Advancing Delirium Treatment Trials in Older Adults: Recommendations for Future Trials From the Network for Investigation of Delirium: Unifying Scientists (NIDUS)

Abstract

OBJECTIVES:

To summarize the delirium treatment trial literature, identify the unique challenges in delirium treatment trials, and formulate recommendations to address each in older adults.

DESIGN:

A 39-member interprofessional and international expert working group of clinicians (physicians, nurses, and pharmacists) and nonclinicians (biostatisticians, epidemiologists, and trial methodologists) was convened. Four expert panels were assembled to explore key subtopics (pharmacological/nonpharmacologic treatment, methodological challenges, and novel research designs).

METHODS:

To provide background and context, a review of delirium treatment randomized controlled trials (RCTs) published between 2003 and 2023 was conducted and evidence gaps were identified. The four panels addressed the identified subtopics. For each subtopic, research challenges were identified and recommendations to address each were proposed through virtual discussion before a live, full-day, and in-person conference. General agreement was reached for each proposed recommendation across the entire working group via moderated conference discussion. Recommendations were synthesized across panels and iteratively discussed through rounds of virtual meetings and draft reviews.

RESULTS:

We identified key evidence gaps through a systematic literature review, yielding 43 RCTs of delirium treatments. From this review, eight unique challenges for delirium treatment trials were identified, and recommendations to address each were made based on panel input. The recommendations start with design of interventions that consider the multifactorial nature of delirium, include both pharmacological and nonpharmacologic approaches, and target pathophysiologic pathways where possible. Selecting appropriate at-risk patients with moderate vulnerability to delirium may maximize effectiveness. Targeting patients with at least moderate delirium severity and duration will include those most likely to experience adverse outcomes. Delirium severity should be the primary outcome of choice; measurement of short- and long-term clinical outcomes will maximize clinical relevance. Finally, plans for handling informative censoring and missing data are key.

CONCLUSIONS:

By addressing key delirium treatment challenges and research gaps, our recommendations may serve as a roadmap for advancing delirium treatment research in older adults.

Delirium, characterized by acute disturbances to attention, awareness, and cognition directly attributable to physiologic effects of illness, injury, or medical treatment, is a common and serious condition largely affecting older adults (patients ≥ 50 yr old), especially those in healthcare settings (1–3). Delirium often initiates a cascade of events culminating in reduced function, increased morbidity and mortality, and high healthcare costs (4). Furthermore, delirium has been strongly associated with long-term cognitive decline, including incident dementia (5–9).

Recent research advancements have improved our understanding of the risk factors, outcomes, diagnosis, and prevention of delirium (10–13). While multimodal delirium prevention interventions have proven effective in reducing delirium in acute care (14, 15) and ICU (16) settings, there is a striking lack of corresponding literature addressing delirium treatments. For the two-thirds of delirium patients who enter the hospital with delirium, prevention strategies would not be appropriate or effective (17). Currently, delirium treatment focuses on addressing precipitating factors for delirium, treating delirium symptoms (e.g., psychotic features), and preventing complications. However, this approach has not been found to be effective, and there is a lack of evidence on what constitutes appropriate treatment for delirium.

To address this major gap in the field, the National Institutes of Health-supported Network for Investigation of Delirium: Unifying Scientists (NIDUS) sponsored a NIDUS Scientific Think Tank to develop expert consensus on recommendations to advance future treatment trials in delirium with an emphasis on older adults (age ≥ 50 yr). NIDUS is a collaborative network to advance scientific research on delirium with a specific focus on the older adult population (13, 18–21). The overall goal of the Scientific Think Tank was to address three objectives: 1) to review published randomized controlled trials (RCTs) of delirium treatment to overview the state of the evidence and identify key gaps; 2) to characterize the unique challenges for delirium treatment trials, and to formulate recommendations to address these challenges; and 3) to identify innovative approaches to consider for future delirium treatment trials. While the focus of this article is on older adults, where delirium is most common, the evidence base includes studies on all adults and the recommendations of the Think Tank were intended to be broadly applicable.

