Detecting Delirium in Hospitalized Elderly Patients: A Review of Practice Compliance

Abstract

Background: The Ontario Senior Friendly Hospital Strategy recognizes delirium prevention and management as a top priority and recommends implementation of delirium screening as well as management protocols. This strategy proposes that hospitals monitor 2 specific indicators: (1) rate of baseline delirium screening and (2) rate of hospital-acquired delirium. Objective: To (1) determine compliance with the Ontario Senior Friendly Hospital Strategy indicators; (2) describe the use of pharmacological and nonpharmacological interventions for management of delirious patients; and (3) identify predictors of screening compliance. Methods: We conducted a retrospective review of patients aged ≥65 years admitted to 4 different inpatient units for ≥48 hours. Data were extracted for 7 two-month time blocks chosen between September 2010 and October 2013, following the implementation of various geriatric and delirium related initiatives at the hospital. Results: A total of 786 patients met study inclusion criteria. Overall, 68.2% had baseline delirium screening (indicator 1), with screening rates increasing over time (P < .001). Inpatient unit and year of study were both statistically significant predictors of delirium screening. Among those screened, the overall rate of hospital-acquired delirium was 17.2% (indicator 2). Early mobilization and device removal were the most common nonpharmacological interventions, while initiation of an antipsychotic and discontinuation of benzodiazepines were the most common pharmacological interventions. Conclusions: Although the rates of baseline delirium screening have significantly increased over the sampled time period, rates are still below the averages referenced in other literature. Our study suggests we need additional efforts to improve compliance with delirium screening in our institution.

Introduction

Delirium, an acute confusional state, is identified in 14% to 56% of hospitalized older adults (≥65 years).¹ The causes of delirium are typically multifactorial and include both predisposing and precipitating factors.¹ This syndrome is more likely to be associated with negative outcomes including higher mortality, longer durations of hospital stay, and an increased likelihood of long-term cognitive impairment compared with those who do not develop delirium.^2,3

Delirium is thought to be a preventable event in approximately 40% of cases.^1,4,5 As such, the Ontario Senior Friendly Hospital Strategy has made it a top clinical priority⁶ and proposes hospitals implement interprofessional delirium screening, as well as prevention and management protocols to improve outcomes.⁷ It further proposed that hospitals monitor and target 2 indicators—(1) rate of baseline screening (ie, screening on admission) and (2) rate of hospital-acquired delirium—to effectively monitor its efforts and better self-identify further areas of improvement surrounding screening, prevention, and management of delirium.⁷

Our institution, an urban acute care teaching hospital, launched a hospital-wide Acute Care for Elders (ACE) Strategy in 2011⁸ that promotes elder-friendly care models including the opening of dedicated care settings such as an ACE unit. The ACE unit is a 28-bed General Internal Medicine (GIM) unit that employs innovative strategies to improve care of frail older patients.⁹ Some examples of these strategies include order-sets that mandate delirium screening, the appropriate use of urinary catheters, bowel and pain management regimens that minimize delirium, a Least Restraint Policy, a Maximizing Aging Using Volunteer Engagement¹⁰ volunteer program, a Hospital Elder Life Program,¹¹ access to online Nursing Improving Care for Health System Elders¹² training modules in geriatrics for all staff, access to geriatric medicine and psychiatry interprofessional consultation services, and hospital rooms that include large clocks and white boards to help with reorientation. All of these strategies enable an interprofessional, team-based approach to patient care.¹³ Most of these strategies implemented on the ACE unit have been implemented across other hospital units. In November 2010, the geriatric medicine and psychiatry interprofessional consultation services became available hospital-wide, and in 2011, an ortho-geriatrics service was established.¹⁴ Delirium screening was further incorporated as a practice standard for all patients admitted to the ACE and non-ACE GIM units and the intensive care unit (ICU) in April 2011, and the orthopedic units in October 2012.

