Skip to main content

A systematic review of implementation strategies for assessment, prevention, and management of ICU delirium and their effect on clinical outcomes

Abstract

Introduction

Despite recommendations from professional societies and patient safety organizations, the majority of ICU patients worldwide are not routinely monitored for delirium, thus preventing timely prevention and management. The purpose of this systematic review is to summarize what types of implementation strategies have been tested to improve ICU clinicians’ ability to effectively assess, prevent and treat delirium and to evaluate the effect of these strategies on clinical outcomes.

Method

We searched PubMed, Embase, PsychINFO, Cochrane and CINAHL (January 2000 and April 2014) for studies on implementation strategies that included delirium-oriented interventions in adult ICU patients. Studies were suitable for inclusion if implementation strategies’ efficacy, in terms of a clinical outcome, or process outcome was described.

Results

We included 21 studies, all including process measures, while 9 reported both process measures and clinical outcomes. Some individual strategies such as “audit and feedback” and “tailored interventions” may be important to establish clinical outcome improvements, but otherwise robust data on effectiveness of specific implementation strategies were scarce. Successful implementation interventions were frequently reported to change process measures, such as improvements in adherence to delirium screening with up to 92%, but relating process measures to outcome changes was generally not possible. In meta-analyses, reduced mortality and ICU length of stay reduction were statistically more likely with implementation programs that employed more (six or more) rather than less implementation strategies and when a framework was used that either integrated current evidence on pain, agitation and delirium management (PAD) or when a strategy of early awakening, breathing, delirium screening and early exercise (ABCDE bundle) was employed. Using implementation strategies aimed at organizational change, next to behavioral change, was also associated with reduced mortality.

Conclusion

Our findings may indicate that multi-component implementation programs with a higher number of strategies targeting ICU delirium assessment, prevention and treatment and integrated within PAD or ABCDE bundle have the potential to improve clinical outcomes. However, prospective confirmation of these findings is needed to inform the most effective implementation practice with regard to integrated delirium management and such research should clearly delineate effective practice change from improvements in clinical outcomes.

Introduction

The problem of delirium is far from an academic one. Not only does the presence of delirium often complicate and render more difficult the treatment of a serious illness, but also it carries the serious possibility of permanent irreversible brain damage’, Engel and Romano [1].

This quote, written over 50 years ago by icons in the field of medicine, would seem to be a clarion call for those caring for humans suffering from serious disease. Elsewhere in the same classic manuscript, Engel and Romano make two statements about inadequacies of the approach taken by healthcare professionals in treating delirium: They seem to have little interest in and, indeed, often completely overlook delirium’ [1,2] and The deficiencies in the education of many physicians will equip them to recognize any but the most flagrant examples of delirium.’ Even when armed with the wealth of information present in the literature over the past decade about the importance of assessing, preventing and managing delirium in the ICU, effecting the needed changes in care through appropriate implementation programs still requires a substantial change in culture and attention to human factors that are often beyond the scope of training of most medical teams.

In the Society of Critical Care Medicine’s recently released Clinical Practice Guideline for the Management of Pain, Agitation, and Delirium (PAD) in Adult Patients in the ICU current evidence is brought together on optimal management of pain, agitation, sedation and delirium [3]. A previously constructed framework to facilitate the implementation of many aspects of the evidence described in the PAD guidelines is the awakening and breathing coordination, choice of sedative, delirium monitoring and management and early mobility (ABCDE) bundle. The ABCDE bundle is specifically aimed at minimizing sedation, encouraging early ventilator liberation, improving delirium assessment and management and facilitating early mobilization in the ICU [4]. Importantly, both the protocols of the trial that established the value of the ABCs [5] and the seminal randomized controlled trial (RCT) that established the positive effects of early mobilization in critically ill patients [6] included routine daily delirium assessments with the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU), with the latter study even establishing a significant reduction in delirium incidence. Therefore, current evidence suggests that: 1) clinical effectiveness of the ABC and E within the ABCDE bundle implies routine delirium assessment with a validated tool, and, inversely, 2) delirium prevention and management requires an integrated multidisciplinary approach with standardized care processes including early mobilization, which in turn is linked to a strategy of minimizing sedation by means of ‘awake(ning) and spontaneous breathing coordination’. As such, ‘brain failure’ (that is, delirium and coma) may be regarded as avoidable and representing an intermediate state on the pathway towards adverse outcomes, such as death and increased length of ICU stay [7]. However, although from the ABCDE bundle or PAD guidelines it may seem evident what to aim for in everyday clinical practice, health care professionals often struggle with how to implement guidelines, especially when these include integrated care covering many domains concurrently and involving multiple care providers.

Therefore, this systematic review of the literature aims at summarizing the implementation strategies and their effectiveness to improve practices of assessment, prevention or management of delirium and clinical outcomes in the critically ill.

Methods

Search strategy and selection criteria

This review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [8]. We searched PubMed, Embase, PsychINFO, Cochrane and CINAHL for studies published between January 2000 and April 2014 with no search filter limits. The year 2000 was chosen because a preliminary Pubmed search with the search terms ‘delirium’, ‘implementation’ and one of ‘ICU’, ‘critically ill’, or ‘critical care’, yielded only one study that year that pertained to the subject of this review and none before [9]. A biomedical information specialist at the medical library of the Erasmus MC - University Medical Center Rotterdam guided the search. Search terms included intensive care and delirium, and were tailored to each database and its indexing system (see Additional file 1). Reference lists of retrieved articles, reviews and books were screened to identify additional papers that met the inclusion criteria.

