Delirium and Associated Length of Stay and Costs in Critically Ill Patients

Purpose Delirium frequently affects critically ill patients in the intensive care unit (ICU). The purpose of this study is to evaluate the impact of delirium on ICU and hospital length of stay (LOS) and perform a cost analysis. Materials and Methods Prospective studies and randomized controlled trials of patients in the ICU with delirium published between January 1, 2015, and December 31, 2020, were evaluated. Outcome variables including ICU and hospital LOS were obtained, and ICU and hospital costs were derived from the respective LOS. Results Forty-one studies met inclusion criteria. The mean difference of ICU LOS between patients with and without delirium was significant at 4.77 days (p < 0.001); for hospital LOS, this was significant at 6.67 days (p < 0.001). Cost data were extractable for 27 studies in which both ICU and hospital LOS were available. The mean difference of ICU costs between patients with and without delirium was significant at $3,921 (p < 0.001); for hospital costs, the mean difference was $5,936 (p < 0.001). Conclusion ICU and hospital LOS and associated costs were significantly higher for patients with delirium, compared to those without delirium. Further research is necessary to elucidate other determinants of increased costs and cost-reducing strategies for critically ill patients with delirium. This can provide insight into the required resources for the prevention of delirium, which may contribute to decreasing healthcare expenditure while optimizing the quality of care.


Introduction
Delirium is defined as an acute, fluctuating disturbance in attention and awareness, with additional alterations in cognition, not explained by a preexisting neurocognitive disorder or generalized medical condition [1]. Delirium often occurs in the context of multiorgan failure and critical illness and therefore is common within the intensive care unit (ICU). Up to 40% of patients in the ICU experience delirium, of which 60-90% are mechanically ventilated [2][3][4][5][6]. Patients that experience delirium within the ICU have worse outcomes, including higher mortality, increased rates of mechanical ventilation, and longer length of stay (LOS) [4,[6][7][8][9].
is likely translates into increased costs, which is supported by previous prospective studies demonstrating delirium is associated with up to 40% higher ICU and hospital costs, compared to patients without delirium [11,12]. erefore, prevention or early identification of delirium in the ICU may represent an area of optimizing healthcare spending and reducing costs. While previous review articles have analyzed the effect of delirium on clinically relevant ICU outcomes including LOS and mortality, no review articles to our knowledge have reviewed the influence of delirium on ICU costs [9,18]. e purpose of this study is to evaluate the influence of delirium on ICU and hospital LOS and associated costs, in a narrative review and cost analysis.

Search Strategy and Selection Criteria.
A narrative review and systematic literature search was conducted. We evaluated prospective observational studies and randomized controlled trials published between January 1, 2015, and December 31, 2020, in addition to studies published from 1966 to 2015 included in a previous review [9]. Databases including PubMed, EMBASE, CINAHL, e Cochrane Library, and PsycInfo were searched.
Studies with the following criteria were included: (1) observational prospective cohort studies or randomized controlled trials; (2) study population of adults (age ≥ 18 years) admitted to an ICU; (3) patients were evaluated for delirium using a validated screening or diagnostic instrument such as the Confusion Assessment Method (CAM), Confusion Assessment Method for the Intensive Care Unit (CAM-ICU), Intensive Care Delirium Screening Checklist (ICDSC), Diagnostic and Statistical Manual of Mental Disorders 4th and 5th edition (DSM-IV and DSM-V), Delirium Observation Screening Scale (DOS), or the Neelon and Champagne (NEECHAM) Confusion Scale; (4) outcomes measured included ICU LOS; and (5) articles were available in full text in English. Studies were excluded if (1) they had no comparison group of patients without delirium; (2) they were retrospective cohort or case series; (3) the largest subgroup of the patients had a primary central nervous system disorder (including stroke, traumatic brain injury, central nervous system infection, brain tumour, or recent intracranial surgery), were undergoing cardiac surgery or organ/tissue transplantation, were experiencing alcohol or substance withdrawal, or were diagnosed with COVID-19, since these are neurological processes considered to be distinct from delirium; or (6) the primary study endpoint was the comparative efficacy or safety of different sedative drugs.

Data Extraction.
Two investigators (CD and CS) independently screened titles, abstracts, and full-text articles based on the above inclusion and exclusion criteria and extracted data from the relevant included studies. Discrepancies were handled through team discussion. Information was extracted using a standardized form, which included eligibility criteria, diagnostic tool used for identification of delirium, patient characteristics, illness severity, organ dysfunction scores, and outcomes (ICU and hospital LOS).

