Cases & Commentaries
Cognitive Overload in the ICU
- Identify the role of cognitive overload—especially interruptions—in compromising quality of care and patient safety.
- List solutions that involve both enhanced team cooperation and decision making along with the prudent use of supportive technology.
- Describe the complexity of the critical care environment and why effective decision making can be challenging in such settings, especially at times of workload imbalance.
- Offer insights based on cognitive science principles that allow readers to understand the inevitable nature of some errors and the need to emphasize recovery, mitigation, and prevention.
Case & Commentary—Part 1:
A 72-year-old woman was admitted to the intensive care unit (ICU) with acute respiratory distress syndrome secondary to severe acute pancreatitis. A central line was placed into the left subclavian vein to administer intravenous fluids and vasopressors. A chest radiograph showed the central line in the proper position, with no evidence of pneumothorax.
Over the next 8 hours, despite maximal treatment, the patient's condition worsened, with progressive hypotension, acidosis, and hypoxia. The intensivist in charge was actively managing her complex care while intermittently updating the family about her worsening condition. At the same time, the intensivist was also managing nine other sick and complicated ICU patients, with nurses and respiratory therapists intermittently approaching her with questions about their ongoing management.
The patient with acute pancreatitis continued to worsen and had progressive bradycardia from acidosis resulting in a pulseless electrical activity (PEA) cardiac arrest. The intensivist led the resuscitation, which lasted for approximately 25 minutes. The patient ultimately regained a pulse but required two vasopressors to maintain her blood pressure and near maximal ventilator support because of severe hypoxia. A new set of laboratory tests, an arterial blood gas, and a chest radiograph were performed.
The intensivist responded to the full set of labs (e.g., repleting electrolytes, etc.), ordered additional fluid boluses, and changed the antibiotics. The patient remained acidotic and hypoxic over the next few hours. The intensivist was not sure why the patient was getting worse and sat down to think more about the clinical course—wondering what she might be missing. At that moment, another nurse in the ICU approached her about a different patient who had new hypotension and altered mental status.
Severe acute pancreatitis threatens life by several mechanisms: respiratory failure, renal failure, and circulatory shock. Standard evaluation includes an estimate of lung water by chest radiograph and of intravascular volume by measurement of central venous pressure.(1) Standard treatment routinely includes intubation and mechanical ventilation, infusion of intravenous fluids, and administration of potent vasopressors if necessary. Bedside cannulation of the central circulation for measurement of central venous pressure and for vasopressor infusion is often indicated.(2) The intensivist in this case took all the appropriate steps: inserting a central line, obtaining follow-up chest radiograph, intubating the patient and placing her on mechanical ventilation, and carefully monitoring fluids, electrolytes, and acid–base status.
Although the intensivist appears to have provided standard care for this patient, she was pulled in several directions by concerns about nine other sick patients in the ICU and had limited time to explore the reasons for the patient's deterioration. This phenomenon is known as cognitive overload, which is a familiar and frequent challenge in fast-paced environments such as ICUs and emergency departments.(3,4) In this case, a major source of cognitive overload was the frequent interruptions that prevented quiet contemplation and interpretation of the existing problems affecting the sickest of her ICU patients.(5)
Interruptions can increase the cognitive load on memory and attention, and they can subsequently lead to errors, compromising patient safety.(6,7) Yet, interruptions are an inescapable component of clinical work. Because they cannot be completely eliminated, it is necessary to understand the circumstances in which interruptions are likely to occur and to consider ways that their effect on cognitive load and patient safety can be mitigated.(8-10)
Studies suggest that interruptions can compromise memory and attention by requiring individuals to switch focus from one task to another.(11) As in this case, returning to a disrupted task requires completion of the interrupting task and then regaining the context of the original task. Multiple variables, including the characteristics of the primary task, the nature and length of interruptions themselves, and the environment itself, may influence the impact of interruptions on clinical tasks and errors.(12) For example, interruption disrupts complex cognitive tasks, which then require almost three times longer to resume effectively than simple tasks.(13) Interestingly, surveys and retrospective accounts of adverse incidents have often implicated interruptions, yet real-world evidence of the relationship between interruptions and clinical errors is scarce.
