Intensive Care Units provide care for the most critically ill patients. Though the work done in ICUs is undeniably important, adverse effects such as ICU delirium can occur. Intensive Care Unit delirium is characterized by an acute fluctuation in mental status, inattention, and disorganized thinking.1,2 It can result in cognitive impairment, behavior changes, or emotional disturbances and has been associated with increased length of hospital stay, increased risk of mortality, impaired cognition, and increased healthcare costs.3–5 Previously thought to be a reversible neuropsychological disorder, recent evidence suggests the effects of delirium could be more enduring.3,5,6 Development of delirium also affects families and caregivers of patients, as it increases distress and fear during hospitalization.7,8 The detrimental effects delirium can have on patient outcomes highlight the necessity for increased focus on caring for the ICU patient with delirium.
Understanding common clinical factors present in ICUs, as well as the patient population most at risk for developing ICU delirium, is essential in laying groundwork to improving care for patients with delirium. Current literature has identified factors that are used to recognize patients at risk for developing ICU delirium, but the strength of evidence and the methods used to identify predictive factors varies. The current study was developed to identify the incidence of ICU delirium and clinical factors associated with the occurrence of delirium in a large sample of ICU patients. The purpose of this descriptive analysis is to provide a landscape of the data obtained.
Clinical Practice Guidelines for the Prevention and Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption in Adult Patients in the ICU recommends that all critically ill adult patients be assessed for symptoms of delirium using a validated screening tool.9 The CAM-ICU has been validated and is considered reliable in mixed ICU populations.10–12 It is the screening tool used at the study site in this analysis.
There is a wide range of reported incidences of ICU delirium, and estimates vary based on patient population and screening practices. Krewulak et al.13 conducted a systematic review and found the pooled prevalence of ICU delirium to be 31%. The reported incidence of delirium in mixed ICUs ranges from 32.3%14 to 45%.15 The incidence of delirium appears to increase in specialized ICUs, with estimates as high as 77% in burn ICUs16 and 75% in neuroscience ICUs.6 Studies evaluating ventilated patients also reported higher incidence, with over 80% of patients developing delirium in some cases.1,17 The wide range of reported occurrence of delirium in individual studies with small sample sizes and various tools to assess for delirium symptoms demonstrates the necessity of using a large data set and a consistent measure for accurately calculating delirium incidence.
Because of the poor patient outcomes associated with the development of ICU delirium, substantial resources have been dedicated to determining risk factors with the potential to predict delirium. Age and severity of illness are frequently studied, with increased age being associated with increased risk of delirium.18–22 Past medical history associated with increased risk of delirium include dementia, schizophrenia, alcohol abuse or withdrawal, smoking or nicotine use, neurological trauma, and hypertension.19,20,22,23 Other risk factors include use of sedative medications, immobility, prolonged pain, mechanical ventilation, metabolic disturbances (including fluid deficit), and use of restraints.19,20,24 Recognition of risk factors associated with ICU delirium has influenced the development of prediction tools and risk stratification models used to identify patients most at risk for developing ICU delirium on admission to the critical care unit. Pagali et al.23 developed a risk stratification model using patient characteristics, including hearing and visual impairment, fall risk score, comorbidities, and laboratory data to place patients into high, medium, and low risk groups. Wassennar et al.25 tested prediction models that assessed age, coma, surgery, infection, metabolic acidosis, sedative use, and emergency surgery prior to admission to the ICU to determine risk for developing ICU delirium.
There is a lack of evidence evaluating possible associations between events and procedures commonly occurring in the ICU and patients who develop symptoms of ICU delirium. Disruptions in patients’ sleep and wake cycles, procedures that cause discomfort, and procedures requiring sedation could provide meaningful information about the patients at risk for developing delirium. This is a report of foundational, descriptive data related to the development of ICU delirium and selected frequently occurring yet understudied clinical events and procedures common in ICUs.
Methods
A retrospective chart review was employed to collect data on all adult patients, aged 18 and older, consecutively admitted to any of four ICUs at a large teaching hospital in Kansas City, Missouri (Medical-Surgical ICU, Cardiovascular Medical ICU, Cardiovascular Surgical ICU, and Neuroscience ICU). Data were extracted by the study site data analytics team from patient electronic medical records spanning the time from January 1, 2016 through April 30, 2022. Patient orders, laboratory results, and nursing documentation were included for data extraction. All data retained were limited to the patients’ ICU stay.
Specific clinical factors collected for this analysis were based on a review of the literature, personal experience, and suggestions from ICU nurses, with emphasis on understudied, but commonly occurring, factors. Presence and frequency of the factors were collected, as well as timing relative to onset of delirium, based on CAM-ICU results. For factors such as laboratory results, all available values of the laboratory test from the patients’ entire ICU stay were extracted.
