• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

National Clinical Guideline Centre (UK). Delirium: Diagnosis, Prevention and Management [Internet]. London: Royal College of Physicians (UK); 2010 Jul. (NICE Clinical Guidelines, No. 103.)

12Treatment of delirium: non-pharmacological (hospital setting)


What are the most clinical and cost effective multicomponent interventions for treating people with delirium in hospital?

12.1. Clinical introduction

Despite the advances in medical science over the last three decades, mortality and morbidity from delirium have remained unchanged and health costs for this syndrome remain high. Current management of delirium relies on early recognition, elimination or correction of underlying causal factors and general symptomatic and supportive measures. However, there is much uncertainty about the effectiveness of various interventions.

Early recognition and investigation of delirium is challenging and studies have repeatedly shown that delirium is missed in two-thirds of patients in hospitals. Moreover, delirium often has multi-factorial causes and multiple potential consequences. This has led to suggestions that multicomponent interventions, including non-pharmacological interventions might be appropriate for the treatment of delirium, and several such interventions have been investigated.

12.2. Description of studies

Details of included, excluded papers together with study design are reported in table 12.1.

Table 12.1. study inclusion, exclusion and design.

Table 12.1

study inclusion, exclusion and design.

One study (Pitkala 2006) had more than one report (Pitkala 2006 and Pitkala 2008); hereafter these studies are referred to by the first name reports, but separately in the results section.

12.2.1. Study Design

None of the studies were conducted in the UK. Information on study sizes, geographical location and funding are described in table 12.2.

Table 12.2. study characteristics.

Table 12.2

study characteristics.

The unit of randomisation in the RCTs was at patient level. In one of the historical controlled trial (Naughton 2005), eligible patients were enrolled at two different time periods. The Naughton (2005) study considered three groups of patients: those studied in the pre-intervention and two groups after the intervention had ceased – these patients were studied 4 and 9 months after the initial education phase of the intervention was completed.

12.2.2. Population

All studies took place in a hospital setting; the intervention in the Rahkonen (2001) study continued after discharge from hospital as it involved support for the patient over 3 years; Patients were all admitted to medical wards, with the exception of one study (Milisen 2001). Patients were included in each of the studies if they had delirium: this was based on screening with CAM, apart from the Rahkonen (2001) study which specified that the diagnosis was based on DSM-III-R but did not specify that CAM was used. In the Pitkala (2006) study, patients found to be positive on CAM screening had their diagnosis confirmed by a physician using DSM-IV criteria.

The Naughton (2005) study reported that for all patients admitted to the Acute Geriatric Unit (AGU) one criterion for admission was cognitive impairment (score less than 25 on the MMSE).

Some patients had dementia in the studies, (Cole 1994; Cole 1992; Pitkala 2006) ranging from 10% to 58% of participants, except in the Rahkonen (2001) study, where patients with dementia were excluded.

Method of assessment of dementia varied and the following methods were reported:

  • SPSMQ; scale scores range from: 0 to 10, from no impairment to severe; score of 5 or more indicative of moderate to severe cognitive impairment) (Cole 1994)
  • Medical record data for the diagnosis of preexisting dementia (Milisen 2001)
  • Clinical Dementia Rating Scale (CDR; scale scores range from 0.5 to 3, from very mild to severe dementia), DSM-IV criteria for dementia or diagnosis by specialist using standard diagnostic tests (no further details were given) (Pitkala 2006).

The mean age across the studies was 81 to 85.5 years; the studies had a mixed gender population with a majority of females (Cole 1994: 65%; Cole 2002: 54%; Milisen 2001: 81%; Naughton 2005: 63%; Pitkala 2006: 74%; Rahkonen 2001: 90%). Ethnicity was not reported in any of the studies.

12.2.3. Interventions

The included studies investigated multicomponent interventions in a hospital (or hospital plus community in the case of Rahkonen 2001) setting for the treatment of delirium (see table 12.3, Appendix D).

Nursing intervention protocol (Cole 1994, Cole 2002),

This intervention comprised of a multidisciplinary team consisting of geriatricians and liaison nurse.