Selection of studies

Our search focused on clinical studies aimed at implementation of delirium screening, prevention or management in the adult ICU setting. Implementation could be focused at single components of delirium care (for example, delirium screening) or could include delirium screening, prevention and/or management as an integral part of a wider bundle or guideline (for example, ABCDE bundle or PAD guideline). We considered the PAD guideline and the ABCDE bundle as similar for the purpose of this review because, next to delirium screening, they share several integrated evidence-based care components (early mobilization, awakening and breathing coordination or targeting light sedation and systematic pain assessment and treatment). We did not limit the search to specific types of ICU. To be included in the review, the study had to contain a clear description of the implementation process (that is, an explanation of what exactly was done to implement it). We excluded studies that concerned delirium related to alcohol withdrawal and/or were focused solely on validation of delirium screening tools. Further, the efficacy of the implementation intervention had to be reported in terms of a clearly defined outcome such as mortality, length of stay, and/or adherence to delirium screening. Reviews, opinion papers, editorials and comments on original articles were also excluded.

Two authors (ZT, EI) independently checked abstracts of retrieved articles on compliance with selection criteria. Relevant full-text articles were checked for final inclusion. Consensus on final selection was achieved by discussion with a third author (MJ).

Data extraction and synthesis

The first reviewer (ZT) extracted data on design, population, implementation strategies, and outcomes and studies were subject to further critical appraisal by two other authors (EI, MJ). The individual implementation strategies were classified into four categories: professional (for example, distribution of educational materials, reminders), organizational (for example, provider-oriented interventions, structural interventions), financial and regulatory (for example, peer review, changes in medical liability) using the Cochrane Effective Practice and Organization of Care group (EPOC) classification system checklist (Table 1) [10]. From these 4 categories, we then distinguished 17 individual implementation strategies (Table 1). The implementation strategies concern all phases of a formal implementation process as has been described before in the literature [11]. For instance, the strategy of ‘marketing/tailored interventions’ includes first performing an analysis of barriers to implementation to be able to design a subsequent implementation strategy addressing these barriers, to enhance implementation effectiveness. As such, the use of more strategies concurrently may indicate a more complete implementation process.

Table 1 Implementation strategy taxonomy according to the EPOC classification system

With regard to the outcomes, we distinguished between clinical outcomes (ICU length of stay (LOS) and mortality) and process outcomes (adherence to screening for the presence of delirium, knowledge of delirium, incidence of delirium, use of antipsychotics) [12]. Changes in these outcomes were assessed before and after implementation (or with and without implementation in the case of the only RCT included). Three authors (EI, ZT, MJ) independently scored the implementation strategies in the implementation studies reporting clinical outcomes. Differences in assessment were resolved afterwards by discussion. The studies that did not report mortality were assessed equally by two authors (ZT, MJ). We tabulated the key features deemed important for this review of all included studies: number and type of implementation strategies, care components (that is, using integrated strategy such as PAD/ABCDE or separate interventions such as screening only), implementation model and the process and clinical outcomes as previously defined.

Methodological quality

We rated the methodological quality of all implementation studies in an effort to ascertain a minimum quality of included studies. We used a rating system adapted from Anderson and Sharpe [13] (see Additional file 2), which evaluated the impact of various types of intervention on the behavior change of health care workers, in line with our review. Two reviewers (ZT/EI) independently assessed each study on quality and differences in quality scores were resolved through discussion. Studies that rated fewer than three points were excluded because of very poor methodological quality.

Statistical analyses

Associations between study characteristics and outcomes were assessed with Pearson’s chi-square or Fisher’s exact test after dichotomization (for example, significant decrease of delirium incidence: yes/no). The number of implementation strategies used in the implementation studies was summarized as median with IQR.

Whenever possible, for meta-analysis we quantitatively pooled the results at the patient level for the included studies when the original data were retrievable. We contacted the authors of the original articles for these data when not provided in the published paper. We expressed the effectiveness of the implementation interventions as a risk ratio (RR) for dichotomous outcomes by using a DerSimonian and Laird random-effect model [14] and as a weighted mean difference (WMD) for continuous outcomes with 95% CIs. The heterogeneity among studies was tested using the Cochran Q-test of heterogeneity, and Higgins and Thompson I2 [15]. The degree of heterogeneity was defined as a value of I2: low (25% to 49%), moderate (50% to 74%), and high (>75%) values [15]. Subgroup analysis was performed for number of implementation strategies (low number = below median, high number = median or higher), and use of either PAD guideline/ABCDE bundle. Analysis was performed with Microsoft Excel 2013 and IBM SPSS 21.0. Statistical significance was defined as a P-value <0.05.

Results

Selection of studies

We reviewed 3,981 hits and after excluding duplicates and studies not meeting inclusion/exclusion criteria, 21 studies were evaluated [16-36] (Figure 1). Mortality and ICU-LOS changes were reported in ten studies [16,20,24,26-28,30,32,35,36] and in one study ICU-LOS was reported but not mortality [33]. One publication was a duplicate with regard to study period and population and was therefore excluded from the analysis of clinical outcomes but included in the assessment of studies that reported process measures [27]. Sixteen of twenty-one included studies were before-after studies; one was an RCT, and the remaining studies were prospective or retrospective cohort studies.

Figure 1
figure 1

Selection of included studies for the review.

