Selection of study households where indoor air was monitored
Previous studies have identified several potential determinants of exposure to IAP: fuel type, time spent in cooking, cooking location, structural characteristics of houses and household ventilation practices (opening of windows and doors etc.) All of these factors may be important in Bangladeshi households, which exhibit significant diversity in cooking fuels, stove types, cooking locations and ventilation characteristics of houses. Discussion with local experts revealed widespread use of gas, electricity, kerosene, firewood, cow dung, rice husks, straw, jute sticks, bagasse and sawdust as fuel; four cooking locations (separate attached, separate detached, outside/open, single room dwelling – no separate kitchen); and thatch, tin, mud and brick as common structural materials of houses.
For this research, we used stratified sampling in urban and peri-urban areas of Narshingdi (Dhaka region) to incorporate representative variations in fuel use, cooking arrangements and structural characteristics that affect ventilation.2 We separated the households into groups defined by cooking fuel, kitchen type and location, and construction material. Then we selected households independently from each group. Our sample size was 236, given cost constraints.
Monitoring of indoor air
At each household, PM concentrations in the cooking and living areas were monitored for a 24-hour period during December 2003 – February 2004. We used two devices for monitoring indoor air: (1) a real-time monitoring instrument, the Thermo Electric Personal DataRAM (pDR-1000) (Thermo Electron 2004), and (2) a 24-hour instrument, the Airmetrics MiniVol Portable Air Sampler (Airmetrics 2004). The pDR-1000 uses a light scattering photometer (nephelometer) to measure airborne particle concentrations. The operative principle is real-time measurement of light scattered by aerosols, integrated over as wide a range of angles as possible. This instrument operated continuously for 24-hour periods, recording PM10 concentrations at 2-minute intervals. The Airmetrics MiniVol Portable Air Sampler, on the other hand, is a more conventional device that samples ambient air for 24 hours. The MiniVols were generally programmed to draw air at 5 litres per minute through PM10 particle size separators. The particles were caught on the filters, and the filters were weighed pre- and post-exposure with a precisely calibrated microbalance at Airmetrics, Inc. The readings of pD-RAM and MiniVol air samplers provide a detailed record of IAP concentration in each household.
Household questionnaire administration
A short questionnaire was administered to each household on the same day as the air monitoring to obtain information on socio-economic characteristics of the household members, fuel type, fuel quantity, stove location, cooking time, number of people cooked for, duration of fire after cooking, the use of iron, mud, thatch and concrete for construction of the house and kitchen, the placement and size of windows, doors and ventilation spaces between walls and roofs, ventilation practices such as opening doors and windows after cooking, smoking practices, and the use of lanterns and mosquito coils. In addition, all members of the households were questioned regarding their time activity pattern: average hours spent in the cooking area, living areas and outdoors in a typical day to get an indication of exposure.
Regression analysis to explore the determinants of IAP
A regression analysis for these 236 households was conducted to explore the determinants of IAP. Households’ PM concentrations were regressed on fuels used during the monitored day, cooking time, duration of fire after cooking, number of people cooked for, stove location, the use of iron, mud, thatch and concrete for construction of the house and kitchen, the placement and size of windows, doors and ventilation spaces between walls and roofs, ventilation practices such as opening doors and windows after cooking, smoking practices, and the use of lanterns and mosquito coils. Among these variables, a small set were found to significantly affect household PM concentrations: fuel type, stove locations, building materials, and opening doors and windows after cooking.
Extrapolation and exposure reconstruction
Since we have analysed the determinants of IAP using a stratified sample of urban and peri-urban households in the Dhaka region, the sample was not intended to represent all Bangladeshi households. In order to assess the broader implications, a representative household survey was conducted in six districts in six geographical regions: Rangpur in the Northwest, Sylhet in the Northeast, Rajshahi and Jessore in the West, Faridpur in the Centre, and Cox's Bazar in the Southeast. Table 2 documents total population, urban population and total number of villages in each district.
n each district, we have attempted to randomly survey 25 peri-urban households, 25 urban households and 50 rural households.6 The same household questionnaire was administered to each household to obtain information on socio-economic characteristics of the household members, fuel, cooking time, structural characteristics, ventilation practices and time activity pattern of each household member. The regression results for Dhaka region were then extrapolated to estimate cooking and living area PM levels for this random sample of 600 households. Outdoor 24-hour PM concentrations were monitored for a number of peri-urban and rural monitoring points. Respondent estimates of exposure duration (in cooking areas, living areas and outside) from the household surveys were then combined with the respective estimates of pollutant concentration to estimate exposure to PM10. We estimated IAP exposure for family members by age/sex group, with a particular focus on young children.