Basic monitoring in cardiac anaesthesia embraces at least 19 different parameters of haemodynamics and blood gas analysis. In special cases additional measurements may be desirable, providing a total of up to 44 variables displayed on various monitors, as depicted in Fig. 2. The recording of such an amount of data is only feasible with automated recording systems. Therefore, in the past 6 years we have introduced three different computer systems to our cardiac anaesthesia workplaces. The experiences in their handling are reported. MATERIAL AND METHOD. Three systems were investigated: (1) System S 4000 (Siemens, Germany), based on a central processor unit (PDP 11, DEC, Japan) connected with 20 bedside input/output terminals and Sirecust 404a monitors (Siemens). The system collected the data in a ring buffer with a capacity for about 24-48 h. (2) Patient Care Manager (PCM; Siemens, Germany), a single workplace system based on an IBM-compatible personal computer (PC) with the operating system environment DOS 5.0/Windows 3.0. In our test configuration it was connected with a Sirecust 1281 monitor (Siemens). (3) Monitor-Data-Manager (MDM) (our own development). This single workplace system is also based on an IBM-compatible PC running under DOS and was connected to four different monitors used in our cardiac surgery operating theatre (Fig. 2). A second computer (Sirecust S 425, Siemens) served as an interface between the two 404 monitors (not featuring a serial output like RS 232) and the PC. The self-developed program for that interface was memory resistant and executable with three key presses when the anaesthetic record was started. The three systems were compared with regard to their ease of use, function and practicability. RESULTS. System S 4000: Because of the older system architecture, the response time to key inputs was fairly long and the menu structure somewhat uncomfortable. A major drawback was the limited data buffer capacity and the lack of a long-term storage medium as well as the lack of compatibility with the industry standard for PCs. Software interfaces to other companies' monitors were not implemented, limiting the system to the Sirecust 404 devices. Patient-Care-Manager: The user interface is a Windows 3.0 application representing an up-to-date graphical environment. Unfortunately some Windows features were not fully used, e.g. the free positioning and sizing of graphic windows and the color options. Drug inputs were somewhat long-winded, limiting the system's suitability for the operating theatre. The main disadvantage, however, was the lack of interfaces to monitors other than those from Siemens. Monitor Data Manager: The system was designed to sample data from all monitors operating in our hospital's heart surgery department. Each parameter was displayed in a digital form to get close control over the recorded data (Fig. 1b); additionally calculated values like total peripheral resistance or oxygen demand could be drawn from a separate window. Furthermore, key inputs were reduced to minimum, making drug inputs faster than the hand-written protocol. The ease of performing calculations of continuous drug infusions (from microgram/kg/min to ml/h pump speed) was particularly appreciated by the users. Since the data were saved as an ASCII file, they could easily be imported by any spreadsheet like Lotus 1-2-3 or Excel, providing the whole variety of their graphical presentation or calculation features. Because of the high sampling rate (3 min), even short-lasting drug effects could be registered, making the system favourable for scientific studies. CONCLUSION. Automated monitor data record systems are considered to be a prerequisite not only for research in anaesthesia but also for quality assurance. A basic requirement for wide acceptance in clinical practice is a user interface that provides fast and convenient key inputs as well as further information about parameters not displayed on other monitors. In our h