Computer-aided DSM-IV-diagnostics - acceptance, use and perceived usefulness in relation to users' learning styles.

Abstract

BACKGROUND: CDSS (computerized decision support system) for medical diagnostics have been studied for long. This study was undertaken to investigate how different preferences of Learning Styles (LS) of psychiatrists might affect acceptance, use and perceived usefulness of a CDSS for diagnostics in psychiatry. METHODS: 49 psychiatrists (specialists and non-specialists) from 3 different clinics volunteered to participate in this study and to use the CDSS to diagnose a paper-based case (based on a real patient). LS, attitudes to CDSS and complementary data were obtained via questionnaires and interviews. To facilitate the study, a special version of the CDSS was created, which automatically could log interaction details. RESULTS: The LS preferences (according to Kolb) of the 49 physicians turned out as follows: 37% were Assimilating, 31% Converging, 27% Accommodating and 6% Diverging. The CDSS under study seemed to favor psychiatrists with abstract conceptualization information perceiving mode (Assimilating and Converging learning styles).A correlation between learning styles preferences and computer skill was found. Positive attitude to computer-aided diagnostics and learning styles preferences was also found to correlate. Using the CDSS, the specialists produced only 1 correct diagnosis and the non-specialists 2 correct diagnoses (median values) as compared to the three predetermined correct diagnoses of the actual case. Only 10% had all three diagnoses correct, 41 % two correct, 47 % one correct and 2 % had no correct diagnose at all. CONCLUSION: Our results indicate that the use of CDSS does not guarantee correct diagnosis and that LS might influence the results. Future research should focus on the possibility to create systems open to individuals with different LS preferences and possibility to create CDSS adapted to the level of expertise of the user.