Users’ Awareness of Alarm Settings

Analyzing the User Interface of a Physiological Monitor
  • Kathrin Lange
  • Armin Janß
  • Siyamak Farjoudi Manjili
  • Miriam Nowak
  • Wolfgang Lauer
  • Klaus Radermacher
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 818)


Whether or not a physiological monitor will issue an alarm in a certain condition depends largely on the particular settings of the device. To be able to safely use the monitor, the user has to be aware of these settings. We addressed a potential contribution of interface design to users’ awareness of devices settings by analyzing a monitor’s user interface. Based on a previous analysis of incident reports, we selected the following functions for further analysis: Deactivating individual alarms (SpO2), changing alarm limits (arhythmia), muting alarm volume and completely disabling the measurement of a particular parameter (blood pressure). We applied two different methods of assessing interface usability: an analytical approach supported by the formal-analytical method mAIXuse and an empirical approach, i.e. observing whether expert users in a simulated care setting were aware of the various causes underlying alarm failures. In the simulation study, seven experienced intensive care nurses took part. The mAIXuse analysis showed that detecting altered alarm limits, the muting of alarms and the disabled measurement should be somewhat less easy than detecting the deactivation of an alarm. With regard to comprehension, alarm limits and disabled measurement were judged to be inferior to muting and blocking of alarms. During the usability-test, not a single user identified the muted alarm and the altered alarm limits as potential causes for the absence of an audible alarm without being prompted. By contrast, three of seven nurses directly recognized that the SpO2 alarm was blocked. The results of the formal-analytical analysis and the simulation study will be compared and conclusions regarding the contribution of the monitor display to users’ awareness of alarm settings will be discussed.


Physiologic monitoring alarms User computer interface Awareness 

1 Introduction

1.1 To Safely Use Alarms, Users Have to Be Aware of Device Settings

Intensive care nurses have to be aware of the vital status of the patients assigned to them. The monitoring of vital functions is assisted by specific devices (physiological monitors) and largely automated. The device measures and interprets vital parameters and provides an alarm to alert the user, if it infers the presence of a critical physiological condition, e.g. insufficient oxygen saturation. In addition, devices monitor for technical conditions that compromise their intended functioning, for instance the detachment of an electrode. Theoretically, the automatic monitoring of vital parameters and use of alarms allow care givers to immediately become aware of critical conditions without watching the patient uninterruptedly. This is necessary, because there are situations, in which nurses have to rely on alarms – at least for a certain while. Examples include the need to care for a second patient or performing other tasks off the patient’s bed. To be able to rely on alarms in these situations, nurses have to be aware of the factors affecting whether or not a certain physiological condition will trigger an alarm.

1.2 Users of Monitoring Devices May Not Always Be Sufficiently Aware of Alarm-Related Settings

There is evidence that awareness for alarms and alarm behavior of physiological monitors is not as straightforward as it should be, as a recent analysis of incident reports regarding “alarm failure” in physiological monitors that were issued to the Bundesinstitut für Arzneimittel und Medizinprodukte [Federal Institute for Drugs and Medical Devices] (BfArM) shows [9]. Between 2009 and 2015, 184 alarm-related incidents with physiological monitors were reported. The alarm-related malfunctions were missing alarms (144/184), ceasing alarms (21/184), wrong alarms (16/184) and alarms not saved (3/184). Crucially, only a small proportion of these malfunctions had a technical cause (16/184). Rather, more than half of the reported problems (101/184) reflected incorrect assumptions or expectations regarding device functioning. Often, users were not aware of the complex conditions that determined the specific alarm behavior or whether the monitor would issue an alarm, at all.

Table 1 gives an overview of the monitor settings mainly involved in the reported incidents. In another 58 cases the root-cause could not be identified, 9 cases were due to failures during implementation or maintenance.
Table 1.

Selected monitor settings that have contributed to missing alarms that were interpreted as device failure and reported to the BfArM, see [9] for details

1.3 Users’ Awareness of Alarm-Related Device Settings Depends on Many Factors

Particularly in a complex work environment such as critical care, users’ awareness of device behavior or device settings may be influenced by a variety of factors (see also Fig. 1). Two of these seem pivotal: Firstly, users need a precise mental model of the device. Secondly, they have to be able to retrieve and actively use this knowledge if necessary. The users’ mental model of the device (i.e. device knowledge) includes, for instance, its functionality, its individual components and their interplay, and the external factors that may influence device functioning [7]. This requires an adequate representation of the respective information in long term memory. For active use, the information has to be retrieved and maintained, which is related to attentional control and working memory functions.
Fig. 1.

