The intensive care unit (ICU) of the future, in which technology enables dramatic improvements to clinical care while reducing costs, has been predicted since the 1970s, but has largely been unrealized.(1) In fact, today we interact with patient data essentially as we did back in the 1950s when ICUs were first introduced. Sensors measure different aspects of patient physiology and typically display the last few seconds of waveform data on a patient monitor. Additional isolated medical devices measure more sophisticated physiological processes and display this information on their own smaller displays. This information may or may not be transmitted to the primary patient monitor. Very little attention has been given to how to best use data in the ICU. New monitoring technologies continue to arrive and are simply placed next to the other devices in already cramped patient rooms; data are added to those already displayed. There is virtually no ability to go back and analyze what has happened over time, no concept of the complex dynamic interactions among all monitored physiological processes. Despite impressive technological advances in the last 60 years, simply determining the average blood pressure over the last few hours is nearly impossible in most ICUs around the world. In this issue of Critical Connections, four outstanding programs are highlighted, each aiming to change how patient data are used to improve intensive care quality, patient outcomes, and patient satisfaction.
These essential efforts are largely overshadowed by the focus on electronic medical record (EMR) adoption and complying with “meaningful use” standards that are tied to receiving maximum financial incentives.(2) These standards include electronic prescribing, health information exchange among clinicians and hospitals, and automated reporting of quality performance.3 Policy makers hope EMR adoption will benefit the healthcare system by promoting efficiency, facilitating the transfer of medical records, reducing medical errors and the resulting lawsuits, improving medical device management, and generally reducing costs.(4-6) Hospitals face many challenges just to meet this level of functionality.(3)
In the outpatient setting in particular, evidence suggests many benefits to using EMRs.(7) However, despite process benefits,(8) EMR adoption has been slow to improve care quality and reduce costs in ICUs,(9) where one in every seven healthcare dollars is spent.(10) One important reason for this is that current EMR models do not address the vast streams of unconnected data. This “fire hose” of information places caregivers in a constant state of data overload. For every patient, clinicians on morning rounds are asked to absorb and understand more than 2000 variables, including vital signs, laboratory tests, clinical assessments, medications, images, and more.(11,12) Cognitive science reveals that people have difficulty understanding how more than two variables relate without tools or assistance.(13) It is no wonder that physicians and nurses continue to scribble data and notes on a piece of paper to try to figure out what is happening with the patient.
In the background, mostly meaningless patient monitor alarms(14) sound continuously, leading to alarm fatigue that results in lower-quality care and sometimes fatal events. The problem is so perverse that The Joint Commission issued a sentinel event alert in April 2013 requiring hospitals to address alarm management.(15) Despite overwhelming amounts of data and alarms, ICU clinicians actually receive a paucity of useful information. A quick glance at a snapshot of vital signs displayed on a bedside monitor, without context, fails to provide the true meaning of a patient’s current physiologic state.
Nurses who are already overloaded with clinical duties are pulled away from treating patients to document information that is already available in digital form. Too often they must record data from a medical device on paper and manually enter it into an EMR or another computer system. Even data transmitted directly to the EMR must be validated by the nurses. Worse still, EMRs have limited functionality in supporting clinical decisions or retrieving data for analysis, either by caregivers or other computer systems, so that clinically meaningful information can be extracted. In preparation for morning rounds, someone – a nurse, nurse practitioner, resident, attending physician – transfers these data back onto paper to note important clinical events or summarize vital signs. In one literature review, eight of 12 studies reported that instituting a clinical information system did not significantly reduce patient charting times and in many cases actually increased it.(16) The situation is unacceptable now and will become untenable if future projections prove correct in the number of patients who will need ICU care compared to the number of available intensivists and nurses.(17)
We do not need further technological breakthroughs to achieve the ICU of the future. Medicine simply needs to apply technologies that have been successfully utilized in other fields, such as finance or aviation. Data availability, both from medical devices and from clinical information systems that capture data, is the biggest barrier to progress. Significant efforts to establish a medical device connectivity standard have been ongoing since the 1980s, but have not been widely adopted by the medical device industry. Concerns regarding compliance with the Health Insurance Portability and Accountability Act have exacerbated the problem by causing many device manufacturers to transmit only anonymous patient data or device management data. The U.S. Food and Drug Administration (FDA) also must continue to revise its model for medical device regulation to accommodate the 21st century data model – which is still evolving. As recently as 2011, the FDA released new regulations regarding data aggregation and analysis systems; expect these regulations to continue to evolve over the next several years.
