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

Trogrlić, Zoran; van der Jagt, Mathieu; Bakker, Jan; Balas, Michele C; Ely, E Wesley; van der Voort, Peter HJ; Ista, Erwin

doi:10.1186/s13054-015-0886-9

Critical Care

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

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

Author, year (design)	Implementation			Process outcomes				Clinical outcomes
Author, year (design)	Number of strategies used	Implemented care components	Implementation model	Screening adherence	Delirium incidence	Use of antipsychotic drugs	Delirium knowledge	Mortality change	ICU length of stay, days
Balas, 2014 [16] (B/A^a study, n = 296)	12	ABCDE^b	CFIR^c	+50% (0 to 50%)^d	−13% (62 to 49%), *P =* 0.02	+12 mg (6 to 18 mg)^e, P = 0.24	-	−8.6% (19.9 to 11.3%), P = 0.04	−1^f (5 to 4), P = 0.21
Van den Boogaard, 2009 [17] (B/A study, n = 1742)	12	Delirium screening	Model of Grol and Wensing	+14% (77 to 92%), P <0.0001	+13% (10 to 23%), P <0.05^g	−12 mg (18 to 6 mg)^e , P = 0.01	+1.2 (6.2 to 7.4), P <0.001	-	−0.3 (1.3 to 1)^f P < 0.05
Riekerk, 2009 [18] (B/A study, n = NA)	10	Delirium screening	Structural implementation pathway	+57% (38 to 95%)^d	-	-	+1^d,h (3–4)	-	-
Hager, 2013 [19] (B/A study, n = 202)	10	PAD^w	4Es frameworkⁱ	0 (90 to 90%)	+18%^j (20 to 38%), P = 0.01	-	-	-	-
Skrobik, 2010 [20] (B/A study, n = 1133)	9	PAD	-	+3^k (89 to 92%), P = 0.055	−0.5% (34.7 to 34.2%), P = 0.9	+0.3% (39.4 to 39.7%), P = 0.7	-	−6.5% (29.4 to 22.9%), P = 0.009	−0.97^l (6.32 to 5.35), P = 0.009
Bowen, 2012 [21] (pilot study, n = 34 nurses)	8	Delirium screening	Diffusion of Innovations theory	+75% (10% to 85%)	-	-	-	-	-
Soja, 2008 [22] (Prospective study, n = 347)	10	Delirium screening	-	+84% (0 to 84)^d	-	-	-	-	-
Gesin, 2012 [23] (B/A study, n = 20 nurses)	7	Delirium screening	-	-	-	-	+2.1 (6.1 to 8.2), P = 0.001	-	-
Mansouri, 2013 [24] (RCT, n = 201)	7	PAD	-	+100%^m (0 to 100%)	-	−2.5 mgⁿ (3.2 to 0.7 mg), P = 0.12	-	−12% (24 to 13%), P = 0.046	−3.1 (7.1 to 4.0)^f , P <0.001
Pun, 2005 [25] (Prospective study, n = 711)	6	PAD	-	+90% (0 to 90)^d +84% (0 to 84)^d	-	-	-	-	-
Radtke, 2012 [26] (B/A^e study, n = 131)	7	PAD	Modified extended training	+1.6 (0 to 1.6), P <0.01	-	-	-	−4.8%^o (9.9 to 5.1%), P = 0.16	−4 (18 to 14)^p, P = 0.40
Radtke, 2012 [26] (B/A^e study, n = 131)	7	PAD	Modified extended training	+1.6 (0 to 1.6), P <0.01	-	-	-	−4.8%^o (9.9 to 5.1%), P = 0.16	−4 (8 to 4)^p , P <0.01
Eastwood, 2012 [27] (B/A study, n = 288 patients/2368 shifts)	4	Delirium screening	-	-	-	+8.5%^q (5.8 to 14.3%), P <0.0001^r	-	+3.2% (5 to 8.2)^s,t P = 0.31	0 (2 to 2), P = 0.34
Kamdar, 2013 [28] (B/A study, n = 285)	6	Multifaceted sleep promotion program	Structured QI model	-	odds ratio 0.46^a , P = 0.02	-	-	−6% (25 to 19%), P = 0.88^s	−1.1^u (5.4 to 4.3), P = 0.60
Scott, 2012 [29] (B/A study, n = 119)	4	Delirium screening	-	+78% (0 to 78%)^d	-	-	+14%^v (71 to 85%), P <0.001	-	-
Dale, 2014 [30] (B/A study, n = 1483)	5	PAD	-	+1.14^x (0.35 to 1.49), P <0.01	odds ratio 0.67, p = 0.01	−1.7 (2.7 to 1.0)^y , P <0.01	-	0 (14 to 14%), P = 1.0	−12.4%^j , P = 0.04
Kastrup, 2011 [31] (B/A study, n = 205)	7	Visual feedback system	-	+37.5% (0.5 to 38%), P <0.01	+4% (25 to 29%), P = 1.0^za	-	-	-	-
Robinson, 2008 [32] (B/A study, n = 119)	5	PAD	-	-	-	+14% (31 to 45%), P = 0.25	-	−2.9% (17.6 to 14.7), P = 0.64	−1.8 (5.9 to 4.1), P = 0.21
Devlin, 2008 [33] (B/A study, n = 601)	6	Delirium screening	SCT^zb	+70% (12 to 82%), P <0.0005	-	-	-	-	-
Page, 2009 [34] (Retrospective study, n = 60)	4	Delirium screening	-	+92% (0 to 92%)^d	-	-	-	-	-
Reade, 2011 [35] (B/A study, n = 288)	4	Delirium screening	-	-	−16% (37 to 21%), P = 0.004	-	-	+3.2% (5 to 8.2)^zc, P = 0.31	0 (2 to 2), P = 0.34
Bryczkowski, 2014 [36] (B/A study, n = 123)	3	Delirium prevention program	-	-	+11% (58 to 47%), P = 0.26	−1% (7 to 6%), P = 0.83	-	−4% (7 to 3%), P = 0.31	−3 (9 to 6), P = 0.04

