in residence for a period of months, and maintaining comfort is important. In a recent report, Memarzadeh and Manning (2000) present a summary of research in which 36 different numerical experiments of a patient room were conducted. These numerical experiments tested different seasons (winter and summer) and different diffusers and ventilation rates. The results suggest that for summer conditions, the diffuser combination, return location, and ventilation rate do not have a significant effect on the overall acceptability of the room. They used ageof-air distributions to assess the efficiency of the ventilation system, thermal comfort to assess the conditions for an occupant,and the Air Diffusion Performance Index (ADPI) as a measure of uniformity of conditions. The ventilation patterns for a winter scenario were more variable, and minimum ventilation rates of 6 ACH were recommended along with baseboard heating. The Role of CFD Modeling The design of air distribution systems in patient isolation rooms can be greatly assisted by computer simulations based on computational fluid dynamics (CFD) modeling. CFD can predict air velocities, temperatures, and contaminant concentrations throughout the room for a range of design challenges. This information can be interpreted in terms of indoor air quality indices that can be compared against health criteria and also thermal comfort indices to assess patient comfort. Supply diffuser locations and types, flow rates, exhaust air vent locations, distributions of heat loads in the room, arrangements of furniture, and other blockages to air movement can be assessed and comparisons made to judge the best design alternatives. The theoretical details are that CFD is a numerical simulation technique in which the standard equations of fluid flow representing the conservation of mass, momentum, and energy are solved. For most practical environmental and engineering flows, equations imposing the effects of turbulence are required. Finally, the transport of a contaminant, in this case representing a cloud of light particles released during a cough, can be included. In fact, even individualparticle trajectories can be predicted if necessary. Many authors have included a review of CFD techniques, including Patankar (1980), Chen and Srebric (2001), and Jiang et al. (2003). The
IAQ 2004

simulations presented here have been conducted using two different commercial CFD packages. Age-of-Air Analysis The use of age-of-air techniques to assess indoor environmental flows was introduced to the ventilation field by Sandberg and Sjöberg (1983). Since then it has been put into practical form by many authors including Sutcliffe (1990). It is an analysis technique that parallels the residence time techniques used in chemical engineering. It permits one to assess how fresh the air is at a location within a ventilated space or the average age of the air leaving the room. If some region of the room has an age-of-air that is greater than the average age-ofair leaving the room, then it is underventilated. If the age-ofair within the room is uniform and equal to that at the exhaust, then the room is well mixed. A number of indices have been developed for age-of-air analysis and these may be found in ASHRAE (2001). These are frequently used to assess comfort or the effectiveness with which the HVAC system distributes the supply air for a variety of ventilation environments including patient rooms. Age-of-air-based ventilation indices typically report how a ventilated environment compares to a space that is either perfectly mixed or one with that has plugdisplacement flow. A notable difficulty in using the indices is that there has been a set of slightly different terminology used for similar indices – the terms ventilation efficiency and ventilation effectiveness have been used interchangeably when they should not be. ASHRAE (2001) provides a definition of air change effectiveness, and it is this index that the authors have used. Age-of-Contaminant The use of age-of-contaminant to assess ventilation flows is less common. It is based on the premise that a contaminant is released at a location in a room and potentially travels about the space and is then removed from the space by the ventilation system. The rate at which the contaminant concentration is reduced at various locations within the space provides information on the ability of the ventilation system to flush the space. The rate of flushing in any particular region provides information about the amount of fresh air entering that region. Brouns and Waters (1991) provide a summary of the technique. In a manner similar to age-of-air analysis, the age-ofcontaminant calculations can be expressed in terms of indices. Brouns and Waters (1991) list the contaminant removal effectiveness as one index that can be used for the assessment of contaminant transport about a room. In a review of different indices for assessing ventilation environments, Novoselac and Srebric (2003) conclude that for rooms in which the contaminant source or strength is unknown, theage-of-air indices are appropriate. However, in cases where the contaminant source location and strength is known, age-of-contaminant indices are more informative in communicating information about contaminant removal effectiveness. Contaminant removal effectiveness is defined
3