Our primary objective was to determine compliance with delirium screening among older patients within 24 hours of a hospital admission (baseline) and any time after the first 24-hour period up until discharge across our general medical, orthopedic, and intensive care units. Our secondary objectives were to (1) determine the rate of hospital-acquired delirium in these settings, (2) describe the use of pharmacological and nonpharmacological interventions for management of delirious patients, and (3) identify predictors of screening compliance.

Methods

Prior approval from the institutions research ethics board (#MSH REB 14-0111-C) was obtained before the study was conducted.

Setting and Patients

We audited all patients aged ≥65 years who were admitted for ≥48 hours to 1 of the following 4 care units between 2010 and 2013: (1) ACE, (2) medical surgical ICU, (3) non-ACE GIM, and (4) orthopedics (hip fracture patients only). We audited 7 distinct 2-month time periods between 2010 and 2013 to cover the various pre- and postimplementation elder-friendly care practices at our institution: September to October 2010, 2011, 2012, and 2013; and March to April 2011, 2012, and 2013. If multiple admissions for the same patient were identified, only the first admission was used. We excluded patients with a documented active palliative status at admission.

Data Collection

A standardized data collection form (Supplementary Figure 1, available in the online version of this article) was developed and piloted by a multidisciplinary team. We extracted the following data: demographics (eg, age, sex), location prior to admission (eg, nursing home), hospital admission and discharge dates, admission unit, admitting diagnosis, past medical history, discharge disposition, delirium screening results, documented adverse events potentially attributable to delirium (eg, harm to self or others, falls, self-removal of indwelling devices), and medications defined as likely attributable to delirium in the literature.^15,16

We reviewed the chart for documentation of the Confusion Assessment Method,^17,18 the Intensive Care Delirium Screening Checklist,¹⁹ or clinical notes stating the patient was assessed for delirium by the unit clinical team or a consulting geriatrician or psychiatrist. (Supplementary Figure 2a, available in the online version of the article) We categorized baseline delirium screening as a documented assessment within 24 hours of admission to the unit of interest. Delirium assessments completed after the initial 24 hours but prior to discharge were categorized as follow-up delirium screening.

Pharmacological and nonpharmacological delirium interventions were collected for the first 48 hours after a documented positive delirium score (Supplementary Figure 2b, available in the online version of the article). Pharmacological interventions collected included the initiation of medications used to manage delirium (eg, antipsychotics) and discontinuation of medications thought to contribute to delirium (eg, benzodiazepines). Nonpharmacological interventions collected were (1) removal of urinary catheters, (2) mobilization, and (3) physical restraint use.

Data Analysis

Rate of hospital-acquired delirium was calculated as the number of patients who screened positive at any point during hospitalization after a negative baseline screen within the first 24 hours of admission, divided by the number of patients with a negative baseline screening on admission and a documented follow-up screening after the first 24 hours before discharge (whether it was positive or negative). Screening compliance was defined as the number of days screening occurred at least once within a 24-hour period, divided by the total number of days admitted to the unit of interest. If the patient was on ≥9 medications (excluding as-needed medications) on admission, it was classified as polypharmacy.²⁰ Demographics, clinical variables, and pharmacological and nonpharmacological delirium management practices used were reported using means (standard deviation [SD] and range), or counts and proportions, where appropriate. Categorical data were compared using χ² tests.

Predictors of screening within 24 hours were identified using multivariate logistic regression. The following variables were included in the model following a literature search and consultation with the research team: age, gender, Katz score 3 or less, year of admission, admission unit, dementia, and delirium. The following variables were only considered for inclusion in the model if the P value was <.1 in the univariate analysis: admission diagnosis, visual or auditory impairment, hypertension, polypharmacy, more than 3 psychoactive medications, alcohol abuse, admission from long-term care facility, or hospital versus from home.

All tests were 2-tailed where a P-value of ≤.05 was deemed significant. All analyses were conducted using SPSS 24.0 IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp, Armonk, NY).

Results

Of the 844 patients who initially met the inclusion criteria, 98 were excluded because of documented palliative status on admission and 4 patients had charts not available for review (Figure 1). We included 786 patients in the final analysis.

Figure 1.