Providers and institutions can take measures that allow a more facile return to a complex task or reduce the likelihood of interruptions. For example, staff can be educated about the need to allow attending clinicians to have uninterrupted time for contemplation. However, in a complex, fast-paced, life-and-death environment, this practice may be challenging. As a result, careful assessment of staffing requirements, as well as individual roles and responsibilities, is crucial. In this case, the attending physician appeared to be the lone intensivist managing nine patients. While not an excessive volume of patients, the workload can quickly become unmanageable if multiple patients become unstable simultaneously. Monitoring and alerting technology may also be helpful. The notion is to use other people and technology to distribute the cognitive tasks (sharing mental efforts) so that one person is not overwhelmed as the sole recipient of the cognitive load in such environments.(14)
There are also lessons related to our growing use of electronic health records (EHRs), within and outside of the ICU setting. The traditional paper chart offered a shared information space; it allowed each member of the team (including nurses and other health professionals) to share their interpretation of patient data, thus leading to negotiation of meaning and consensus building.(15) In contrast, most current EHR systems provide clinicians with an individual narrow view of a patient case, not allowing multiple clinicians to create and share common views of the available patient data. New tools for collaborative annotation and access to decision histories can enrich traditional EHR systems and help the entire team engage in collective sensemaking over patients' care.(16) Some have proposed something akin to a "Facebook wall" for patients where all providers engaged in a patient's care can access the wall and add information as deemed necessary.(17)
Case & Commentary—Part 2:
Over the next few hours, the intensivist responded to new clinical data for the patient with pancreatitis, but she never found time to reflect on the clinical presentation. At 4:00 AM (3 hours after the chest radiograph had been performed following the PEA arrest), the radiologist called to inform the intensivist that there was a large pneumothorax. A thoracic surgeon was called to place a chest tube. The patient immediately responded, with improved oxygenation and blood pressure. The intensivist felt the pneumothorax likely explained most of the worsening clinical course overnight. It was not clear when or why the pneumothorax had occurred. Despite the temporary improvement, the patient experienced recurrent hypotension and progressive acidosis. Ultimately, after discussions with the family, care was withdrawn and the patient died peacefully. In thinking about the case, the intensivist recognized that she simply didn't have the time to think more broadly about why the patient was worsening despite treatment, and that she had felt overwhelmed by numerous required tasks and multiple interruptions.
In retrospect, the patient's initial deterioration and subsequent cardiac arrest were due to a pneumothorax not yet evident on the initial film obtained when the central line was placed. Two common risks to insertion of central lines include (i) misdirection of the catheter (cranially up a jugular vein or laterally into the axillary vein) and (ii) pneumothorax resulting from puncture of the lung.(18) The rate of pneumothorax following central venous insertion has declined from its initially reported rate of 2% to below 1%. Technical improvements to cannulation procedures (including cannula-over-wire ["Seldinger"] techniques and ultrasound guidance) have reduced, but not eliminated, pneumothorax as a complication.(19) It therefore remains common practice to obtain and interpret a chest radiograph after central line placement. In this case, the initial radiograph provided reassurance that the central line was in the correct position without evidence of lung injury. The latter is especially important because the combination of a punctured lung ventilated with positive pressure can lead to hemodynamic compromise due to a tension pneumothorax.
That the initial film did not demonstrate a pneumothorax is reassuring but not definitive. Several reports spanning decades suggest that a small fraction of pneumothoraces—fewer than one in five—that arise following central line insertion are not visible on initial chest radiograph.(20) This means that an occult pneumothorax occurs quite rarely—fewer than once in every 500 attempted insertions. Most clinicians will never encounter an occult pneumothorax following central line placement—or its potentially deadly consequences—during their professional lifetimes.
We cannot know for certain whether the pneumothorax was the result of the central line insertion. Mechanical ventilation of stiff acute respiratory distress syndrome–affected tissue can also lead to pneumothorax. Moreover, the clinician followed standard care when, following successful resuscitation, laboratory studies and a repeat chest radiograph were ordered. What is of interest in this case—and the narrative is silent on these matters—is (i) whether the clinician had a bedside ultrasound device immediately available to evaluate for pneumothorax, and (ii) whether a protocol was in place that a critical chest radiograph (such as one obtained after a cardiac arrest) would be interpreted by a radiologist as a priority.
