For the implementation of time-critical decision support al- gorithms in a clinical information system (CIS) a precise relation between medical interventions and effects needs to be established. We evaluated for selected drugs and infusions the relation in time between charted dose and effect on on-line hemodynamic variables. The time of the intervention was compared with the onset of the change of the hemodynamic variables as determined by new time series methods. The average time difference between intervention and calculated hemodynamic effect was 13.23 min (0-29) which did not differ significantly between different interventions. The marked lag between intervention and effect and the great variance of this lag pose an important problem for time-critical decision support. Even after optimizing data acquisition important factors will remain unaccounted for. Therefore, decision support systems may need extensive testing with real-world data before they are released into clinical practice. (supported by the Deutsche Forschungsgemeinschaft, Sonderforschungs-bereich 475 “Complexity Reduction in Multivariate Data Structures”)

Intelligent alarm systems are needed for adequate bedside decision support in critical care. Clinical information systems acquire physiological variables online in short time intervals. To identify complications as well as therapeutic effects procedures for rapid classification of the current state of the patient have to be developed. Detection of characteristic patterns in the data can be accomplished by statistical time series analysis. In view of the high dimension of the data statistical methods for dimension reduction should be used in advance. We discuss the potential of statistical techniques for online monitoring.

We present a robust graphical procedure for routine detection of isolated and patchy outliersin univariate time series. This procedure is suitable for retrospective as well as for onlineidentification of outliers. It is based on a phase space reconstruction of the time series whichallows to regard the time series as a multivariate sample with identically distributed butnon independent observations. Thus, multivariate outlier identifiers can be transfered intothe context of time series...

As high dimensional data occur as a rule rather than an exception in critical care today, it is of utmost importance to improve acquisition, storage, modelling, and analysis of medical data, which appears feasable only with the help of bedside computers. The use of clinical information systems offers new perspectives of data recording and also causes a new challenge for statistical methodology. A graphical approach for analysing patterns in statistical time series from online monitoring systems in intensive care is proposed here as an example of a simple univariate method, which contains the possibility of a multivariate extension and which can be combined with procedures for dimension reduction.

Objectives: Time series analysis techniques facilitate statistical analysis of variables in the course of time. Continuous monitoring of the critically ill in intensive care offers an especially wide range of applications. In an open clinical study time series analysis was applied to the monitoring of lab variables after liver surgery, and to support clinical decision making in the treatment of acute respiratory distress syndrome.

Objectives: Time series analysis techniques facilitate statistical analysis of variables in the course of time. Continuous monitoring of the critically ill in intensive care offers an especially wide range of applications. In an open clinical study time series analysis was applied to the monitoring of lab vari- ables after liver surgery, and to support clinical decision making in the treatment of acute respiratory distress syndrome. Patients and Results: For the analysis of lab variables (blood lactate) in 19 patients after liver resec- tions ARIMA (Auto Regressive Integrated Moving Average) models were developed for an estima- tion period of at least 14 measurements. Prediction values from these models for the following data points were then compared to the actual lab values. With these models in all cases of hepatic compli- cations pathological changes in the lab values could be differentiated from random variance.