The review, led by Yanlin Niu of the Chinese Center for Disease Control and Prevention and colleagues, examined research on the development and validation of heat vulnerability indices. The investigators searched PubMed, Web of Science, ScienceDirect, China National Knowledge Infrastructure and Wanfang Data for peer-reviewed articles published in English or Chinese between January 2010 and October 2020. Their initial search returned 941 records. After duplicates, irrelevant studies and papers that did not test their index against observed health data were removed, only 13 studies met all the criteria. This distinction is crucial: dozens of studies had created or applied an HVI, but far fewer had checked whether the index actually corresponded to real-world outcomes such as mortality, hospital admissions, emergency visits or ambulance callouts. Most of the included studies came from the United States, with additional work from the United Kingdom, Canada, China and South Korea, revealing how heavily the evidence base depends on a small number of countries with detailed health and census data.

An HVI is usually built around a broad definition of vulnerability. In climate science, vulnerability is not simply the amount of heat a person experiences. It combines exposure, sensitivity and adaptive capacity. Exposure describes the heat hazard itself: temperature intensity, the number of hot days, duration, humidity or other measures of thermal stress. Sensitivity refers to how strongly a person or community may respond to that hazard, including age, chronic illness or disability. Adaptive capacity describes the ability to avoid or reduce harm, through air conditioning, access to health care, income, social support, transportation, information and effective public institutions. The reviewed studies grouped their indicators into five major categories: hazard exposure, demographic characteristics, socioeconomic conditions, the built environment and underlying health. In practice, however, researchers often selected variables according to local data availability, previous publications or informed judgment, rather than a universally accepted scientific framework.

Across the studies, the number of indicators included in an index ranged from four to 19. Demographic, socioeconomic and built-environment variables appeared in almost every analysis, while health-related indicators were included in seven of the 13 studies. Heat exposure itself was represented in only five. The factors most frequently used were social cohesion, race or ethnicity, landscape, age and economic status. Social cohesion commonly appeared through measures such as living alone, which can become especially important when an older person is isolated during a heat emergency. Age captures physiological vulnerability: infants and older adults can have less effective thermoregulation, while older people are also more likely to live with cardiovascular, respiratory or metabolic disease. Income and poverty reflect the ability to pay for cooling, relocate temporarily or obtain transportation and medical care. Race and ethnicity may act not as biological explanations, but as markers of structural inequality, unequal access to services, discrimination and historical patterns of housing and environmental disadvantage.

The physical shape of a neighborhood can amplify or reduce heat exposure. Dense areas dominated by asphalt, concrete and dark roofs absorb solar energy and release it slowly, creating urban heat islands that remain hot after sunset. Vegetation can cool surroundings through shade and evapotranspiration, the process by which plants move water into the atmosphere and consume energy as that water evaporates. Several reviewed studies therefore used land cover or vegetation as proxies for environmental protection. Air conditioning was another important adaptation indicator, because it can lower indoor temperatures and reduce heat stress. Yet the researchers emphasized that environmental measurements are not interchangeable. Land-surface temperature, often estimated from satellites, describes the temperature of the ground or roofs, whereas air temperature measures the atmosphere surrounding people. These quantities can diverge because of wind, shading, clouds, surface materials, solar radiation and the angle from which a sensor observes the landscape. Human health is influenced by both radiative heat from hot surfaces and convective heat from the air, so relying on only one can distort the exposure component of an index.

The dominant statistical technique used to construct the indices was principal component analysis or factor analysis, applied in 11 of the 13 studies. These methods transform many correlated variables into a smaller number of mathematical components. For example, poverty, unemployment and low car ownership may load onto a common component interpreted as socioeconomic disadvantage. The resulting component scores can then be combined into an index and mapped across neighborhoods. This approach is attractive because it reduces dimensionality and can limit arbitrary decisions about how much weight each variable should receive. But it also creates interpretive problems. The mathematical components may be difficult to explain in practical terms, and the final index can change substantially when researchers alter the input variables, geographic scale or study population. A neighborhood may receive a high score not because it has one dominant risk, but because several indicators happen to align statistically.

The geographic unit used to calculate an HVI also matters more than it may appear. Eight of the 13 studies relied on census-based areas, including census tracts, block groups and their British equivalents. Other analyses used counties, postal codes or administrative districts. One Shanghai study used a 500-meter grid, offering a much finer spatial resolution. These choices can produce different conclusions because conditions within a single neighborhood are rarely uniform. A census tract may contain both tree-lined streets and heat-exposed apartment blocks, or residents with very different access to cooling and health care. This is a classic problem in spatial analysis known as the modifiable areal unit problem: patterns can change when the same data are grouped into different boundaries or scales. A map constructed at county level may be useful for broad resource planning but miss small pockets of extreme risk. A high-resolution map may locate vulnerable blocks more precisely, but it can also create a false impression of certainty if the underlying health data are sparse.

