Health interview surveys in developing countries: a methodological review

Abstract

This review covers cross-sectional health interview surveys in which respondents were asked about their recent illness and their use of preventive and curative health services. Country survey results differed widely, for both morbidity prevalence estimates and health services utilization, reflecting major methodological differences as well as any true differences that may exist between the population groups studied. Comparison of morbidity and utilization rates is thus thwarted by the absence of standardization in survey methodology, methods of analysis, and the classification of results. A lack of theoretical studies upon which to base methodology choices combined with a general lack of rigor in applying the methodologies chosen has limited the interpretation of many surveys carried out to date. The review concludes with a series of specific recommendations for improving the survey methodology of future health interview surveys in less developed countries.

PIP: This review covers cross-sectional health interview surveys in which respondents were asked about their recent illness and their use of preventative and curative health services. Any attempt to compare morbidity rates between studies is thwarted by the lack of standardization of survey methodology and methods of analysis and classification. This review of the literature has highlighted the most important methodological issues to consider when planning health interview surveys. The primary objective of health interview surveys should be to produce useful results for health service planners. The study design should be tailored to fit the detailed objectives of the survey, which should be explicitly stated in quantitative terms. Once the questionnaire has been finalized, multistage, random sampling techniques should be used to determine the sample population. Interviewers should reside in the study area. The order in which the interviewer asks the respondent questions can affect the results substantially. Surveys must clearly specify whether events that started prior to the recall period, but extended into it, should be included. Some index of illness severity should be included in the questionnaire. Morbidity results should be reported by a symptom or grouped into functional systems. The internal consistency of completed questionnaires should be checked within 24 hours.