Methodological quality

One study was of very low methodological quality (2 points) and was excluded [37] (see Additional file 3: methodological quality rating of included studies and Figure 1). This study was a randomized trial but details on randomization, interventions and assessment of delirium were insufficient with regard to reproducibility.

Implementation strategies

Implementation strategies that were used in the 21 included studies reporting process and clinical outcomes are shown in Table 2 (strategies are explained in Table 1). These studies were published between 2005 and 2014. Professional-oriented strategies (that is aimed at changing professionals’ behavior) and organizational strategies (that is aimed at changing the structure of care delivery) were the most frequently used categories of implementation strategies. Of the professional-oriented strategies, education (meaning one or both of the following strategies: distribution of educational material (81%) and/or educational meetings (100%), was used in all studies (Tables 1 and 2). Patient-mediated interventions, corresponding with implementation of screening for delirium with a validated tool such as CAM-ICU, was applied in 86% of the studies, whereas outreach visits, audit and feedback and local consensus processes were applied in 67%, 62% and 57% of the studies respectively (Table 2). Three of the seventeen implementation strategies were not used at all (that is mass media, changes in medical liability and management of patient complaints). Three strategies were used in only one or two studies (provider-oriented interventions/financial compensation [24], licensure [16] and patient-oriented interventions [16,31]). Tailored interventions were used in 33% of the studies [16,20].

Table 2 Summary of implementation strategies

Implementation characteristics, process outcomes and clinical outcomes

The number of implementation strategies used varied from 4 to 12 per study (Table 3). The overall median number of implementation strategies used per study was 7.0 (IQR 4.5 to 9.5). In the studies reporting clinical outcomes (n = 9) versus only process outcomes (n = 12) the median number of used strategies was 6.0 (IQR 4.5 to 8.0) and 7.0 (IQR 7.5 to 10.0) respectively (P = 0.46) (Table 2). Within the nine studies with clinical outcomes, the following implementation strategies were reported only in studies with significant mortality reduction (that is, the studies by Mansouri, Skrobik, Balas): tailoring, encouragement for implementation by means of financial incentives, licensure, and audit and feedback (Table 2). Audit and feedback was used in all studies showing significant mortality reduction but in none without significant reduction of mortality (P = 0.012). In contrast, these and other strategies were used frequently in studies that reported process outcomes without clinical outcomes. The number of strategies per study belonging to the domains of organizational, financial or regulatory implementation strategies (that is, not aimed at the professional) (Table 2) in the clinical outcome versus the process outcome studies did not differ (P = 0.92). However, within the nine clinical outcome studies the studies with significant mortality reduction after the implementation [14,24,31] used more of these non professional-oriented strategies (median 2, IQR 2 to 3) than studies without a significantly reduced mortality [26,28,30,32,35,36] (median 0.5, IQR 0.0 to 1.0, P = 0.024).

Table 3 Implementation characteristics and changes in important process and clinical outcomes before versus after implementation

Delirium screening adherence was assessed in 15 of the 21 studies, of which 13 showed a significantly increased adherence (Table 3) [16-18,21,22,24-26,29-31,33,34]. In studies specifically focused on implementation of delirium screening (n = 10), improvements in adherence to screening ranged from 14% to 92%, but the definition of adherence varied widely. These studies with focus on delirium screening typically did not report clinical outcomes (1 of 10 studies), whereas process outcomes were assessed in all of these studies (Tables 2 and 3). Significant improvement of screening adherence after the implementation was reported in 82% (9/11) of the studies that did not report on clinical outcomes, versus 56% (5/9) of the studies that assessed clinical outcomes. Use of integrated delirium management (PAD/ABCDE) was reported in 18% (2/11) of studies without clinical outcome assessment versus in 67% (6/9) of studies with clinical outcome. Knowledge improvement was reported in 4 of 21 studies and varied both in magnitude and definition [17,18,23,29]. Knowledge improvement was reported in 36% of studies (4/11) without clinical outcome data, versus 0% in studies with only process outcome data. Changes in reported delirium incidence [16,17,19,20,28,30,31,35,36] and use of antipsychotic drugs [16,17,20,24,27,30,32,36] after implementation varied between studies (some showed increased and some showed decreased incidence, Table 3). No significant associations existed between changes in the process measures (delirium incidence, use of antipsychotic drugs or screening adherence) and mortality before and after the implementation. Likewise, no significant associations were found between the process measures and ICU LOS.

In pooled analysis, we did not find differences in delirium incidence (n = 8) before versus after the implementation when comparing the studies using PAD/ABCDE versus those that did not use these frameworks, or comparing those with high versus low number of implementation strategies, and high inconsistency existed in such pooled analyses (see Additional file 4: Figures S4a,b). Implementation studies focusing on delirium screening tools did not report increased delirium incidence after the implementation compared with studies that used other frameworks (for example, PAD/ABCDE, that is, a more integrated program, see Additional file 4: Figure S4c). Pooled analysis of relations between implementation strategies and adherence rates for screening or knowledge were not possible due to highly variable definitions for the process outcomes, and irretrievable original data.

