In public health, we often talk about efficacy and effectiveness as if they are two stops on the same train line. Efficacy is the controlled trial: the intervention works under close supervision, with trained staff, regular follow-up, and participants who meet strict inclusion criteria. Effectiveness is the messy real world: the same intervention, now handed to an understaffed urban primary health centre, a construction workers’ settlement, or a heat-exposed street market in Ahmedabad or Delhi. The distance between those two stops is not a technical glitch. It is a landscape of caste, migration status, water access, informal labour, and the daily arithmetic of survival. For those of us working on environmental health equity in Indian cities, that distance is where the real work happens.
This article is about that gap. It is about why a respirator distribution programme can reduce silica exposure in a pulmonary clinic study but fail in a stone-cutting cluster in Rajasthan. It is about why a heat-health early warning system can look excellent in a pilot and then go silent in a slum resettlement colony where phones are shared, electricity is intermittent, and outdoor work cannot stop. And it is about what participatory health data, collected with rather than about communities, can reveal when standard evaluation metrics cannot.
I write this as someone who has spent years moving between research sites and neighbourhood health camps, between data tables and door-to-door conversations. I have learned to be suspicious of any intervention that looks too clean on paper. Cleanliness, in our context, often means that someone has quietly excluded the hardest cases: the migrant worker without an Aadhaar-linked address, the woman who cannot leave her stall to attend a follow-up, the adolescent who breathes diesel fumes for ten hours a day and is counted as a “non-responder” when his peak flow reading does not improve.
Efficacy, Effectiveness, and the Missing Third Term
Let us start with definitions, because the words matter. Efficacy asks: does this intervention work under ideal conditions? Effectiveness asks: does it work in the real world? But in Indian cities, I would add a third term: equity of reach. Does the intervention work for the people who bear the heaviest environmental burden, or only for those who are easiest to reach?
Consider a study on improved cookstoves. A randomised controlled trial in a rural district may show a 40% reduction in household air pollution among women who receive the stove, training, and monthly maintenance visits. The stove is effective in the trial. But in an urban informal settlement, the same stove may fail because the household buys fuel in small daily quantities, the roof is too low for the chimney, the landlord forbids structural changes, or the woman’s cooking schedule is dictated by a domestic worker’s shift that starts before dawn. The stove is not the problem. The problem is that the intervention was designed for a household that can control its own space, time, and fuel supply. That household is not the norm in the settlements where I work.
This is not a failure of the community. It is a failure of the study’s external validity: the assumption that findings from one population can be generalised to another without accounting for structural differences. External validity is not a statistical nicety. It is the difference between a health intervention that works in a study and one that works in the world.
Why the Gap Is Wider in Environmental Health
Environmental health interventions are especially vulnerable to the efficacy-effectiveness gap because exposure is not a single event. It is continuous, cumulative, and shaped by infrastructure that no individual can control. A tuberculosis patient can take a six-month course of drugs and be cured. A waste picker cannot take a six-month course of clean air. Her exposure to particulate matter, heavy metals, and biological hazards is renewed every morning when she reaches the landfill or the sorting shed.
This means that environmental health interventions often depend on behavioural compliance in contexts where the behaviour is not freely chosen. A study may show that wearing a properly fitted N95 mask reduces particulate exposure by 80%. But the study does not account for a construction worker in Chennai who earns ₹450 a day, works in 40-degree heat, and finds the mask unbearable after two hours. The mask works in the study. It does not work in the world because the world includes sweat, dust, humidity, and the need to shout instructions to a co-worker over the noise of a concrete mixer.
There is also the problem of measurement. In a study, we can place a personal air monitor on a worker’s collar and collect eight hours of continuous data. In the world, that monitor may be removed during a tea break, shared with a colleague, or left at home because the worker fears it will mark him as a troublemaker. The data we collect in studies is often cleaner than the reality it claims to represent. That cleanliness is seductive. It makes us think we understand more than we do.
What Participatory Methods Reveal
This is where community-led health data becomes essential. When residents of an industrial neighbourhood in Howrah map their own symptoms, water sources, and factory emissions, they produce a different kind of evidence. It is not a randomised controlled trial. It is a situated account of how exposure is experienced, not just measured. And it often reveals what standard studies miss.
I remember a participatory mapping exercise in a peri-urban area near a chemical storage facility. The official health survey had found no significant increase in respiratory symptoms. But when women in the community drew their own maps, marking where children played, where water pooled, where the wind carried odours, and where family members slept, a pattern emerged. The official survey had sampled households by random selection. The participatory map showed that the most affected households were clustered along a drainage line that the official sampling frame had treated as a boundary, not a pathway of exposure. The intervention that followed, a community-designed drainage cover and a revised school schedule to avoid peak odour hours, would never have emerged from the official data alone.