Average rate of baseline delirium screening across 4 hospital unit settings.

Abbreviations: ACE, acute care for elders; GIM, general internal medicine; ICU, intensive care unit; Ortho, ortho-geriatrics.

Participants

Demographics are presented in Table 1. The mean age of patients was 80.2 years (SD = 8.6) and 59.4% female. The orthopedics unit was the oldest, with a mean age of 83.2 years (SD = 8.3); the medical-surgical ICU had the youngest population at 75.0 years (SD = 7.9). On admission to hospital, the mean total number of daily medications per patient was 8.0 (SD = 4.0), and the mean total number of psychoactive medications per patient was 1.2 (SD = 1.1). Polypharmacy (use of ≥9 medications) on admission was common (40.0% of all patients).

Table 1.

Demographics of Study Patients (N = 786)^a.

Characteristics	All Patients
Age (years)	80.2 (8.57)
Female	467 (59.4)
Admission nursing units
ACE GIM	324 (41.2)
Non-ACE GIM	227 (28.9)
ICU	150 (19.1)
Orthopedics	85 (10.8)
Admitting diagnosis
Respiratory	156 (19.8)
Surgery/trauma	147 (18.7)
Neurological	146 (18.6)
Gastrointestinal/genitourinary	105 (13.4)
Metabolic/hematologic	72 (9.2)
Infection	56 (7.1)
Functional decline/weakness	51 (6.5)
Cardiovascular	29 (3.7)
Otherb	24 (3.1)
Baseline KATZ score
Past medical historyc	4.6 (1.61)
Hypertension	527 (67.0)
Dementia/movement disorders	217 (27.6)
Diabetes	192 (24.4)
Mood disorders	145 (18.4)
Visual impairment	133 (16.9)
TIA/stroke	116 (14.8)
COPD	108 (13.7)
Auditory impairment	48 (6.1)
Delirium	21 (2.7)
Medication and substance history
Antidepressants/antipsychotics	289 (86.8)
Benzodiazepines	182 (23.2)
Opioids	155 (19.7)
Acetylcholinesterase inhibitors	70 (8.9)
Anticholinergics	70 (8.9)
Anticonvulsants	52 (6.6)
Non-benzodiazepine sedatives	36 (4.6)
Non-psychoactive medicationsd	825 (105.0)
Current smoker/marijuana use	75 (9.5)
Alcohol abuse	57 (7.3)

Abbreviations: ACE, acute care for elders; GIM, general internal medicine; ICU, intensive care unit; TIA, transient ischemic attack; COPD, chronic obstructive pulmonary disease.

^aData are presented as n (%) or mean (standard deviation).

^bOther = dermatological, dysphagia, cancer.

^cAdds up to >100% because of multiple past medical history conditions per patient.

^dNon-psychoactive medications include nonopioid analgesics, β-blockers, antiarrhythmic drugs, digoxin, oral corticosteroids, and H2 receptor antagonists.

Delirium Screening

Of the 786 patients, 536 (68.2%) were screened for delirium within 24 hours of admission (indicator 1) and 651 (82.8%) were screened at least once during hospitalization after the first 24 hours. Baseline screening rates increased over time with 35.9%, 61.6%, 73.3%, and 83.6% of patients screened in each subsequent year (P < .001; Table 2).

Table 2.

Delirium Screening According to Year and Unit (%).

	2010		2011		2012		2013a		Overall
Units	<24 hoursb	>24 hoursc	<24 hours	>24 hours	<24 hours	>24 hours	<24 hours	>24 hours	<24 hours	>24 hours
ACE	45.5	15.9	66.1	91.5	82.1	99.1	94.8	98.3	78.7	86.1
Non-ACE GIM	42.3	13.5	40.5	70.3	62.1	100	83.3	100	60.8	75.3
ICU	9.1	40.9	70.0	96.0	81.1	97.3	73.2	97.6	64.7	88.7
Ortho	20.0	10.0	66.7	72.2	52.2	95.7	58.8	94.1	54.1	80.0
Totald	35.9	18.8	61.6	86.0	73.3	98.7	83.6	98.1	68.2	82.2

Abbreviations: ACE, acute care for elders; GIM, general internal medicine; ICU, intensive care unit; Ortho, ortho-geriatrics.