Validation was the most revealing part of the review. Ten of the 13 studies used mortality data, while others examined morbidity, ambulance callouts, hospital admissions or emergency-department visits. The health datasets covered periods ranging from one to 17 years, with an average duration of eight years, and the number of geographic units ranged from 159 to 4,765 where sample sizes were reported. Researchers commonly used Poisson, logistic or negative-binomial regression to test whether higher HVI scores were associated with more adverse outcomes. Most studies found the expected direction of association: as vulnerability scores rose, health risks tended to rise as well. But the strength of the relationship varied and was usually not strong. In one Dallas analysis, a coefficient of determination, or R², of 0.03 indicated that the index explained only about 3 percent of the variation in total deaths. Other work using supervised and unsupervised principal component methods also produced very low R² values when tested against deaths occurring on extreme-heat days.

That weakness may result from several scientific and practical problems rather than from one failure of the HVI concept. Heat-related mortality is influenced by weather duration, nighttime temperatures, humidity, air pollution, public warnings, access to transportation, social contacts, medication use, housing quality and individual behavior. These factors can vary rapidly and may not be visible in annual census statistics. Health outcomes may also occur at a different location from a person’s home, while deaths can be recorded without enough detail to identify the precise circumstances of exposure. Short validation periods may capture unusual events, and data from one city may not transfer to another with different housing, climate, health systems or patterns of social inequality. The index itself may also be circular if variables are selected because they are already known to correlate with past deaths. A map can therefore appear scientifically sophisticated while still failing to predict how a particular heatwave will affect a particular community.

The review points toward a more demanding future for heat-risk mapping. New indices should incorporate better measurements of actual heat exposure, including both air temperature and land-surface temperature, as well as humidity and the duration and timing of heat events. Underlying health conditions and access to medical services deserve greater attention, because the same temperature can produce very different consequences in populations with different levels of chronic disease or care access. Governance is another overlooked factor. The capacity of local authorities to issue warnings, open cooling centers, conduct outreach, maintain electricity and water supplies and respond rapidly may determine whether vulnerability becomes illness or death. Public awareness also matters: people need to recognize danger, understand protective guidance and have the means to act on it. The researchers argue that validation should use longer periods, more locations and higher-quality health data, while protecting privacy. For now, an HVI is best understood as a decision-support tool—a way to identify where preventive action may be needed—not as a crystal ball. Its maps can save lives when combined with local knowledge and flexible emergency planning, but their predictions must be tested continuously against what happens on the ground.

The study addresses a problem that is frequently obscured by citywide averages. A single temperature reading can suggest that an entire city is experiencing the same conditions, while residents in different neighborhoods may face dramatically different levels of heat exposure. Informal settlements often contain dense housing, limited tree cover, narrow pathways, extensive metal or concrete surfaces and inadequate ventilation. Homes may be built with materials that absorb solar radiation during the day and release heat slowly at night, preventing the body from recovering between consecutive hot days. In many communities, unreliable electricity, limited access to cooling and insufficient water infrastructure further increase exposure. These factors can transform an extreme-weather event into a prolonged physiological stress test.

Heat affects the human body through several interacting mechanisms. Under normal conditions, the body sheds excess heat through radiation, convection and evaporation, especially sweating. During a heatwave, high air temperatures reduce the efficiency of radiation and convection, while high humidity slows the evaporation of sweat. The result is a rising core temperature that can impair cardiovascular, renal and neurological function. The danger is particularly acute at night, when persistently warm indoor conditions interfere with sleep and prevent thermal recovery. Older adults, infants, people with chronic illnesses, outdoor workers and residents taking medications that alter hydration or cardiovascular responses may be especially vulnerable. The study’s focus on infrastructure recognizes that health outcomes are shaped not only by individual behavior, but also by the physical environments in which people live.

Urban infrastructure can influence heat exposure at multiple scales, from an individual dwelling to an entire neighborhood. Roofing materials, wall construction, window placement and ventilation determine how quickly heat enters and leaves a home. At the street level, pavement and unshaded surfaces can absorb solar energy and intensify the urban heat island effect, in which built-up areas become warmer than surrounding rural or less developed landscapes. Vegetation can cool the environment through shade and evapotranspiration, the process by which plants release water vapor into the atmosphere. Water access, drainage systems, sanitation and reliable electricity can also affect whether residents can cool themselves safely, remain hydrated and avoid secondary health threats during extreme weather. An infrastructure upgrade, therefore, is not a single intervention but a network of changes that may alter exposure, vulnerability and the ability to recover.