The presence or absence of symptoms of delirium were collected from CAM-ICU data. At the study site, ICU nurses are required to screen each patient for delirium at least once during each 12-hour shift using the CAM-ICU. Often, patients are screened more frequently due to acuity and fluctuations in mental status. All available documented features for each CAM-ICU screening and the timing of the screening for each patient were collected.
Results
Data were collected from 20,241 unique adult patient electronic medical records admitted to any of four ICUs at the study site; 54.5% of the patients were male. The mean age was 63.0 years (SD=17.4), with a range of 18 years to 110 years. The mean hospital length of stay (LOS) was 10.1 days (SD=10.8), with a range of 1.00 days to 238.6 days. The mean ICU LOS was 4.1 days (SD=6.4), with a range of 0.01 days to 130.5 days. A majority of patients (83.9%) were admitted to an ICU at the study site once during the data collection period; 12.7% were admitted twice; and 3.3% were admitted three separate times. Additionally, 97.6% of the patients were admitted under “Emergency” criteria, versus 2.1% admitted under “Elective” criteria.
Patients in the study sample were screened for delirium an average of 22.5 times during their ICU admission, as determined by documentation of “Feature 1: Acute Onset or Fluctuating Course” of the CAM-ICU. Level of consciousness, evaluated with Richmond Agitation Sedation Scale scores (RASS), ranged from -5 “Unarousable” to +3 “Very Agitated,” with an average RASS score of -0.5. Of the 20,241 patients in the sample, 5,082 patients had at least one positive CAM-ICU screening, indicating about 25.1% of patients experienced symptoms of delirium during their ICU admission. Of the 5,082 patients that experienced symptoms of delirium, approximately 38.9% had a positive documented CAM-ICU screening within 24 hours of admission to the ICU, and 48.9% of the 5,082 patients had a positive documented CAM-ICU screening within 48 hours of admission to the ICU.
Table 1 summarizes the percentage of patients with specific past medical history documented in the EMR, with comparison of those patients that did experience symptoms of delirium with those patients that did not experience symptoms of delirium. International Classification of Diseases, Tenth Revision (ICD-10) codes were used to extract past medical history. A greater proportion of patients that experienced symptoms of delirium (delirious patients) had a past medical history of AIDS, brain damage or closed head trauma, dementia, heart damage or failure, kidney damage or failure, leukemia, liver damage or failure, lung damage or failure, schizophrenia, substance abuse or withdrawal, and solid tumor, compared with the proportion of patients that did not experience symptoms of delirium (non-delirious patients). A slightly greater proportion of non-delirious patients had a past medical history of stroke, compared with delirious patients.
Table 2 summarizes the descriptive data and compares the patients that did experience symptoms of delirium (n=5,082), as evidenced by at least one positive CAM-ICU screening, with the patients in the sample that did not experience symptoms of delirium while hospitalized in the ICU (n=15,159). Time, location, and frequency of line placements, including arterial access lines, central venous catheters, and nasogastric and orogastric tubes were collected from nurses’ documentation. Events that frequently occurred in the ICU included electroencephalograms (EEGs), dialysis, echocardiograms, physical therapy evaluations and treatments, and transfer events. At the study site, EEGs, dialysis, echocardiograms, and rehabilitation and therapy evaluations are conducted in the ICU at the bedside. The transfer events referred to any event that required the patient to leave the ICU and return, including computed tomography (CT) scans, magnetic resonance imaging (MRI), and surgeries. These ICU events were collected from patient orders and nurse documentation, including the date, time, and frequency of the events. Additionally, because management of patient temperature can be pivotal in preserving brain and organ function, data were collected related to documented temperatures greater than 37.80 c, as well as the use of cooling and heating devices.
Table 2 describes the percent of patients in each group that had at least one occurrence of the clinical factor while hospitalized in the ICU. For example, approximately 32.2% of the 5,082 patients that experienced symptoms of delirium had at least one occurrence of a radial arterial line placed, while approximately 11.4% of the 15,159 patients that did not experience symptoms of delirium had at least one occurrence of a radial arterial line placed. Across many of the clinical factors, there were higher proportions of delirious patients that were subjected to the procedure or event compared to non-delirious patients. This was true for all arterial lines, all central venous catheters, nasogastric tubes, orogastric tubes, general EEGs, echocardiograms, dialysis, physical therapy evaluations, red blood cell and plasma transfusions, temperatures greater than 37.80 c, use of cooling and warming devices, administration of all doses of nicotine, and administration of steroids. For example, 40.6% of the 5,082 patients that experienced symptoms of delirium had at least one nasogastric tube placed during their ICU admission, while only 9.4% of the 15,159 patients that did not experience symptoms of delirium had at least one nasogastric tube placed during their ICU admission. There was a slightly higher proportion of non-delirious patients (96.4% of the 15,159 patients) that had at least one occurrence of a transfer event, compared to the delirious patients (95.9% of the 5,082 patients).