METHODS

Three NIDUS leaders convened a 39-member interprofessional expert working group made up of clinicians (physicians, nurses, and pharmacists) and nonclinicians (biostatisticians, epidemiologists, neuroscientists, and trial methodologists). The clinicians represented broad, interdisciplinary expertise, including anesthesiology, critical care, emergency medicine, geriatric medicine, internal medicine and subspecialties, neurology, nursing, palliative care, and psychiatry. Guided by an experienced medical librarian, we reviewed English-language RCTs on delirium treatment enrolling patients 18 years old or older published between 1993 and 2023; cohort studies were excluded. Key search terms included delirium, RCT, and treatment. Studies were included regardless of the specific management or treatment strategy (drug or a nonpharmacologic intervention) and whether the comparison group was another treatment, standard practice, or placebo. These studies were abstracted for the following information: specific type of intervention (pharmacologic, nonpharmacologic, or both), enrollment of only older adults (≥ 50 yr), clinical setting (ICU, postoperative, general medical unit, palliative care unit, or long-term care facility), and primary study outcomes (e.g., delirium duration, delirium severity, agitation, or other).

Based on the literature review, 43 RCTs of delirium treatment were identified. The evidence was summarized by the NIDUS leaders, and key evidence gaps were identified. To delve deeper, four expert panels were formed to further conceptualize gaps and propose recommendations to address the gaps, focused on four specific themes: pharmacological treatments, nonpharmacological treatments, methodological challenges, and novel research designs. Each expert panel received the literature summary pertinent to their area. After identifying the unique challenges of researching delirium treatments in their assigned theme area, each expert panel discussed proposed recommendations to address the challenges in an initial virtual meeting. Consensus on recommendations was achieved through panel presentations and moderated discussions at a full-day working group meeting during the 2023 American Delirium Society Annual Meeting (June 11, 2023). Each panel’s discussion topics and faculty presenters are presented in Supplemental Table 1 (http://links.lww.com/CCM/H614). Each panel summarized its main recommendations from the meeting in a follow-up summary report.

Following the live meeting, the three NIDUS leaders and four expert panel leaders met several times to review the summary reports and further refine the recommendations. Given the overlap of recommendations across panels, all were combined across groups and synthesized into a final set of recommendations for this article, with final consensus through iterative rounds of virtual meetings and draft reviews by all expert panel members.

RESULTS

Review of Published Delirium Treatment Randomized Controlled Trials

Our search identified 728 intervention RCTs for delirium prevention or treatment over the past 30 years. Only 43 (5%) focused solely on delirium treatment (22–64). Table 1 provides an overview of the evidence from the 43 RCTs, along with the key gaps identified. These gaps and challenges formed the basis for further discussion of the expert panels.

TABLE 1.

Overview and Gaps Identified From Review of Published Delirium Treatment Trials (n = 43)

Overview

Of the 43 trials, studies primarily enrolled ICU (17, 40%) or hospitalized general medicine or geriatric (16; 37%) patients; only 5 (12%) enrolled either postoperative or palliative care patients. While not the exclusive focus for most trials, older adults (age 50+) were the highest risk group for delirium across all studies.

Thirty-five trials (81%) evaluated a pharmacologic intervention, 10 (19%) evaluated a nonpharmacologic intervention; only one trial (2%) evaluated both.

Only one (10%) of the ten trials evaluating a nonpharmacologic intervention evaluated it in the context of a multicomponent intervention strategy.

Delirium duration 19 (44%), delirium severity 16 (37%), or resolution of delirium-associated agitation 5 (12%) were the most common primary study outcomes.

Less than half (18; 42%) of the trials reported the study intervention(s) to significantly improve the primary study outcome; significant improvement was observed more commonly in nonpharmacologic (50%) than pharmacologic (26%) trials.

Significant improvement was twice as likely in trials evaluating the primary outcome of delirium severity compared with delirium duration.

Key gaps identified

The potential underlying causes for delirium rarely considered.

The underlying risk for delirium (i.e., baseline vulnerability) infrequently considered and stratification by baseline risk or by presence of cogent delirium risk factors not considered.

