What do we know about wellbeing?November 2016

Written by Roderic Bunn, Building Performance Analyst, BSRIA Sustainable Construction Group

The health and wellbeing bandwagon is well and truly rolling. But what is truly important and how can it be measured? And having measured it, will the numbers make sense, and will they inform better design, construction and facilities management decisions? Roderic Bunn plots a course through the maze.

Improved wellbeing. Greater functionality and effectiveness. A healthier workplace and higher staff productivity. Wouldn’t it be wonderful to put firm numbers to these factors? To be able to determine causes and effects? To say, categorically, that a particular design decision, operational setpoint, or space planning choice will lead to a particular beneficial outcome. To know, for sure, that any given set of built environment decisions will result in happier and more productive people. Maybe lower absenteeism and sickness rates, reduced staff turnover, and less haemorrhaging of expensively procured staff skills and expertise.. “Where do I sign up?,” you might ask.

Absolutely, and who wouldn’t, when wellbeing consultants are claiming they can measure a vast array of environmental factors, do some sums, and present you with a ‘wellness’ rating for your building that you can promote to prospective occupiers? For a price, of course – nothing as radical, extensive and as demanding of expertise as a wellbeing assessment will come cheap. The question is, can it be done? Can the complex cocktail of variables that combine to determine occupant comfort be poured into a statistical centrifuge and separated out to the extent that one can say that doing this, this, and that, will equate to a 15 per cent improvement in wellbeing and make an office 20 per cent more productive than the one next door?

The sobering answer is that it’s probably not very likely. Decades of research into the relationships between physical environments and occupant health, comfort, and productivity have found many correlations with possible sources of that discomfort, such as poor indoor air quality, low air change rates, poor thermal conditions, and so on. But, as for proven and repeatable causal links? No, nothing that can lead to reliable metrics.

Most significant academic work in this field over the last 30 years has been based on large cohort studies, where data gathered from, say, 600 buildings have been concatenated, and statistical calculations conducted to correlate relationships between environmental conditions and self-reported comfort levels. In some cases absenteeism or sickness rates are gathered and found to closely correlate against some factors but not others. When such studies are repeated, usually with different research methodologies, smaller samples, possibly more sophisticated modelling and often increasingly esoteric statistical tests, then correlations between some factors lose their strength while others that were loosely or poorly correlated in previous studies suddenly gain statistical strength. All this is fair enough in the research world: developing hypotheses and testing them is vital to inch the knowledge forward, but don’t look for certainty in cause and effect. You’ll find it in one paper but not in another. Even where strong correlations are found you’re still left wondering “Yes, but how important is it, and does it really matter?”

Relying on findings from laboratory studies is also risky. Laboratory research where mental agility, reasoning and accuracy exercises are conducted in different environmental conditions can lead to interesting productivity results, but those results do not create outputs that translate to the vastly more complex real-world office. Real offices possess many confounding factors that the laboratory researcher has diligently stripped out in order to get purer answers, only to find that the laboratory results then don’t export back to the real world for exactly that reason. Customers of such findings need to be wary of claims that this is possible without loss of fidelity.

Figure 1: BSRIA staff definitions of wellbeing

Most non-clinical research also tends to rely on self-assessed surveys where building occupants report their comfort perceptions via web or paper surveys, and/or in structured interviews. Done well this can be informative, but questions have to be ones that respondents can easily answer. Most people can differentiate between hot and cold conditions or whether they have enough daylight, but abstract concepts like ‘morale’ or wellbeing’ can confuse them. For example, this writer conducted a straw poll of BSRIA staff to define – without them conferring – what wellbeing meant to them. Figure 1 shows a word cloud of all the responses, with some redundant verbs and conjunctive terms removed to better highlight the defining characteristics of the department’s wellbeing perceptions.

It’s immediately evident that ‘health’ was mentioned most, but it’s also interesting that the (admittedly small and not scientific) sample mentioned a wide variety of other factors. Many related wellbeing to a mixture of light, air, space and noise, while others included fabric, infrastructure and how the building services are managed.