  • consultation by a geriatrician or geriatric psychiatrist (completed within 24 hours after referral)
  • follow-up by a liaison nurse
    • follow up included daily visits during the patients’ stay (up to a maximum of 8 weeks), liaising with family members, recording information on patient’s metal status and discuss management with the patient’s nurses with the use of the protocol
    • assess compliance with consultant recommendations. Where appropriate, the nurse discussed management problems with the geriatrician or geriatric psychiatrist and where necessary patient was reassessed by the specialists.
  • the intervention protocol targeted the following risk factors:
    • environment (not having excessive, inadequate or ambiguous sensory input, medication not interrupting sleep, presenting one stimulus or task at a time);
    • orientation (room should have a clock, calendar, and chart of the day’s schedule; evaluate need for glasses, hearing aid, interpreter)
    • familiarity (objects from home, same staff, family members staying with patient, discussion of familiar areas of interest),
    • communication (clear, slow, simple, repetitive, facing patient, warm, firm kindness, address patient by name, identify self, encourage verbal expression)
    • activities (avoid physical restraint, allow movement, encourage self care and personal activities).

The intervention in the later trial (Cole 2002) was described as more intensive than in the earlier study (Cole 1994) and the following components were added to the intervention:

  • consultant not only assessed initially but also followed up the patients;
  • the study nurse visited the patient 5 days per week;
  • the intervention team (2 geriatric psychiatrists, 2 geriatric internists and the study nurse) met after every 8 to 10 patients were enrolled to discuss delirium management problems; and
  • the study investigator met the nurse weekly to discuss problems of diagnosis, enrolment and interventions.

Multicomponent geriatric intervention (Pitkala 2006)

Patients received a comprehensive geriatric assessment, which included history taking, interview with caregiver, physical examination, assessment of cognition and physical functioning, screening for depression, nutrition, and medication review. Other aspects of the intervention included:

  • recognising delirium and any underlying conditions
  • orientation (with calendars, clocks, photographs)
  • physiotherapy
  • general geriatric interventions (calcium and vitamin D supplements; nutritional supplements for those at risk of malnutrition or malnourished; hip protectors)
  • comprehensive discharge planning (including consultation of a social worker, occupational therapist’s home visit, involvement of caregivers).
  • medical management (avoiding neuroleptics; administering atypical antipsychotics for hyperactive/psychotic symptoms; use of cholinesterase inhibitors if patient’s cognition did not improve to MMSE score above 23).

The intervention group received significantly more atypical antipsychotic drugs than the control group (69.0% versus 29.9%, p<0.001), more acetylcholinesterase inhibitors (58.6% versus 9.2%, p<0.001), vitamin D and calcium supplements (77.0% versus 9.2%, p<0.001), nutritional supplements (92.0% versus 0.0%, p<0.001) and fewer conventional neuroleptics (8.0% versus 23.0%, p=0.006).

Nurse-led interdisciplinary intervention (Milisen 2001)

This intervention involved nurse education to identify high-risk patients which included:

  • education: a poster was developed to educate all nurses on the essential aspects of delirium, depression and dementia. This poster included the core symptoms of delirium according to the CAM criteria, comparative features and differences between delirium, dementia and depression and the relevance of correct and early recognition of delirium;
  • systematic screening of cognitive function using the NEECHAM Confusion Scale following training;
  • pain management: scheduled pain medication to provide effective post-operative pain control; and
  • consultative service: access to a resource nurses who were given training in identifying patients by a geriatric nurse specialist in the identification and management of older hip-fracture patients. If necessary, the resource nurses could consult with a geriatric nurse specialist or psycho geriatrician; resource nursed to help the primary nurses in implementing appropriate antidelirium interventions.
  • the nurses were provided with ‘A nursing guide for the evaluation of causes of delirium in elderly hospitalised patients’ (as reported in Milisen 1998). The guide advised a nurse to report to the attending physician of any changes in patient’s status on the following: medication, pain, hypoxemia, dehydration, electrolyte and metabolic disturbances, and infection. The interventions are briefly described below:
    • medication: to be vigilant of polypharmacy, especially anticholinergics, antiparkinsonian drugs, histamine H2-receptor antagonists;
    • pain: inquire systematically about pain; observe verbal and nonverbal expressions; use of as many possible analgesics based on nonopiod drug (e.g. paracetamol) and where required minimum dose of opioids combined with non opioid drug;
    • hypoxemia: monitor abnormalities in rate, depth and quality of respiration, cyanosis, PO2 ≤ 32; administer oxygen as ordered; determine source of hypoxia; low respiration (<10 l/min) due to opioid intoxication; consult attending physician for treatment with naloxone as antidote; in patients undergoing surgery: monitor hypothermia and postoperative shivering; maintain optimal patient temperature by applying warming [fluids and blood; gowns and blankets; humidified oxygen]; be alert for nocturnal desaturation during the first 3 days postoperatively and especially in obese patients; administer 2 l of O2 (unless contraindicated);
    • dehydration: encourage patient to drink water regularly and when necessary prepare for blood or fluid replacement;
    • electrolyte and metabolic disturbances: monitor abnormalities of blood and urine chemistry; give frequent small meals and add nutritional supplements, such as calorie/protein rich drink;
    • infection: be alert for urinary tract, respiratory, mouth and feet infections; stimulate patient for adequate water intake (2 l/day) (unless contraindicated); observe for abrupt onset for fever (rectal temperature >100°F) and apply cooling techniques as needed.