ICU-LOS

Nine of the ten studies that reported LOS showed reduced ICU LOS after implementation (the study by Eastwood and Reade were the same population); of which five were statistically significant (Tables 2 and 3). The study by Radtke et al. [26] included populations from three different ICUs and were stratified according to standard or more intensive implementation strategies. Pooling all studies that reported ICU LOS and of which data were retrieved (n = 7) showed a reduction of −1.26 days (95% CI −1.84, −0.69) after the implementation (Figure 2a). Pooled data from four studies reporting ICU LOS after implementation of PAD or the ABCDE bundle approach yielded significantly shorter LOS after implementation compared with not using these approaches (WMD = −1.71; 95% CI −2.45, −0.98 versus WMD −0.55, 95% CI −1.48, 0.38) (Figure 2a). Using a high (≥6) number of strategies showed a reduced ICU LOS (−1.51, 95% CI −2.16, −0.86) versus no change when using fewer strategies (−0.36, 95% CI −1.61, 0.89) (Figure 2b). Within the studies using PAD or ABCDE (n = 4) the signal that using more strategies reduced ICU LOS was less evident (Figure 2c). None of the individual strategies were used more often in studies with versus without statistically significant ICU LOS reduction.

Figure 2
figure 2

Pooled analysis of determinants of changes in ICU length of stay (days) in implementation studies (n = 7) that included delirium-oriented interventions. Determinants of ICU length of stay reduction that were studied were: use of either the guideline for the management of pain, agitation and delirium (PAD) or the awakening and breathing coordination, choice of sedative, delirium monitoring and management and early mobility (ABCDE) bundle (a) or use of high or low number of implementation strategies (b). (c) Impact of high or low number of strategies within the studies reporting ICU length of stay and using PAD/ABCDE (n = 4). See text for more details. Study by Radtke reported multiple populations and these were separately assessed.

Mortality

Seven of the nine studies with mortality data before versus after implementation showed a reduction in mortality ranging from 2.9% to 12% (Table 3). Mortality was most often defined as hospital mortality (n = 6), but sometimes as ICU mortality [24,36] and 30-day mortality [20]. Three of these studies reported a statistically significant decrease in mortality between 6.5% (P = 0.009) and 12% (P = 0.046, Table 3) [16,20,24]. In the pooled analysis of all (n = 9) studies with mortality data, the mortality rates after implementation declined overall (RR = 0.82; 95% CI 0.71, 0.96 (Figure 3a). There was no inconsistency between the studies for this association (I2 = 0%, P = 0.526). Studies using PAD/ABCDE reported reduced mortality, whereas studies that did not use these frameworks did not (RR = 0.81; 95% CI 0.69, 0.96 versus RR = 0.93; 95% CI 0.61, 1.42). However, this difference in mortality risk reduction between the pooled data in studies with and without PAD/ABCDE did not reach statistical significance (P = 0.531). Mortality risk reduction was significantly higher (P = 0.0424) in studies that used high number of implementation strategies (RR = 0.73; 95% CI 0.60, 0.88) compared with studies with low number (Figure 3b). Further, in the studies that used the PAD guideline or ABCDE approach (n = 6) (Figure 3c) mortality reduction was higher (P = 0.0478) in studies that used a higher number of implementation strategies (RR = 0.73; 95% CI 0.59, 0.88 versus RR = 0.98; 95% CI 0.74, 1.30).

Figure 3
figure 3

Pooled analysis of determinants of changes in mortality (risk ratio) in implementation studies (n = 9) that included delirium-oriented interventions. Determinants of mortality reduction that were studied were: use of either the guideline for the management of pain, agitation and delirium (PAD) or the awakening and breathing coordination, choice of sedative, delirium monitoring and management and early mobility (ABCDE) bundle (a) or use of high or low number of implementation strategies (b). (c) Impact of high or low number of strategies within the studies reporting mortality and using PAD/ABCDE (n = 6). See text for more details.

Discussion

This systematic review and structured analysis of the literature aimed to summarize the implementation strategies and their effectiveness to change practice with regard to delirium assessment, prevention and management in the ICU and clinical outcomes. To accomplish this goal we tried to address both the why and the how questions regarding implementation. With regard to the why, an important finding of this review indicating that multi-component implementation that included delirium-oriented interventions in critically ill patients can be useful, is that many studies reported improvements of both process outcomes (delirium screening adherence, knowledge) and clinical outcomes (short-term mortality and ICU LOS). With regard to the how, several results of this review are worth highlighting: 1) some individual strategies such as audit and feedback and tailored interventions may be important to establish clinical outcome improvements, but otherwise robust data on effectiveness of specific implementation strategies are scarce; 2) using implementation strategies targeted not only at the health care professional but also at organizational, financial or regulatory domains is associated with better clinical outcomes; 3) using a higher number of implementation strategies (that is, six or more) concomitantly and delirium management being integrated according to the PAD guidelines or the ABCDE care bundle, are associated with positive effects of implementation efforts on clinical outcome, and 4) in contrast, a high number of implementation strategies and PAD/ABCDE use were not associated with reductions in delirium incidence. With regard to the third finding, it is imperative to note that the association between the use of six or more implementation strategies and mortality reductions should be regarded as a hypothesis-generating finding with regard to the effectiveness of implementation interventions for clinical outcome improvement, and therefore, does not imply that using more implementation strategies will definitely result in improved outcomes.