The multifactorial nature of delirium was seldom fully addressed.

Treatments not targeted to underlying causal delirium pathophysiologic pathways; biomarkers rarely evaluated.

Intervention potency, tolerance, feasibility, and fidelity were seldom pilot-tested.

The use of a standard approach to delirium reduction across both the intervention and control groups was rarely used.

Many trials used delirium duration as the primary outcome; delirium severity and long-term outcomes were often not evaluated.

Many studies were small and potentially under-powered.

Missing data were often not anticipated and planned for, or handled appropriately in analyses.

Recommendations to Address the Unique Challenges for Treatment Trials in Delirium

The expert panels made many recommendations to address challenges in future delirium treatment trials. Many of these recommendations were well-known best practice recommendations for clinical trials in any field, including selecting appropriate comparison groups (e.g., treatment vs. placebo), randomization, blinding, assuring potency of interventions, pilot-testing interventions for feasibility and tolerance, involving patient and families in processes to improve recruitment, using well-validated outcome measures, assuring adequate sample size, and developing a well-defined analysis plan. Collaboration with a clinical trials methodologist and biostatistician is highly recommended. Given that many treatises exist on these important areas (65–69), we will not address these areas in the current article. We will focus instead on the challenges considered particularly unique for treatment trials in delirium in older adults.

The complex, multifactorial nature of delirium, its fluctuating nature, and its occurrence across multiple settings of healthcare in heterogeneous patient populations poses many challenges for development and testing of effective treatments. The panel identified eight unique challenges and made recommendations to address each challenge (Table 2).

TABLE 2.

Recommendations to Address the Unique Challenges for Treatment Trials in Delirium

Challenge	Recommendation
The multifactorial nature of delirium not considered	Delirium is inherently multifactorial; thus, effective treatments will require a multifactorial approach to target multiple vulnerability and precipitating factors.
Nonpharmacologic and pharmacologic treatment approaches each important but rarely combined	Pharmacologic treatment approaches should be combined with non-pharmacologic approaches.
Pathophysiologic mechanisms remain poorly understood	Advancing the understanding of mechanistic pathways will be essential to develop targeted and effective treatments.
Useful intermediary treatment targets are lacking	Biomarkers may serve as useful intermediary targets in treatment trials and may help to inform intervention/drug development, pathophysiologic and mechanistic understanding, and evaluation of intervention effectiveness.
Baseline vulnerability to delirium varies between patients and may affect treatment response	Select patients with moderate baseline vulnerability to delirium to optimize treatment effectiveness since this group may be most likely to benefit from treatment.
Not all delirium leads to poor outcomes; in some cases, delirium is transient and reverses quickly, with minimal adverse effects	Identify delirium episodes associated with adverse sequelae, particularly delirium of moderate or high severity that persists or is predicted to lead to deleterious long-term outcome(s) (e.g., cognitive decline) and target delirium treatments of these episodes.
Lack of measurement of outcomes of maximal clinical relevance	To best assess outcomes with maximal clinical relevance, clinical trials should use validated measures of delirium severity as the primary outcome and also measure relevant clinical outcomes, such as cognitive and functional decline, length of hospital stay, readmissions, nursing home placement, and death.
Patients with delirium often drop out of studies or have missing data	Investigators should anticipate informative censoring and missing data, and should build in plans to address at each step of the trial.

Address the Multifactorial Nature of Delirium.

Challenge.

Delirium involves a complex interaction between baseline vulnerability (e.g., cognitive impairment, multimorbidity) and precipitating factors (e.g., new infection), yet this multifactorial nature is rarely considered in the design of current treatment trials. Previous research has demonstrated the importance of addressing multiple factors, including baseline cognitive factors, when treating delirium, given its complex and multifactorial etiology (14), yet this multifactorial nature is rarely considered in the design of current treatment trials.

Recommendation.

Delirium is inherently multifactorial; thus, effective treatments will require a multifactorial approach to target multiple vulnerability and precipitating factors. While addressing these multifactorial contributors will be key to a successful treatment strategy, it remains unclear how to prioritize and target multiple vulnerability and precipitating factors in a treatment trial.