Statistical Analysis.
e primary analysis compared the mean differences in hospital and ICU LOS and costs between patients with and without delirium. First, mean hospital and ICU LOS were obtained based on the summary statistics reported in the studies. For studies that reported median and interquartile ranges for the LOS, means were calculated using the method proposed by Wan et al. [19]. ese were calculated for patients with and without delirium.
Hospital and ICU costs were derived by multiplying the mean LOS for delirious and nondelirious patients (across all included studies) by their respective costs per day, using the methodology by Kahn et al. and applied in other studies [20][21][22]. For estimated ICU costs, daily direct variable costs were as follows: day 1 $3,678, day 2 $1057, day 3 $839, day 4 $834, and day 5 $690 onward. Estimated hospital costs were calculated by using $249/day in addition to the total ICU cost [20]. Mean LOS was rounded up to the nearest day for these calculations. Direct variable costs, which exclude equipment, salaried labor, and other fixed costs, were used because they best reflect direct and immediate economic impact associated with reducing LOS [23,24]. Costs were reported in USD with standard error and inflated to December 2020 prices according to the Consumer Price Index [25].
Means and mean differences in LOS and costs were calculated using the DerSimonian-Laird random-effects model with OpenMeta [Analyst] (version Yosemite Build, Centre for Evidence-Based Medicine, Brown School of Public Health, Providence, RI). p values <0.05 were considered statistically significant.
All studies reported ICU LOS. ere were two studies for which mean LOS was combined from two patient groups: Ouimet et al. and Yamada et al., "No delirium" and "Subsyndromal delirium" groups were combined [26,63]. e mean ICU LOS for patients with delirium was 9.40 ± 0.47 days, compared to a mean LOS of 3.39 ± 0.07 days for patients without delirium. e mean difference of the ICU LOS between patients with and without delirium was significant at 4.77 days (95% CI 3.94 to 5.60, p < 0.001). Of these studies, the hospital LOS was available for 27 studies. For patients with delirium, the mean hospital LOS was 22.3 ± 2.78 days, as compared to 16.0 ± 4.00 days for patients without delirium. e mean difference of hospital LOS between patients with and without delirium was significant at 6.67 days (95% CI 5.51 to 7.82, p < 0.001). ese data are displayed in Table 2.
We calculated costs data for the 27 studies in which both ICU and hospital LOS were available, given that hospital costs were obtained directly from ICU costs. e mean ICU cost for patients with delirium was $12,935 ± $556, compared to a mean cost of $9,013 ± $61 for patients without delirium. e mean difference of the ICU costs between patients with and without delirium was significant at $3,921 (95% CI $2,973 to $4,869, p < 0.001). For patients with delirium, the mean hospital cost was $20,236 ± $1,361, compared to a mean cost of $14,300 ± $1,267 for patients without delirium. e mean difference of the hospital costs between patients with and without delirium was significant at $5,936 (95% CI $4,663 to $7,209, p < 0.001).

Discussion
Delirium occurs frequently within the ICU and impacts the outcomes of critically ill patients, contributing to increased length of stay and mortality [4,[6][7][8][9]. We found that delirium in critically ill patients is associated with significantly higher ICU and hospital LOS, which has been supported by previous studies [4,[6][7][8][9]29]. Patients with delirium often require prolonged mechanical ventilation and take longer to reach a cognitive and physical state that enables discharge from acute care [2][3][4][5][6]65]. Taken together, this results in increased LOS and may explain the significantly higher LOS in the ICU and hospital found in our study for patients with delirium. Our cost analysis demonstrated that delirium is also correlated with significantly increased costs of approximately $5000 per admission, both within the ICU and  Critical Care Research and Practice hospital, which is a more novel addition to the literature. While the previous single-center, prospective studies have demonstrated that delirium is associated with increased ICU costs, this study reveals this on a larger scale and integrates hospital costs as well [11,12]. ese costs are likely to be cumulatively significant given the pervasiveness of delirium in the ICU [2][3][4][5][6]. In our study, increased costs are predominantly driven by prolonged LOS in the ICU, as hospital costs were derived from ICU LOS [66,67]. In addition, patients with delirium often require increased interventions such as numerous investigations, increased nursing care, pharmacological and physical restraints, and treatments aimed at managing the underlying cause of delirium, which are all costly [2][3][4][5][6][7][8][9]68]. We found that delirium prolonged LOS for patients by nearly one week, both within the hospital and ICU. Prior studies have demonstrated that there are especially high costs in the first week of developing delirium in the ICU, likely reflecting the increased need for procedural care and invasive mechanical ventilation in this timeframe [3,4,12].
e high cost of delirium should prompt evaluation into its prevention and early identification, as an opportunity to reduce healthcare expenditure. Recent studies have outlined recommendations for the prevention and management of delirium [69][70][71]. Measures that include optimization of sleep, mobility, and extended family visitation may reduce the risk of developing delirium while accruing minimal cost [70,72,73]. Screening tools and prediction models may identify delirium promptly and enable the implementation of early intervention and management [74,75]. For example, there is evidence that adequate pain control and avoidance of certain triggers such as benzodiazepines promote the resolution of delirium [69]. is may shorten the duration of delirium and thereby reduce LOS and associated costs. ere is no strong evidence supporting the use of pharmacologic agents to treat delirium in critically ill patients, and this may actually prolong delirium and increase costs [70]. Although dexmedetomidine was previously thought to reduce the duration of mechanical ventilation in patients with delirium, which could have reduced costs, recent literature suggests there is no difference in ventilator-free days or length of time without delirium, when compared to propofol [70,76]. Given the high prevalence of delirium in critically ill patients, the above strategies may contribute to significantly reduced costs [2][3][4][5][6]. Furthermore, by enhancing the resolution of delirium, these methods can also reduce the risk of long-term cognitive impairment and mitigate the emotional burden of family members [77,78].
ere are several limitations to this study. Firstly, ICU and hospital costs were represented in USD and estimated and derived exclusively from LOS. Costs may vary by patient demographics, country, hospital protocols, and severity of illness, although there is evidence that delirium increases LOS even when adjusted for severity of illness [4,9,29]. Furthermore, increased short-term interventions may drive up costs independently of the LOS, which may underestimate total hospital costs. However, LOS has been previously found to be the greatest predictor of ICU costs, suggesting using this method is valid [79]. Finally, our inclusion criteria necessitated the use of delirium screening tools for diagnosis of delirium, which may have excluded some studies of patients with delirium.

Conclusions
Delirium in critically ill patients results in increased ICU and hospital LOS and costs. In this study, increased costs are largely driven by ICU LOS. Further research is required to determine other factors influencing ICU and hospital costs in patients with delirium, including increased investigations, monitoring, and treatments utilized. Taken together, these findings should prompt investment in the resources necessary for the prevention, early identification, and mitigation of delirium, which may contribute to a substantial reduction of healthcare expenditure.

Data Availability
All data generated or analyzed during this study are included in this published article.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this article.