^aB/A = before-after. ^bABCDE = awakening and breathing coordination, delirium monitoring/management and early exercise/mobilization bundle. ^cCFIR = Consolidated Framework for Implementation Research. ^dStatistical significance not reported or assessable from data in article but presumed to be statistically significant because of strong effect (difference before-after shown in parentheses). Significant changes are shown in bold letters. ^eTotal dose of haloperidol per patient. ^fMedian. ^gChi-square test. ^hIncrease in median level of agreement on a scale of 5 (1 = totally disagree, 5 = totally agree, with 3 = neutral about statement and 4 = agree) with true statements about delirium, signifying increased knowledge. ⁱ4Es framework = Engage, Educate, Execute and Evaluate. ^jPercent of ICU days delirium present per patient. ^kAdherence calculated by dividing delirium assessments judged to be possible by total number of patients in Table 1 in reference. Adherence data to screening not explicitly provided in text. ^lMean. ^mNo explicit mention of screening adherence, but after CAM-ICU implementation as part of the PAD guideline the authors mention strict adherence surveillance to the PAD protocol: 15 patients in protocol group excluded from analysis because of noncompliance with PAD guideline. ⁿMean dose of drug (haloperidol) used per patient. ^oMortality calculated from numbers given in Table 1 in original article for combined data of ICU 1 and 2 (n = 131, before-after comparison made with chi-squared test, degrees of freedom (df) = 2). ^pThis study reported different interventions (standard training versus extended training and implementation) in different ICUs. Numbers given here are those from the B/A study in two ICUs that received modified extended training. ^qPercentage is total number of administered doses of either haloperidol (5 mg), olanzapine (5 to 10 mg) or quetiapine (25 mg) divided by the total number of 8-hour shifts in pre- and post-CAM-ICU implementation period. Study of Eastwood is duplicate report of study by Reade, therefore, data were combined for analysis. ^rChi-squared statistic = 47, df = 1. ^sUnstructured delirium screening versus CAM-ICU screening. ^tData on change in mortality were not included for analysis of all mortality data because these data are same as those of Reade, 2011 [35]. ^uCalculated for survivors, median, frequency of delirium monitoring per day per patient. ^vCalculated agreement with true statements about delirium and its importance increased with 14% after the implementation, signifying increased knowledge (chi-squared statistic = 14, df = 1). ^wPAD = integrated pain, agitation/sedation and delirium monitoring and management; ^xNumber of CAM-ICU assessments/day (mean). ^yMean daily haloperidol dose (mg). ^zaFisher’s exact test. ^zbSCT = script concordance theory. ^zcPercent patients ever receiving haloperidol.

Back to article page

ISSN: 1364-8535

Contact us

Submission enquiries: journalsubmissions@springernature.com