To achieve the promises of high-quality personalized medicine while reducing healthcare costs, we must shift how clinicians interact with data. Clinicians should be making medical decisions, instead of moving patient data from system to system while trying to understand what the information means. Instead of focusing efforts on how to document and store ICU patient data, energy should be directed to moving data automatically, without human intervention or validation. Information instead should shift directly from devices to computer systems designed to analyze it and display essential information clinicians need to make the best treatment decisions for their patients. All patient data need to be seamlessly integrated, converted into actionable information, and delivered to caregivers in a form that makes it easy to review trends, chart progress, and answer questions.
Intensivists make hundreds of decisions a day. The essence of critical care comes down to monitoring the consequences of these decisions, changing plans in a timely manner when the results are not satisfactory. To accomplish this, clinicians must be informed about who is changing physiologically or starting to deteriorate. Algorithms that analyze patient data to provide preclinical detection of life-threatening complications, such as sepsis,(18) are essential to improving care quality and patient outcomes. Clinicians need actionable information, such as whether a patient will benefit from fluid administration(19) or is ready to be extubated,(20) as well as feedback about the effectiveness of treatments to facilitate adjustments toward goal-directed therapy.(21) In the future, life-threatening and time-sensitive alarms (e.g., ventricular fibrillation) will remain relatively unchanged, but smarter alarms will be able to weigh all data and improve the specificity of less-critical alarms.(22-24) Automatic video annotation of actions taken bedside(25) – such as how often patients are turned and suctioned, or the angle of the bed – will be leveraged to improve care quality. Infusion pump data will be compared to medication orders to ensure that patients are receiving the correct medications, when they should, at the correct dose.(26) All of this information should be presented to clinicians intelligently on a single integrated display to improve the speed and accuracy of clinical decisions.(27) These systems should be designed by applying user-centered design engineering methods to ensure that technology works around people instead of people working around the technology.(28)
The programs highlighted in this issue of Critical Connections are a step towards the ICU of the future. Significant challenges must be overcome to promote data availability and innovation, while continuing to protect patients against misapplications of technology, fraud, and violations of privacy. Meeting these challenges will ultimately determine whether technology can be leveraged to improve ICU patient care and reduce costs, or its promises remain unmet.
1. Amaral ACK-B, Rubenfeld GD. The future of critical care. Curr Opin Crit Care. 2009;15(4):308-313.
2. Blumenthal D. Launching HITECH. N Engl J Med. 2010;362(5):382-385.
3. Jha A. Meaningful use of electronic health records. JAMA. 2010;304(15):1709.
4. Adhikari N, Lapinsky S. Medical informatics in the intensive care unit: overview of technology assessment. J Crit Care. 2003;18(1):41-47.
5. Martich G, Waldmann C, Imhoff M. Clinical informatics in critical care. J Intensive Care Med. 2004;19(3):154.
6. Clemmer TP. Computers in the ICU: where we started and where we are now. J Crit Care. 2004;19(4):201-207.
7. Jamoom E, Beatty P, Bercovitz A, et al. Physician adoption of electronic health record systems: United States, 2011. NCHS Data Brief. 2012;Jul (98):1-8.
8. Holroyd-Leduc JM, Lorenzetti D, Straus SE, et al. The impact of the electronic medical record on structure, process, and outcomes within primary care: a systematic review of the evidence. J Am Med Inform Assoc. 2011;18(6):732-737.