Our results seem to be consistent with the premises of the Society of Critical Care Medicine (SCCM) guideline on management of Pain, Agitation and Delirium (PAD) [3] and the ABCDE care bundle, that: 1) integrated management of pain, agitation/sedation and delirium together with early mobilization should be a component of the plan of ICUs to improve patient safety and comfort, and 2) complying with these components of evidence-based critical care has the potential to improve clinical outcomes depending on the baseline practices of any individual ICU and the patient population. Of the evidence-based interventions mentioned, early mobilization is the only intervention that has been shown to improve both delirium and clinical outcomes, but regrettably the integrated nature of both PAD and ABCDE precluded us from studying early mobilization implementation in isolation. Establishing such integrated management on a daily basis in all patients and by all ICU health care professionals is not an easy task, as it requires consideration of an intense amount of human factors and cultural adaptations. The data from this review support that putting effort into implementation may be worthwhile, while at the same time confirming that not all programs will meet with the same success. Importantly, we cannot exclude that the positive effects of using a high number of implementation interventions on mortality may in part be explained by the Hawthorne effect, meaning that using many implementation strategies at the same time may have improved quality of care due to improved attention for specific aspects of care, which may not always have been linked directly to delirium [38]. Another explanation may be that local ICU culture in these studies - which typically is unmeasured and thus unaccounted for - may have promoted successful implementation of changes into clinical practice. For instance, an ICU team consisting of professionals who are capable of adopting new practices within a limited time frame and that has acquired effective communication and collaboration across different types of health care professionals is probably more likely to implement multiple strategies successfully compared to a team that lacks these characteristics. The number of implementation strategies used may then confound the true causal association between local ICU culture and improved clinical outcome.

Although this review focused on delirium in the ICU, targeting delirium alone would not suffice to establish outcome improvements. Therefore, we argue that delirium screening alone would not likely establish mortality reduction when not embedded in an ABCDE bundle, for instance [7]. In other words, it is the circumstances leading to or sustaining brain dysfunction that should be dealt with in the first place. This view, that exclusively dealing with delirium may not suffice to improve clinical outcomes, is supported by a recent study showing that the attributable mortality caused by delirium in ICU patients is questionable and that long-term sequelae may be better clinical outcome measures for delirium-related outcomes than short-term mortality [39]. On the other hand, it is perceivable that delirium-focused management embedded in PAD or ABCDE may establish outcome improvement in spite of the fact that delirium may not be causally linked to mortality directly, analogous to lactate-guided management that may improve outcome in critically ill patients, in spite of lactate not being causally linked to mortality [40].

Several methodological limitations of this review need to be addressed. First, the included studies showed strong heterogeneity with regard to design, focus of implementation (prevention, assessment or management of delirium as primary focus or delirium-oriented interventions being part of the implementation program but not the main focus), applied implementation strategies and model and whether the study was primarily aimed at studying the implementation itself or not. Definitions of process and clinical outcomes varied between studies. For instance, delirium measures varied importantly between studies ranging from delirium incidence after admission to ICU to percentage of ICU days with delirium present per patient, which hampered comparability. Second, although early mobilization seems to be the only intervention within PAD/ABCDE that has been shown to affect both delirium and clinical outcomes, we could not isolate studies specifically reporting an implementation intervention that linked delirium and early mobilization implementation with clearly defined process or outcome measures, as per our inclusion criteria. Third, in spite of rigorous assessment of the implementation strategies that were used in included studies according to predefined EPOC definitions, a potential limitation hampering interpretation of the association between improved outcome and number of strategies is that the effort put in to execute these implementation strategies could not be assessed. For instance, two studies using the same number of strategies may still differ with regard to the efficacy of the implementation due to ongoing educational efforts in one but only a single educational session in the other study. We speculate that when more effort is put into the implementation it may be more successful even with the same number of implementation strategies used. Fourth, there is some evidence that suggests that uncontrolled pre-post test studies as included in this review may overestimate the effects of implementation or quality improvement studies [41]. Fifth, the results on ICU LOS should be considered cautiously because concurrent changes in mortality may affect ICU LOS, instead of the implementation intervention itself being responsible for lower ICU LOS, as censoring by death may bias and (theoretically) even reverse the associations found. On the other hand, strengths of this review include the systematic assessment of the implementation strategies by three independent investigators based on the description of strategies provided by EPOC, the focus on the clinical endpoints and the systematic assessment of methodological quality. Furthermore, inconsistency of the pooled analysis with regard to the clinical outcomes was low, which supports the generalizability of our findings.

Summarizing the current status of implementation work that has been done to date with regard to ICU delirium reveals which implementation strategies have not yet been studied extensively in this field. For example, reminders and computerized support have mostly been previously found to be effective strategies [11], whereas these strategies did not stand out in this review; assessment of these strategies in future work aimed at ICU delirium should therefore be considered. We think that our work may encourage health policy makers to invest in multifaceted implementation efforts to improve care for delirious ICU patients.

More research is necessary to elucidate which types of individual strategies and/or which combination of strategies used in implementation programs are most successful in establishing mortality reduction in delirious critically ill patients. Further, several aspects of implementation deserve further evaluation, as this review shows that these issues in implementation have lacked attention, such as cultural aspects pertaining to the medical ICU team, nurse-physician interaction and establishing sustainability of practice changes [12]. Prospective, adequately powered before-after studies may be most suitable for evaluation of practice changes and cluster-randomized trials are conceivably the best study designs to evaluate the effect of implementation strategies on outcome improvements [42]. Therefore, an important issue to be considered is the distinction between successful practice change and clinical outcome improvements in implementation research. In our study successful implementation was evident in most studies on delirium screening implementation that showed improved adherence, even without known benefit for clinical outcomes. On the other hand cumbersome implementation may result in improved outcomes.