All of which is absolutely fine, but consensus on what constitutes wellbeing is clearly difficult to achieve if different people have different ideas swirling around in their heads when they answer the question. That’s because experience is forged by many factors: gender, age, whether they are trapped at their desk all day or free to move, whether they have access to natural light. They may have the ability to control conditions or they may be reliant on centralised control. They might be sitting in a corridor zone being assailed by others walking up and down talking on mobile phones. They may not have understood the survey question anyway, and if they didn’t, the wellbeing surveyor may not be aware of that.

Figure 2: Problems in hypothetical work area

Figure 2 shows the problems inherent in making sense of self-assessed comfort data. Take a hypothetical work area. In this corner of a naturally-ventilated office, occupants have been characterised by their location to openable windows (dark blue seats), those who are part-time (pink seats) those assigned hot desks (yellow seats), those in deep-plan office space with their backs to nominated circulation routes (green seats), and colleagues who enjoy daylight via a rooflight (the yellow box). People represented by bright blue seats sit adjacent to a management office and meeting room and have little visual or acoustic privacy, nor do they have access to windows or environmental controls. They may be frequently disrupted by colleagues using the printer hub, and people wandering to and from the cellular spaces.

Let’s consider ventilation issues. Satisfaction with natural ventilation may depend on the ability to manage the trade-offs between ventilation and draughts, room temperature, radiant temperature, natural light and glare, and views out to the planted external landscape. Arguably such trade-offs can be managed best by those seated in the dark blue seats and closest to the control devices (windows and blinds). A democratic consensus over the position of these devices can usually be easily reached between two to four people – what one can call the control group – but gets progressively more difficult as the control group increases in size to include the orange seats. It may break down entirely with those in the red seats who may perceive conditions very differently to those in the blue seats, but are too far away to exercise control. A group of eight seats perpendicular to the window may be the limit at which democratic consensus can be achieved. Rows of desks longer than this may make consensus impossible. People may start to suffer. They may even become resentful of the building in other ways, which might breed intolerance of noise or other factors that other staff can tolerate.

Dissatisfaction with thermal conditions and indoor air quality may depend on exposure and dose rather than a particular instrumented level. Those only in the building for short periods (the part-timers) may be able to tolerate moderately adverse conditions. The hot-deskers may be also moderately more satisfied if they can choose their desk, or perhaps rotate position on a daily basis to share the benefits (and drawbacks) of a window location. Those sitting under the skylight may suffer downdraughts, but this will be offset by the (greater?) benefits of direct and indirect daylight, and possibly emotionally uplifting views of the sky.

Noise disturbance is often a major issue, particularly in open-plan offices. Those by an openable window are closest to external noise but have the ability to manage its volume. Road works can be suppressed by trading-off against ventilation, while birdsong may be welcomed. Again, those furthest away from the window will have less say over that control. Furthermore, those seated by the perimeter will be interrupted less by cross-talk and general movement in circulation zones, which will be suffered more by those in centrally-located desks. On top of that, noise conditions change during the working day. In open-plan, we all know hubbub can be useful in cancelling out irritating sources of noise. Those sources become a problem once the hubbub recedes.

Arguably, those seated in the green seats have the worst of all worlds. But, their jobs may mean they are less tied to their desks than other workers. They may be nearer the printers or the beverage points – both of which have advantages and drawbacks. Here, they have plants which soften the surroundings. Such characteristics won’t overcome shortcomings in the physical environment, but they might make those people more tolerant of noise, or lack of views out, or whatever.

The point of this example is that occupant wellbeing is highly variable and context-dependent. If one adds the effects of age and gender (particularly relevant for thermal criteria) the situation gets even more complicated. Then there is the availability of personal storage, or whether meeting rooms are readily available or always booked up, forcing irritated staff to improvise. Then add the issue of utilisation – how often all seats are occupied. At normal times occupant density may hover above the British Council for Offices recommendation of 1 person/8 m2, but at peak times it may fall below 1 person/6 m2, which might – might – start to hamper productivity and cause people to exhibit what one might call ‘escape behaviours’ to coffee points, atriums or even a home office. Will this typically human response be healthy one, or reduce productivity? How would you know? It all becomes statistically debateable.

