Design and Health World Health Design

New approaches to the creation of healthy environments

Technologies are emerging that can reveal the reactions of mind and body to specific features of the designed environment. This paper reviews a selection of these innovations, which can provide the means to conduct pre-design evaluations.

Eve A Edelstein MArch PhD (Neuro) EDAC Assoc AIA F-AAA

The impact of building-design strategies on non-communicable disorders, unhealthy behaviour and global ecological conditions has recently been recognised in studies by the United Nations, the World Health Organization and the Institute of Medicine.

Such research acknowledges a compelling need to improve our cities and buildings for the benefit of human health and wellbeing. Revelations surrounding the frequency at which unhealthy and adverse events occur in healthcare environments have motivated architects to create design strategies that improve human and sustainable outcomes. These strategies, however, are not always as informed as they could be, because “too often, the form and function of architectural environments neglect to take into account the influence of the built setting on human responses and, indeed, on human health itself”.1

In response to these needs, a new generation of designers, architects and engineers is embracing an interdisciplinary approach and employing novel technologies to inform salutogenic design. The incorporation of findings from neuroscience, physiology and psychology, along with architectural research models based on philosophical constructs and sociological and ethnographic methods, offer the means to relate human responses to design in a more powerful fashion.

A ‘neuro-architectural’ approach
The neuro-architectural process informs design by correlating specific measures of the built environment (input) with quantified measures of the brain’s and body’s responses (neural, physiological and psychological responses), as well as sociological, behavioural and economic outcomes (output). This process gives weight to scientific methods of research, linking input, response and output, and allowing for statistical and critical evaluation of design outcomes. 

With the advent of recently developed neuroscientific instrumentation, a higher degree of objectivity enables measurement of both conscious and sub-conscious responses without relying on the subjects’ understanding or ability to articulate their cognitive, sensory or emotional response to design. In this way, resulting findings therefore offer greater potential to reveal the needs and preferences of the growing number of individuals with developmental or physical conditions, or those with dementia or cognitive impairment.

Figure 1: Neuro-architectural design process
The goals of this integrated approach are to understand better how the brain’s and body’s responses to the environmentinfluence health and wellbeing; and to define and quantify the human, environmental and financial returns on investment in design. The complexity of architectural environments suggests it may be impossible to reduce human interaction with built settings to measurable elements. But the combination of scientific studies with rigorous social and cultural observations can yield greater analytic confidence in the applicability of findings based on physical, physiological, psychological and social (PPPS) measures. These objectives are consistent with the creation of private and social spaces, with natural and innovative features that improve the quality of a place and how users function within it, as well as including salutogenic factors to enhance wellbeing.

Translating the hierarchy of design
A ‘hierarchy of design principles’ is proposed, translating Maslow’s ‘hierarchy of needs’2 into a design process that includes iterative feedback loops to each component of the body, brain and mind that interacts with design features, and consequently drives behaviour. Outcomes that serve salutogenic and economic goals can thus be analysed using the construct of a scientific method applied to design inquiry.

The design process is reversed relative to many evidence-based projects; the first step translates clinical and neurophysiological evidence of the impact of physical design components into principles that reflect how the brain and body respond to place. Design that minimises harm in terms of physical health and safety, error and injury is the first priority and considered fundamental to all design decisions. In an iterative feedback loop, the psycho-physiological impact of each design feature is then assessed to ‘do no harm’ to either mind or body of patients and providers. At this stage, the differing needs of those with specific disabilities or medical conditions are considered. Psycho-physiological responses are considered in terms of their ability to support the quality of care, as well as the quality of work and creativity.

Innovations to enhance outcomes cover all aspects of mind and body. These aspects include sensory, perceptual and cognitive functions for problem-solving and critical decision-making. Socio-cultural constructs inform considerations of private and social spaces to enhance function, engagement and meaning. In this way, design is explored in relation to mind and body as well as health and wellbeing.