Combine Pharmacologic and Nonpharmacologic Approaches.

Challenge.

Nonpharmacologic approaches (e.g., reorientation, cognitive stimulation, physical activity, sensory aids, sleep hygiene, and family engagement) have been most effective for delirium prevention and likely have an important role in delirium treatment (10, 14, 15, 70–77). In our literature review, the efficacy of a pharmacologic delirium treatment intervention in combination with nonpharmacologic interventions was evaluated in only one study (27).

Recommendation.

Pharmacologic treatment approaches should be combined with nonpharmacologic approaches. Along with pharmacologic treatments, nonpharmacologic protocols will need to be fully implemented with high fidelity by skilled staff and monitored for adherence. Dose-response analysis for both treatment types will be required.

Clarify and Target Pathophysiologic Mechanisms.

Challenge.

The pathophysiology of delirium is complex and its mechanistic pathways remain poorly understood. This has made the development and testing of pharmacologic interventions particularly challenging. Pharmacologic interventions may lack efficacy if they address only a single mechanism or pathway in the complex syndrome of delirium (e.g., dopamine excess [48, 55, 56, 64], neuroinflammation [47, 54], or gamma-aminobutyric acid-ergic burden [58, 78]). Furthermore, some treatments have not been well-tolerated by frail older adults (e.g., cholinergic blockade [37]). To date, there have not yet been pathophysiologic targeted treatments for delirium tested in clinical trials.

Recommendation.

Advancing the understanding of mechanistic pathways will be essential to develop targeted and effective treatments. Given the current gaps in understanding delirium pathogenesis, it is crucial to direct resources toward basic and translational science approaches to clarify mechanistic pathways contributing to delirium (13). This approach is key to developing appropriate and targeted treatments for delirium in older adults.

Use Biomarkers As Intermediary Targets.

Challenges.

For a complex syndrome like delirium, it is likely a causal mediating factor(s) exists in certain subgroups that may influence treatment response (79). For example, the magnitude of the postoperative inflammatory responses varies among individuals with delirium (80–82), anti-inflammatory treatments may be more effective in individuals with a greater inflammatory burden. However, the identification of these mediating factors is often challenging, and biomarker measurements have rarely been integrated into delirium treatment trials. While biomarkers hold promise for identifying subgroup effects and/or mediating factors in delirium treatment trials, they face challenges related to protocol complexity and cost.

Recommendation.

Biomarkers may serve as useful intermediary targets in treatment trials and may help to inform intervention/drug development, pathophysiologic and mechanistic understanding, and evaluation of intervention effectiveness. The identification of a mediating biomarker (e.g., blood, cerebral spinal fluid) can help inform future trials evaluating the intervention, to select patients at high risk, or to serve as markers of intervention response (82, 83). Collaboration with experts in preanalytical sample handling and bioassay performance will ensure the accuracy, precision, and validity/interpretability of biomarker data (84, 85). Recent publications on this topic (86–88), including a NIDUS white paper on delirium preclinical and translational models (13), may help to inform the selection of specific biomarkers for use in treatment trials.

Address Baseline Vulnerability to Delirium.

Challenge.

The degree of underlying baseline vulnerability for delirium in patients who develop delirium may influence the success of treatment interventions (3), yet careful selection of patients based on baseline risk has not been a priority in published studies. Patients with delirium having moderate baseline vulnerability are more likely to benefit from treatment compared with patients with low baseline vulnerability. Furthermore, patients with high baseline vulnerability, who often have more severe delirium, may have a lower response to treatment.

Recommendation.

Select patients with moderate baseline vulnerability to delirium to optimize treatment effectiveness since this group may be most likely to benefit from treatment. Identifying baseline risk level may involve development and validation of predictive models for delirium risk (89–91).

Target Delirium Associated With Poor Outcomes.

Challenge.

Not all delirium leads to poor outcomes. In some cases, delirium is transient and reverses quickly, with minimal adverse effects. To optimize efficiency and effectiveness, clinical trials should target delirium of at least moderate severity and duration. Patients with delirium of at least moderate severity and duration are most likely to have adverse outcomes of delirium, such as long-term cognitive decline.