The intensivist failed to verify the absence of a pneumothorax or other chest radiograph problem immediately after the code. The error was cognitive in nature and resulted from external factors common in critical care medicine. In general, common cognitive errors include failure to (i) receive or perceive data; (ii) comprehend data; (iii) communicate or transfer data; and (iv) accurately anticipate the consequences of decisions based on the data. Any of these failures can result from a loss of situational awareness, i.e., lack of intuitive understanding of key elements in the environment, including insights into the actions of team members at any given time and the projection of their likely status in the near future.(21,22) Such loss of situational awareness predisposes to a poor decision or sequence of actions that can lead to adverse events. In this case, stress and interruptions seemed to result in diminished situational awareness regarding the patient's deterioration and consequently a failure to review (i.e., receive) the data necessary to make the appropriate decision.
We may be tempted to blame the physician for the errors that occurred in this case. However, one can argue that in a setting like the ICU, all team members had a role in preventing the error: the clinician who was too busy to check on the postcode film, the nursing staff who interrupted the intensivist and pulled her in multiple directions, and the radiologists who evidently failed to note the problem immediately on the patient's postcode film and to notify the clinician urgently about the pneumothorax. And one may argue that the complexity of the ICU environment makes such occasional errors inevitable.
Yet, in complex settings, teamwork can help to reduce errors. We know that characteristics of teamwork are intimately tied to the occurrence of errors and to their prevention or mitigation.(23) Although much lip-service is paid to the role of teamwork in clinical care, health care systems have tended to assume that providers will learn and become proficient in such teamwork skills through experience alone. But opportunities exist to enhance training so that team-based care is better understood as a cognitive collaboration, one that requires joint discussion and communication to ensure errors are recognized early while there is still time to intervene. Simulation has long been used to train teams to work more effectively together, and those methods are becoming more sophisticated. Complex tasks in complex environments (e.g., active full resuscitation efforts) can be replicated using simulation to train individuals and teams. Evidence suggests that realistic and high-fidelity simulation programs can improve knowledge and confidence among critical care providers.(24) Simulation methods should accordingly be used to train members of ICU teams to work together more effectively and to learn how to enhance their joint ability to recover from the inevitable errors that will occur in such environments.(24) Recovery from error is arguably more important and realistic than trying to prevent all errors at the outset.(25)
In this case, multiple interruptions, stress, and patient complexity led a critical care physician to experience cognitive overload and an inability to maintain situational awareness to recover from interruptions. Technology could have been harnessed to help alert the physician to results in a more timely manner. Better staffing and protocols might have lessened the impact of workload and interruptions. Lastly, enhanced teamwork, facilitated through specific training and the use of simulation, may have created an environment where the team could have collaborated to ensure safe and effective care.
- In the complex ICU environment, teams must recognize that their work is cognitively distributed across all members and the support systems used.
- With effective approaches to balancing and sharing tasks, well-functioning teams can help to reduce the cognitive load and interruptions that are often a source of errors in such settings.
- Technology can support busy clinicians?monitoring for problems, alerting clinicians when appropriate, and providing intuitive and accurate displays of the information.
- In critical care environments, emphasis should be on team communication, prompt recognition of errors when they occur, and steps to recover from such errors or to mitigate their adverse effects.
- Effective teams are built through effective training, including the use of simulation methods.
- Teams should learn how to distribute cognitive loads so as to complement each other, offload tasks when appropriate, and contribute to increased patient safety.
Vimla L. Patel, PhD Professor of Biomedical Informatics Director, Center for Cognitive Studies in Medicine and Public Health New York Academy of Medicine New York, New York
Timothy G. Buchman, PhD, MD Professor of Surgery and Anesthesiology Director, Critical Care Center Emory University School of Medicine Atlanta, Georgia
Faculty Disclosure: Drs. Patel and Buchman have declared that neither they, nor any immediate members of their families, have a financial arrangement or other relationship with the manufacturers of any commercial products discussed in this continuing medical education activity. In addition, the commentary does not include information regarding investigational or off-label use of pharmaceutical products or medical devices.