Embedded in this approach is the inclusion of guidelines on sustainable design, so that materials, systems and building performance ‘do no harm’ to the environment or inhabitants. The merger of healthy and sustainable design is evident in the recent development of sustainability guidelines for healthcare facilities, by institutions such as BREEAM for Health, United States Green Building Council, LEED for Healthcare, AIA Facilities Guidelines Institute, the Joint Commission, and the Global Health and Safety Initiative, among others. Most guidelines focus on exposure to pollutants and toxins in air and water, and via physical contact with materials. But design strategies to improve clinical outcomes are equally important and sustainability guidelines should be directed at improving both human and environmental outcomes.1

Figure 2:  Visuo-acoustic simulations in the CAVE evaluate speech intelligibility of doctors reciting ‘sound-alike’ medication orders when competing with clinical recordings of conversations, ambient noise, equipment alarms and ventilation sounds

This scientific research paper reviews the development of several techniques that combine novel biological and environmental sensors in real and virtual simulation settings to test design hypotheses and allow subjects to see and hear the consequences of design. Recent developments in visualisation and acoustic rendering enhance the realism of immersion in virtual mock-ups.

The application of this flexible simulation environment is evaluated below in regard to the impact of sound on medication error and the quality of care; spatial cognition and preference; wayfinding; and the influence of light on human outcomes.

The development of visuo-auditory simulation environments, such as the Cave Automatic Virtual Environments (CAVE), provide controlled settings in which design hypotheses can be tested before design is finalised. Viewers interact with the virtual images using a 3D joystick and a head-tracking infrared sensor system, which registers the subject’s location and orientation in space, and moves 3D visual fields according to their point of view. The viewer’s head and joystick locations are logged over time, dynamically tracking their first-person perspective, position and interactions within the virtual setting. Use of collaborative-CAVE software also allows visualisation in many environments to be distributed in synch across many locations around the world.3

A novel computer-aided design system named CAVECAD allows users to alter dynamically the virtual environment while subjects stand within the stereoscopic model itself. This approach eliminates the need to create a 3D model at a desktop computer before importing it to a virtual environment. Thus, a number of design conditions can be tested without necessitating the building of, or change to, mock-ups prior to further testing. By  logging subject responses over a sequence of trials, multiple design changes can be tested, according to controlled protocols and during synchronous recording of brainwave responses.

In CAVE simulation environments and dedicated listening environments, acoustic simulation systems have been developed to test, predict and improve the impact of acoustic design on human responses and function. Using dynamic audio-rendering software (SoniCAVE), scenes of sound can integrate databases of materials, audio samples from real healthcare environments and equipment, and geometric reverberation computations to create accurate predictions of real-life scenarios.4

Table 1: Innovation-team survey results. The hierarchy of design priorities for acoustic modulation of healthcare environments

Sound design
An acoustic simulation was designed to model the consequences of noise conditions on work errors. The Center for Quality Improvement and Patient Safety report, and analysis of 26,000 records in a US-wide anonymous error-reporting system (MEDMARX), revealed an increased rate of error in medications with similar-sounding names.5 The consequences of this issue were demonstrated at a workshop where speech intelligibility and error were presented in three conditions: a dedicated sound lab, an auditorium, and in CAVE virtual simulations.

Sound-alike medications from the National Patient Safety ‘sound-alike’ medication list6 were recorded in the presence of, and without, competing sounds (eg recordings of medical-instrument alarms, nearby conversations and HVAC noise). In all conditions, the participants were unable to discriminate between ‘sound-alike’ names from the list if the competing noises were greater than 15dB above the medication list (played at 65dB(A), the approximate level of conversational speech).
Although the participants were not clinicians, and would therefore be unfamiliar with the medication names, these results are consistent with the body of research that shows intelligibility improves as speech levels rise 15dB, or more, above the background noise level. Even greater separation between speech and noise is required in order to achieve equivalent intelligibility scores for those listening in a second language, and for those with hearing disorders and hearing that has diminished with age.

Unfortunately, the majority of critical-care environments far exceed these levels, dramatically increasing intelligibility and associated error. Figure 3 plots sound levels in a variety of hospital conditions. Using standard protocols for evaluating averaged noise levels, the intensity of sound ranged from 75-85dB(A) Leq in critical-care units such as emergency and intensive-care departments. These findings are consistent with recent studies demonstrating that noise in healthcare environments has been steadily increasing over the past 50 years, with no single facility operating within the sound levels recommended by the World Health Organization.7 When impulse-sound peaks are measured, using time constants capable of recording sounds from alarms and equipment, a near constant impulse level is seen (in green lines), ranging from 100 to 120dB peak during shift changes.