^aP < .001 within screening years.

^b<24 hours, screened within less than 24 hours of hospital admission to the unit studied, classified a baseline screening.

^c>24 hours—screened within greater than 24 hours of hospital admission to the unit studied, classified as follow-up screening.

^dP < .001 within units.

The rates of screening within 24 hours and after 24 hours progressed similarly among the 4 units (Table 2; P < .001). In addition, 11.1% of patients were never screened for delirium at any point during their hospital stay. This was highest in the non-ACE GIM units (27.3%) followed by the ortho (9.4%), ICU (9.3%), and ACE (8.3%) units. The proportion of patients not screened was the highest in 2010, with 54.3% of patients never being screened followed by 2011 with 10.6%. This value dropped close to 0% in 2012 and 2013. The rate of baseline screening after delirium screening was implemented as a standard of practice in 2011-2012 was 78.5%.

In a multivariate logistic regression analysis (Table 3), statistically significant predictors of screening within 24 hours of admission included year of admission and unit. Age, gender, Katz score 3 or less, dementia, and delirium were not significant predictors of being screened within 24 hours of admission.

Table 3.

Predictors of Delirium Screening within 24 Hours of Admission^a.

Variables	OR	95% CI	P
Age	0.991	0.969-1.014	.441
Female gender	1.030	0.719-1.476	.872
Year of admission
2013	1
2012	1.929	1.206-3.085	.006
2011	3.355	2.047-5.498	<.001
2010	9.958	5.854-16.936	<.001
Katz score of 3 or less	1.484	0.902-2.443	.120
Admission units
ACE	1
Non-ACE GIM	2.085	1.348-3.224	.001
ICU	1.887	1.149-3.099	.012
Ortho	4.184	2.338-7.487	<.001
History of dementia	1.234	0.789-1.929	.357
History of delirium	0.836	0.294-2.378	.737

Abbreviations: OR, odds ratio; CI, confidence interval; ACE, acute care for elders; GIM, general internal medicine; ICU, intensive care unit; Ortho, ortho-geriatrics.

^aThe following variables were considered in univariate analysis but not retained (P > .1): admission diagnosis, visual or auditory impairment, hypertension, polypharmacy, more than 3 psychoactive medications, alcohol abuse, admission from long-term care facility or hospital versus from home. Hosmer-Lemeshow test (P = .652). Nagelkerke R² = 0.214.

Hospital-Acquired Delirium

The rate of hospital-acquired delirium over the entire study period was 17.2% (74/429). As screening increased, the overall incidence of delirium also increased from 2010 to 2013, reaching its highest in 2013 at 32%. This trend was seen in all 4 of the units of study, with increased incidence of delirium in the latter half of the study period. The highest average incidence of hospital-acquired delirium over the 4-year study period was in the ACE unit, averaging 9.5%, followed by in the ICU, at 4.75% (Figure 2). A third of the study population (36.6%) did not have a confirmed delirium status because screening was either never completed or screening was not completed for both at baseline and after the first 24 hours up until discharge.

Figure 2.

Incidence of hospital-acquired delirium by year and by unit.

Abbreviations: ACE, acute care for elders; GIM, general internal medicine; ICU, intensive care unit; Ortho, ortho-geriatrics.

Delirium Management

Table 4 summarizes the utilization of nonpharmacologic and pharmacologic interventions for delirium. Of the 74 patients (17.2%) who had hospital-acquired delirium, 80.0% were mobilized after delirium detection, 13.5% had their urinary catheters removed within 48 hours of screening positive, and 14.9% were physically retrained at least once. We found an antipsychotic was more frequently initiated for patients with hospital-acquired delirium (17.6% of patients), compared with those screened negative (2.5% of patients), while initiation of a benzodiazepine was similar between those screened positive and those screened negative (6.8% vs 7.9%, respectively). Benzodiazepines were stopped within 48 hours of screening positive for delirium in 9.5% of patients with hospital-acquired delirium.