Ishikawa diagram showing different factors influencing awareness of device settings (not exhaustive)

Obviously, long-term memory, working memory and attentional control are primarily associated with the individual. However, many characteristics of the work environment impact on these cognitive functions. For instance, there are situational factors, all of which contribute to mental workload and working memory demands, such as the complexity or difficulty of the current task(s), the frequency of distractions, the multitasking demands, or the presence of time pressure. Organizational influences include training opportunities, staffing levels, standards and protocols regarding device settings or the choice of equipment (regarding both individual devices and the composition of the device pool). Examples for team factors are communication (e.g. of changes to device settings) or the safety culture on the ward (how strictly does the personnel adhere to protocols). Even characteristics of the patient may have an effect, e.g. the specific appearance of the measured bio-signals or the recent alarm-history. Finally, the device itself has to be considered, for instance the number and complexity of functions, the precise algorithms used, the design of the user manual or – crucial for the present paper - the design of the user interface.

Evidently, there are many interdependencies between the individual factors depicted in Fig. 1 (not shown). For instance, it is safe to assume that an individual’s device knowledge (a feature of the individual) is influenced (amongst other things) by the number and complexity of functions of the device (a device property), by the frequency and quality of the training (an organizational factor), and by whether only a single device model is used for the purpose in question (also an organizational decision). For related discussions of these topics see also [8, 9]. Likewise, users’ awareness of specific alarm settings may depend on the interaction between device knowledge, attentional orienting and the design of the user-interface (a device property) and may be expected to be further mediated by the current demands on working memory. In the present paper, we focus on the user-interface and its suitability for enhancing users’ awareness of alarm-settings.

1.4 Potential Contribution of the Interface to Users’ Awareness of Device Settings

The display of the user interface is the means of a device to convey information regarding, amongst others, its current state. This is particularly important for highly automated functions. Even automated functions require user interactions at some point (particularly, when automation breaks down and things go wrong). To take over responsibility, the user must be able to quickly get an overview of the critical parameters of the system. A well-designed user interface facilitates the integration of this information in the user’s momentary mental representation of the system– their “situation awareness” [3]. It provides the information necessary to develop adequate situation awareness in the most effective way. What is most effective depends on the demands of the use-case. Information on certain parameters may be presented in a relatively raw format that leaves the interpretation to the user, for instance displaying the heart rate value or a value indicating current alarm volume. Alternatively, the display may include an interpretation of the information presented, such as a sign Bradycardia or a crossed bell, indicating that the heart rate is lower than a pre-defined value and that there will be no audible output from the device, respectively. In principle, both formats allow the user to become aware of the same information: A specific heart rate and a specific volume setting. However, only the latter format contributes to comprehension, i.e. understanding the meaning of these values. If only the raw data are given, they have to be combined with knowledge stored in long-term memory to attain the same level of awareness or understanding. While this may be effortlessly achieved by expert users in standard situations, greater difficulty may be experienced by novices or if mental demands are particularly high, thus loading on working memory. Another important aspect, which may be particularly important for complex systems, is the challenge associated with the need to deal with large numbers of variables and the interdependencies between these variables. Of course, users have to be provided complex information. However, overly complex displays compromise the perceptual salience of the information that is actually presented [2, 4].

The user interface of a physiological monitor is only one facet of the entire “device-patient” system a critical care nurse has to keep in mind. The display of a physiological monitor includes graphical and numerical representations of the bio-signals measured, together with their respective alarm settings (e.g. status of alarm or alarm limits). Complexity results from the fact that a typical device monitors a large number of different vital functions. Moreover, for each of these functions, there are many parameters influencing alarm behavior, only some of which are explicitly displayed. Our goal was to investigate how well displays of currently used physiological monitors support users’ awareness of alarm-related device settings. We started by investigating a particular monitor model of a manufacturer frequently used in German hospitals, for which several alarm-related incident reports were issued. Based on an analysis of these reports [9], we chose four settings for evaluation: muting alarm volume, disabling the measurement of a bio-signal, blocking the alarm for a particular bio-signal, and changing alarm limits of a bio-signal. These settings differ with respect to whether they affect all or only selected bio-signals and alarm-conditions and how the current setting is displayed by the device (Fig. 2 and Table 2). We evaluated how well the monitor’s user interface supports awareness for these functions, both using a formal analytical approach (mAIXuse) and by empirically assessing how easily users become aware of the individual settings if they search for the reason why an expected alarm did not sound.
Fig. 2.