Exposure to such sound levels increases the risk of noise-induced hearing loss, as well as the likelihood of physiological and psychological changes. Increased cardiovascular risk has been observed when daytime noise levels exceed moderate levels, and stress reactions, such as cortisol disturbances, have been observed in children who are exposed, for long periods, to low-frequency traffic noise averaged at less than 55dB(A).8 Unwanted noise exposure and lack of speech confidentiality and privacy further diminish performance, communication, satisfaction and the healing quality of healthcare environments. 

Findings from surveys of 118 medical practitioners and administrators at the California Institute for Telecommunications and Information Technology (Calit2), at the University of California, San Diego, confirmed concerns regarding acoustic conditions in healthcare settings. The list in Table 1 reveals the priorities for improving sound conditions, among those surveyed. Ordered in a ‘hierarchy of care’, acoustic modification should first consider the likelihood that unwanted noise may lead to death, or severe adverse events, such as medical or medication error as a result of miscommunication related to competing or high-level sound. The list of needs also considers enhancement of provider performance and the quality of the care environment, in addition to patient needs.

While today’s guidelines call for greater attention to acoustic optimisation and acoustic consultation with healthcare design teams, the high sound levels and atypical sound profile generated by equipment and people during the provision of critical care exceed the conditions that minimum acoustic performance standards are designed to address. For example, the recommended wall systems for privacy in standard office spaces are insufficient to ensure confidential-speech privacy – especially in healthcare settings, where voices are often raised to command attention, or to express urgency.
The currently available computational, digital acoustic-modelling systems are based on algorithms with greater predictive accuracy for large theatre and concert spaces, yet low accuracy for small spaces such as patient rooms or emergency bays. New acoustic modelling software is being developed to overcome these limitations; promising new tunable beam-forming speaker-array systems are being developed to enhance communication more effectively where needed, and masking only where desired, using narrow beams to avoid unwanted masking effects. These techniques will afford designers the means to control unwanted sound distribution without use of walls or physical barriers that impede access, or obscure views.9

Figure 3: Unacceptable sound levels in critical-care settings. Sound-level equivalent (Leq) averages in decibels (dB(A) = blue lines) and impulse sounds (green lines) as a function of time and clinical-care function Figure 4: The National Patient Safety list of ‘sound-alike’ and ‘look-alike’ medications associated with serious adverse events reveals a great need to create visual (pink labels) and auditory environments to reduce error

A view through space
Designers have typically considered the visual domain as the primary stimuli driving the human response to design. Several authors hypothesise that innate responses based on evolutionary pressures may account for design preference. A preference for places of prospect and refuge is thought to derive from adaptation to survival in a savanna ecosystem, where sightlines to predators are essential.10

It is further suggested that places of refuge, characterised by visual occlusion, appeal where safety is desired; however, places lacking permeability that limit escape or foresight of oncoming intrusion may induce a sense of fear. Completely enclosed spaces such as rooms with no windows, or confined spaces such as an MRI tube, can elicit a strong sense of discomfort, and evoke agoraphobia or claustrophobia.11 In contrast, places of prospect, with a broad visual access from a single vantage point, may reduce stress.

It is also hypothesised that the design of spaces with high visibility and connectivity with multiple vantage points may promote environmental comprehension, and entice curiosity and engagement with a place.12 Varying levels of visual and locomotive permeability have been found to determine probabilistic movement. Biederman and Vessel13 suggest that the neurobiological system that rewards learning, via endomorphin release in parts of the brain involved in memory of place (parahippocampal cortex), may encourage exploration. Therefore, views that hint of the presence of concealed information and locomotive accessibility may, in fact, entice learning and increase preference of environmental experiences.

Analytic programmes, such as Space Syntax, have been widely used to map the relationship between spatial interconnectivity, visual vantage point and architectural form to the inclination to travel in a particular direction, and predict the aspects of space most likely to be learned.14 Such methods, however, do not strongly consider the significance of vertical dimension and sense of volume, or the ease with which an observer may acquire spatial knowledge as they move their head position through several vantage points. In order to address these issues, systems have been developed to test the response of subjects to specific features of design while they move through immersive, stereoscopic CAVE visualisations of full-scale photo-realistic buildings.15 Head-tracking systems allow for more natural exploratory movement and multiple vantage points, in comparison with computer navigation of a digital model on a desktop screen. Several spatial prototypes can be used, each one expressing a different balance between visual permeability and occlusion, and variation in depth layers evaluated.