Recommendation.

Identify delirium episodes associated with adverse sequelae, particularly delirium of moderate or high severity that persists or is predicted to lead to deleterious long-term outcome(s) (e.g., cognitive decline), and target delirium treatments for these episodes.

Measure Outcomes With Maximal Clinical Relevance.

Challenges.

Delirium measurement instruments vary widely by purpose, population focus, users, characteristics, and domains (3, 92, 93). While the primary outcome for delirium prevention trials has been incident delirium, the outcomes used for published delirium treatment trials have focused around delirium prevalence and duration. These may not represent the most useful or sensitive outcomes to evaluate treatment effectiveness.

Recommendation.

To best assess outcomes with maximal clinical relevance, clinical trials should use validated measures of delirium severity as the primary outcome and also measure relevant clinical outcomes, such as cognitive and functional decline, length of hospital stay, readmissions, nursing home placement, and death. Studies have documented that delirium severity, or a combination of severity plus duration, is the outcome that best predicts important clinical outcomes of delirium (94). Ideally, both short-term (e.g., hospital-based outcomes) and long-term outcomes (e.g., outcomes at one to several years follow-up) should also be assessed (95, 96). The instrument(s) chosen to measure an outcome should effectively measure the targets of treatment. Instrument sensitivity and specificity should be reported.

Delirium severity should be measured on a quantitative scale; the optimal severity instrument to use in treatment trials depends on multiple factors including the patient population being studied (86–98). While guidance on specific outcome measures is beyond the scope of this article; the NIDUS website (99) provides a wealth of information on relevant delirium measures, including delirium severity and delirium core outcome sets (100–104).

Anticipate Informative Censoring and Missing Data.

Challenge.

Delirium is associated with poor clinical outcomes, such as nursing home placement and death, and can persist long after hospital discharge. Thus, patients with delirium often drop out of studies or have missing data. Informative censoring (i.e., loss of data due to the condition under study) and missing data may threaten the validity of the trial results.

Recommendation.

Investigators should anticipate informative censoring and missing data, and should build in plans to address at each step of the trial. Follow-up plans for trials in older adults should include rigorous methods to gather as much data as possible on every patient in the long-term (including remote assessments, proxy interviews, or chart review when patient interview is not possible). Analytic plans should use formal methods (such as imputation) to account for missing data and incorporate sensitivity analyses. Worst- and best-case assumptions should be made to assess the impact of missing values.

Innovative Approaches to Consider for Future Delirium Treatment Trials

The four expert panels identified innovative approaches to consider in the design of future delirium treatment trials (Table 3). The application of these approaches will be novel to delirium trials and will be important to advance future clinical trials in delirium.

TABLE 3.

Innovative Approaches to Consider for Future Delirium Treatment Trials

Innovation	Recommendation
Precision approaches	Individualized treatment will be required according to the patient’s risk factors, and clinical and biomarker profiles.
Adaptive trial designs	Consider use of an adaptive trial design to evaluate multiple delirium treatments across heterogeneous groups.
Pragmatic trials	Use a pragmatic trial design to test whether the treatment approach works in routine clinical practice.
Pragmatic outcomes	Consider pragmatic outcomes for delirium treatment trials, which might include application of artificial intelligence or natural language processing approaches to large claims databases or electronic health records.
Delirium subtyping	Delirium phenotype should be carefully measured and accounted for in the statistical analysis, and its association with response to therapy documented.

Precision Approaches.

Delirium is a complex syndrome, and every patient has a unique delirium risk factor and biomarker profile. At present, a well-validated biomarker (or biomarker signature) for delirium has not yet been identified that will serve as a treatment target for future trials (3). As previously recommended, patients with at least a moderate vulnerability to delirium should be prioritized for treatment. Precision approaches that include individualizing treatment to each patient’s risk factors and their clinical and biomarker profile holds potential promise.

Recommendation.