Finally, detailed information on extent, form and content of implementation interventions, especially education, was often lacking in studies on implementation (data not shown). Therefore, reproducibility of delirium implementation research should also be taken into account in future investigations.

Conclusion

This review and meta-analysis shows that multifaceted implementation programs that included assessment, prevention and management of ICU delirium have been shown to effectively change adherence to delirium screening and delirium knowledge. Implementation programs may enhance their effectiveness when not only health care professionals are targeted for behavioral change but also organizational changes are employed. Although using more rather than fewer implementation strategies simultaneously and delirium management being integrated with structured pain and agitation management (PAD), awakening and breathing coordination and early mobilization (ABCDE bundle) were associated with improved clinical outcomes, these results should be regarded as preliminary and hypothesis-generating with regard to the link between implementation practice and outcome improvement. Therefore, to determine whether these associations are causal our findings require confirmation and further study is needed on the most effective implementation strategies and the importance of focusing on delirium as an important form of organ failure within implementation programs aimed at practice change.

Key messages

  • Implementation programs can effectively improve delirium screening adherence or knowledge, but have had varying effects on delirium incidence and use of antipsychotic drugs.

  • There seems to be no easy way out implementing delirium-oriented interventions, especially when combined with related care components as described in the PAD guidelines or ABCDE bundle: to implement these inclusive, integrated management frameworks, use of multiple implementation strategies concurrently that are targeted both at the care providers and at organizational aspects seems to be necessary.

  • Successful implementation, meaning effective practice change, should be clearly delineated from the effect of such practice changes on clinical outcomes.

  • Robust data on effectiveness of specific implementation strategies with regard to the care of delirious critically ill patients are scarce and there is a lack of data on the association between specific practice changes (for example, delirium screening) and improvements in clinical outcomes.

Abbreviations

ABCDE bundle:

awakening and breathing coordination, choice of sedative, delirium monitoring and management and early mobility bundle

EPOC:

Cochrane Effective Practice and Organisation of Care group, classification system checklist

LOS:

length of stay

PAD guideline:

guideline for the management of pain, agitation and delirium

PRISMA:

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

RCT:

randomized controlled trial

WMD:

weighted mean difference

References

  1. Engel GL, Romano J. Delirium, a syndrome of cerebral insufficiency. J Chronic Dis. 1959;9:260–77.

    CAS  PubMed  Google Scholar 

  2. Hipp DM, Ely EW. Pharmacological and nonpharmacological management of delirium in critically ill patients. Neurotherapeutics. 2012;9:158–75.

    CAS  PubMed  PubMed Central  Google Scholar 

  3. Barr J, Fraser GL, Puntillo K, Ely EW, Gelinas C, Dasta JF, et al. Clinical practice guidelines for the management of pain, agitation, and delirium in adult patients in the intensive care unit. Crit Care Med. 2013;41:263–306.

    PubMed  Google Scholar 

  4. Vasilevskis EE, Pandharipande PP, Girard TD, Ely EW. A screening, prevention, and restoration model for saving the injured brain in intensive care unit survivors. Crit Care Med. 2010;38:S683–91.

    PubMed  PubMed Central  Google Scholar 

  5. Girard TD, Kress JP, Fuchs BD, Thomason JW, Schweickert WD, Pun BT, et al. Efficacy and safety of a paired sedation and ventilator weaning protocol for mechanically ventilated patients in intensive care (Awakening and Breathing Controlled trial): a randomised controlled trial. Lancet. 2008;371:126–34.

    PubMed  Google Scholar 

  6. Schweickert WD, Pohlman MC, Pohlman AS, Nigos C, Pawlik AJ, Esbrook CL, et al. Early physical and occupational therapy in mechanically ventilated, critically ill patients: a randomised controlled trial. Lancet. 2009;373:1874–82.

    PubMed  Google Scholar 

  7. van der Jagt M, Trogrlic Z, Ista E. Untangling ICU delirium: is establishing its prevention in high-risk patients the final frontier? Intensive Care Med. 2014;40:1181–2.

    PubMed  Google Scholar 

  8. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JP, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009;6:e1000100.

    PubMed  PubMed Central  Google Scholar 

  9. Slomka J, Hoffman-Hogg L, Mion LC, Bair N, Bobek MB, Arroliga AC. Influence of clinicians’ values and perceptions on use of clinical practice guidelines for sedation and neuromuscular blockade in patients receiving mechanical ventilation. Am J Crit Care. 2000;9:412–8.

    CAS  PubMed  Google Scholar 

  10. EPOC:2002. http://epoc.cochrane.org/sites/epoc.cochrane.org/files/uploads/datacollectionchecklist.pdf.

  11. Grol R, Grimshaw J. From best evidence to best practice: effective implementation of change in patients’ care. Lancet. 2003;362:1225–30.

    PubMed  Google Scholar 

  12. Sinuff T, Muscedere J, Adhikari NK, Stelfox HT, Dodek P, Heyland DK, et al. Knowledge translation interventions for critically ill patients: a systematic review*. Crit Care Med. 2013;41:2627–40.

    PubMed  Google Scholar 

  13. Anderson LA, Sharpe PA. Improving patient and provider communication: a synthesis and review of communication interventions. Patient Educ Couns. 1991;17:99–134.

    Google Scholar 

  14. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:177–88.

    CAS  PubMed  Google Scholar 

  15. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557–60.