Participatory methods are not a substitute for rigorous epidemiology. They are a corrective to its blind spots. They remind us that health data is not neutral. It is collected by someone, for someone, with assumptions about who counts and what matters. When those assumptions are made visible, the gap between study and world begins to close.
Three Examples from Indian Cities
1. Heat-Health Early Warning Systems
Ahmedabad’s heat action plan is often cited as a success. It includes early warnings, hospital preparedness, and public awareness campaigns. Studies have shown a reduction in heat-related mortality since its introduction. But the plan works best for people who have access to the warnings: those with smartphones, television, or radio, and those who can alter their work schedules. For a street vendor who must stand at a traffic signal from noon to 8 p.m., the warning is information without power. She cannot close her stall. She cannot afford to lose a day’s income. The intervention works in the study of mortality statistics. It works less well in the world of her daily heat exposure.
What would a more equitable heat intervention look like? It might include shaded rest stations at major intersections, drinking water kiosks that are open during peak heat hours, and a cash transfer programme that compensates informal workers for lost income on extreme heat days. These are not clinical interventions. They are structural ones. But they address the actual mechanism of harm: not a lack of information, but a lack of options.
2. Silica Dust and Stone Cutting
In stone-cutting clusters, silicosis is a well-documented occupational disease. Wet cutting and local exhaust ventilation can reduce silica dust exposure dramatically. In a controlled workshop, these measures work. But in the informal sector, where a worker may cut stone in a small shed behind his house, the cost of water, the lack of electricity for ventilation, and the absence of enforcement mean that the intervention never reaches him. A study may show that wet cutting reduces silica levels by 90%. The world shows that wet cutting is not used because water is expensive, the stone dries too slowly, and the buyer wants the product delivered the same day.
Here, the gap is not about knowledge. The workers know that dust is harmful. They have seen their colleagues die. The gap is about economic structure. An intervention that ignores the price of water, the pressure of deadlines, and the power imbalance between worker and contractor will remain a study finding, not a lived reality.
3. Household Air Pollution and Urban Cooking
Improved cookstove programmes have a long history in India. Many have shown reductions in kitchen smoke in controlled settings. But in urban slums, the kitchen is often a corner of a single room, the fuel is whatever is cheapest that day, and the stove must be portable enough to move when the landlord threatens eviction. A fixed chimney stove is useless. A stove that requires dry biomass is useless when the only available fuel is damp scrap wood from a construction site. The intervention works in the study of rural households with stable tenure and predictable fuel supplies. It does not work in the world of urban precarity.
What Would a World-Ready Intervention Look Like?
I do not have a simple answer. But I can describe the features that seem to matter, based on years of watching interventions succeed and fail in Indian cities.
First, a world-ready intervention is co-designed. It is not designed in a university and then adapted in the field. It is designed with the people who will use it, from the first question to the final evaluation. This is slower. It is messier. It produces interventions that look less elegant in a journal article. But it also produces interventions that people actually use.
Second, a world-ready intervention accounts for time and money. If a health behaviour requires an extra hour a day, that hour must come from somewhere. If it requires an extra ₹50 a week, that money must come from somewhere. Interventions that ignore these costs are not neutral. They are quietly regressive, asking the poorest people to pay the highest price for health.
Third, a world-ready intervention is embedded in infrastructure, not just in individual behaviour. A mask is a behavioural intervention. A properly ventilated workplace is an infrastructural one. A health warning is behavioural. A shaded bus stop is infrastructural. We need both, but we have spent far too much effort on the former and far too little on the latter.
Fourth, a world-ready intervention is evaluated by the people it is meant to serve. This does not mean abandoning quantitative methods. It means adding qualitative and participatory methods that ask: did this help you? What made it hard? What would you change? The answers are often uncomfortable. They reveal that our elegant intervention was actually a burden, or that the benefit we measured was not the benefit people wanted.
The Role of Community Health Data
Community-led health data is not just a tool for evaluation. It is a way of shifting power. When a residents’ association in an industrial neighbourhood collects its own air quality data, it can challenge the official narrative that “levels are within permissible limits.” When waste pickers document their own injuries and illnesses, they can demand that the municipality recognise them as workers with rights, not as invisible hands in the city’s waste stream.