Prosdocimi, Klima and DeYoreo frame slum upgrading as a potential pathway to climate resilience rather than solely as a housing or economic-development policy. The distinction matters because adaptation measures can produce unequal outcomes if they ignore how neighborhoods actually function. A new road may improve emergency access but also replace permeable ground with heat-retaining pavement. A roof replacement may reduce indoor temperatures, while poorly planned construction could restrict airflow. Tree planting can provide shade, but only if species are suited to local conditions and maintenance is sustained. Electrification may allow households to use fans or cooling systems, yet it can also increase costs or place pressure on fragile power networks. Evaluating infrastructure through a health lens helps reveal these trade-offs before they are treated as evidence of resilience.

The technical challenge is connecting physical changes in the built environment to measurable health effects. Researchers studying heat risk commonly combine meteorological data with indicators such as air temperature, humidity, wind speed and radiant heat. Measures including the heat index, wet-bulb globe temperature and physiological heat-balance models can provide more realistic estimates of human stress than air temperature alone. Indoor conditions may require separate monitoring because walls, roofs and ventilation can produce temperatures that differ substantially from outdoor weather-station readings. Health analysis can then consider outcomes such as heat exhaustion, heatstroke, dehydration, cardiovascular strain, kidney stress, hospital admissions or excess mortality. By evaluating infrastructure in relation to these pathways, the study speaks to a growing field of climate-health research that treats the city itself as part of the exposure system.

The paper is also significant because informal settlements are not interchangeable. Their risks depend on local climate, topography, housing materials, population density, access to services and social networks. A community in a humid tropical region may face dangerous nighttime heat and limited evaporative cooling, while a dry-climate settlement may experience intense daytime solar exposure and severe water scarcity. Hillside neighborhoods may have different airflow patterns from settlements built in low-lying basins. Residents may develop local strategies—sharing water, opening communal spaces or checking on vulnerable neighbors—that reduce risk but are difficult to capture in conventional datasets. Any meaningful assessment of upgrading must therefore account for social capacity as well as concrete, asphalt and building envelopes.

The research arrives as cities and governments search for adaptation strategies that can be implemented before the next record-breaking summer. Large-scale cooling centers, heat-warning systems and emergency medical responses remain essential, but they are often activated only after dangerous conditions have already emerged. Infrastructure upgrades offer the possibility of reducing exposure in advance. Reflective or insulated roofs, improved ventilation, shaded public areas, permeable surfaces, expanded vegetation, dependable water supplies and better access to electricity could function as layers of protection. The most effective interventions may be those that deliver multiple benefits at once, improving thermal comfort while also reducing energy demand, supporting public health and strengthening everyday living conditions.

Yet upgrading carries a political risk: improvements intended to protect residents can trigger displacement. When infrastructure raises land values without guaranteeing secure tenure or affordable housing, existing communities may be pushed out and replaced by wealthier populations. Climate resilience can then become a mechanism of climate exclusion. A health-centered approach must ask not only whether a neighborhood becomes cooler, but also who remains able to live there, who receives the benefits and who bears the cost. The study’s emphasis on slum upgrading therefore connects heat adaptation to environmental justice. Protecting residents from extreme heat requires investment in physical systems, but it also requires policies that preserve community stability and ensure that resilience is not measured solely through property values or redevelopment.

The broader message is that heatwaves expose the hidden architecture of inequality. Climate change raises the temperature hazard, but infrastructure determines how much of that hazard reaches the body. By evaluating the relationship between urban upgrades and health during heatwaves, Prosdocimi, Klima and DeYoreo contribute to a debate that cities can no longer postpone. The future of heat adaptation may depend less on isolated technological fixes than on coordinated improvements to housing, streets, vegetation, water, electricity and public services. If upgrading informal settlements can reduce dangerous exposure while protecting residents from displacement, it could become one of the most consequential climate-health strategies available to rapidly growing cities. As heat records continue to fall, the question is no longer whether urban infrastructure affects health, but whether cities will redesign it quickly enough to prevent extreme heat from becoming a routine cause of illness and death.

Subject of Research: Urban infrastructure, slum upgrading, climate resilience and health risks during heatwaves

Article Title: Upgrading slums as pathway to climate resiliency: evaluating urban infrastructure impact on health during heatwaves

Article References: Prosdocimi, D.O.B., Klima, K. & DeYoreo, M. “Upgrading slums as pathway to climate resiliency: evaluating urban infrastructure impact on health during heatwaves.” npj Urban Sustainability (2026). https://doi.org/10.1038/s42949-026-00442-w

Image Credits: AI Generated

DOI: 10.1038/s42949-026-00442-w

Keywords: Heatwaves, informal settlements, slum upgrading, urban infrastructure, climate resilience, public health, urban heat, environmental justice, climate adaptation