Systematic intervention (Rahkonen 2001

The intervention consisted of a case manger (nurse specialist) and an annual one-week rehabilitation period at a Brain Research and Rehabilitation Centre. Patient’s rehabilitation team included the study physician, the nurse specialist, physiotherapist, neuropsychologist and occupation therapist.

  • a nurse specialist trained in geriatrics and care of the elderly acted as the case manager. Patients received continuous and systematic support provided by the case manager with responsibility in supporting the patients during community care through out the 3 year follow-up acting as a counsellor and advocate and in the rehabilitation unit (as the primary care nurse);
  • care in the community: arranged in consultation with relatives and health and social care services, and continuity of care was achieved with regular follow-ups, including in-home visits and ‘phone calls by the case manager. Study physician was also available for consultation and medical care throughout the follow up; and
  • rehabilitation period: individually structured physiotherapy once or twice daily; mobility and other special aides for daily living (e.g. hearing aids and special shoes) were arranged when needed; patients were encouraged to participate in occupational therapy and free-time events.

Education and management intervention (Naughton 2005)

The intervention was designed to improve the recognition of delirium in medically ill older adults evaluated in the emergency department [ED triaged these patients with delirium specifically to the acute geriatric unit (AGU)]. This was achieved by addressing the following factors:

  • education:
    • The charting procedures in ED were changed and physicians were reminded to evaluate adults aged 75 years and older for cognitive impairment and delirium and direct the admission to the AGU. Nurses and physicians were trained to triage patients using yes/no answers to four questions from the history and mental status examination. A study nurse periodically reported the proportion of older adults correctly admitted to the AGU from the ED.
  • the education component for the AGU nurses (provided by geriatricians and geriatric nurse) involved:
    • educating on prevalence and outcome of delirium;
    • training on methods of mental status assessment;
    • guidelines on medication management of cognitive impairment and delirium.
    • small group consensus process used to develop assessment and charting procedures; and
    • AGU physicians were provided with information on cognitive impairment and delirium in the elderly, recommended metal status assessment procedures, and review of the intervention guidelines.
  • treating underlying medical factors;
  • treating precipitating factors (removing precipitating medications; addressing immobility);
  • providing family support;
  • using non-pharmacological support for: physically non aggressive behaviour and episodes triggered with ADL care;
  • medication management: reduce the use of psychotropic medications (benzodiazepines and anticholinergics); consider using synergistic agents such as neuroleptics or antidepressants that supplement behaviour treatment; sleep medication: trazadone 50 to 100 mg; zolpidem: 5 mg;
  • fewer patients in the AGU received benzodiazepines (22.6% compared with 30.9% at baseline); antihistamines (6% compared with 15.5%; p<0.02); increased use of antidepressants (22.7% compared with 10% at baseline; p<0.02); and neuroleptics (27.4% compared with 10.9% at baseline; p<.01)
  • simplifying pain regimen (minimise p.r.n.); and
  • environmental stimuli: addressing problems with environmental stimuli for example, noise, sleep disruption, disruptive room mate,
  • None of the studies included more than two study arms, and the comparator in all studies was ‘usual medical care’ (no further details given).

