(Taken from application abstract): Efforts at reducing health care costs have led to pressure on primary care physicians to be more judicious in the use of ancillary services, hospitalization, and referrals, thus increasing the need for effective diagnostic decision support. Obstetrician/gynecologists who now can be considered primary care providers are likely to feel this pressure even more intensely. This project will examine the impact of using QMR a diagnostic decision support system, on primary care obstetrician/ gynecologists' diagnostic skills. We will examine (1) the effect of using QMR on diagnostic accuracy and (2) the effects of QMR training and amount of QMR use on physicians' diagnostic performance and skill in using QMR. The project contains three highly interconnected investigational studies. Study I is a randomized controlled trial of the effect Of Using QMR on diagnostic accuracy. Subjects will be randomized into two groups, one that is trained on using QMR and uses it with their own patients for four months (Group 1). The other group will not have access to QMR during that time period (Group 2). Using a set of test cases, we will estimate the unaided diagnostic accuracy and will then compare the diagnostic accuracy of physicians using either QMR in addition to texts for decision support or using only traditional text sources of information. Study 2 is also a randomized controlled trial that will examine the effects of extensive QMR training compared with the usual way that QMR is distributed. Group I described above will receive extensive training and Group 2, who will receive QMR after the first four months and without special training, will be the usual distribution group. We will compare the diagnostic accuracy, as well as frequency of use, skill in using QMR and perceived usefulness of QMR between these two groups. In Study 3 we will examine whether the frequency of use of QMR in the clinical setting affects skill in using QMR perceived usefulness of QMR, and diagnostic accuracy in both groups.