1. Prout WG. Relative value of central-venous-pressure monitoring and blood-volume measurement in the management of shock. Lancet. 1968;291:1108-1112. [go to PubMed]
2. Guidelines for standards of care for patients with acute respiratory failure on mechanical ventilatory support. Task Force on Guidelines; Society of Critical Care Medicine. Crit Care Med. 1991;19:275-278. [go to PubMed]
3. Patel VL, Zhang J, Yoskowitz NA, Green R, Sayan OR. Translational cognition for decision support in critical care environments: a review. J Biomed Inform. 2008;41:413-431. [go to PubMed]
4. Sweller J. Cognitive load theory, learning difficulty, and instructional design. Learn Instr. 1994;4:295-312. [Available at]
5. Beyea S. Interruptions and distractions in health care: improved safety with mindfulness [Perspective]. AHRQ PSNet [serial online]. February 2014. [Available at]
6. Trafton JG, Altmann EM, Brock DP, Mintz FE. Preparing to resume an interrupted task: effects of prospective goal encoding and retrospective rehearsal. Int J Hum Comput Stud. 2003;58:583-603. [Available at]
7. Bailey BP, Konstan JA. On the need for attention-aware systems: measuring effects of interruption on task performance, error rate, and affective state. Comput Human Behav. 2006;22:685-708. [Available at]
8. Chisholm CD, Collison EK, Nelson DR, Cordell WH. Emergency department workplace interruptions: are emergency physicians "interrupt-driven" and "multitasking"? Acad Emerg Med. 2000;7:1239-1243. [go to PubMed]
9. Alvarez G, Coiera E. Interruptive communication patterns in the intensive care unit ward round. Int J Med Inform. 2005;74:791-796. [go to PubMed]
10. Westbrook JI, Woods A, Rob MI, Dunsmuir WTM, Day RO. Association of interruptions with an increased risk and severity of medication administration errors. Arch Intern Med. 2010;170:683-690. [go to PubMed]
11. Monsell S. Task switching. Trends Cogn Sci. 2003;7:134-140. [go to PubMed]
12. Li SYW, Magrabi F, Coiera E. A systematic review of the psychological literature on interruption and its patient safety implications. J Am Med Inform Assoc. 2012;19:6-12. [go to PubMed]
13. Dodhia RM, Dismukes RK. Interruptions create prospective memory tasks. Appl Cogn Psychol. 2009;23:73-89. [Available at]
14. Cohen T, Blatter B, Almeida C, Shortliffe E, Patel V. A cognitive blueprint of collaboration in context: distributed cognition in the psychiatric emergency department. Artif Intell Med. 2006;37:73-83. [Available at]
15. Mamykina L, Hum S, Kaufman D. Investigating shared mental models in critical care. In: Patel VL, Kaufman DR, Cohen T, eds. Cognitive Informatics in Health and Biomedicine: Case Studies on Critical Care, Complexity and Errors. London, UK: Springer; 2014:91-112. ISBN: 9781447154891.
16. Albolino S, Cook R, O'Connor M. Sensemaking, safety, and cooperative work in the intensive care unit. Cogn Technol Work. 2007;9:131-137. [Available at]
17. Wachter RM. The Digital Doctor: Hope, Hype, and Harm at the Dawn of Medicine's Computer Age. New York, NY: McGraw-Hill; 2015. ISBN: 9780071849463.
18. Tarbiat M, Manafi B, Davoudi M, Totonchi Z. Comparison of the complications between left side and right side subclavian vein catheter placement in patients undergoing coronary artery bypass graft surgery. J Cardiovasc Thorac Res. 2014;6:147-151. [go to PubMed]
19. McGee DC, Gould MK. Preventing complications of central venous catheterization. N Engl J Med. 2003;348:1123-1133. [go to PubMed]
20. Tyburski JG, Joseph AL, Thomas GA, Saxe JM, Lucas CE. Delayed pneumothorax after central venous access: a potential hazard. Am Surg. 1993;59:587-589. [go to PubMed]
21. Schulz CM, Endsley MR, Kochs EF, Gelb AW, Wagner KJ. Situation awareness in anesthesia: concept and research. Anesthesiology. 2013;118:729-742. [go to PubMed]
22. Endsley MR. The role of situation awareness in naturalistic decision making. In: Zsambok CE, Klein G, eds. Naturalistic Decision Making. Mahwah, NJ: LEA; 1997:269-283. ISBN: 9780805818741.
23. Patel VL, Cohen T, Batwara S, Almoosa KF. Teamwork and error management in critical care. In: Patel VL, Kaufman DR, Cohen T, eds. Cognitive Informatics in Health and Biomedicine: Case Studies on Critical Care, Complexity and Errors. London, UK: Springer; 2014:59-90.
24. Boling B, Hardin-Pierce M. The effect of high-fidelity simulation on knowledge and confidence in critical care training: an integrative review. Nurse Educ Pract. 2016;16:287-293. [go to PubMed]
25. Patel VL, Kannampallil TG, Shortliffe EH. Role of cognition in generating and mitigating clinical errors. BMJ Qual Saf. 2015;24:468-474. [go to PubMed]