12.2.4. Comparisons

The following comparison was carried out:

Two studies (Naughton 2005; Pitkala 2006) reported concurrent medications:

  • opiates (42.7%); benzodiazepines (30.9%); antihistamines (15.5%); antidepressants (10.0%); neuroleptics (10.9%)
  • conventional neuroleptics (22%); atypical antipsychotics (14%) and cholinesterase inhibitors (6%) (Pitkala 2006).

12.2.5. Outcome measures

The following primary and secondary outcome measures were reported:

12.3. Methodological quality

12.3.1. RCTs

The method of sequence generation was adequate in two RCTs in which a computer-generated sequence was employed (Cole 2002, Pitkala 2006), and was not stated in one RCT (Cole 1994).

One RCT reported adequate allocation concealment - central randomisation with details of a retained schedule (Pitkala 2006). One RCT was partially adequate (with independent allocation but no further details, Cole 2002). In the third RCT, allocation concealment was not stated (Cole 1994).

Outcome assessors were stated to be blinded in two RCTs (Cole 1994, Cole 2002) and this was not stated in the other RCT (Pitkala 2006). Patients were not blinded in any of the RCTs.

Two RCTs reported an a priori sample size calculation. One RCT (Cole 1994) reported that a sample of 30 or more was required for 80% power to detect a difference of at least 1SD in the change in the measures used (p=0.05). One RCT (Pitkala 2006) reported that 58 to 91 patients per group were needed to show a 20% difference in the combined endpoint (discharge to permanent institutional care or death) with 80% power (p=0.05). The third RCT did not report a sample size calculation (Cole 2002).

All three RCTs included in the review demonstrated baseline comparability of the groups on measures such as age, gender and baseline scores measuring delirium or mental state.

All RCTs used an intention to treat analysis for at least some outcome measures. One RCT reported no missing data in either group (Pitkala 2006). In one RCT (Cole 2002), 7 patients withdrew in the intervention group (6.2%) versus 2 (1.8%) in the control group. In the third RCT (Cole 1994), 33% of patients died in the intervention group versus 37% in the control group; mean scores for some of the outcome measures SPMSQ and Crichton Geriatric Behavioural Rating Scale [CGBRS] were given for surviving patients only (i.e. fewer than 70% of the number randomised), although all patients were included in some outcome measures (length of stay, discharge to new long-term care, mortality).

Overall, one RCT was considered to have the potential for bias (Cole 1994). This study did not state randomisation or allocation concealment methods, and some outcome measures had missing data due to patients who had died (Cole 1994). This study was considered in sensitivity analyses.

12.3.2. Non-RCTs

In the Rakhonen (2001) study, the control group was formed by matching pairs of patients on age and gender from patients fulfilling the inclusion criteria from the earlier time period; in the remaining two studies patients were not individually matched but the groups were comparable on age and gender. The Milisen (2001) study reported that the non intervention cohort had significantly greater comorbid conditions (e.g. cardiac, vascular and abdominal problems).

One study reported that the investigator was blinded to the data of the main outcome measure of the study in the control patients (Rahkonen 2001: information was collected from registers for the control patients) and unclear in the other two studies.

One study (Rahkonen 2001) reported not all eligible patients were included (10%) and it was unclear in the other two studies.

Overall, we considered the three non-RCT studies to be of low quality because of the study design.

12.4. Results

12.4.1. Multicomponent intervention versus usual care

Primary outcomes of the review

Duration of delirium

Only one study reported the duration of delirium (Milisen 2001). This was significantly shorter in the intervention cohort (median = 1 day, interquartile range [IQR] = 1) compared with the non-intervention cohort (median = 4 days, IQR = 5.5, p=0.03, Mann-Whitney U test).

Number of patients recovered from delirium (complete response)

Two RCTs (Cole 2002; Pitkala 2005) reported complete response. The Pitkala (2006) study defined the response rate as a permanent improvement of at least 4 points on the MDAS (severity of delirium scored 0 to 30, with 30 being the worst) at 8 days; although no data or references were supplied to justify the use of this score as the measure for improvement, and the GDG considered this to be a poor measure of complete response.

