The effect of multitasking on the communication skill and clinical skills of medical students
Mental workload is an abstract concept that perceives cognition as the brain having a small and finite capacity to process information, with high levels of workload associated with poor performance and error. While an individual may be able to complete two different tasks individually, a combination of tasks may lead to cognitive overload and poor performance. In many high-risk industries, it is common to measure mental workload and then to redesign tasks until cognitive overload is avoided. This study aimed to measure the effect of multitasking on the mental workload and performance of medical students completing single and combined clinical tasks.
Medical students who had completed basic clinical skills training in a single undergraduate Medical School completed four standardised tasks for a total of four minutes each, consisting of: inactivity, listening, venepuncture and a combination of listening and venepuncture. Task performance was measured using standard binary checklists and with mental workload measured using a secondary task method.
The tasks were successfully completed by 40 subjects and as expected, mental workload increased with task complexity. Combining the two tasks showed no difference in the associated mental workload and performance at venepuncture (p = 0.082) However, during the combined task, listening appeared to deteriorate (p < 0.001).
If staff are expected to simultaneously complete multiple tasks then they may preferentially shed communication tasks in order to maintain their performance of physical tasks, leading to the appearance of poor communication skills. Although this is a small-scale study in medical students it suggests that the active assessment and management of clinician workload in busy clinical settings may be an effective strategy to improve doctor-patient communication.
KeywordsMental workload Clinical skills Communication skills
In 2008 Wickens outlined ‘multiple resource theory’ as a way of investigating the effect of increasing workload on the performance of operators in complex environments . This theory defines mental capacity as an abstract concept expressing the maximum rate at which an individual can process information. Mental workload is defined as “the percentage of our mental capacity in use at any one time” . A key prediction of this theory is that if the cognitive demand of any task exceeds the capacity of the operator, then there is an increased risk that the operator will become cognitively overloaded and therefore be prone to error and/or poor performance. The point at which overload occurs is often seen as a balance between three aspects of workload: the demands of the task, the effort exerted by the operator and the expertise of the operator . That is, even where a task is normally completed successfully, performance may deteriorate if the task becomes more difficult, the operator becomes tired, or an expert operator is replaced by a novice.
A review in 2011 found that published studies involving the measurement of the mental workload of clinical staff to be small scale, use widely varying techniques and lack follow up studies to validate their initial findings. However they did identify high mental workload as a common problem,  with two areas have a more extensive literature: laparoscopic/robotic surgery [3, 4, 5, 6, 7, 8, 9, 10, 11] and the delivery of anaesthesia [12, 13, 14, 15, 16, 17, 18, 19]. Within medical education, there are few studies which have used mental workload as an outcome [20, 21, 22], although it appears to becoming more common as an outcome in, for example, simulation based training .
As mental workload is an abstract concept, it is not possible to measure it directly and a variety of techniques have been developed to provide subjective and objective estimates. Secondary task methods use an additional, simple task completed in addition to the main or primary task. The principle is that the secondary task response will remain constant as long the operator has spare capacity (mental workload < mental capacity). If the workload of the primary task exceeds capacity, there will be no spare capacity to compete the secondary task and secondary task performance will deteriorate . Examples of a secondary task include responding to a vibrating stimulus or performing mental maths, with any delay in response or error used as outcomes [21, 24].
When an operator is required to complete more than one task, the effect on overall workload is may be difficult to predict. Wickens’ multiple resource theory  posits that the brain is made up of a variety of different areas which process different types of information. That is, an operator may complete multiple tasks, as long as they do not compete for the same cognitive resource. For example, we may be able to examine a patient (visual) and listen to what they are saying (aural), but we cannot simultaneously listen to a patient (aural) and listen for heart sounds (aural).
Studies in fields such as aviation show that when faced with excessive mental workload, we subconsciously process only some information, so that subjects ignore or ‘shed’ some of the workload in order to maintain their performance in at least one task . Task shedding, can therefore allow an individual to maintain some function in a high workload situation and avoid complete cognitive overload [27, 28].
Many high-risk industries, such as aviation, use mental workload to assess performance, but the techniques are rarely used in medicine, perhaps because the complex nature of the clinical environment makes it hard to define the mental workload of specific clinical tasks . Understanding and measuring mental workload among clinicians is important as excessive mental workload appears to be a common problem within the clinical environment and may be a significant factor in clinical error . A better understanding of the effect of task complexity on mental workload may help us to understand why errors occur and provide new ways of improving patient safety .
The aims of this study were to determine whether a series of tasks of increasing complexity would be associated with an increase in measured mental workload and to determine the effect of combining two different tasks on mental workload and performance. A secondary outcome was to further validate the secondary task methodology by measuring mental workload during a period of inactivity.
