Reliability of reporting on lifestyle and agricultural factors by a sample of participants in the agricultural health study from iowa

PURPOSE: Information on agricultural practices has been obtained by questionnaire in several epidemiologic investigations. This project evaluated the reliability of self-reported information on pesticide use and various demographic and lifestyle factors among a group of farmers from Iowa. METHODS: 2,921 Iowa farmers participating in the Agricultural Health Study completed enrollment questionnaires approximately one year apart. Responses on the two questionnaires were compared for percent agreement and by Kappa statistics to evaluate reliability. RESULTS: Percent agreement for ever/never use of specific pesticides and application practices was quite high and generally ranged from 70% to over 90% and did not vary by age or educational level. Kappas were typically in the 0.50 to 0.60 range. Agreement was lower (typically 50% to 60%) for duration or frequency of use of specific pesticides. Agreement on lifestyle and non-agricultural factors was comparable to that reported in other studies. CONCLUSIONS: Level of agreement regarding pesticide use in this population is similar to that found for diet, physical activity, and medical conditions, which have been successfully evaluated in many epidemiologic studies. Information on agricultural practices from self-completed questionnaires has sufficient reliability for use in epidemiologic investigations.