Cole (2002) reported the number of patients with an improvement in cognitive status, as defined by the MMSE, during the hospital stay (mean length of stay 19 days). “Improvement” was defined as an increase in MMSE of 2 or more points; with no decrease below baseline plus 2 points thereafter. If the MMSE score at baseline was 27 or more, improvement was no decrease below 27; MMSE ranges from 0=poor to 30=excellent; a score of 23 or less indicates cognitive impairment) or ’not improved‘. The GDG decided that ‘the number improved’ was an unsatisfactory definition of recovery from delirium, so the study was not included in the analysis for this outcome.

In the Pitkala (2006) study, the intervention significantly increased the number of patients who had recovered from delirium at 8 days after admission (RR 2.00, 95% CI 1.30 to 3.08) This corresponds to a number needed to treat of 5 (95% CI 3 to 10); figure 12.1(Appendix K). The GDG debated whether a change of 4 points on the MDAS scale would clearly show improvement and considered that any conclusions drawn from the Pitkala (2006) study should be treated with caution.

Secondary outcomes of the review

Cognitive impairment

Three studies (Cole 1994; Milisen 2001; Pitkala 2006) reported cognitive impairment.

The Cole (1994) study reported scores for the SPSMQ, a 10-item questionnaire that evaluates orientation, memory and concentration (0=no impairment to 10=severe impairment) at 8 weeks. There was no difference between the intervention and usual care groups (figure 12.2, Appendix K), although the result is imprecise (MD −1.10 (95% CI −4.95 to 2.75).

The Pitkala (2006) study measured cognitive impairment with the MMSE at 6 months (Pitkala 2006). The study reported a mean score of 18.4 in the intervention group versus 15.8 in the usual care group, but no standard deviations were given (p=0.047 for repeated measures analysis of variance (ANOVA); baseline scores used as covariates). This was just significant.

The Milisen (2001) study reported the mean MMSE scores for the delirious patients in the intervention group and the non intervention group (mean MMSE scores: intervention group (delirious): 15.5; non intervention group (delirious): 9.5); the study reported that although the intervention group showed a higher overall cognitive function this difference was not statistically significant; p values or standard deviations were not reported.

Length of stay

Length of hospital stay was reported by all three RCTs (Cole 1994; Cole 2002; Pitkala 2006). The result for the Pitkala (2006) study is presented as a subgroup as the intervention differed from the other two studies (Cole 1994; Cole 2002).

The Cole (1994) study did not report standard deviations, so the study’s contribution to the meta-analysis of the two studies was not estimable. There was no significant difference between intervention and usual care groups in Cole (2002), although the result is imprecise [MD 0.60 (95% CI −3.90 to 5.10)] (figure 12.3, Appendix K).

In the Pitkala (2006), length of stay appeared shorter in the usual care group. We note that the distribution of lengths of stay was skewed (median 21 days in the intervention group, range 2 to 110 days; median 16 in the usual care group, range 1 to 90 days; mean 29.3 days, SD 25.6 in intervention group and mean 22.4 days, SD 18.4 in control group; means are less than twice SD so data likely to be skewed). The result is imprecise [MD 6.90 (95% CI 0.28 to 13.52)].

Two non RCTs also reported length of hospital stay (Milisen 2001; Naughton 2005). The Milisen (2001) study reported a median of 13.5 days (IQR 3.75 days) for the intervention cohort and 14 days (IQR 5 days, p=0.6) for the non-intervention cohort. The Naughton (2005) study reported that following intervention, a mean of 3.3 days was saved in length of stay following each episode of delirium.

Discharge to long-term care

All three RCTs reported discharge of patients who had become more dependent since their admission. Two studies reported that patients were discharged at a greater level of dependency: Cole (1994) reported the percentage of patients discharged required more care (numbers were calculated as the proportion of patients remaining alive at the end of the study); Cole (2002) reported that living arrangements were arranged hierarchically from least dependent (e.g. home alone) to most dependent (e.g. nursing home); living arrangements at discharge were compared with those at admission and were rated as more dependent, same, or less dependent.

The Pitkala (2006) study reported the number of patients discharged to permanent institutional care, and these represented new admissions to such care as patients already in permanent institutional care at admission were excluded from the study.

The results are presented as subgroups in figure 12.4 (Appendix K). There was no significant difference in effect of the intervention on discharge to higher care [OR 1.04 (95% CI 0.19 to 5.65)], a more dependent living arrangement at discharge [OR 0.77 (95% CI 0.31 to 1.92)] or to new long-term care [OR 0.69 (95% CI 0.38 to 1.26)], although the results for all three studies are imprecise.

