Design of tooltips for data fields: A field experiment of logging use of tooltips and data correctness

Abstract

Many health professionals in developing countries carry out tasks which require a higher level of education than they have. To help such undereducated health workers filling correct data in patient information systems, data fields were furnished with tooltips for guiding users. In a previous study with questionnaires and interviews, health workers preferred tooltip contents being normal values of the data with medical explanation as the second best. The experiment reported in this paper set out to test these content alternatives and also aimed at finding health workers’ use of tooltips and possible effects on data correctness. In order to resemble the work setting, each of the 15 undereducated health workers participating was given a tablet PC with the patient information system and booklet of 22 cases to be entered over a period of two weeks. They were given a one hour introduction to the system. Their use of the tablet was recorded, and after completing, the participants were interviewed. The health workers opened tooltips frequently for the first cases, and thereafter the use dropped. Reasons given were that they learnt the data field during the first cases, and thereafter they did not need the tooltips so often. The number of correct data entries increased over time. The group with medical explanation tooltips performed better than the group with normal value tooltips, thus the preferred tooltip in the questionnaire gave a lower performance than the second alternative. While the experiment demonstrated that tooltips improved performance, it did not quantify the effect.