9. Wu S, Chaudhry B, Wang J, et al. Systematic review: impact of health information technology on quality, efficiency, and costs of medical care. Ann Intern Med. 2006;144(10):742-752.
10. Halpern NA, Pastores SM. Critical care medicine in the United States 2000-2005: an analysis of bed numbers, occupancy rates, payer mix, and costs. Crit Care Med. 2010;38(1):65-71.
11. Herasevich V, Litell J, Pickering B. Electronic medical records and mHealth anytime, anywhere. Biomed Instrum Technol. 2011;Fall Suppl:45-48.
12. Morris G, Gardner R. Computer Applications. In: Hall J, Schmidt G, Wood L, eds. Principles of Critical Care. New York: McGraw-Hill; 1992: 500-514.
13. Imhoff M, Fried R, Gather U. Detecting relationships between physiological variables using graphical modeling. Proc AMIA Symp. 2002; 340-344.
14. Welch J. An evidence-based approach to reduce nuisance alarms and alarm fatigue. Biomed Instrum Technol. 2011;45(S1):46-52.
15. Sentinel Event Alert: Medical Device Alarm Safety in Hospitals. In: The Joint Commission; 2013. p. 1-3.
16. Mador RL, Shaw NT. The impact of a Critical Care Information System (CCIS) on time spent charting and in direct patient care by staff in the ICU: a review of the literature. Int J Med Inform. 2009;78(7):435-445.
17. Kelley MA, Angus D, Chalfin DB, et al. The critical care crisis in the United States: a report from the profession. Chest. 2004;125(4):1514-1517.
18. Moorman JR, Carlo WA, Kattwinkel J, et al. Mortality reduction by heart rate characteristic monitoring in very low birth weight neonates: a randomized trial. J Pediatr. 2011;159(6):900-906.
19. Bendjelid K, Romand J-A. Fluid responsiveness in mechanically ventilated patients: a review of indices used in intensive care. In: Pinsky MR, Brochard L, Mancebo J, eds. Applied Physiology in Intensive Care Medicine 2. New York, NY: Springer; 2012:3-11.
20. Seely AJ, Green GC, Bravi A. Continuous multiorgan variability monitoring in critically ill patients—complexity science at the bedside. Conf Proc IEEE Eng Med Biol Soc. 2011;2011:5503-5506.
21. Puskarich MA, Marchick MR, Kline JA, et al. One year mortality of patients treated with an emergency department based early goal directed therapy protocol for severe sepsis and septic shock: a before and after study. Crit Care. 2009;13(5):R167.
22. Graham KC, Cvach M. Monitor alarm fatigue: standardizing use of physiological monitors and decreasing nuisance alarms. Am J Crit Care. 2010;19(1):28-34.
23. Blum JM, Tremper KK. Alarms in the intensive care unit: too much of a good thing is dangerous; Is it time to add some intelligence to alarms? Crit Care Med. 2010;38(2):702-703.
24. Scalzo F, Liebeskind D, Hu X. Reducing false intracranial pressure alarms using morphological waveform features. IEEE Trans Biomed Eng.2013;60(1):235-239.
25. Siebig S, Kuhls S, Imhoff M, et al. Collection of annotated data in a clinical validation study for alarm algorithms in intensive care—a methodologic framework. J Crit Care. 2010;25(1):128-135.
26. Trbovich P, Pinkney S, Cafazzo J, Easty AC. The impact of traditional and smart pump infusion technology on nurse medication administration performance in a simulated inpatient unit. Qual Saf Health Care. 2010;19(5):430-434.
27. Koch S, Staggers N, Weir C, et al. Integrated information displays for ICU nurses: field observations, display design, and display evaluation. Proceedings of the Human Factors and Ergonomics Society Annual Meeting. 2010;54 (12):932-936.
28. Zhang J. Human-centered computing in health information systems. Part 1: analysis and design. J Biomed Inform. 2005;38(1):1-3.