Poor urban environments: use of paraffin and other fuels as sources of indoor air pollution.

Method A stratified random sample of 75 houses was used, with stratification into four zones being based on housing type and availability of electricity. Air sampling was conducted at times to include peak fuel use periods in each household using electrochemical Exotox Model 75 continuous monitors (calibrated daily) for sulphur dioxide (SO2), nitrogen dioxide (NO2), carbon monoxide (CO) and temperature and humidity; and Gill Air model 224-XR pumps for total suspended particulates (TSP). Data on fuel utilisation, household ventilation (including opening of doors or windows) during the monitoring period and household characteristics (including chimneys) were collected on the day after the air monitoring using an interviewer administered questionnaire. Maximum hourly average values of SO2, NO2 and CO were modelled against fuel use patterns using linear regression models. Log transformations of the pollutant levels were used to represent the positively skewed distributions of air pollutant levels. Households were categorised according to the patterns of use of paraYn as follows: (a) no paraYn use; (b) paraYn used for any one of the following: cooking, heating, lighting; (c) paraYn used for any two of the above; (d) paraYn used for all three of the above activities. Adjustment was made for crowding (number of people per room in the house) and smoking in the home at the time of monitoring.