Excluding the Cole (1994) study due to its possible bias did not materially alter the results (a forest plot showing sensitivity analysis is not presented).

The Rahkonen (2001) study reported the duration of long-term care in the three years of the study. This was a mean of 441 days (SD 366) in the intervention group compared with 535 days (SD 308) in the control group [MD −94 days (95% CI −225.28 to 37.28)] (figure 12.5, Appendix K). The mean age was comparable (82.1 years in both groups) and the study excluded patients with confirmed or suspected dementia, however, individuals with mild cognitive impairment were included.

Health related quality of life (HRQoL)

One report (Pitkala 2008) of the Pitkala (2006) study reported health related quality of life along the following dimensions: mobility, vision, hearing, breathing, sleeping, eating, speech, elimination, usual activities, mental function, discomfort and symptoms, depression, distress, and vitality. Patients were assessed with the 15D questionnaire at baseline and discharge [range 0 (poor HRQoL) to 1 (excellent HRQoL)].

There was a small significantly higher HRQoL for the intervention group (MD 0.06 (95% CI 0.02 to 0.10); figure 12.6 (Appendix K). The study reported that there were significant differences for the intervention and usual care group on the following dimensions on the 15D questionnaire: mental function corresponding to cognition and alertness (p<0.001), usual activities corresponding to functioning in activities of daily living (p<0.001), vitality (p= 0.004), depression (p=0.044), and speech (p=0.024).


Three RCTs (Cole 1994; Cole 2002; Pitkala 2006) and one non-RCT (Rahkonen 2001) evaluated the number of patients who died: two RCTs at 8 weeks (Cole 1994; Cole 2002) and the other RCT at 1 year (Pitkala 2006) and the non-RCT at 3 years (Rahkonen 2001).

The Cole (1994) study reported that overall 35% (31/88) patients died in 8 weeks (33% [14/42] and 37% [17/46] deaths occurring in the intervention and control groups, respectively) [OR 0.90 (95% CI 0.51 to 1.60)]; the causes of death were not given.

The Cole (2002) study reported that overall 21% (47/227) of patients died (22% [25/113] and 19% [22/114] deaths occurring in the intervention and control groups, respectively) [RR 1.15 (95% CI 0.69 to 1.91)]; and the Pitkala (2006) study reported that overall 32% (56/174) patients died over 1 year (34% [30/87] and 30% [26/87] deaths occurring in the intervention and control groups, respectively) [RR 1.15 (95% CI 0.75 to 1.78)]; the causes of death were not reported in either study.

There was no significant difference between the interventions and usual care in the mortality rates, but the results were very imprecise (figure 12.7, Appendix K).

The non-RCT study (Rahkonen2001) reported that during the three-year follow up, a total of 42% (43/102) patients died, the causes of death were not reported [RR 0.87 (95% CI 0.55 to 1.37)] (figure 12.8, Appendix K).

Overall summary

Summary of results for the multicomponent intervention is reported in table 12.4.

Table 12.4. summary of results: multicomponent non pharmacological treatment of delirium in hospital setting.

Table 12.4

summary of results: multicomponent non pharmacological treatment of delirium in hospital setting.

12.5. Health economic evidence

12.5.1. Multicomponent interventions for the treatment of delirium in a hospital setting

One economic evaluation study was included as evidence (Pitkala 2008). This was a Finnish RCT of 174 consecutive delirium patients aged above 69 years who were admitted to the general medicine services and whose life expectancy was predicted to be above 6 months. The study aimed at assessing the effects of multicomponent geriatric treatment on costs of care and HRQoL in delirious in-patients. Patients in the intervention group received a comprehensive geriatric assessment at baseline for good detection of delirium, as well as careful diagnosis of the underlying etiological conditions. They received atypical antipsychotics if necessary and effective general treatments were implemented for all patients. After the acute phase of delirium, all patients not recovering from impaired cognition underwent detailed diagnostics for dementia and thereafter, received acetyl cholinesterase inhibitors. Patients in the comparator arm received usual care and this was not exactly described.