Students studying medicine in 4 different year groups, within a single Undergraduate Medical School, were recruited via social media and poster advertising, with the first 10 volunteers in each year group of 300 students accepted. Data collection took place over a 10 week period. A sample size of ten participants from each year group was chosen empirically as there were no data available to perform a formal power calculation. Only students who had already completed at least one formal teaching session on venepuncture were included. We were not able to identify further opportunities for formal teaching in venepuncture in the later years with students practicing their skills during clinical placements. After ethical approval from the Cardiff School of Medicine Research and Ethics Committee (SMREC ref.: 15/68), forty participants were recruited and provided informed, written consent.
As a secondary task, each participant had an iPhone™ (Apple, Inc., Cupertino, CA, USA), strapped to their non-dominant upper arm by a neoprene phone holder. The iPhone™ used an application called Workload Insight™ (Swansea, UK) programmed to vibrate at random intervals between 10 and 90 s. Participants were asked to tap the screen, as soon as they felt the vibration. Tapping the screen terminated the vibration and the time delay from the onset of the vibration to the response (milliseconds) was recorded by the app.
Summary of the four tasks completed by each subject
Control – Inactivity
Mental workload and four post task questions.
Mental workload and 16 point checklist
Combined Listening and Venepuncture
Mental workload, 16 point checklist and four post task questions.
Mental workload measurements used a previously validated methodology, with the upper limit of a normal response time as 1350 ms (average + 3SD) . Any response time of less than 1350 ms was regarded as normal with a recorded delay of 0 ms. Any response times greater than 1350 milliseconds were regarded as delayed, with the delay recorded as the response time minus 1350 ms. As described in previous studies, the measure of mental workload used was the average delay, calculated as the total delay in the measurement period, divided by the total number of responses .
The normality of distribution of results was determined using a Sharipo-Wilk test. The significance of differences between groups were estimated using Spearman’s rank correlation, with a probability of < 0.05 used as to define significance.
Forty medical students, a total of 5 males and 5 females from each year, were recruited and all sessions were completed successfully with data available for analysis. Initial analysis of the results showed that neither response time (Sharipo-Wilk, n = 160, p < 0.0001) nor performance scores (n = 160, p < 0.0001), were normally distributed, so further analysis used non-parametric methodology.
The statistical significance of pairwise comparisons between the mental workload during different tasks
p-values (N = 40 in all cases)
Ven + Listen
When performing venepuncture on its own (Task 3) subjects scored 13.5 (4.5 - 16) out of 16 (median, range) and 13 (6 - 16) when performing venepuncture at the same time as the listening task (Task 4). The difference was not statistically significant (n = 40, p = 0.082).
When performing the listening task on its own (Task 2) subjects scored 3 out of 4 (0.5 – 4) and when listening at the same time as venepuncture (Task 4) scored 2 out of 4 (0 – 4). This difference was statistically significant (n = 40, p < 0.0001).
Subgroup analysis was performed to identify trends in performance between students in different years of study and of different gender, but no significant differences were identified.
This study builds on previous publications using the same methodology and suggests that the secondary task method provides a practical method of measuring the mental workload of those performing clinical tasks .
The observed relationship between measured workload and subjective task complexity suggests that it is possible to estimate the mental workload required for specific clinical tasks. However, the nonlinear relationship observed suggests that subjective estimates are inadequate to assign specific levels of workload to individual tasks. As expected, inactivity and listening were confirmed as low workload tasks. The significant rise in average delay during venepuncture however, confirms venepuncture as a high workload task. In particular, the observed level of average delay in some subjects suggests significant cognitive overload consistent with poor performance or error. This was not unexpected as no participant had been formally certified as competent to perform venepuncture and their poor performance was confirmed by the checklist performance scores.
While the performance of venepuncture was expected to be a high workload task, the addition of the listening task was expected to further raise the mental workload of subjects and this was expected to be evident in increased average delay. The observation that the combination of tasks did not increase workload can be explained by two different mechanisms. Firstly, that the subjects improved their level of expertise and were therefore able to perform both tasks with the same level of mental workload (which would seem unlikely given that the tasks were sequential) or that the task itself had changed.
The low measured workload and high listening scores indicate that in the single task the subjects’ had expertise in listening (they were all native speakers of English). The most logical explanation of the significant fall in listening performance in the combined task is that, when faced with cognitive overload, subjects ‘shed’ the listening task  and used almost all their available cognitive capacity to complete the venepuncture task. So, while subjects were still overloaded during the combined task, they were at least able to maintain their performance in venepuncture, evidenced by the unchanged checklist scores.
A relevant criticism of this form of research is that the increased response times may simply represent a physical task preventing subjects completing the secondary task. We would reject this hypothesis on two grounds. Firstly, a previous study involving clinical work addressed this question by allowing subjects to identify any period when they were ‘hands full’ and exclusion of these data did not remove the effect of workload on average delay . Secondly, mental workload theory would predict that cognitive overload will impair all aspects of performance including sensory perception, processing as well as motor response. This is supported by previous studies which have shown both delayed response times and impaired recall of events [15, 31].