by Brouns and Waters (1991) to be the ratio between the nominal time constant for the ventilation air and the nominal time constant for the contaminant. The contaminant removal efficiency is then a manipulation of the effectiveness. These parameters for a particular room configuration are typically compared to values for perfect mixing and plug flow. The air change effectiveness is defined as a ratio between the nominal time constant and the room average age-of-air. Local air change effectiveness can be calculated if the local age-of-air is used instead of the room average. Thermal Comfort To predict thermal comfort conditions, different models have been developed or adopted by various recognized authorities. ASHRAE (2001) promulgates the Fanger (1972) Predicted Mean Vote (PMV) comfort index to predict thermal comfort. This is consistent with ISO 7730 (ISO 1984). This numerical manipulation of air speed, temperature, humidity, occupant clothing, work output, activity, incident radiation, and turbulence intensity predicts a thermal sensation scale that can then be interpreted. Thermal comfort assessment of ventilated environments is common, including those for patients. Other comfortindices such as the Air Diffusion Performance Index (ADPI) (ASHRAE 1990) may be used to assess comfort; however, the drawback with this index is that it does not include activity level, clothing, radiant conditions, or relative humidity as part of the assessment. For the simulations presented here, thermal comfort plots using the PMV scale have been generated assuming that the patient’s activity level is 0.7 (ASHRAE 2001), which is equivalent to sleep, and work level was assumed to be nil. Furthermore, the patient was assumed to be wearing a short-sleeved hospital gown with a light blanket cover. The clothing level was then adjusted by 0.25 due to the presence of the bed, yielding a total value of 1.3. It is also assumed that the relative humidity is 50%. Summary of Background One objective of the background section was to discuss means by which patient isolation rooms can be assessed. Some of the key points from this review include: 1. Patient isolation rooms may be assessed to determine whether the patient and/or other occupants will perceive the room to be comfortable or express discomfort. It is possible to use modeling tools to assess the potential risk of exposure to occupants in or around patient isolation rooms. Some assessments of isolation rooms have evaluated the efficacy of UVGI as part of a contaminant containment strategy. Patient isolation rooms have also been assessed from the perspective of age-of-air. This allows a determination of howwell the ventilation air is distributed within the room.



A more appropriate means of assessing the ventilation system in a TB isolation room could be to assess the contaminant removal efficiency. This could also be applied to a positive pressure isolation room where the contaminant source is untreated outside air through a temporarily open door or another source such as window leakage.

CASE STUDIES
Case 1: TB Isolation Room The ventilation assessment of the TB isolation room presented here takes the reader through details about the room configuration, CFD model simulation process, age-of-air and age-of-contaminant calculations, thermal comfort assessments, and the predicted dilution of a patient cough. The primary conclusion that the reader may find interesting is that the standard age-of-air analysis has notable limitations and that the prediction of age-of-contaminant is helpful in comparing proposed ventilation strategies. Room Description, Loads, and Ventilation System Configuration Three simulations are presented here; they show the predicted conditions representing a laminar diffuser arrangement and two different throw patterns for the square diffusers. The first has the diffusers acting as four-way diffusers. The second has the diffusers acting as two-way diffusers. The HVAC arrangements are compared on the basis of the age-ofair and relative concentrations of aerosol particles expressed as dilution ratios at the face of the healthcare worker following the cough. The layout of the TB isolation room is presented in Figures 1 and 2 for an HVAC configuration using laminar and square diffusers, respectively. The room itself had a floor plan of approximately 16.26 m2 (175 ft2) and was approximately 46.72 m3 (1650 ft3) in volume. The heat loads and boundary conditions applied to the room for the different HVAC configurations were the same and are summarized in Table A1 in Appendix A for the summer design case. The HVAC flows into and out of the room are summarized in Tables A2 and A3 for the laminar diffusers and square diffusers (two- and four-way), respectively. The flows in the tables represent a global ventilation rate of approximately 15 ACH. Assessment Methodology. The CFD process is briefly described in Appendix B. The results of this process were a prediction of the steady-state flow field for each of the diffuser arrangements from which age-of-air, age-of-contaminant, and thermal comfort analyses were conducted. An additional method of assessing the ventilation efficiency in the room was to model a cough. This was achieved by implementing a release of contaminant from the mouth of the patient at 3.56 m/s (700 fpm), which emulated a cloud of light particles. These transient (time-varying) simulations were conducted by starting with a solved steady-state flow
IAQ 2004

2.