The average cost per patient in the intervention arm was €19,737 while the average cost per patient in the usual care arm was €19,557. The extra cost attributable to intervention was €446 per patient. This included the cost of atypical antipsychotics, acetycholinesterase inhibitors, vitamin D-calcium supplements, hip protectors, and nutritional supplements. Average unit costs in Finland were used. Health related quality of life was measured using the 15D questionnaire but the question on sexual activity was omitted. Subjective health was assessed using an ordinal scale at discharge. An unadjusted mortality rate of 35% and 30% were reported in the intervention and usual care groups respectively. The patient’s measure of health status was 0.68 and 0.62 in the intervention and control groups respectively. The dimensions of HRQoL showing significant differences favouring intervention were mental function, usual activities, vitality, depression and speech.

The results of this study could be used to estimate the cost per unit of improvement in health status of delirium patients. However, patient’s measure of health status was based on 15D which elicited health status scores from a Finnish general population. It was reported only at the point of discharge from hospitalisation for delirium and quality-adjusted life years were not reported. Furthermore, there was no sensitivity analysis to test the effect of the uncertainties surrounding the cost and health outcome measures. Costs were not assessed from a UK NHS and PSS perspective. The results of this study were judged to be not directly applicable to this guideline.

12.6. Clinical evidence statements

  • There is very low quality evidence which showed that a multicomponent intervention targeting six modifiable risk factors (orientation, sleep, sensory impairment improvement, early mobilisation, environmental, medication) following a consultation with a geriatrician or geriatric psychiatrist and follow up by a liaison nurse showed no significant difference in:
    • cognitive impairment (measured at 8 weeks). However, there is much uncertainty around this result
    • the number of patients discharged with a greater level of dependency; there is much uncertainty around this result
    • mortality rates at 8 weeks; there is uncertainty around this result

Additional follow up assessment by geriatirican and a liaison nurse showed no significant difference.

  • There is very low quality evidence to show that a multicomponent intervention targeting three modifiable risk factors (dehydration/nutrition, pain management, medication management) with training showed:
    • significantly shorter duration of delirium in patients in the intervention group
    • no significant difference in the median length of stay in hospital
  • There is moderate quality evidence to show that a multicomponent geriatric intervention based on targeting four modifiable risk factors (orientation, dehydration/nutrition, early mobilisation, medication management) with comprehensive geriatric assessment showed a:
    • significant number of patients recovered from delirium (at 8 days) in the intervention group; however, there is much uncertainty around this result
    • significant difference showing a decreased length of stay in the usual care group;
    • small significant improvement in the health related quality of life (mental function, daily functioning, depression, vitality, and speech) for the intervention group at discharge
    • borderline significant difference showing a lower level of cognitive impairment at 6 months for the intervention group
    • non significant difference in the number of patients discharged to long-term care; there is much uncertainty around this result.
  • There is very low quality evidence to show a multicomponent intervention targeting two modifiable risk factors (orientation, early mobilisation) with training, continuous nursing support and annual one-week visits to a rehabilitation unit showed no significant difference in:
    • long-term care stay over the duration of the study (3 years); there is much uncertainty around this result
    • mortality rates at 3 years

12.7. From evidence to recommendations

The evidence suggested that enhanced treatment strategies for people with delirium are more effective than usual care however the GDG did not feel confident in recommending a particular multicomponent intervention because of the low quality evidence. Instead the GDG drew on the principles of the multicomponent interventions and their clinical expertise to inform the recommendations. The GDG agreed treatment of delirium should comprise the following:

  • initial management for all people with delirium,
  • second line management for those who are distressed or are considered a risk to themselves or others,
  • management for people whose symptoms do not resolve, either following initial or second line management.

Initial management

The multicomponent treatment review showed some indication of clinical effectiveness in one study (Pitkala 2006). The GDG considered the measure of delirium to be too unreliable to support this in economic modeling or recommendations. The GDG did draw on the components comprising the multicomponent interventions, and used them, together with information from the risk factors review to make a consensus recommendation on treating possible underlying causes of delirium (recommendation 1.6.1). The GDG recognised that sometimes there was more than one underlying cause.

The GDG also considered evidence from the non-pharmacological risk factors review and the patient information review, and drew on their clinical experience.

The GDG recognised the importance of listening and talking to the person experiencing delirium. The GDG specifically took into account the messages conveyed by the GDG patient representatives describing how difficult it was for them to tell relatives and staff about their changes in cognition.