Table 4.

Pharmacological and Non-Pharmacological Practices for Management of Delirium (%).

	Hospital-Acquired Deliriuma (N = 74)	Delirium Negativeb (N = 355)	Never Screenedc (N = 87)
Pharmacological practices
Initiation of an antipsychotic	17.6	2.5	30.0
Initiation of benzodiazepine	6.8	7.9	10.3
Nonpharmacological practices
Mobilization	80.0	84.9	77.0
Physical restraint use	14.9	5.1	8.0
Urinary catheter removal	13.5	N/A	N/A

^aDocumented negative delirium screening within first 24 hours and screened positive thereafter.

^bPatients who received a documented negative delirium screening both at baseline and after 24 hours up until discharge.

^cAt any point during the entire hospital admission.

Discussion

Current research supports screening older hospitalized patients for delirium using a validated screening tool.⁷ As such, several Ontario hospitals have implemented delirium prevention, screening, and management algorithms into their standards of practice.²¹ According to recent literature, the average reported baseline screening across other Ontario hospital units range from 86% to 97%.⁷ This is much higher than the average baseline screening revealed in our 4-year study. However, in looking at the baseline screening rates after implementation of the protocols in 2011-2012, our rates (78.5%) are closer to the Ontario reported rates.

The inpatient unit of study was shown to be a statistically significant predictor of baseline screening rates at our hospital, likely reflecting the adopted standards of practice on each unit. Screening rates were highest in the ACE unit, which was expected, given the strong geriatric care focus of this unit, by specifically incorporating standards of practice that enforce preventative measures for consequences of hospitalization, such as delirium. The integration of delirium screening and reporting into the hospital’s electronic medical record likely offers an additional reminder to clinicians to screen their patients. The orthopedics unit in this study implemented this change to patient charting at a later time; therefore, this could contribute to their overall screening rates being lower in this setting. The non-ACE GIM unit had triple the number of patients who went through their entire hospital admission without being screened once for delirium, compared to the other 3 units of study, despite the fact that the baseline screening protocol was implemented around the same time as in the ACE unit and the ICU. This may possibly have reflected staff assumptions that the GIM patients most prone to delirium were likely admitted to the ACE GIM unit and therefore not a priority for the other GIM patients admitted to the non-ACE GIM units.

Other published literature informs us that delirium is commonly underrecognized and underreported.^22,23 The implementation of the ACE unit was a response to an acknowledged gap in care to meet clinical needs to address delirium prevention in hospitalized older adults. The results of our retrospective chart review indicate that since the implementation of the hospital’s ACE strategy, over time the overall screening and detection of delirium has significantly improved. As such, year of study was noted to be a statistically significant predictor of delirium screening. In 2010, more than half of the patients were never screened for delirium during their entire hospital stay. This number decreased to 10% in 2011 and close to 0% in 2012 and 2013. This trend is thought to be a reflection of the changes to the standards of practice surrounding delirium prevention in 2011, with the launching of the hospitals ACE strategy and subsequent opening of the ACE unit, and the spreading of ACE practices to the other units and services throughout 2011 and 2012. The detection of hospital-acquired delirium increased 10-fold (from 3% to 30%) from 2010 to 2013. This trend was consistent for all 4 units of study and most likely reflects a better level of recognition and therefore, documentation of the presence of delirium.

The overall rate of hospital-acquired delirium detected in the study is at the lower end of the range reported in other studies (14% to 56%). However, as our study shows, the rate of hospital-acquired delirium significantly varies depending on the unit. Other research identifies that the highest incidence of delirium in hospital is usually seen in the ICU, with at least one third of patients admitted to the ICU developing delirium.²⁴ Our study found that the incidence of delirium in the ICU was 4.75% on average over the 4-year study duration, with the highest incidence reaching 7.0% in 2011. This is much lower than other studies report. This could be due, in part, to the missed opportunities for detection because of noncompliance with screening or use of deep sedation for high acuity of illness. Among patients admitted to GIM units, the incidence of delirium cited in other recent literature ranged from 10% to 20%.²⁵ Our results showed that in the ACE GIM unit, the rate of hospital-acquired delirium is on par with current literature; however, the non-ACE GIM units had rates much lower than those reported in other studies. The non-ACE GIM units had the lowest screening compliance and the highest number of patients never screened, which likely contributed to the low incidence of hospital-acquired delirium identified.