4.

4


IAQ 2004

Figure 1 TB isolation room layout and boundary conditions for laminardiffuser configuration.

5


6

Figure 2 TB isolation room layout and boundary conditions for square 2 and four-way diffuser configurations.
IAQ 2004


Figure 3 Region of TB isolation room with age-of-air > 300—laminar diffuser configuration.

Figure 4 Region of TB isolation room with age-of-air > 270—square four-way diffuser configuration.

field for the room. Using a transient solver, the patient’s mouth was then turned into a source of air laden with particles. At 0.15 second after the start of the cough, the flow from the mouth was stopped. The particle concentrations were then monitored at various locations in the room. The most important locations can be considered to be the health care worker and locations near the door. The pulsed release of a tracer representing a cloud of aerosol particles permits one to assess the contaminant removal efficiency of the different ventilation configurations. Age-of-Air and Thermal Comfort Figures 3, 4, and 5 present iso-surfaces of age-of-air at 300, 270, and 300 seconds for the laminar and square diffusers, four-way and two-way, respectively. The square four-way diffuser room configuration resulted in an air flow pattern that had a maximum age-of-air within the room of approximately 280 seconds. Regions inside the iso-surface have relatively stale air. Those outside the surface have air that is predicted to be fresher than the threshold value identified for the isosurface. The figures show that thelaminar (Figure 3) and square diffusers configured for a two-way throw (Figure 5) do not meet the ventilation efficiency criterion of having the ageof-air at all locations within the room at 300 seconds or less. In fact, it is very difficult to meet this form of ventilation criterion unless the room is either very well mixed with an air change rate greater than the target criterion or employs a displacement ventilation system. The thermal comfort in the two rooms for the patient was essentially the same. The PMV values were 0.65 (slightly warm) and 0.46 (comfortable) for the laminar and square fourway diffusers, respectively. The reason that the displacement diffuser setup resulted in a sensation of slightly warm was that the temperature over the patient is predicted to be slightly higher. The four-way diffusers do a better job of penetrating the cooler supply air into the patient region.
IAQ 2004

Figure 5 Region of TB isolation room with age-of-air > 300—square two-way diffuser configuration.

Dilution of Cough Figures 6 and 7 present the time-varying concentration of particles at various locations for the laminar and four-way square diffusers for the first 60 seconds. The concentrations have been scaled to a reference concentration of 10,000 at the mouth of the patient. Thus, a concentration of 10 at any given location represents a dilution of 103. A monitor was established in the path of the cough approximately 0.12 m (4.75 in.) from the patient’smouth. The locations identified as near the door region are 0.61 m (2 ft) from the face of the door, 0.91 and 1.52 m (3 and 5 ft) high. The plots show that the concentration peaks at the mouth of the health care worker at approximately 5 seconds of the cough and rapidly tails off for both ventilation strategies. The peak magnitude of the cough does not change significantly for the two strategies. The inset graphs in Figures 6 and 7 show the concentrations for a 10-minute duration. In addition to the concentrations at each location, the concentration for a hypo7


8

Figure 6 Particle concentration for laminar diffuser configuration.

IAQ 2004


IAQ 2004

Figure 7 Particle concentration for square four-way diffuser configuration.

9


Table 1. Peak Concentrations and Accumulated Dosages for the Laminar and Four-Way Square Diffuser
Location Cough path 6 in. above desk 3 ft above floor near door 5 ft above floor near door At health care worker’s mouth 1 ft above patient Perfectly mixed room
* †

Peak Concentration* Laminar Square 4-way 6190.0 0.126 0.017 0.017 43.6 26.2 0.097 6187.8 0.120 0.307 0.483 46.3 3.1 0.097

Accumulated Dosage † Laminar Square 4-way 1275.8 26.3 7.5 7.9 212.6 155.2 22.9 1275.8 (not a typo) 22.3 26.0 27.5 182.2 56.1 22.9

The units of concentration are [cough particles/cough volume]. The units of accumulated dosage are [cough particle seconds/cough volume].