Evidence from the multicomponent treatment review and the GDG’s clinical expertise highlighted the importance of reorientation in people diagnosed with delirium. The GDG felt that reorientation could be addressed by communicating the role of the healthcare professional, who the person is and the day, date, time and place. Familiar faces of family, friends and carers may also help with orientation (evidence underpinning this came from the patient information review, chapter 14).

Hospital environments, artificial lighting and time loss through disturbed sleep patterns and periods of unconsciousness can easily lead to disorientation with the potential to aggravate delirium. The GDG therefore also considered that an important part of reorientation included maintaining a suitable care environment (recommendation 1.3.1) for people diagnosed with delirium, and therefore included this within the recommendation.

Recommendation 1.6.2 should be carried out for all people diagnosed with delirium.

Distressed people

The GDG discussed the care of people who are distressed or those considered at a risk to themselves or others and this was informed by the NICE Violence guideline (clinical guideline25) which provides information on how to calm down an escalating situation. The GDG recognised that the NICE violence guideline was restricted to short-term management of violent and disturbed behaviour in psychiatric settings and emergency departments; however, they agreed that the principles of effective communication and de-escalation techniques could be extrapolated to this guideline. The GDG considered that non-pharmacological de-escalation approaches should be tried before resorting to pharmacological treatment (recommendation 1.6.3). This was partly on the basis of their clinical experience and partly in view of their reservations about the evidence on medication (section 13.8). The GDG also noted that identifying distress in people who had hypoactive delirium can be more difficult than in people with hyperactive delirium. Although they often appear to be calm, they may be distressed by psychotic symptoms and this may not be intuitive. The GDG decided to add a statement to the recommendation to this effect.

The GDG recommended that when de-escalation techniques had not worked pharmacological interventions should be considered (see recommendation 1.6.4 in the pharmacological treatment chapter 12).

The GDG also made a recommendation for people whose symptoms remain unresolved following first or second line treatment. Persisting delirium could be due to underlying causes remaining unaddressed. Alternatively the person might have dementia rather than delirium. The GDG made a recommendation to capture this and cross-referred to the ‘Dementia’ guideline (NICE clinical guideline 42) for advice on diagnosing and managing dementia (see recommendation 1.6.6 in the pharmacological treatment chapter 12). This recommendation was consistent with the GDG’s strategy for patients in whom it was difficult to distinguish between delirium, dementia and delirium on dementia. Failure to resolve delirium should lead the health care professional to consider dementia.

The GDG agreed that there was limited evidence relating to health care professional effective communication skills when caring for people with delirium. In addition, the recognition and recording of delirium are important factors. The GDG agreed a future research recommendation for staff education which include effective communication strategies (see below and Appendix H):

Future research recommendation

Does an education programme for staff improve the recovery from delirium in patients in hospital compared with an education leaflet or usual care?

12.8. Recommendations

Initial management

In people diagnosed with delirium, identify and manage the possible underlying cause or combination of causes. [1.6.1]

Ensure effective communication and reorientation (for example explaining where the person is, who they are, and what your role is) and provide reassurance for people diagnosed with delirium. Consider involving family, friends and carers to help with this. Provide a suitable care environment (see recommendation 1.3.1). [1.6.2]

Distressed people

If a person with delirium is distressed or considered a risk to themselves or others, first use verbal and non-verbal techniques to de-escalate the situation. For more information on de-escalation techniques, see ‘Violence’ (NICE clinical guideline 25). Distress may be less evident in people with hypoactive delirium, who can still become distressed by, for example, psychotic symptoms. [1.6.3]

Image appkf127
Image appkf128
Image appkf129
Image appkf130
Image appkf131
Image appkf132
Image appkf133
Image appkf134
Copyright © 2010, National Clinical Guideline Centre - Acute and Chronic Conditions.

Apart from any fair dealing for the purposes of research or private study, criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, no part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

The rights of National Clinical Guideline Centre to be identified as Author of this work have been asserted by them in accordance with the Copyright Designs and Patents Act, 1988.

Cover of Delirium
Delirium: Diagnosis, Prevention and Management [Internet].
NICE Clinical Guidelines, No. 103.
National Clinical Guideline Centre (UK).

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...