Indoor gyms near home tracked lower obesity in 164,649 Chinese preschoolers

The claim that neighbourhood design shapes childhood weight has been made for two decades on evidence the researchers themselves describe as inconsistent [s1]. A study published on 28 August brings unusual scale to it — 164,649 Chinese preschool children across 551 cities — and finds one result specific enough to be worth arguing about [s1].

Indoor, not outdoor

Facility accessibility was measured using point-of-interest data, weighted by road-network distance rather than straight-line distance, within residential buffers of 500, 1,000 and 1,500 metres. Multilevel regression accounted for clustering at both the city and kindergarten levels [s1].

Greater accessibility to indoor physical activity facilities was consistently associated with lower odds of overweight or obesity (OR 0.89 to 0.98) and of obesity specifically (OR 0.84 to 0.98) across nearly all buffer distances, each at P<0.05 [s1]. Outdoor facilities showed protective associations too, but primarily at the 1,500-metre buffer after covariate adjustment [s1].

These are small odds ratios. An OR of 0.89 at the strong end of the range is not a public health intervention on its own; across 551 cities it is a population-level signal that survived multilevel adjustment.

The subgroup that gives it away

The stratified results are more interesting than the main effect. The inverse association for indoor facilities was more pronounced among children living in areas with higher PM2.5 exposure, and among children whose mothers had lower educational attainment [s1].

The pollution finding has an obvious reading: where outdoor air is bad, indoor space is where activity actually happens, so proximity to it matters more. That is a coherent mechanism and it is also, strictly, a post-hoc interpretation of a subgroup — the authors report the pattern, and the study design cannot confirm the pathway [s1]. The maternal education finding points the same way as the greenness literature: environmental resources matter most to households with the fewest substitutes.

The design's limits are the standard ones and they are not small. This is cross-sectional. It measures proximity to facilities, not use of them, and not physical activity — no accelerometry, no activity diaries. Families who prioritise activity may choose to live near facilities. And point-of-interest data records that a facility exists at an address, not whether it is open, affordable, or open to preschoolers.

What is happening to the facilities

A US surveillance study published on 19 August supplies the other half of the picture, and its findings run in an uncomfortable direction.

Using adult proxy-reported data from the National Survey of Children's Health across four two-year periods from 2016-2017 to 2022-2023, covering 215,634 youth aged 6 to 17, the authors tracked four neighbourhood built-environment supports: sidewalks or walking paths, parks or playgrounds, libraries, and recreation centres [s2].

Over that span, the proportion of US youth with neighbourhood sidewalks or walking paths and with parks or playgrounds increased, while the proportion with neighbourhood libraries or recreation centres decreased [s2].

The subgroup breakdown is the finding. Increases in sidewalks and parks or playgrounds occurred only among high-income youth. Decreases in libraries and recreation centres occurred only among low-income youth [s2].

Set that against the Chinese result. The facility type associated with the largest reduction in odds of preschool obesity, and associated most strongly among the most disadvantaged children, was the indoor one [s1]. The facility type disappearing from American low-income neighbourhoods over seven years is the indoor one [s2].

The limits on the comparison

These are two different countries, two different age ranges, two different outcomes. The Chinese study measures weight status against facility proximity in preschoolers; the US study measures the prevalence of neighbourhood amenities among 6-to-17-year-olds and does not measure weight at all [s1][s2]. Nothing here shows that closing a recreation centre raises a child's BMI.

Both studies also rely on data with known softness. The US figures are adult proxy report of whether an amenity exists in the neighbourhood — a perception, not an audit [s2]. The Chinese exposure is a commercial points-of-interest database [s1].

What follows

Not advice for a parent, who cannot choose the neighbourhood's facility mix. The relevance is to planning: the Chinese authors' own conclusion is that strategies should optimise the distribution of both indoor and outdoor infrastructure while addressing the social inequalities that limit effective use of it [s1], and the US surveillance suggests that on the indoor side the distribution is currently moving the wrong way for the children with the least [s2].

What to watch is whether any longitudinal study links a specific facility opening or closure to subsequent activity and weight in the same children. Every study in this literature is currently describing a correlation that could be reversed causation, and the natural experiments — closures, openings — are sitting there unanalysed.

Sources

• [s1] Residential Accessibility to Indoor and Outdoor Physical Activity Facilities and Overweight/Obesity among Chinese Preschool Children in 551 Cities. Environmental Research, 28 August 2026. https://doi.org/10.1016/j.envres.2026.125587
• [s2] Changes in US Youth With Neighborhood Built Environment Supports for Physical Activity by Sociodemographic Characteristics, 2016-2017 to 2022-2023. Journal of Physical Activity & Health, 19 August 2026. https://doi.org/10.1123/jpah.2026-0197