Using microcomputers for rapid data collection in developing countries

Abstract

The use of microcomputers during a study and closer to the point of data collection is discussed. Examples are given for microcomputer use for computerized morbidity surveillance in Gambia and management of community health in PCEA Chogoria Hospital Kenya. Also examples of field use for anthropometric data collection with a Reitmaier system are provided. The Systeme de Collecte (SIC) project is described in its usefulness in preventing error. A field data collection system (FDCS) on a Psion Organiser XPII is described as it is used in a survey of malaria morbidity in Gambia. It is concluded that there are many solutions to data management but the use of microcomputers is invaluable for longterm surveillance. The potential has not been exhausted. For shortterm large-scale cross sectional studies the use of portable laptop computers with several handheld models linked up has powerful potential. With the new programs being developed it is becoming increasingly easier to analyze long verbatim responses or frequency of words and phrases for analysis of anthropological data. Computers and integrated packages are available for data preparation (preparation of interview schedules and administration interview supervision and management survey analysis and postsurvey data management) and for decreasing the time between questionnaire response and analysis of results. A variety of different computers are mentioned including the BBC microcomputers Compaq Deskpro machines the Sharp PC 1600 computer with a 4-color printer and a cassette recorder or a disk drive for storage IBM compatibles and the handheld Psion Organiser XPII. In the example of the FDCS the computer can store 128K worth of data on each storage device which is slotted in from the rear. Software includes a basic questionnaire specification additional specification and administration which can be accessed with different passwords. Several questionnaires can be stored simultaneously with a limit of 16 characters per questions plus a 100-character supplemental scrolled help text (includes skip options consistency checks and automatic date). Predefined options are available such as date number string or coded. Questions are displayed sequentially and input is checked for errors which are displayed immediately if the response is invalid. In a trial field run in the Gambia the FDCS was compared with paper. The results showed errors of transcription but these were fewer with the Psion by about 50%. Variability between interviewers as well a time were also reduced. The final interviewer feedback was the preference for the Psion.