How much of nature's benefit is really the nature?

This quarter brought seven systematic reviews on nature, indoor air and sound. They point one way. Contact with nature lowers anxiety and depression, and poor air and noise cost people health, most in the buildings they cannot leave. The direction is old news. What is new is the doubt inside the best studies, over how much of the measured benefit is the nature at all, and how much is the exercise, the setting or the way the trial was built.

How much is really the nature

Esmaeel Saeedy Robat and colleagues built the largest synthesis of the set (Saeedy Robat et al., 2026, Nature Human Behaviour). They pooled 116 systematic reviews, drew 30 into a second-order meta-analysis and worked from a base of 3,870 primary studies and an estimated ten million people. Nature-based interventions cut anxiety by a standardized mean difference of 0.83 and depression by 0.72, both large. Saeedy Robat marks the limit in the same paper. Most of the underlying trials compared nature against doing nothing, so the number almost certainly runs larger than the real effect. Until more trials pit nature against an active alternative, its true size stays open.

That comparison arrived this quarter. Yun Liang and colleagues ran a network meta-analysis of 15 randomized trials covering 809 adults, ranking exercise in nature against the same exercise indoors and against no exercise at all (Liang et al., 2026, Frontiers in Public Health). Green exercise beat the no-exercise control by a standardized mean difference of 0.96, a large effect. Indoor exercise did not: its effect was small and short of significance. Head to head, green exercise beat indoor by 0.75 and ranked first with near certainty. Liang reads the outdoor setting as carrying part of the gain on its own, from a thin base of fifteen trials and 809 people, and writes it as a direction to test rather than a settled dose.

If the setting carries part of the effect, the next question is whether a copy of it does. Yuhan Zhou and colleagues examined portable digital nature, meaning a screen, a recording or a headset in place of the real thing (Zhou et al., 2026, Clinical Psychology Review). Across 36 randomized trials and 2,925 people, the substitute still worked: stress fell by a Hedges g of 0.51, anxiety by 0.67, depression by a weaker 0.26. The channel mattered more than the immersion. For anxiety, sound alone beat both video and full multisensory setups. A recording of birdsong or moving water may carry more of the benefit than an expensive immersive room, which matters for any budget that assumed the reverse.

The fourth review widens the frame to policy. Diya Chakravorty and colleagues reviewed 115 European studies of nature-based solutions and health (Chakravorty et al., 2026, Environments). Almost 88 percent of the quantitative studies reported at least one positive association, and mental health was studied nearly twice as often as physical health. Two gaps sit under the headline. The evidence clusters in the United Kingdom and Spain and leans on short, low-intensity interventions, and almost none of it tracks equity, whether a new park reaches the residents who carry the most risk or the ones already living near green space.

The envelope the building controls

The same quarter carried three reviews on the physical envelope a building sets, its air, its noise and its temperature. These are the levers an owner controls directly, and the evidence behind them is thinner than the confidence with which they get specified.

Jing Liu and colleagues pulled together 196 studies of subway particulate matter from 24 countries (Liu et al., 2026, Journal of Hazardous Materials). Concentrations ran highest in tunnels, then on platforms and during the commute, and station offices reached peaks of 405 micrograms per cubic meter of coarse particulate and 212 of fine, an occupational exposure that sits largely unwatched. Subway air is not ordinary city air. Iron ran up to 66 times its outdoor level, ground off wheels and rails. Station depth, platform doors, ventilation and service frequency set how bad it gets, each of them at the drawing board.

Jingyi Mu and colleagues reviewed 180 studies of how acoustic conditions affect the people most exposed to them, older adults, children, people with hearing loss and hospital patients (Mu et al., 2026, Journal of Multidisciplinary Healthcare). Poor acoustics tracked with worse health, weaker cognition and lower mood across all four groups. Mu then rated the confidence behind most of those outcomes as low to very low under GRADE, the standard scale for how much a body of evidence can bear. The direction holds and the confidence behind it is low, which is worth stating plainly when the conclusion points at spending money on quieter schools, hospitals and care homes.

Fatemeh Paridokht and colleagues reviewed 42 studies of thermal comfort and ventilation in schools by Cochrane method, the one building type whose users cannot leave (Paridokht et al., 2026, Journal of Health and Safety at Work). Carbon dioxide swung widely between classrooms, some rooms above recommended limits, and naturally ventilated rooms ran hotter and stuffier than the guidance allows. Demand-controlled ventilation, which adjusts airflow to how many people are in the room, cut pollutant levels where it was installed. The air improved only where the ventilation actually ran. A system on the drawings changed nothing.

Across the seven the reading is narrow. The direction of the evidence is not in doubt. The size is smaller than the raw figures wherever the comparison was weak, and the new active-comparison and dose studies give a designer something to specify from, which channel and which dose for whom. On the envelope, the certainty is lowest where the stakes are physical and the occupants cannot leave, in the tunnel, the classroom and the ward.

What the industry is doing with it

The trade press spent the month treating the brain-building link as settled enough to price, further than the reviews above will yet go. The Centre for Conscious Design, an advocacy group, published a model of beauty as a health intervention, casting good design as a lever on the body’s stress chemistry. Harvard’s Healthy Buildings group put a number on the case: for a 40-person firm at a 75,000-dollar average salary, it estimates that full healthy-building investment lifts net income by 10.7 percent, against 0.36 percent from energy savings. Productivity carries that estimate, and it rests largely on the group’s own CogFX study of 600 workers across six countries.

The same evidence base also indicts buildings, and the writers who said so this month were specific. In Psychology Today, a researcher on neuroarchitecture documented rooms that make people ill, among them a part of MIT’s Stata Center where a reported third of visitors to one conference room feel dizzy, the off-square geometry setting the inner ear against the eye. In Metropolis, a critic on the autism spectrum walked through Casa Batlló, where 85 percent of the front-of-house staff are now neurodivergent, and reported acute sensory overload in the same Gaudi rooms sold elsewhere as healing. Not one of these pieces links a specific design change to a measured health outcome in a controlled study. The market is pricing a causal chain no one has yet demonstrated, the same gap the new reviews flag from the other side.

The wellbeing evidence also entered the argument the industry most wants settled, the return to the office, and this quarter it came with harder numbers than the usual survey. Science reported on 588,000 US workers from 2011 to 2024: fully remote staff spent an extra hour a day alone and showed sharper rises in mental distress, a real cost on the pro-office side. The counter-evidence was as concrete. Researchers at King’s College London found roughly one in six fathers working fully from home would quit under a full return mandate, up from 3 percent in early 2021, and Wharton researchers spent six years tracing leaders’ resistance to remote work to narcissism rather than to any measured drop in output. Gallup put 71 percent of Gen Z on the side of hybrid, the highest share of any generation. The office now has to earn its case on the evidence, and the people it has to convince have brought evidence of their own.

One physical stake runs under all of it and draws almost no capital. Writing in Metropolis, a critic on the case for shade noted that the United States sets no federal heat standard for outdoor workers, and put the cost of that gap at a projected 55 billion dollars in lost earnings by midcentury as heat rises. Almost none of it gets built.

For anyone deciding where to spend, the quarter narrows the choices. The large nature effects are real in direction, but assume the size is smaller than the headline once an active control is in the picture, and favor what that active work now supports: moving outdoors over the same movement inside, and a sound recording tested before an immersive suite is bought. The envelope money belongs where the evidence is thin and the exposure is high, the air in the tunnel and the classroom, the sound on the ward. And the wellness business case is worth reading closely, a real productivity signal built around a causal claim no controlled study has yet closed.