E.I. Shulman - Ph.D. (Biol.), General Director, Scientific-innovative company «Medical information technology» (Novosibirsk)
A.E. Vlasenko - Engineer-mathematician, MBUZ OT «Sectional medical information and analysis center» (Novokuznetsk)
N.M. Gilina - Ph.D.(Eng.), Professor, Chair of medical cybernetics and informatics, GBOU DPO «Novokuznetsk State Institute of Improvement of Doctors» (Novokuznetsk)

The problem of patients safety is particularly acute in the developed countries. By estimates of Institute of Medicine, in hospitals of the USA annually dies from 44 to 98 thousand patients owing to errors of medical staff. About 50% of adverse drug events it is caused by pharmacotherapy. Use of the clinical information systems (CIS) with decision-making functions leads to reduction of pharmacotherapy errors and adverse drug events. Reduction of errors is reached owing to generation of CIS of the pro-active messages which have received in foreign literature the name "trigger tools". Such messages are addressed to physicians and warn them about existence of risks of ordered pharmacotherapy. Drugs interactions is the most frequent reason of adverse drug events. The objective consisted in studying of influence of pro-active messages of CIS DOCA+ on frequency of ordering interacting drugs to patients. In seven medical organizations of various status and the geographical position (Kemerovo region and Novosibirsk region, Primorsky Krai and Khanty-Mansi autonomous region), more than three years using CIS DOCA+, quarterly dynamics of number of pro-active messages which are received by physicians on the monitor at order of pharmacotherapy is analyzed. Two hospitals are the teaching hospitals, three are city hospitals and two are rural hospital. The message are generated by pro-active function in each case of detection in the list of ordered drugs of interacting couple and warns the physicians about of risk of interaction. Each such event and reaction of the physician were registered in a system database. As it was found earlier, the number of pro-active messages about order of interacting drugs has certain dynamics - decreases over time. Approximation of dynamics received in each organization, showed that two of them are in the best way described by the linear function, two - logarithmic and three - modified exponenty. In all seven organizations the number of messages decreases over time statically significantly. Results of the analysis confirm the assumption that physicians remember received messages and further avoid of interacting couples of drugs ordered, reducing risks of adverse drug events.