New technology in anaesthesia

New technology in anaesthesia

  • Issue 76

Closed-loop Systems and Automation within Anaesthesia

Brenton Alexander MD, Joseph Rinehart MD, Jacques Duranteau MD PhD, Alexandre Joosten MD

joosten-alexandre@hotmail.com
Alexandre_Joosten@erasme.ulb.ac.be

Closed-loop technologies, defined as systems that use input variables to adjust output interventions without human interaction, have snuck into almost every sector of our lives and will only continue to expand in the coming years. Their ability to decrease costs and human workload while also increasing accuracy, consistency, and safety have propelled these technologies forward at an exponential rate. The thermostat is an easy to understand example of a closed-loop system that was first described 400 years ago by Cornelis Drebbel as he automated the temperature control of an egg incubator. The modern thermostat in air conditioning systems uses the ambient air temperature as an input variable to control either a heating or cooling unit as an output variable. A ‘supervising’ user defines the desired end point (in this case the temperature) and the system is then autonomous in achieving that goal. At the other end of the complexity spectrum – the cutting edge of technological integration into a closed-loop system – is the autonomous vehicle, which some project will be 25% of all vehicles by 2040 and may prevent 90% of all current car accidents.1 The inputs and output interventions of this system, and ones like it, are incredibly complex and have multiple inputs and outputs as well as multiple control objectives that must be carefully considered in both design and testing.

Within medicine, and specifically within Anaesthesiology, these closed-loop techniques have been primarily adopted to curb therapeutic errors and inconsistencies that occur with manually delivered care. The problem with manual titration of IV drugs, for example, is that despite their best intentions and efforts, human beings are quite limited when it comes to repetitive attention-requiring tasks, especially when those tasks are cognitively trivial (like adjusting an infusion rate up or down).

This topic is unfortunately quite relevant as preventable errors are some of the deadliest threats to the modern patient and are the third leading cause of mortality in the United States at a cost to the U.S. economy of $20 billion dollars per year.2 Compounding these errors is the fact that protocols standardizing a wide variety of perioperative interventions known to decrease patient morbidity are often poorly or not at all implemented.3,4 When appropriate compliance with goal-directed fluid therapy protocols is achieved, the benefits are profound, ranging from decreased length of stay to decreased morbidity.5,6

The logical next question is whether automation can be the solution that increases best-practice implementation during the perioperative period, thus increasing compliance with protocols known to be effective at improving patient outcome while decreasing medical errors. With demonstrated decreased perioperative variability and increased time in target ranges, closed-loop systems are uniquely posed to achieve these objectives without significant increases in provider workload, although the initial expense and long development time are realistic initial obstacles.

In looking at current research systems, the lowest hanging fruit in the anaesthesiologist’s workday are systems with clear physiologic endpoints and well-established protocols to optimize such endpoints. This has led to a significant amount of research into the automation of goal-directed hemodynamic interventions, including fluid and vasopressor therapies.7-9 Other aspects that are also being discussed and explored are closed-loop anaesthesia, analgesia, drug delivery, and ventilation. Specifically regarding automated vasopressor administration, the concept has demonstrated practicality and effectiveness in experimental models10,11 and in the management of post-neuraxial hypotension in healthy obstetric patients.12,13 There are quite a few developments in both fluid and vasopressor automation coming in the next few years, as it is an exciting area of active research for multiple groups around the globe.14 Outside of hemodynamic control, other automated systems have recently demonstrated improved maintenance of ventilation, glucose levels, and depth of anesthesia.15-17

These various techniques will continue to improve in their efficacy, cost, safety, and usability over the upcoming decades. Beyond refinement of these individual systems, another significant future direction will be the use of multiple closed-loop systems simultaneously and, ultimately, the integration of multiple different closed-loop systems together into one multisystem controller.18,19 This is important because the impact of one system’s output on another system’s input variables may be relevant and could be destabilizing in some circumstances. Additionally, these various systems should be investigated with clearly identified clinical end-points (include neurologic, cardiac, pulmonary, gastrointestinal, and renal), as this is the evidence that will be needed to propel such technology into mainstream acceptance.

 

References

  1. Perez-Marin AM, Guillen M. Accid Anal Prev 2018;123:99-106.
  2. Makary MA, Daniel M. BMJ 2016;353:i2139.
  3. Miller TE, Roche AM, Gan TJ. Anesth Analg 2011;112:1274-6.
  4. Joosten A, Rinehart J, Cannesson M. Rev Esp Anestesiol Reanim 2015;62:181-3.
  5. Habicher M, Balzer F, Mezger V, et al. Perioper Med (Lond) 2016;5:31.
  6. Cannesson M, Ramsingh D, Rinehart J, et al. Crit Care 2015;19:261.
  7. Joosten A, Rinehart J. Anesth Analg 2017;125:20-2.
  8. Joosten A, Alexander B, Delaporte A, et al. Anaesthesiol Intensive Ther 2015;47:517-23.
  9. Rinehart J, Ma M, Calderon MD, et al. J Clin Monit Comput 2018;32:5-11.
  10. Libert N, Chenegros G, Harrois A, et al. Ann Intensive Care 2018;8:89.
  11. Marques NR, Whitehead WE, Kallu UR, et al. Anesth Analg 2017;125:110-6.
  12. Ngan Kee WD, Tam YH, Khaw KS, et al. Anesth Analg 2017;125:117-23.
  13. Ngan Kee WD, Khaw KS, Tam YH, et al. J Clin Monit Comput 2017;31:617-23.
  14. Jin X, Bighamian R, Hahn JO. IEEE Trans Biomed Eng 2018 Nov 19. doi: 10.1109/BME.2018.2880927
  15. Uemura K, Kawada T, Zheng C, et al. BMC Anesthesiol 2017;17:145.
  16. Brogi E, Cyr S, Kazan R, et al. Anesth Analg 2017;124:446-55.
  17. Bally L, Thabit H, Hartnell S, et al. N Eng J Med 2018;379:547-56.
  18. Joosten A, Jame V, Alexander B, et al. Anesth Analg 2018 May 9. doi: 10.1213/ANE.0000000000003433
  19. 19. Joosten A, Delaporte A, Cannesson M, et al. A A Case Rep 2016;7:260-5.