Where does this leave the construction professional seeking greater certainty about the factors that determine wellbeing in any particular building? And, indeed, to know whether any of it really matters? Maybe we can always rely on the innate human ability to adapt to our environmental conditions, rising above all but the worst conditions to do a job of work well and happily.

In their influential 1997 paper Productivity: The Killer Variables1 (recently revised), Bordass and Leaman warned that the cats’ cradle of causality and association [of comfort factors] differs from one building to the next, making it dangerous to be overassertive about causation without careful appreciation of context. They identified five variables that have a critical bearing on the performance of occupied spaces, particularly offices:

  • Personal control: People are generally happier with opportunities for more control and control over personal conditions and greater freedom of movement.
  • Responsiveness: When people want conditions to change, they want that to happen quickly. That can be temperature levels, or a caretaker replacing a flickering fluorescent light.
  • Building depth: Greater depth correlates with greater complication, use of air-conditioning with less personal control, and more dependence on facilities management, which has to be proportionately better.
  • Workgroup size: Smaller and more integrated workgroups tend towards higher perceived productivity. Increased occupant density tends to break down perceptions of being in a smaller workgroup, with benefits of ease of communication with colleagues compromised by higher levels of unwanted noise2.
  • Design intent: How well a building’s features work for its occupants. Features that are invisible, hidden from view, or not easily controllable will not be appreciated.

The inverse of the killer variables are buildings which are comfortable and controllable, with clear and well-communicated design intent, with usable systems that respond quickly to need. Shallow plan is generally better than deeper plan spaces, preferably with some form of natural ventilation. Careful management of zoning and control of density helps with control of many environmental variables, particularly ventilation, lighting and noise.

Will a greater focus on wellbeing in building design and management improve occupants’ actual experience? Maybe yes – it seems reasonable and much evidence suggests so, but don’t expect to be able to guarantee the optimisation of building design by applying generic, rule-based wellbeing metrics. For many contexts, elimination of fundamental causes of discomfort might be the best that can be achieved. Performance might be dictated more by the numbers of people subsequently piled into the building, and, for example, whether those people are happy queuing for unisex toilets at peak times. In an organisation with 60% women, it’s reasonable to say “probably not”, for many reasons too graphic to describe here. Use your imagination. But, does it matter? I mean, really? Definitive statistics or absolute values don’t exist. If the current focus on wellbeing leads to, for example, regulatory rules on occupant density and health and safety limits on persons per toilet, it’s a faintly scary prospect even if well-intentioned. Current data are not yet definitive enough.

My doctoral research, involving longitudinal studies of buildings over periods of up to 20 years, is generating some interesting statistics. In two notable office buildings (one a large multi-tenanted office), people reported spending up to more than an hour extra per day in the building, and more than two hours a day more on screen, than they had reported in the mid-1990s. That’s 10 hours more on screen, per week. This might include tablets and mobiles, but it is more likely to be sitting down in front of a screen (or screens) as these offices are still desk-based environments. How does this one factor inform ‘wellbeing’ assessments? Are people being more productive at the expense of their health? How does increased time on screen relate to their tolerance of the building’s environmental systems? Is it important, or is it not? Maybe they escape to Facebook or dating websites when they’re fed up with Excel. I’m not sure a “wellbeing consultant” would be able to put a weighting to that.

I asked a well-known social scientist what tools he thought most important when researching occupant satisfaction in buildings. Instrumentation? Physical measurements? Occupant surveys? “A functioning set of eyeballs,” he said. He had a point. Buildings aren’t nearly as mysterious as some people like to make out. Most problems are right in front of you, if you bother to look properly. The big question is whether what you’ve found actually matters. And that, as is often the case, is a matter of judgement.

Roderic Bunn is a building performance analyst at BSRIA and a postgraduate doctoral researcher at the Bartlett School of Architecture, University College London.


Explore BSRIA's range of wellbeing services