This data is often dismissed as unscientific. But that dismissal misses the point. Community data is not trying to replace epidemiological surveillance. It is trying to make visible what surveillance cannot see: the lived experience of exposure, the daily trade-offs, the symptoms that never reach a clinic, the deaths that are never recorded as occupational. In a city where official data is often incomplete, outdated, or inaccessible, community data is not a supplement. It is a necessity.
I have seen this data change conversations. A group of women in a resettlement colony kept a simple diary of when the nearby factory’s emissions were worst, when their children coughed, when the water tasted strange. The diary was not a peer-reviewed study. But when they presented it to the pollution control board, the board could not dismiss it as anecdote. The pattern was too consistent. The women were not asking for a study. They were asking for action. And the diary, humble as it was, gave their demand a weight that no external researcher could have provided.
Why This Matters for Climate Adaptation
Climate change is widening the efficacy-effectiveness gap. Interventions that worked in a stable climate may fail in a hotter, more erratic one. A heat action plan designed for a city that experiences 10 extreme heat days a year may be inadequate for a city that now experiences 40. A drainage intervention designed for a 1-in-50-year flood may be overwhelmed by a 1-in-10-year flood. The assumptions embedded in our studies are becoming less valid, faster.
This means we need to be humble about what we know. A study is a snapshot, not a prophecy. It tells us what worked, for whom, under what conditions, at a particular moment. It does not tell us what will work next year, in a hotter city, with a more precarious workforce. The only way to know that is to keep listening: to the workers, the residents, the women who cook over smoky stoves, the men who cut stone without water, the children who play near drainage lines. They are not the subjects of our studies. They are the experts on their own exposure.
What I Have Learned
I have learned to be suspicious of certainty. I have learned that a p-value is not a promise. I have learned that the most important question is not “does this intervention work?” but “for whom, under what conditions, and at what cost?” I have learned that the people who live with environmental hazards are not waiting for us to save them. They are already collecting data, already organising, already demanding change. Our job is not to lead. It is to listen, to share what we know, and to be honest about what we do not.
The gap between a health intervention that works in a study and one that works in the world is not a gap in knowledge. It is a gap in power. Closing it requires more than better study designs. It requires a different relationship between researchers, communities, and the state. It requires us to treat local knowledge as evidence, not as anecdote. It requires us to design interventions that fit the world as it is, not as we wish it were.
That is the work. It is slow, and it is humbling, and it is the only kind of work that has a chance of mattering in the cities where I live and work.
Frequently Asked Questions
What is the difference between efficacy and effectiveness in health interventions?
Efficacy refers to how well an intervention works under ideal, controlled conditions, such as a clinical trial with careful supervision and selected participants. Effectiveness refers to how well it works in real-world conditions, where resources are limited, populations are diverse, and many factors cannot be controlled. In Indian cities, the gap between the two is often wide because of informal labour, unstable housing, and uneven infrastructure.
Why do environmental health interventions often fail outside of studies?
Environmental health interventions often fail outside of studies because exposure is continuous and shaped by structural factors that individuals cannot control. A mask or a cookstove may work in a trial, but in the real world, heat, cost, time pressure, and lack of enforcement can make the intervention impractical. The intervention itself may be sound, but the conditions that made it work in the study are not present in the community.
How can community-led health data help close the gap?
Community-led health data helps close the gap by revealing what standard studies miss: the daily patterns of exposure, the trade-offs people make, and the symptoms that never reach a clinic. When residents collect their own data, they can challenge official narratives, demand action, and shape interventions that fit their actual lives. This data is not a replacement for rigorous research, but a corrective to its blind spots.
What makes a health intervention more likely to work in the real world?
A health intervention is more likely to work in the real world if it is co-designed with the people who will use it, accounts for their time and money constraints, is embedded in infrastructure rather than relying solely on individual behaviour, and is evaluated by the community it is meant to serve. These features make interventions slower and messier to develop, but far more likely to be used and sustained.
A Note on Where This Leads
This article is part of a longer conversation on this blog about environmental health equity in Indian cities. In future pieces, I want to look more closely at specific participatory methods: how to run a community exposure mapping exercise, how to design a worker-led occupational health survey, and how to translate community data into demands that municipal bodies will actually hear. I also want to explore the ethics of community data: who owns it, who can use it, and what happens when it contradicts official numbers. If you have questions or experiences to share, I would like to hear them. The comment section is open, and I read everything.
Dr. Meera Iyer is a public health researcher and writer based in India. She works with community groups, municipal health teams, and occupational health clinics to document environmental exposures and design interventions that fit the realities of urban life.