    PubMed  PubMed Central  Google Scholar 

  16. Balas MC, Vasilevskis EE, Olsen KM, Schmid KK, Shostrom V, Cohen MZ, et al. Effectiveness and safety of the awakening and breathing coordination, delirium monitoring/management, and early exercise/mobility bundle. Crit Care Med. 2014;42:1024–36.

    PubMed  PubMed Central  Google Scholar 

  17. van den Boogaard M, Pickkers P, van der Hoeven H, Roodbol G, van Achterberg T, Schoonhoven L. Implementation of a delirium assessment tool in the ICU can influence haloperidol use. Crit Care. 2009;13:R131.

    PubMed  PubMed Central  Google Scholar 

  18. Riekerk B, Pen EJ, Hofhuis JG, Rommes JH, Schultz MJ, Spronk PE. Limitations and practicalities of CAM-ICU implementation, a delirium scoring system, in a Dutch intensive care unit. Intensive Crit Care Nurs. 2009;25:242–9.

    PubMed  Google Scholar 

  19. Hager DN, Dinglas VD, Subhas S, Rowden AM, Neufeld KJ, Bienvenu OJ, et al. Reducing deep sedation and delirium in acute lung injury patients: a quality improvement project. Crit Care Med. 2013;41:1435–42.

    PubMed  Google Scholar 

  20. Skrobik Y, Ahern S, Leblanc M, Marquis F, Awissi DK, Kavanagh BP. Protocolized intensive care unit management of analgesia, sedation, and delirium improves analgesia and subsyndromal delirium rates. Anesth Analg. 2010;111:451–63.

    PubMed  Google Scholar 

  21. Bowen CM, Stanton M, Manno M. Using Diffusion of Innovations Theory to implement the confusion assessment method for the intensive care unit. J Nurs Care Qual. 2012;27:139–45.

    PubMed  Google Scholar 

  22. Soja SL, Pandharipande PP, Fleming SB, Cotton BA, Miller LR, Weaver SG, et al. Implementation, reliability testing, and compliance monitoring of the Confusion Assessment Method for the Intensive Care Unit in trauma patients. Intensive Care Med. 2008;34:1263–8.

    PubMed  PubMed Central  Google Scholar 

  23. Gesin G, Russell BB, Lin AP, Norton HJ, Evans SL, Devlin JW. Impact of a delirium screening tool and multifaceted education on nurses’ knowledge of delirium and ability to evaluate it correctly. Am J Crit Care. 2012;21:e1–11.

    PubMed  Google Scholar 

  24. Mansouri P, Javadpour S, Zand F, Ghodsbin F, Sabetian G, Masjedi M, et al. Implementation of a protocol for integrated management of pain, agitation, and delirium can improve clinical outcomes in the intensive care unit: a randomized clinical trial. J Crit Care. 2013;28:918–22.

    PubMed  Google Scholar 

  25. Pun BT, Gordon SM, Peterson JF, Shintani AK, Jackson JC, Foss J, et al. Large-scale implementation of sedation and delirium monitoring in the intensive care unit: a report from two medical centers. Crit Care Med. 2005;33:1199–205.

    PubMed  Google Scholar 

  26. Radtke FM, Heymann A, Franck M, Maechler F, Drews T, Luetz A, et al. How to implement monitoring tools for sedation, pain and delirium in the intensive care unit: an experimental cohort study. Intensive Care Med. 2012;38:1974–81.

    PubMed  Google Scholar 

  27. Eastwood GM, Peck L, Bellomo R, Baldwin I, Reade MC. A questionnaire survey of critical care nurses’ attitudes to delirium assessment before and after introduction of the CAM-ICU. Aust Crit Care. 2012;25:162–9.

    PubMed  Google Scholar 

  28. Kamdar BB, King LM, Collop NA, Sakamuri S, Colantuoni E, Neufeld KJ, et al. The effect of a quality improvement intervention on perceived sleep quality and cognition in a medical ICU. Crit Care Med. 2013;41:800–9.

    PubMed  PubMed Central  Google Scholar 

  29. Scott P, McIlveney F, Mallice M. Implementation of a validated delirium assessment tool in critically ill adults. Intensive Crit Care Nurs. 2013;29:96–102.

    PubMed  Google Scholar 

  30. Dale CR, Kannas DA, Fan VS, Daniel SL, Deem S, Yanez 3rd ND, et al. Improved analgesia, sedation, and delirium protocol associated with decreased duration of delirium and mechanical ventilation. Ann Am Thorac Soc. 2014;11:367–74.

    PubMed  PubMed Central  Google Scholar 

  31. Kastrup M, Nolting MJ, Ahlborn R, Braun JP, Grubitzsch H, Wernecke KD, et al. An electronic tool for visual feedback to monitor the adherence to quality indicators in intensive care medicine. J Int Med Res. 2011;39:2187–200.

    CAS  PubMed  Google Scholar 

  32. Robinson BR, Mueller EW, Henson K, Branson RD, Barsoum S, Tsuei BJ. An analgesia-delirium-sedation protocol for critically ill trauma patients reduces ventilator days and hospital length of stay. J Trauma. 2008;65:517–26.

    CAS  PubMed  Google Scholar 

  33. Devlin JW, Fong JJ, Howard EP, Skrobik Y, McCoy N, Yasuda C, et al. Assessment of delirium in the intensive care unit: nursing practices and perceptions. Am J Crit Care. 2008;17:555–65. quiz 566.

    PubMed  Google Scholar 

  34. Page VJ, Navarange S, Gama S, McAuley DF. Routine delirium monitoring in a UK critical care unit. Crit Care. 2009;13:R16.