Overview of the alarm-related monitor settings addressed in the present study before and after changes were performed

Table 2.

Overview of alarm-related settings selected for evaluation


Bio-signals affected

Alarm conditions affected

Other indicator of alarm condition?

What indicator on display?

Absolute indicator?

Alarm volume






Measurement of bio-signal




Absence of readings/values

Not applicable

Blocking of alarms




Symbol next to parameter value

No (position on screen)

Alarm limit



Not applicable

Limit values next to parameter value

No (position on screen, semantic relation)

2 Methods

2.1 Formal-Analytical Approach: mAIXuse

In the framework of the usability evaluation of the graphical user interface (GUI) of the patient monitor, which has later been used for the interaction-centered usability tests, the mAIXuse method has been conducted initially, in order to get further impressions regarding the usage deficits. In a two day workshop the GUI of the patient monitor has been evaluated with the mAIXuse methodology at the Chair of Medical Engineering.

The formal-analytical mAIXuse methodology [5] for usability evaluation and human risk analysis is based on a two-folded approach. On the one hand, HiFEM provides a task-specific modelling structure (user- and system-tasks) with additional integrated temporal relations and on the other hand the methodology enables to analyze the use process regarding human-induced risks and to document the results in a FMEA-conform (FMEA - Failure Modes and Effect Analysis) data sheet. Within the task modelling process, the investigator gains a systematic overview of the high-level operations of the user, the system and especially the Human-Machine-Interaction. These tasks are subdivided into “Perception”, “Cognition” and “Action”. In contrast to traditional task analyses [1], within the mAIXuse modelling not only hierarchic dependencies are presented, but with the help of temporal relations (e.g. sequence, concurrency, disabling, sharing, choice, etc.) coequal tasks are linked with each other [6]. The methodology is based on several standards for the development of medical devices (e.g. ISO 14971:2013 - Application of risk management to medical devices and IEC 62366:2015 - Application of usability engineering to medical devices). The approach can be used from early developmental stages up to the validation process.

In the present study, we assessed perceptual, cognitive, and motor qualities of display elements underlying selected alarm-related monitor settings within the framework of the mAIXuse analysis. In order to compare the results of the formal-analytical approach and the usability-evaluation, the same device settings were chosen for both approaches: turning off alarm volume, deactivating measurement of blood pressure, blocking the SpO2 alarm and changing alarm limits.

The mAIXuse analysis provides failure causes (in the perceptional and the cognitive level of the human information processing) and potential failures (in the motoric action) within use process steps and classifies these potential failures and causes according to various human error taxonomies. In the subsequent human risk analysis the parameters causes, failures and consequences (which characterize the risk priority number) are documented in a FMEA standard sheet. Finally, countermeasures for the different parameters of the risk priority number are systematically identified and documented as well.

2.2 Empirical Approach: User Performance in a Simulated Use Scenario

To empirically investigate, how awareness of alarm settings is supported by the monitor’s user interface, we simulated a use scenario, in which participants were asked to identify reasons for the absence of audible alarms, given a deterioration of the patient and certain device settings. Seven nurses from two different intensive care units of University Hospital Aachen took part after giving their informed consent. All had experience with the monitor that was used (between 1 month and 10 years) and no one had used models by other manufacturers before.

The setting consisted of two rooms at the Chair of Medical Engineering, simulating patient rooms. Each was equipped with a physiological monitor, syringes, and dressing material. Room 1 was equipped with a manikin that could be used for performing the task “changing dressings”. Note, however, that a fully functional manikin, allowing for the simulation of alarm conditions, was not available (this will be discussed below).

During the test scenario, participants were asked to perform a selection of tasks typically performed on an ICU. They started with the initial control of the patient and the monitor, which is mandated at the hospital at the beginning of a shift. Performance of the initial control served as a manipulation to evaluate, whether the participants engaged in the simulation. As a next step, participants took blood, performed a blood gas analysis and started to change the dressings on patient A (see also Fig. 3). Triggered by a syringe pump alarm (played from a recorded sound), they were required to switch to room 2 to perform a different task, there (calculate medication dose, prepare syringe, administer drug).
Fig. 3.

Procedure of use simulation

While the participant was absent from room 1, the experimenters changed several alarm-related monitor settings (see also Fig. 3):
  1. a.

    Alarm volume was changed from 7 to 0

  2. b.

