Skip to main content
Learn · Air Quality & Health

Environmental Health Research Methods

Designing an exposure assessment study — from sensor data to epidemiological evidence

Grades UG 90 min research live IESH data

Undergraduate students design a rigorous exposure assessment study using the sensor hub, grapple with the gap between sensor readings and health-relevant concentrations, and write a study protocol suitable for ethical review.

Learning objectives

  • Design a personal exposure assessment study using wearable or fixed sensors.
  • Apply WHO air quality guidelines to interpret sensor readings in a health context.
  • Identify and mitigate the main sources of bias and confounding in an air quality study.
  • Write a study protocol including background, objectives, methods, analysis plan and ethics.

Background

Epidemiological studies of air pollution use one of two exposure assessment approaches: ambient monitoring (measuring pollution in the outdoor environment, typically at fixed stations) and personal exposure monitoring (measuring what an individual actually inhales over a 24-hour period). Personal exposure is always different from ambient because people move through different environments — home, kitchen, road, office — each with different pollution levels, and they breathe at different rates during different activities.

The WHO 2021 Global Air Quality Guidelines set the following 24-hour mean limits: PM2.5 15 µg/m³; PM10 45 µg/m³; CO 4 mg/m³ (approximately 3.5 ppm); NO₂ 25 µg/m³. Our MQ sensors do not directly measure these concentrations — they produce relative resistance changes that can be converted to approximate ppm values only after careful calibration against certified gas standards. This is a critical limitation that must appear in any study protocol using this device.

Confounding is the central methodological challenge in environmental health research. People who live near roads (high pollution exposure) are also more likely to be poor, to smoke, to have less access to healthcare and to work in physically demanding jobs. Attributing health differences to air pollution alone requires controlling for these other factors — through study design (matching, restriction) or statistical analysis (regression).

Procedure

  1. Read today's MQ-7 and MQ-135 values plus their calculated sensor voltages. Calculate approximate relative CO level as a fraction of the sensor's typical maximum response.
  2. Identify the key gap: this sensor measures resistance changes, not calibrated ppm concentrations. Write a 100-word "Sensor Limitations" section for a study protocol.
  3. Define a specific research question your study would answer — e.g. "Does CO exposure in rural kitchens using biomass fuel exceed WHO 24-hour guideline levels during cooking?" Make it specific, measurable and answerable.
  4. Design the study: what population, how many participants, what measurement frequency, how long, what comparison group, what confounders to measure alongside air quality.
  5. Write a consent and ethics section: what are participants' risks? What data will be collected? Who has access? How will data be stored and anonymised?
  6. Write a complete 800-word study protocol with sections: Background, Research Question, Methods, Analysis Plan, Limitations, Ethics.

Discussion

  1. Your sensor reads a raw ADC value of 600 during cooking. Without calibration data, what can you legitimately claim about CO exposure? What cannot you claim?
  2. A community members asks: "Is the air in my kitchen safe?" How do you answer responsibly given the sensor's calibration limitations?
  3. If you found that kitchen CO levels exceeded WHO guidelines during cooking, what would be the appropriate pathway from sensor data to public health action?

Worksheet

1. Write a 100-word "Sensor Limitations" section for a study protocol using this device. Include at least three specific limitations. [5 marks]

Answer guide (for teachers)

Should include: no calibration against certified gas standards; MQ sensors respond to multiple gases (cross-sensitivity); readings affected by temperature and humidity; no direct PM2.5 measurement; voltage divider correction required; sensor drift over time without recalibration. Award marks for specificity and scientific accuracy.

2. What is confounding? Give one example of a potential confounder in a study comparing air pollution exposure and respiratory health in Nepal. [4 marks]

Answer guide (for teachers)

Confounding occurs when a third variable is associated with both the exposure and the outcome, creating an apparent relationship that is not causal. Example: poverty is associated with both high pollution exposure (biomass cooking, proximity to roads) and poor respiratory health (poor nutrition, limited healthcare access) — failure to account for poverty would confound a study of pollution and respiratory disease.

Vocabulary

Exposure Assessment
The process of estimating the concentration of a pollutant and the duration of contact experienced by study participants.
Confounding
A situation where an observed association between exposure and outcome is distorted by a third variable associated with both.
Ambient Monitoring
Measurement of outdoor air pollution at fixed stations, representing average community exposure rather than individual exposure.
DALY
Disability-Adjusted Life Year — a measure of disease burden combining years of life lost to premature death and years lived with disability.

Extension

Read one peer-reviewed paper on indoor air pollution in South Asia or Nepal (search Google Scholar for "household air pollution Nepal"). Critically evaluate its exposure assessment methods: what sensors or methods did they use, how did they handle calibration, what confounders did they control for, and what limitations do they acknowledge?

← Air Quality & Health