    PubMed  PubMed Central  Google Scholar 

  35. Reade MC, Eastwood GM, Peck L, Bellomo R, Baldwin I. Routine use of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) by bedside nurses may underdiagnose delirium. Crit Care Resusc. 2011;13:217–24.

    PubMed  Google Scholar 

  36. Bryczkowski SB, Lopreiato MC, Yonclas PP, Sacca JJ, Mosenthal AC. Delirium prevention program in the surgical intensive care unit improved the outcomes of older adults. J Surg Res. 2014;190:280–8.

    PubMed  Google Scholar 

  37. Khalifezadeh A, Safazadeh S, Mehrabi T, Mansour BA. Reviewing the effect of nursing interventions on delirious patients admitted to intensive care unit of neurosurgery ward in Al-Zahra Hospital, Isfahan University of Medical Sciences. Iran J Nurs Midwifery Res. 2011;16:106–12.

    PubMed  PubMed Central  Google Scholar 

  38. van Zanten AR, Brinkman S, Arbous MS, Abu-Hanna A, Levy MM, de Keizer NF, et al. Guideline bundles adherence and mortality in severe sepsis and septic shock. Crit Care Med. 2014;42:1890–8.

    PubMed  Google Scholar 

  39. Klein Klouwenberg PM, Zaal IJ, Spitoni C, Ong DS, van der Kooi AW, Bonten MJ, et al. The attributable mortality of delirium in critically ill patients: prospective cohort study. BMJ. 2014;349:g6652.

    PubMed  PubMed Central  Google Scholar 

  40. Jansen TC, van Bommel J, Schoonderbeek FJ, Sleeswijk Visser SJ, van der Klooster JM, Lima AP, et al. Early lactate-guided therapy in intensive care unit patients: a multicenter, open-label, randomized controlled trial. Am J Respir Crit Care Med. 2010;182:752–61.

    PubMed  Google Scholar 

  41. Eccles M, Grimshaw J, Campbell M, Ramsay C. Research designs for studies evaluating the effectiveness of change and improvement strategies. Qual Saf Health Care. 2003;12:47–52.

    CAS  PubMed  PubMed Central  Google Scholar 

  42. Ista E, Trogrlic Z, Bakker J, Osse R, van Achterberg T, van der Jagt M. Improvement of care for ICU patients with delirium by early screening and treatment: study protocol of iDECePTIvE study. Implement Sci. 2014;9:143.

    PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

The authors thank Professor E Boersma for statistical advice. Wichor Bramer, Biomedical information specialist at the medical library of the Erasmus MC - University Medical Center Rotterdam guided the literature search. The research was funded by The Netherlands Organization for Health Research and Development (ZonMw), ZonMw grant number 171203008.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Zoran Trogrlić.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

ZT, MJ, JB and EI participated in designing and coordinating the manuscript, conceived the paper, performed statistical analysis and drafted the manuscript. MB, WE and PV participated in designing the manuscript, and drafting the paper. All authors have critically read and approved the final manuscript.

Authors’ information

Z. Trogrlić, MSc, nurse scientist, ICU nurse, investigator-PhD student, Erasmus MC; M van der Jagt, MD, PhD, neurologist-intensivist, ICU Erasmus MC; Professor J Bakker MD PhD, internist-intensivist, vice-chair, ICU Erasmus MC; Michele C Balas, PhD, RN, APRN-NP, CCRN, nurse scientist, The Ohio State University, College of Nursing, Center of Excellence in Critical and Complex Care, Columbus; E Wesley Ely, MD, MPH, Department of Medicine, Division of Pulmonary and Critical Care, Health Services Research Center, Vanderbilt University Medical Center, and Associate Director of Aging Research, Veteran’s Affairs Tennessee Valley Geriatric Research Education Clinical Center (GRECC); Peter HJ van der Voort, MD, PhD, intensivist, Department of Intensive Care, Onze Lieve Vrouwe Gasthuis; E Ista, RN, PhD, implementation fellow, nurse scientist, Erasmus MC.

Additional files

Additional file 1:

The search strategy used in the systematic review. This file provides details of the search strategy.

Additional file 2:

Adapted rating system from Anderson and Sharpe. Explanation of the quality assessment tool that was used for included studies.

Additional file 3:

Quality rating. Table showing quality assessment details of included implementation studies in the systematic review.

Additional file 4:

Pooled analysis of determinants of change in delirium incidence (risk ratio) in implementation studies (n = 8) that included delirium-oriented interventions. of delirium incidence reduction that were studied were: use of either the guideline for the management of pain, agitation: and delirium (PAD) or the awakening and breathing coordination, choice of sedative, delirium monitoring and management, and early mobility (ABCDE) bundle (a) or use of high or low number of implementation strategies (b). (c) Both studies that focused on delirium screening implementation and studies that did not (but, for example, implemented the ABCDE bundle), found no changes in delirium incidence after the implementation. Only two studies (van den Bogaard and Reade) on delirium screening implementation were included of which individual patient data could be retrieved from authors. See text for more details.

Rights and permissions

This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Trogrlić, Z., van der Jagt, M., Bakker, J. et al. A systematic review of implementation strategies for assessment, prevention, and management of ICU delirium and their effect on clinical outcomes. Crit Care 19, 157 (2015). https://doi.org/10.1186/s13054-015-0886-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s13054-015-0886-9

Keywords