Figure 5: Virtual simulation of the Salk Institute is modified to test preference for permeability and depth
Hamilton12 tested this methodology in a pilot evaluation using a virtual model of the Salk Institute for Biological Studies, La Jolla, California, designed by Louis I Kahn (Figure 5). In this case study, specific design attributes were modified to offer different degrees of permeability and occlusion. A forced-choice assessment of paired spaces around a single vantage point indicated initial preference, followed by free movement, a post-test survey and open-ended questions to reveal changes in preference after exploration. Spatial-quality ‘measurands’ [quantities intended to be measured] were then analysed to determine if qualities other than permeability and occlusion were primary determinants of visual preference. This study found spaces that achieved a volume of visual permeability from 20 to 40% were preferred. No participant preferred the most occluded space, in which the total volume of visual permeability was only 13%.  

There was also a preference for visual volume and permeability that permitted the greatest depth of view. In the post-test survey, trends revealed preference of both visual complexity and order, which may be consistent with the reward for learning in a complex environment, and the desire tocomprehend an ordered, permeable and easy-to-learn environment.

Greater understanding of visual attributes may be gained by emerging technologies using eye-tracking systems that can test visual attention to design features, presented as a factor of depth and visual tracking. For example, Jansen et al.16 showed that saccadic eye movements tend to follow depth cues. Visual-attention tracking that includes depth, as well as the dimension of time, have been tested in virtual-reality CAVE simulations, where carefully controlled calibration can begin to map visual attention to specific architectural features.17 Such measures are more accurate than tracking head movements alone, as eye and head movement are not always in alignment. Ongoing studies combining these technologies will add clarity to initial findings and enable more specific exploration of the preferred balance between occlusion and permeability, complexity and order, and a vast range of design attributes.

Memory in real and virtual places
A user’s attention to many architectural features (visual, auditory, tactile, kinesthetic, etc.) can be studied by recording the subconscious ‘experience’ of design. Wearable and wireless bio-sensors, combined with environmental sensors, can track the body, brain and behaviour, as subjects are exposed to controlled elements of real or virtual buildings. For example, technologies that record heart rate and electroencephalographic (EEG) brainwaves can reveal and correlate cognitive reactions to specific design features, during wayfinding strategies, attention, concentration, relaxation or stress. Edelstein et al17 used a 256-electrode array to reveal activity of individual components of the brain’s cortical areas, as subjects navigated through CAVE simulations of real environments. Tracking systems logged the subjects’ head position and view angle, as well as movement within the full-scale digital mock-up. Differences were observed in the frequency spectrum and intensity of responses during spatial navigation when the subject was in a space absent of cues to location, versus a setting rich with navigational cues.

In this early study, significantly stronger synchronisation in theta brainwaves and stronger desynchronisation of the lower alpha brainwave frequencies were observed in areas of the cortex that play a role in spatial and visual orientation (parietal and occipitotemporal areas). The parietal cortex uses visuo-spatial information from a first-person perspective, along with parietal and occipitotemporal areas involved in processing changes in direction and planning of future paths. Disorientation associated with increased alpha-wave desynchronisation likely reflected increased demands on attention.

In contrast, most psycho-physiological studies use desktop screens to show small architectural visualisations of 3D digital buildings. In such simulations, the subjects must imagine themselves interacting within the building – a difficult task for many clients, students and professional designers, even. In this way, virtual CAVE simulations, in which micro-sensors monitor subjects as they move through full-scale architectural spaces, clearly offer more ‘ecologically relevant’ simulations.

Integrating light
The human visual system is not merely engaged in sight but also in the integration of light to assess the body’s exposure to diurnal and nocturnal patterns. A body of neuroscientific and clinical research, dating from before the 17th century, reveals that exposure to light has significant impact on mental state, cognitive function, behaviour and physical health. Recent epidemiological studies suggest that elevated cancer rates in nurses, night-shift workers and flight crew may be related to unnatural patterns of light or dark exposure. Measures of neuroendocrine levels reveal the correlation between daily fluctuations in melatonin, which modulates sleep and wakefulness, and elevated cortisol levels, which prepare the body for activity.