    Measurement (and display) of blood pressure was deactivated

  3. c.

    The alarm indicating oxygen de-saturation was blocked (i.e. neither visual nor acoustic alarms would be presented)

  4. d.

    Alarm limits for heart rate were altered (lower limit from 50 to 30, upper limit from 120 to 190)


Having addressed the alarm in room 2, the participants returned to room 1 and resumed changing dressings. They were told that they had been away for 15 min to indicate the possibility that a third person might have altered device settings in the meantime. Shortly before finishing their task and while facing the patient rather than the device, the participants were told that the patient’s status had deteriorated, but that the device had not alarmed. At this point, they were asked to search for potential causes of this absence of alarms.

3 Results

3.1 mAIXuse

The formal analytical assessment separately evaluates the contribution of perceptual, cognitive, and motor functions to using a particular display element for a selected task. For the present purpose (awareness of alarm settings), we focused on the perceptual (detection) and the cognitive (comprehension) evaluation.

The mAIXuse analysis came to the conclusion that the symbols indicating alarm volume settings are difficult to detect. Likely causes for the low salience are the small symbol size as well as the insufficient color coding (light grey) of the symbol associated with volume settings, per se. As for the qualitative transition implied by a volume setting of “zero”, this is not associated with a prominent symbol change (e.g. in color or contrast) to capture the user’s attention. This may be regarded a hazard, because when alarm volume is turned off, only visual alarms are available. Not becoming aware of this fact implies a major impairment to the system, as a whole. By contrast, symbol comprehension at a cognitive level was not regarded much of a problem. Because the clinical users are training accordingly, they may be expected to understand that the icon (a crossed bell, IEC 60417-5576) together with the empty volume indicator represent volume “0”.

As for the deactivated blood pressure measurement, the graphical user interface (GUI) does not provide explicit feedback on the absence of the parameter, i.e. the blanking of the blood pressure signals on the screen is the only indication. This implies that the user’s attention is not actively called to the disabled measurement. Becoming aware of the blanked blood pressure signals requires the users to know that this measurement had been active before, thus involving top-down processing. Hence, even detection may not be regarded purely perceptual but rather involves cognitive processes like memory or attention. As an additional problem at the cognitive level, the user has to understand all the implications of the disabled measurement signals (“no measurement”, “no alarms”). Again, because of the complete absence of any explicit symbol, the GUI does not provide any cue to facilitate awareness. At a basic perceptual level, the system’s feedback regarding the blocking of the SpO2 alarm can be easily detected. However, the mAIXuse analysis yielded some difficulty understanding the meaning of the symbol at a cognitive level. Two causes may be responsible for this. Firstly, the symbol representing deactivation of alarms was judged to be ambiguous. Secondly, the mode of alarm deactivation for the respective parameter/signal presentation is not distinct. Confusingly, the vital signs presentation is continuously displayed in an unmodified way. Countermeasures e.g. would be to visually mark the signal itself (e.g. with a different signal color or different background color), the blocked parameter itself or the complete display area (with a border strip), where the signal is presented. However, similar to the icons indicating the muted volume, trained users may be regarded familiar with the icons indicating “blocked alarms” (crossed alarm symbol, IEC 60417-5319) and may, therefore, effortlessly decode its meaning.

Perception of alarm limits was evaluated to be difficult for the user due to the small size and the light green color. The combination of the selected color with the black background is not recommended according to usability design guidelines. Additionally, the system allows setting alarm limits to unreasonable values and does not provide visual feedback if this is the case, i.e. the user interface does not offer any cue to facilitate comprehension. As a consequence, in order to understand the implication of the specific alarm limits, users have to actively relate the displayed value with their stored knowledge.

3.2 Device-Related Performance in the Use Simulation

Participants engaged in the simulation situation somewhat differently: The majority (4/7) thoroughly checked the initial settings of the monitor and even changed certain values. 2 of the 7 participants extensively viewed the settings, but did not alter anything. One participant only performed a cursory control.

With respect to the alterations in alarm-related monitor settings, detection differed between functions: Almost all participants detected the blocking of the SpO2 alarm either directly or upon being prompted and still a majority noticed directly or after a prompt that blood pressure was no longer measured (Table 3). By contrast, not a single participant detected the altered limits of the arhythmia alarm or the muted alarm volume – and 4 of 7 participants never found these potential causes for missing alarms.
Table 3.