thetically perfectly mixed room was also included onthese plots. The inset plot in Figure 7 shows how the distribution of cough aerosols for the four-way diffuser room design approaches the perfectly mixed result at approximately 3 minutes for the locations monitored. However, the inset plot in Figure 6 highlights how the aerosol in the laminar flow room requires up to 8 minutes to approach conditions that would reflect a well-mixed room. These plots also permit one to evaluate the time it takes for aerosol to reach the door region of the two rooms. For the fourway square diffuser, the cough aerosol reaches the door in approximately 35 seconds (for a particle concentration of 0.01). The results for the laminar diffuser configuration highlight how the lack of induced mixing significantly increases the time it takes for the aerosol particles to reach the door— approximately 120 seconds are required for the same concentration to reach the door. Table 1 summarizes the maximum cough particle concentrations, as well as the accumulated dosage over 10 minutes, for individuals located in the regions monitored during the course of the cough dilution. In addition, the peak concentration and accumulated dosage for the perfectly mixed room are presented for reference. It has been assumed that the individuals are not moving during the period of dosage integration. While it may be unrealistic to assume people are motionless for 10 minutes, the integration of dosage does provide a useful measure for exposure to coughparticles at each location monitored. It is clear from Table 1 that there are consequences to the selection of the diffuser type that are not evident until one assesses the time-varying particle concentrations. For example: • The rapid mixing of the square four-way diffusers does not reduce the concentration of particles at the health care worker’s head. However, the peak concentration near the door jumps up by an order of magnitude.

•

•

Neither the peak concentration nor accumulated dose along the cough path is influenced by the diffuser selections modeled here. The effect of having laminar diffusers in the near door region appears to help to prevent the particle laden air from penetrating into the door region – the peak concentrations and accumulated dosages are less than those for the perfectly mixed room. However, the region is poorly ventilated by fresh air, as indicated by Figure 3.

Comparison of both the laminar and square four-way diffuser to the perfectly mixed room conditions suggests that the room HVAC layout could be improved. One additional observation of the time-varying concentrations is that the room exhaust did not start to evacuate the cough particles in any great concentration for approximately 27 seconds for the laminar diffuser configuration and 40 seconds for the fourway square diffusers. Clearly, this is a specific result of the room configuration; however, it does highlight the importance of the interaction between thediffusers and the exhaust. This would explain the reason why the accumulated dosages are larger for the two rooms than the perfectly mixed equivalent. The cough particles in a perfectly mixed room would start to leave the room as soon as they had been generated, reducing the total number of particles in the room available for inhalation. Table 2 compares the contaminant removal efficiency and the air change effectiveness for the two different TB room configurations. It highlights how age-of-air indices alone can misrepresent the performance of a patient isolation room ventilation system. The results in Table 2 show that while the air exchange efficiency indices for both rooms is close to that for a perfectly mixed room, the contaminant removal efficiency is far from it. This indicates that while the room approaches well-mixed behavior, the cough location, along with the locations of the receptors of interest as well as the acute nature of the exposure, suggests that the air exchange efficiency is not a good measure of the performance of a patient room ventilation system.
IAQ 2004

10


Table 2.
Location Cough path 6 in. above desk 3 ft above floor near door 5 ft above floor near door At health care worker’s mouth 1 ft above patient Room average Perfectly mixed room

Contaminant Removal Efficiency and Air Change Effectiveness
Contaminant Removal Laminar Square 4-Way 0.019 0.468 0.723 0.716 0.104 0.136 0.019 (not a typo) 0.509 0.474 0.460 0.120 0.301N/A 0.500 0.500 Air Change Effectiveness Laminar Square 4-Way 1.10 1.01 1.32 1.08 1.09 1.09 1.13 1.00 1.25 1.12 1.29 1.33 1.20 1.18 1.25 1.00

