A longitudinal health interview survey in rural Kenya: potentials and limitations for local planning.

A twelve-month longitudinal household health interview survey in Machakos District (now Makueni District), Kenya, during 1991 covered 390 households randomly selected from 12 village clusters. The survey focused on recent disease symptoms and signs, illness severity, temporary disability and care-seeking behaviour. The total number of reported disease episodes was 9,393, (4.4 per person) with more episodes reported by adult females than by adult males. The disease pattern was dominated by malaria (39.6% of all reported episodes) and respiratory tract diseases (23.1%), followed by gastrointestinal illness (10.7%), joint/muscle disorders (6.4%), injuries (5.4%) and skin conditions (4.8%). Self-medication took place in 39.9% of episodes, while care was sought at hospitals or clinics in 32.1%. This survey generated information useful for local health care planning and management, especially regarding local perception of illness episodes and health care utilization. Respondents developed signs of interview fatigue, however, and the completeness and accuracy of symptom descriptions by the lay interviewers are uncertain. Survey costs were about USD 24,700, one third of which was spent on field work, another third on computerized data processing. Utility in relation to costs is likely to be modest. Improvement of the health information system for local planning and management may be equally or better served by selective improvement of the existing routine reporting system combined with occasional cross-sectional household surveys.