Research in other fields such as aviation suggest that the process of task shedding in the face of high cognitive workload is an effective mechanism to avoid cognitive collapse . While aviation studies suggest that pilots are more likely to shed tasks perceived as low priority, the processes which direct task shedding are poorly understood . It is therefore not clear whether subjects consciously prioritised venepuncture as the more important task or whether the selection was entirely subconscious. We are not aware of any research which indicates whether physical or verbal tasks are more likely to be shed in response to increased workload and believe that this study provides the first experimental evidence of task shedding during a clinical task.
These data suggest that during clinical tasks, if cognitive workload increases towards capacity, listening tasks may be preferentially shed. Importantly, as the shedding process may be subconscious, the individuals may not be aware that they have effectively stopped listening and missed crucial information. This conclusion is supported by a previous study using the same methodology that has suggested that within simulated consultations, medical student cognitive workload is excessive . Many patients complain about the ability of doctors to listen to their concerns  and while communication skills training is an obvious solution, these data suggests that such training may be ineffective if high mental workload is the cause of the failure to listen.
We recognise that this study included only 40 participants, within one UK medical school and that the opportunistic sampling of participants may also have led to self-selection bias. We therefore recognise that these findings may not generalise to other settings. In particular, studies involving clinical staff in a more realistic environment, at different stages in their training would be required to determine whether the problem identified has a significance to clinical practice. We also recognise that randomisation of the tasks would have reduced the effect of any learning during the study period, however, previous studies have demonstrated no learning effect on this particular secondary task method .
This study has confirmed that the secondary task method appears to be a valid measure of cognitive workload during clinical tasks and may be a useful tool to estimate cognitive workload. The most important insight provided by this study is that if individuals are expected to simultaneously engage in multiple tasks, then their performance may be impaired. In particular, communication skills may be particularly susceptible to cognitive overload. While education and training may improve the performance of single tasks, the active assessment and management of clinician workload in busy clinical settings may be an effective strategy to improve doctor-patient communication.
We gratefully acknowledge the support of the staff and students in the School of Medicine, Cardiff University.
No external funding was used to support this study.
Availability of data and materials
The data collection software used in this study is available for free download via the apple app store under the heading ‘workload insight free’.
This study was designed by BW and AB, data was collected by BW with both authors contributing to the analysis of data and writing the manuscript. OB designed the main statistical analysis, contributed to the choice of presentation of results and the conclusions. All authors contributed to and approved the final manuscript.
Ethics approval and consent to participate
This study had institutional ethical approval from the School of Medicine Ethics Committee, Cardiff University and all participants gave informed, written consent prior to participation.
Consent for publication
No individual data is included.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 1.Farmer, E. and Brownson, A. (2003). Review of workload measurement, analysis and interpretation methods. European Organisation for the Safety of Air Navigation: 33.Google Scholar
- 4.Abdelrahman AM, Bingener J, Yu D, et al. Impact of single-incision laparoscopic cholecystectomy (SILC) versus conventional laparoscopic cholecystectomy (CLC) procedures on surgeon stress and workload: a randomized controlled trial. Surg Endosc. 2016;30:1205.Google Scholar
- 7.Klein M, Warm J, Riley M, Matthews G, Gaitonde K, Donovan J, et al. Performance, stress, workload, and coping profiles in 1st year medical students interaction with the EndoscopicLaparoscopic and robot-assisted surgical techniques. Hum Factors Ergon Soc Annu Meet Proc. 2008;52:885–9.CrossRefGoogle Scholar
- 19.Gaba DM, Lee T. Measuring the workload of the anesthesiologist. Anesth Analg. 1990;71(4):354–61.Google Scholar
- 24.Owen R. Clarifying the simple assumption of the secondary task technique. In: Holman R, Solomon M, editors. NA - advances in consumer research, vol. 18. Provo: Association for Consumer Research; 1991. p. 552–7.Google Scholar
- 27.Parasuraman R, Hancock P. In: Hancock P, Paula D, editors. Adaptive control of mental workload. New York: CRC Press; 2001.Google Scholar
- 28.Salas E, Maurino D. Managing workload, performance and situation awareness in aviation systems. Human factors in aviation. London: Elsevier; 1988.Google Scholar
- 29.Oliver M, Davis H, Jones P, Rowe C, Byrne A. Use of a secondary task paradigm to measure medical student’s mental workload during a simulated consultation. Int J Commun. 2010;4(2).Google Scholar
- 33.Data on Written Complaints in the NHS 2014-15 v1.0. Workforce and Facilities Team, Health & Social Care Information Centre, August 2015.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.