Finally, the results presented for the cough simulations represent one conceivable scenario of patient and health care worker location. If the source of the cough were in another location in the room, then the exposure would be somewhat different. However, the magnitude of the change for the fourway diffuser room configuration would be less than that for the room with the laminar diffusers for equivalent exposure locations. This is because the time required for the concentration conditions in the laminar diffuser room to approach wellmixed conditions is approximately eight minutes, while that for the four-way diffuser case is three minutes; therefore, for most locations in the room, the time to mix a release of cough particles around the rest of the space is three minutes for the four-way diffuser and eight minutes for the laminar diffuser. The cough location selection for the laminar diffuser room will play a larger role in the resulting exposure. The results discussed above demonstrate that a TB patient isolation room should be assessed using age-of-contaminant analysis techniques rather than age-of-air indices. It is possible for the age-of-air calculations to be misleading at the very least for the circumstances presented here. Case 2: Immuno-Suppressed Patient Room In the second case study, a positivepressure patient isolation room was designed in a health care facility located in the U.S. Southwest. Here local codes require that there be a minimum of 15 ACH provided to the room and that the supply and return be arranged to minimize stagnation and short-circuiting. Part of the objective for the design of the patient room described here was to change the environment patients experience when subjected to long hospital stays. Patient room pods were designed with a 32.05 m2 (345 ft2) multipurpose work alcove and nurse and physician work areas that serve as the common air lock for six patient rooms. This area is a negatively pressured space with respect to the isolation rooms. The anticipated effect of this configuration would be that the patients would feel more connected to other hospital occupants and not be behind two panes of glass as is typically the case when the room has an anteroom.
IAQ 2004

Figure 8 Positive pressure isolation room configuration with adjacent work alcove.

As part of the design effort, a full-scale mockup of a typical patient room was built. The mockup helped the design team assess the look, ergonomics, and acoustics of the room, test building practices, and test working equipment and systems. In order to better understand the details of the airflow patterns and how these patterns may affect protection and thermal comfort of the patient, a CFD computer model was developed. The CFD computer model was set up to test conservativeworst-case conditions in the room including solar loads on a warm summer day and equipment, lighting, and occupant heat loads. The simulations were conducted to assess conditions with the isolation room doors open and closed to evaluate the likelihood of air entering the room from adjacent spaces and the potential for drafts or dead flow zones in the room. Figure 8 presents the room configuration with the alcove attached. For the door-closed scenarios, a surface was placed across the opening between the alcove and patient room. The room was approximately 18.6 m2 (200 ft2). There were two HEPA filtered radial diffusers located to either side of the
11


patient’s bed in the ceiling. The exhaust vent was wallmounted about 0.15 m (0.5 ft) above the floor, diagonally opposite the patient’s bed near the door. As described above, the ventilation rate was 15 ACH. Blockages representing the patient and a health care worker were incorporated into the model. Additional blockages representing equipment and furniture were also included. The two model studies in this project contained approximately 215,000 and 345,000 nodes for the door-closed and door-open scenarios, respectively. In advance of running the patient room model, calibrations were carried out to adjust momentum sources applied near the diffusers until a good match was achieved between the model and the manufacturer’s information on throw and spread of the air in a separate, open room test model.This effort was similar to the work that is reported in Srebric and Chen (2003) and was conducted before that report was available. The loads implemented in the room were the patient and doctor sensible load (72 W/245 Btu/h each), equipment (495 W/1690 Btu/h), a television (150 W/510 Btu/h), lighting in the room (175 W/600 Btu/h) and in the bathroom (140 W/480 Btu/ h), a radiative solar load (1000 W/3415 Btu/h), and a convective solar load at the window (420 W/1435 Btu/h). The total heating load for the door-closed simulation was 2524 W/ 8620 Btu/h. The supply air temperature was 20.5°C. For the door-open scenario with the work alcove included in the CFD model, additional loads for lighting, occupant, and equipment heat (1045 W/3570 Btu/h total) were added; the loads within the patient room remained the same. There was an additional four-way diffuser located within the work alcove, and the supply air temperature for this diffuser was set to balance the loads from the alcove and meet the same target room temperature.

RESULTS Figure 9 presents the prediction of thermal comfort for the door-closed configuration and indicates that the patient would be slightly cool. However, the temperature at this level is reasonably uniform at 22.0-24.5°C (71.6-76.1°F) with the warmer temperatures caused by the presence of some portable medical equipment adjacent to the bed. The velocity plot, not included in this paper, shows that air currents are low (