Table 3 describes the laboratory values extracted from the patients’ EMR. The values represent the averages of each blood count from the patients’ entire ICU stay, and compare patients that experienced symptoms of delirium (delirious patients) with those that did not experience symptoms of delirium (non-delirious patients). Because many laboratory values in the ICU are collected based on patient diagnosis and course of treatment, the sample sizes for each laboratory value vary. Both delirious and non-delirious patients had average white blood cell counts, red blood cell counts, and blood glucose values that were outside of normal range based on the study site standards. Pre-albumin and albumin levels for both delirious and non-delirious patients were within the normal ranges per the study site standards.
Discussion
The descriptive data presented here contributes to better characterization of patients that are most at-risk for developing incident delirium during their ICU stay. The calculated incidence of ICU delirium in this sample was 25.1%, which is consistent with other studies with smaller sample sizes evaluating the presence of delirium in mixed ICU patient populations. Almost half (48.9%) of the 5,082 patients who developed symptoms of delirium did so within 48 hours of admission to the ICU. The relatively quick development of delirium symptoms suggests nurses and clinicians need to have prevention and management plans in place upon the patients’ arrival to the ICU. Additionally, comparison of delirious and non-delirious patient groups provides context for clinicians and researchers working to understand development of delirium. For example, the greater proportion of delirious patients that had at least one occurrence of NG or OG tube placement, compared to non-delirious patients, prompts clinicians to think about reasons ICU patients might require NG or OG tube placement and foster thoughts about other treatment modalities that would not place these patients at risk for developing delirium. Similarly, further work is needed to understand why there were greater proportions of delirious patients that had at least one occurrence of specific medications administered, such as steroids, compared to non-delirious patients. Studying medications commonly administered in the ICU could contribute to a better understanding of delirium and patients who develop delirium. Comparison of the prevalence of clinical factors in patients that experienced symptoms of delirium with the prevalence of clinical factors in patients that did not experience symptoms of delirium suggests there could be significant relationships between occurrence of clinical factors and development of delirium, requiring further exploration of each of the individual factors collected and reported here. One small limitation of this study is that all data retained for analysis were from the patients’ ICU stay, rather than the patients’ entire hospitalization.
A strength of this study is the purposeful inclusion of all patients admitted to all ICUs at the study site. This allowed us to obtain a large, varied sample that captures the full breadth of ICU patients. Additionally, this study included several commonly-occurring clinical factors that may have previously been overlooked in relation to the development of delirium. Though our calculated incidence of delirium is consistent with other smaller studies, it is important to note that the presence of delirium symptoms was extracted from nursing documentation using the CAM-ICU screening instrument. Though the lack of medical diagnosis for delirium could limit the accuracy of our calculated incidence of delirium, the CAM-ICU is considered the gold standard for screening for incident delirium in clinical practice. Specifically at the study site, the CAM-ICU has been used in practice since 2008 and ICU nurses complete required annual educational updates related to the use of the CAM-ICU to detect ICU delirium, indicating the ICU nurses have comprehensive clinical knowledge of delirium.
Future interests under investigation within this dataset include the timing of clinical factors relative to the onset of delirium symptoms as indicated by positive CAM-ICU screenings. Further work needs to be done to specify the associations between clinical factors and development of ICU delirium. Our goal was to evaluate and describe the clinical factors often present in ICUs thereby providing the foundation for determining which clinical factors might have predictive ability in determining patients at-risk for developing symptoms of delirium. By analyzing factors common in ICUs, we sought to understand ICU patients more fully, specifically comparing those who develop delirium with those who do not.
Acknowledgements
The authors would like to acknowledge the data analytics team at the study site for their time and expertise related to data extraction.
Author Contributions
Dr. Sue Lasiter conceptualized the study and was the primary investigator, overseeing development of study design, methodology, data collection and analysis, and interpretation. Dr. Steven Chesnut was the statistician for the study and was involved in study design, data collection, and data analysis. Katie Callahan is a PhD student at the University of Missouri-Kansas City and was the research assistant for the study; she contributed to study design, data collection and analysis, interpretation, and provided the initial draft of the manuscript. All authors revised and approved the final manuscript.
Ethics Statement
The Institutional Review Board (IRB) at the University of Missouri-Kansas City (UMKC) reviewed the study protocol and determined that it is exempt, with a reliance agreement established with the study site IRB (IRB # 2016859).
Funding Sources
This work was supported by a grant from Funding for Faculty Excellence at the University of Missouri-Kansas City.
Declaration of Interests
The authors have nothing to declare.