Computer-aided personal interviewing. A new technique for data collection in epidemiologic surveys.

Area of Science:

• Epidemiology
• Health Informatics
• Survey Methodology

Background:

• Traditional epidemiologic studies rely on paper-based interviews, involving manual data coding, transcription, and keypunching.
• This manual process is time-consuming and prone to errors during data entry and analysis.
• Developing efficient and accurate data collection methods is crucial for large-scale public health research.

Purpose of the Study:

• To evaluate the feasibility and acceptance of computer-aided personal interviewing (CAPI) in a community-based health survey.
• To assess the impact of CAPI on data quality and the interview process compared to traditional methods.
• To identify potential challenges and benefits of implementing CAPI in field research.

Main Methods:

• Development of a structured interview system on a portable computer for personal interviewing.
• Implementation of immediate error-checking, automatic skip-patterns, and direct data entry to minimize manual transcription.
• Pilot evaluation using a test/re-test methodology with tape-recorded interviews.
• Large-scale deployment of the CAPI system in a community blood pressure survey, completing 2,505 interviews.

Main Results:

• The CAPI system demonstrated feasibility in collecting data during 2,505 home-based interviews for a blood pressure survey.
• Both interviewers and respondents generally accepted the computer-assisted interviewing method.
• A small percentage of interviews (5%) could not be completed on the computer, with paper questionnaires serving as a backup.
• Pilot testing indicated a slightly higher initial error rate, attributed to system and training limitations, and interviews were slightly longer.

Conclusions:

• Computer-aided personal interviewing is a feasible method for data collection in community-based epidemiologic studies.
• While initial pilot evaluations showed some challenges, the system was well-accepted and effectively managed in a large-scale survey.
• Further research is necessary to fully ascertain the impact of CAPI on overall data quality and to optimize its implementation.