Hot Topics: Implementation Science, Augmented Intelligence, and Social Determinants of Health

Implementation science, augmented intelligence, and social determinants of health are at the forefront of critical care and will ultimately change how clinicians provide the best care in ICUs.

When thinking about what is at the forefront of our work in critical care—what we are talking about, what we are advancing, and what we need to learn more about—I keep landing on implementation science (IS), augmented intelligence, and social determinants of health (SDOH).
 
Implementation Science
In the time of research utilization, it took approximately 15 years to integrate research findings into practice. It is no wonder we now have an entire new strategy to do this work to improve outcomes. IS, originally defined in 2006 as identifying a specific strategy to accelerate the uptake of research in clinical care, is gaining in significance as we seek more efficient ways to integrate research as well as evidence-based practice to advance outcomes.¹ It is how we execute processes or bring research to patients and families in a more expeditious way that fits the specific setting. While IS has been around since 2006, not all critical care units use it, and not all critical care clinicians are skilled in this type of research. What works in one intensive care unit (ICU) may not work in a different ICU, even within the same hospital.
 
In a 2020 scoping review, McNett et al elucidated the variability in nomenclature and heterogeneity in IS work resulting in the inability to pool and synthesize findings.² As with any new process, bringing the field together to speak the same language is critical. The Society of Critical Care Medicine’s work on a research data dictionary will most certainly facilitate advancement of IS; however, more work will need to be done to refine IS nomenclature in the future.
 
Augmented Intelligence
There has been a lot of discussion about augmented intelligence as we become increasingly proficient in data science and machine learning. Thinking ahead to the future, we can imagine something very different from today’s reality. Using ICU Liberation as an example, we can envision a future with feedback loops between the critically ill patient and hemodynamic monitoring or between the ventilator and the patient’s muscles such that ventilator weaning is initiated automatically using an algorithm ICU clinicians set at intubation.
 
Other future improvements may include implementing algorithms that include EEG monitoring to notify staff of early stages of delirium so that strategies for improvement can be initiated. Families could be notified of a patient’s wake/sleep state to increase family engagement in their loved one’s care. We already know that integration of the ICU Liberation Bundle (A-F) into ICU care results in decreased ventilator days, delirium days, physical restraint use, ICU readmission, and death.³ Imagine if we could integrate the entire bundle into patient monitoring with an augmented intelligence algorithm. In coordination with the multiprofessional healthcare team, we could advance these results to improve outcomes of patients with critical illness. This future is not so far away!
 
Social Determinants of Health
SDOH has become increasingly recognized as an explanation for admission or readmission to the ICU and associated outcomes for both individual patients and groups of patients. Race/ethnicity, healthcare accessibility, ability to afford prescriptions, exposure to allergens and pollution, and food deserts are just some of the components of SDOH being evaluated for impacting admission or readmission to the ICU. Outcomes associated with race/ethnicity are now emphasized as social constructs within research to advance findings that eliminate health inequities.⁴
 
Aside from research, what can we do in the ICU when faced with healthcare inequities? Collecting this information is a good place to start. Understanding a patient’s situation creates awareness in the healthcare team and can result in discussions to maximize outcomes. For example, if a patient in the ICU has an infection requiring a lengthy course of antibiotics post discharge, knowing the patient’s home setting and health literacy helps determine how to prescribe antibiotic treatment after discharge. Discharging patients who are dependent on technology to a home without electricity is problematic. Understanding the patient from the perspective of their reality facilitates implementation of processes to improve their outcomes.
 
While we look to patients’ biology, pathophysiology, and social constructs to understand more about them and their illnesses, categorizing these factors as biomarkers may elevate SDOH to a more prominent role in healthcare delivery. Imagine identifying an unhoused patient with sepsis as having a hyperinflamed phenotype and using rapid diagnostic testing to discover antibiotic resistance. Is it possible that each of these biomarkers—unhoused patient, hyperinflamed phenotype, and difficult-to-treat infection—would impact the outcome in a different, additive, or synergistic way? Considering these as biomarkers expands our ability to provide personalized healthcare in the ICU. With the imperative inclusion of SDOH, we may alter ICU care delivery.
 
Reflections
IS, augmented intelligence, and SDOH could plausibly be related. While IS typically applies to ICU-specific processes, ICU characteristics are embedded in not only the types of illnesses but also the population served in the ICU. The population reflects the SDOH. While not all ICUs have a homogenous population, we know that certain populations are at risk for certain diseases and death, so could we use that information in our IS approach?
 
Using our population in developing augmented intelligence algorithms may provide even better personalized healthcare. For example, in those at high risk for death due to either their disease or their biomarkers, the algorithm can be altered based on that information. Using our example of the unhoused, hyperinflamed patient with sepsis, this patient could belong to a population at high risk for kidney failure. By inputting these data into the algorithm, we may be able to improve this patient’s outcomes.
 
How we integrate research and evidence into our daily processes, understand SDOH, and integrate augmented intelligence into care delivery will ultimately change the ICU as we know it today. The expert clinician is the key and the pivotal foundation of this work both now and in the future.
 
References

1. Eccles MP, Mittman BS. Welcome to implementation science. Implementation Sci. 2006 Feb;1(1).
2. McNett M, O’Mathúna D, Tucker S, Roberts H, Mion LC, Balas MC. A scoping review of implementation science in adult critical care settings. Crit Care Explor. 2020 Dec 16;2(12):e0301.
3. Pun BT, Balas MC, Barnes-Daly MA, et al. Caring for critically ill patients with the ABCDEF bundle: results of the ICU Liberation Collaborative in over 15,000 adults. Crit Care Med. 2019 Jan;47(1):3-14.
4. Zurca AD, Suttle ML, October TW. An antiracism approach to conducting, reporting, and evaluating pediatric critical care research. Pediatr Crit Care Med. 2022 Feb 1;23(2):129-132.

Posted: 9/2/2024 | 0 comments