Number of participants, who detected changes at the different stages

Detection of …




Muted alarm volume




Disabled blood pressure measurement




Blocked SpO2 alarm




Changed limits of arhythmia alarm




4 Discussion

We aimed at assessing how well the interfaces of currently used physiological monitors support users in obtaining awareness of the device’s alarm-related settings. These settings are important, because they determine whether or not a particular physiological or technical condition will result in an audible alarm. Full awareness of these settings determines the users’ situational expectations for alarms and, therefore, their capability to safely use the alarm function. Triggered by device issues identified in incident reports, we performed an initial assessment of how selected alarm-settings are displayed by one particular model of a physiological monitor. The procedure presented here may serve as a blueprint for following-up on universal issues regarding device use identified by incident reports.

4.1 Comparing the Results of the Formal Analytical Assessment and the Simulation Study

Perceptual Aspects of Selected Display Parts: Detection of Relevant Settings

For three of the four settings, the formal analytical assessment yielded problems with salience: the muted alarms, the alarm limits and the disabled measurement. For muted alarms and alarm limits, reasons are the small size, the low luminance and the low color contrast of the respective elements of the display. For the disabled measurement, information is not actively presented on the display, i.e. attention cannot be exogenously drawn to the absence of measurement. Only the display element indicating the blocking of alarms was considered to be salient, particularly because of its high contrast and luminance. In line with this assessment, the blocked alarms were almost always detected by the participants of the simulation study, whereas more than half of the participants never recognized that alarms may not have sounded because the alarms were muted or because alarm limits had been changed.

Of course, display salience is not the only factor to determine visual information sampling. After debriefing, several participants indicated that volume settings are never changed at their hospital. Therefore, they did not expect this parameter to vary. It may be hypothesized that this affected their top-down attentional control and information sampling strategy, i.e. they simply did not check for the possibility of a muted alarm. This is a particularly adverse constellation, because attending to a potentially critical element of the display is supported by neither bottom-up nor top-down factors. Whereas the volume settings may not be of critical value, per se, at least the complete muting of sound may be indicated with greater salience (e.g. by changing its contrast or color), to overcome the hazard associated with missing this critical information. In a similar vein, the high detection rate for “blocked alarms” may not have been caused by a particularly salient display but rather by the participants’ expectations, based on their previous experience: In fact, the only participant, who did not identify that alarms were blocked, had the least experience as an intensive care nurse.

Cognitive Aspects of Selected Display Parts: Comprehension of Relevant Settings

The degree to which a display element captures the user’s attention is just one aspect of display usability. To achieve full situation awareness, users not only have to perceive the relevant elements of the environment – they have to be able to effortlessly draw the correct conclusions, i.e. comprehend the meaning of their percepts. As discussed above, this should be particularly easy, when the display element in question includes the complete information necessary for interpretation. Comprehension may require more effort, if the displayed element has to be related to other information on the display or to knowledge from long term memory. We regard the general muting of alarms and the blocking of visual and acoustic alarms of a particular parameter to be examples of the former: Both are indicated by a single symbol - a crossed bell (IEC 60417-5576) and a crossed alarm-symbol (IEC 60417-5319), respectively. Once the meaning of these symbols is learned, it may be almost automatically derived. This is an intuitive process for the crossed bell, whereas some learning may be necessary for the crossed alarm symbol. However, because both signs are defined in IEC 60417 and are used frequently on medical devices, professional clinical users should not have difficulty with comprehension.

Comprehending the actual meaning of a particular pair of alarm limits (or changes to them) may not be achieved as easily. These values have to be interpreted relative to, for instance, default values or patient-specific criteria. To conclude that the current settings of the device are the reason why an expected alarm does not sound, users may not rely on a single element of the display but rather have to relate the current parameter value to the alarm limits. In a similar vein, to conclude that no information at all is available for a parameter, which is not displayed on the screen, is not directly evident from this absence: Users need to possess a well-developed mental model of this function and they must be able to recall the respective information from long-term memory. The latter is a question of their current mental workload. Notably, for both the disabled measurement and the changes to alarm limits, there were several participants (3 and 4, respectively), who never recognized these settings as potential causes for the absence of an alarm, although they had been asked several times, if the settings were safe for further monitoring. Whether this was due to the limited salience of the respective display parts (see above) or suggests difficulties in comprehending the associated displays cannot be concluded from the present data.