Current research indicates delirious patients should undergo changes to their medication regimens to stop or substitute deliriogenic medications where possible.²⁶ Medication changes can also include the initiation of drugs thought to help manage delirium, such as antipsychotics. The study results show that patients who developed hospital-acquired delirium underwent more changes to their drug regimens than patients who screened negative for delirium. Antipsychotics were initiated in 17.5% of patients with delirium versus only 2.5% of patients who screened negative for delirium. Although it was noted that more pharmacological practices were initiated in the hospital-acquired delirium group, nonpharmacological management of delirium remains the preferred first-line therapy.²¹ Among the many strategies identified in other studies, early mobilization and the creation of a safe environment by the removal of harmful objects or unfamiliar devices, such as urinary catheters, were 2 of the most common nonpharmacological management strategies found in this study. Early mobilization was noted to occur almost as frequently in the hospital-acquired delirium group and the delirium negative group. This could be due in part to the adopted ACE strategy across all hospital units to ensure positive senior friendly outcomes. Mobilization could be seen as a treatment and a prevention strategy for delirious and nondelirious patients. Therefore, in this case it could have had an impact on the prevention of hospital-acquired delirium in those that screened negative.

Strengths and Limitations

The study was able to capture screening practices and outcome data on 4 different hospital units over a period of 4 years permitting the review of changes in practice over time. It also allowed us to evaluate the impact that creating a delirium screening protocol had on the different hospital units and to determine how this was translated into a measurable outcome over time (delirium screening rates and incidence rates). An important limitation of the study to consider was that the management of delirium only reflects patients who received a positive screening. We are not able to comment on patients who were not documented as a delirium positive screen at baseline and/or at follow-up. Therefore, a third of the study population was not included in the analysis of the management of hospital-acquired delirium because they were either not screened at all or did not have 2 documented screens: one at baseline and at least one in the follow-up period (after the first 24-hour before discharge).

Conclusion

The results of our audit indicate that changes in the standards of practice at our institution as part of its overall ACE strategy resulted in increased rates of baseline delirium screening over the sampled time period. Screening rates were noted to increase significantly after the implementation of these changes, which strongly affected the ability to recognize, diagnose, and initiate management strategies in delirious older hospitalized patients. As such, the study highlights opportunities to improve compliance with delirium screening. The adoption of admission order sets that mandate delirium screening allows clinicians an opportunity to include this in their diagnostic workup at the time of admission. This helps provide rapid recognition of patients with delirium to allow prevention and treatment strategies to be initiated at the start of their hospital stay. With the implementation of screening algorithms into the electronic documentation, this also affords nurses a reminder during their shift that they should be screening for delirium as part of their routine assessments. Our study shows that by adopting the ACE strategy across the other hospital units, more reminders have been put in place to ensure delirium screening is being completed.

Despite the sizable change in screening rates and delirium recognition, the rates are still lower than that demonstrated in other literature, and there was still more than a third of the overall study population left unscreened. Implementation of education around delirium screening for all clinicians can improve accurate recognition and management of delirium to enhance the provision of elder-friendly care and improve hospital performance as a whole on these indicators. Since many patients present to hospital on medications that can exacerbate delirium, perhaps pharmacists can play a valuable role in collecting best practice medication history to try and make appropriate adjustments to help facilitate better care. As prevention processes are implemented and refined, and as screening compliance continues to increase, a decrease in delirium incidence may further indicate overall success. We believe further studies around the barriers to screening for delirium might be beneficial to further increase compliance.