Individualized treatment will be required according to the patient’s risk factors and clinical and biomarker profiles. Ideally, these strategies will address the patient’s underlying pathophysiologic triggers, as well as vulnerability and precipitating risk factor profile. Future clinical trials will be needed to develop and test these personalized approaches for the precision treatment of delirium. Application of basic and translational research approaches (13, 105, 106), including advanced molecular techniques and biomarkers, combined with clinical assessment will be needed to design, target, and appropriately monitor these treatment approaches.

Adaptive Trial Designs.

Nearly all RCTs for delirium treatment have used traditional statistical approaches to provide evidence against the null hypothesis. The design of these trials precludes the ability to modify or adapt trial features (e.g., sample size, treatment arms) based on accumulated data (107). Because delirium treatment typically involves multiple complex interventions targeted to heterogeneous patient populations (3), consideration of more flexible, novel designs may help to advance the field. Adaptive clinical trials often use Bayesian statistics to allow investigators to intermittently compute a posterior probability of the benefit of delirium treatments as trial data is collected and to use these results to make informed revisions to improve trial design (e.g., altering randomization probabilities, intervention doses, dropping or adding treatment arms, altering sample size, and/or refining the study population) (108).

Recommendation.

Consider use of an adaptive trial design to evaluate multiple delirium treatments across heterogeneous groups. These trials are well-suited to investigate delirium treatments given they can reduce the number of participants exposed to less effective (or less safe) treatments, are responsive to heterogeneity of treatment effect, and can improve efficiency in the face of uncertainty (109). This approach can facilitate informed decisions about delirium trial futility (either efficacy or safety) or continuation, and thus, improve trial efficiency and reduce costs (110, 111). With response-adaptive randomization, the probability of assignment to the best interventions can be increased, and new study arms can either be activated or dropped (112).

Pragmatic Trials.

Traditional, explanatory delirium treatment trials often suffer from low external validity. Outcomes in explanatory trials are usually labor-intensive, requiring trained assessors. Patient consent is routinely needed before enrollment in explanatory trials, yet often must be obtained from the legally authorized representative in patients with delirium. While explanatory trials are a necessary first step to establish efficacy of a new treatment, external validity and real-world applicability can be better established with more pragmatic approaches.

Recommendation.

Use a pragmatic trial design to test whether the treatment approach works in routine clinical practice. In contrast to explanatory trials, pragmatic trials focus on whether the delirium treatment works in a real-world clinical setting (113). The high external validity of pragmatic trials results from the fact that they are conducted in regular clinical settings, enroll a heterogeneous population (due to a small number of exclusions), institute a waiver of informed consent, have clinicians (rather than researchers) administer the intervention consistent with routine practice (113, 114), and evaluate outcomes collected as part of routine clinical care. Evaluation of many treatment approaches may be amenable to pragmatic clinical trials, such as deprescribing benzodiazepines or other delirium-associated medications (115). Stepped wedge, cluster-randomized trials, which are often pragmatic, can also be considered to evaluate the effectiveness of nonpharmacologic, multicomponent delirium reduction interventions (14, 116). Researchers should also consider hybrid trial designs incorporating elements of an explanatory trial (e.g., rigorous gold standard outcome measurements for delirium) or effectiveness-implementation trial approaches (e.g., mixed methods to evaluate implementation effectiveness) in pragmatic trials (117, 118).

Pragmatic Outcomes.

Most traditional clinical trials for delirium use outcomes measured by research staff, such as bedside interviews with application of delirium instruments. However, these approaches can be labor-intensive, and limit the size of the trial, and the generalizability of the results. Furthermore, the advancement of artificial intelligence approaches will facilitate development of pragmatic measures in delirium. Approaches to collect the outcomes with maximal clinical relevance in clinical trials are evolving.

Recommendation.

Consider pragmatic outcomes for delirium treatment trials, which might include applications of artificial intelligence or natural language processing approaches to large claims databases or electronic health records (119–122). Wearable or remote digital technologies hold tremendous possibilities as pragmatic outcomes for trials (123, 124). Furthermore, the use of delirium screening in routine nursing practice (3, 16, 19) may also enhance approaches, and provide healthsystem level data on delirium rates.