Consistent with these findings, Edelstein et al18 found significant differences in heart-rate variability (HRV) – a well-established indicator of health risk and stress – during performance of memory tasks when subjects were exposed to less than 15 minutes of red light, versus bright white (with a blue peak) light. Whereas many studies have focused on the influence of blue and bright white light on melatonin responses, this experiment demonstrated that red light is associated with changes in cardiac responses. In red light, HRV relaxation was appropriate during rest, and activated only during the memory task. In contrast, bright white light with a blue peak was associated with a constantly active heart rate throughout the experiment. In a parallel study, brainwaves recorded via a 256-electrode EEG array tended to be different during red-light conditions versus bright white-light conditions in a single-subject self-control study.19

Such research suggests that sustainable-design guidelines should include findings that reveal the spectral range, intensity and pattern of light important to human health and function, as well as vision. Rather than guidelines that suggest average light levels across an entire building, or propose percentages of exposure, lighting design should respond to the specific needs of the users, in addition to the uses of a space. In this manner, programming of spaces for night-shift workers, such as clinicians, factory workers, aircrew or business travellers, could be prioritised when planning access to spaces with natural light.

Furthermore, spaces for control of light and access to darkness would also drive design decisions. The unwanted distribution of light into places occupied by others should be a primary consideration in lighting design for healthy places. Rather than a ‘one size fits all’ approach, thoughtful lighting strategies should provide for safety and egress, as well as individual controls to modulate light exposure according to clinical needs, functional tasks and individual circadian status.19

Figure 6: Dr Edelstein navigates through the virtual-reality stereoscopic model of the Calit2 Atkinson Hall in the StarCAVE
Synchronous measurements of human responses, including both biosensors and, environmental sensors, and behavioural-tracking techniques, now offer the means to explore architectural issues that have, to date, remained as design hypotheses. Using a variety of emerging technologies in pre-design studies and post-occupancy evaluations, rigorous research can inform the design of environments that support, rather than impede, health and wellbeing.
An interdisciplinary neuro-architectural framework for design thinking employs such technologies, merging scientific methods with observational, ethnographic, sociological, psychological, physiological and medical results. This approach is particularly relevant to healthcare facilities, which serve the most fragile and most gifted. Moreover, healthcare environments represent all architectural types, being, as they are, places of healing and health, teaching and learning, and business and rest. Similarly, as places that encounter birth, death, discovery and recovery, healthcare facilities must meet the broadest of human challenges.

As healthcare design increasingly incorporates sustainable-design guidelines, we can apply the evidence derived to address human needs that go beyond reduction of noxious and toxic exposures. Architectural, technical and medical knowledge can, in this manner, accelerate such best practice to enhance human experience, performance and health itself. These applications of new technologies sit at the interface between neuroscience and architecture, and enables the provision of more rigorous data for research-based design. The ultimate goal is to support the design of healthy places for all: the healthy, the infirm, the gifted, and those with special needs, and to promote and enhance health and wellbeing across all peoples.

Dr Eve Edelstein MArch PhD (Neuro) EDAC Assoc AIA F-AAA is a research affiliate at the University of California San Diego and teaches neuro-architecture, healthcare design, healthy cities and universal design studios at the College of Architecture & Planning and Landscape Architecture, the Institute of Place and Wellbeing, at the University of Arizona, Tucson.

The author thanks Ramesh Rao, Eduardo Macagno, Randall Peterson FAIA, John Eberhard FAIA, Norman Koonce FAIA and the AIA Latrobe Fellowship for support and seed funding. The author gratefully acknowledges the many contributors to the development of the systems, technologies and studies, including the sonic-simulation team of Peter Otto, Nicholas Echols, Tashiro Yamada, and Suketu Kamdar; the visualisation team, including Thomas DeFanti and Jurgen Schulze; the CAVECAD team, including Lelin Zhang, Cathy Hughes and Daniel Reirden; the virtual bio-engineering team, including Gert Cauwenberghs, Michael Chi, Lelin Zhang and Cory Stevenson; the wayfinding team, including Klaus Gramman, Elke van Erp, Andrey Vankov and Nimo Bigley; and, finally, the circadian HRV-studies team, including Julian Thayer, John Sollers III, Robert Ellis, Tzyy-Ping Jung and Reason Song.

©2018 All Rights Reserved. Website Design Graphic Evidence