4.2 Limitations of the Current Study and Future Directions

Our current study was only a first step towards investigating the different factors that contribute to users’ alarm-specific situation awareness. While our study revealed some interesting points, several aspects will have to be addressed by future research. For instance, it would be interesting to compare different device models of the same or different manufacturers to identify, which specific features of display design may particularly contribute to awareness of settings. Another aspect relates to the ecological validity of the simulation. There is some evidence that participants did engage in the situation, since all but one thoroughly conducted the initial control. The simulated tasks were typical for those performed on an ICU, as was the interruption and call to a different patient, such that one patient could no longer be monitored visually (see also [10]. Similarly, four of the seven participants indicated that situations, where colleagues changed device settings without notification, are conceivable, one had even personal experience. Nevertheless, future studies should increase the ecological validity of the simulation. This may include using several networked monitors, increasing overall workload and time pressure or presenting typical ICU-noise to approximate perceptual and cognitive demands of the simulation to those of a typical work environment [10]. Noticing the changed settings was measured using a graded system of verbalized conscious detection (direct identification, prompted identification, no identification). Future studies should make use of additional measures to less obtrusively assess awareness of device settings, e.g. by analyzing eye movements or pupil dilation to assess the distribution of users’ attention and their information sampling behavior, respectively, see also [11, 12].


  1. 1.
    Diaper D, Stanton NA (2004) The handbook of task analysis for human–computer interaction. Lawrence Erlbaum Associates, LondonGoogle Scholar
  2. 2.
    Endsley MR (1996) Automation and situation awareness. In: Parasuraman R, Mouloua M (eds) Automation and human performance: theory and applications. Lawrence Erlbaum, Mahwah, pp 163–181Google Scholar
  3. 3.
    Endsley MR (1988) Design and evaluation for situation awareness enhancement. Proc Hum Factors Soc Annu Meet 32:97–101CrossRefGoogle Scholar
  4. 4.
    Endsley MR (2011) Designing for situation awareness: an approach to user-centered design. CRC Press, Boca RatonCrossRefGoogle Scholar
  5. 5.
    Janß A, Lauer W, Radermacher K (2009) Bewertung sicherheitskritischer Systeme im OperationssaalEvaluation of Risk-Sensitive Systems in the OR. i-com Zeitschrift für interaktive und kooperative Medien 8:32Google Scholar
  6. 6.
    Janß A, Radermacher K (2014) Usability first. Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz 57:1384–1392CrossRefGoogle Scholar
  7. 7.
    Kluwe RH (2006) Informationsaufnahme und Informationsverarbeitung. In: Zimolong B, Konradt U (eds) Ingenieurpsychologie. Hogrefe, Göttingen, pp 35–70Google Scholar
  8. 8.
    Lange K, Brinker A, Nowak M et al Patientengefährdungen durch Gerätediversität? Diskussion eines Risikofaktors anhand der Ergebnisse zweier Befragungen an deutschen Kliniken. Der Anästhesist (in press)Google Scholar
  9. 9.
    Lange K, Nowak M, Neudörfl C et al (2017) Umgang mit Patientenmonitoren und ihren Alarmen: Vorkommnismeldungen liefern Hinweise auf Probleme mit Gerätewissen. Zeitschrift für Evidenz Fortbildung und Qualität im Gesundheitswesen 125:14–22CrossRefGoogle Scholar
  10. 10.
    Lange K, Nowak M, Zoller R et al (2016) Boundary conditions for safe detection of clinical alarms: an observational study to identify the cognitive and perceptual demands on an Intensive Care Unit. In: de Waard D, Brookhuis KA, Toffetti A, Stuiver A, Weikert C, Coelho D, Manzey D, Ünal AB, Röttger S, Merat N (eds) Human factors & user experience in everyday life, medicine, and work. Proceedings of the human factors and ergonomics society Europe chapter 2015 annual conference. ISSN 2333-4959 (online), pp 195–208Google Scholar
  11. 11.
    Steelman KS, Mccarley JS, Wickens CD (2011) Modeling the control of attention in visual workspaces. Hum Factors 53:142–153CrossRefGoogle Scholar
  12. 12.
    Wickens CD (2015) Noticing events in the visual workplace: the SEEV and NSEEV models. In: Szalma JL, Scerbo MW, Hancock PA, Parasuraman R, Hoffman RR (eds) The Cambridge handbook of applied perception research. Cambridge University Press, Cambridge, pp 749–768CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Federal Institute for Drugs and Medical Devices [BfArM]BonnGermany
  2. 2.Chair of Medical Engineering in the Helmholtz-Institute for Biomedical EngineeringRWTH Aachen UniversityAachenGermany

Personalised recommendations