Delirium Subtyping.

Delirium can present with many different phenotypes, such as motoric subtypes (hyperactive, hypoactive, normoactive, and mixed) or symptom subtypes (presence of delusions and hallucinations vs. not). Importantly, these phenotypes may reflect differing etiologies or the involvement of differing neural and/or mechanistic pathways (68). Phenotypes have rarely been considered in published treatment trials.

Recommendation.

Delirium phenotype should be carefully measured and accounted for in the statistical analysis, and its association with response to therapy documented. Advancing the understanding of phenotypic response to treatment may help to more appropriately target treatments, and also to clarify underlying pathophysiology.

DISCUSSION

Our white paper elucidates the current evidence gaps in the design and conduct of delirium treatment trials in older adults, highlighting eight unique challenges of these trials and proposing recommendations to address these challenges. We did not address well-known best practice recommendations for clinical trials in any field; rather, we focused on eight challenges considered particularly unique for treatment trials in delirium. Focusing on these unique challenges will enhance the effectiveness and utility of future trials.

The recommendations focus around foundational areas, starting with intervention design—considering the multifactorial nature of delirium; including both pharmacological and nonpharmacologic approaches; and targeting pathophysiologic pathways and measuring biomarkers where possible. Selecting appropriate at-risk patients with moderate vulnerability to delirium is likely to make the treatment trial more effective and efficient since low-risk patients may not need the intervention, and high-risk patients may not respond. Similarly, targeting patients with at least moderate delirium severity and duration will include those most likely to experience adverse outcomes and long-term cognitive decline. Delirium severity should be the primary outcome of choice, and well-validated measures should be used. Measurement of additional short-term and long-term clinical outcomes will also maximize clinical relevance of the trial. Finally, robust analytic plans should anticipate and plan for handling informative censoring and missing data.

Clinical trials methodology is constantly advancing, and we address several innovative approaches to be considered for future delirium trials in older adults. While not addressed in detail, the areas of individualized precision medicine, adaptive trials, pragmatic trials, and pragmatic outcome assessment may truly help to advance delirium treatments. Furthermore, targeting treatments to delirium subtypes may help enhance effectiveness and advance our pathophysiologic understanding of delirium.

We must acknowledge potential limitations of our approach. We were not able to provide detailed guidance on how each recommendation can be applied but provide additional resources where possible. We did not specifically address the ethical challenges in delirium treatment trials, particularly concerning vulnerable populations and the frequent need for proxy consent. We acknowledge that many areas were incompletely explored such as biomarker validation, outcome standardization, implementation science, specific drug types, and trial population inclusivity. We do not discuss the important area of assuring racial and ethnic diversity in delirium trials (125, 126) We focused on older adults in highly resourced countries, the challenges and evidence gaps in other populations (e.g., pediatrics) or resource-poor countries may be different. While we held an in-person consensus-building conference and requested iterative input from our expert panels in a collaborative process, we did not use formal Delphi consensus-building processes. Given the goals of our white paper, evidence-based graded recommendations were not considered appropriate or feasible.

In conclusion, the NIDUS Scientific Think Tank has provided a set of practical recommendations to advance the design and conduct of delirium treatment trials in older adults. Ultimately, we hope that these recommendations will lead to the development and testing of effective treatments for the common and devastating condition of delirium, which will improve clinical outcomes, and provide substantial benefits across healthcare settings and populations.

Supplementary Material

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal).

Question:

Delirium affects one in five older hospitalized adults, yet treatment strategies remain limited. An interprofessional and international expert working group identified current challenges in researching delirium treatments and designing treatment trials in older adults and made recommendations on how each can be addressed.

Findings:

The working group reviewed the published evidence on delirium treatments, and identified gaps based on 43 clinical trials to date. The group further identified eight unique challenges in delirium research and made recommendations to address each of these areas. Finally, innovative approaches for future research were identified.

Meaning:

Important research gaps and challenges remain in delirium treatment trials. The proposed recommendations may serve as a roadmap for advancing delirium treatment trials in older adults.