Bhatnagar · American journal of epidemiology 2026 · Non-randomized matched-cluster controlled trial (protocol) · n=735

The Green Heart Project: objectives, design, and methods.

Cited 7 times in the scientific literature.

Level 3 - non-randomized controlled study

Protocol/methods description for a non-randomized matched-cluster controlled intervention study.

PubMed 39691946 · doi:10.1093/aje/kwae458 · record verified 2026-08-27

What was done

This paper describes the design, objectives, and methods of the Green Heart Project, a controlled community-based trial initiated in 2018 in Louisville, Kentucky. The study established 16 clusters consisting of 8 matched pairs of target and adjacent control areas within a residential neighborhood of nearly 30,000 residents. The intervention comprised planting over 8,000 mature trees and shrubs in target clusters after Wave 2. Environmental measurements included ultrafine particles, ozone, and nitrogen oxides. Participant exams, biological sampling (blood, urine, hair, nails), psychosocial surveys, and cardiovascular testing (arterial stiffness, flow-mediated dilation) were conducted across Wave 1 (n = 735; 2018–2019), Wave 2 (n = 545; 2021), and Wave 3 (n = 561; 2022, post-planting).

What was found

No clinical or environmental outcome effect sizes are reported in the abstract. The abstract reports operational metrics: designation of 16 matched clusters, enrollment of 735 participants in Wave 1, follow-up of 545 participants in Wave 2, planting of >8,000 trees and shrubs in target clusters, and follow-up of 561 participants in Wave 3.

Why it matters

Urban greening is widely promoted for public health, but high-quality interventional trials with biological and vascular endpoints are rare. This project provides a large-scale, prospective matched-cluster design to evaluate whether neighborhood greening directly alters pollution exposure and cardiovascular disease risk.

Limits

The abstract describes study design and protocol without reporting comparative outcome findings. Matched cluster designation was used instead of true random assignment, which leaves room for residual geographic confounding. Participant sample size declined from 735 at baseline to 561 at Wave 3.

Cited by