Abstract

Introduction: there is evidence to suggest that delirium incidence can be reduced in older people admitted to medical services using multi-component interventions that target delirium risk factors. The cost-effectiveness of this approach is uncertain. We therefore developed a novel cost-effectiveness model for delirium prevention.

Method: we compared multi-component delirium prevention intervention with usual care using a model based on a decision tree analysis. The model was used to estimate the incremental net monetary benefit (INMB). The robustness of the cost-effectiveness result was explored using deterministic and probabilistic sensitivity analyses.

Result: the multi-component prevention intervention was cost-effective when compared with usual care. It was associated with an INMB of £2,200 using a cost-effectiveness threshold of £20,000 per quality-adjusted life year (QALY). It remained cost-effective in the majority of the deterministic sensitivity analyses and was cost-effective in 96.8% of the simulations carried out in the probabilistic sensitivity analysis.

Discussion: our analysis has shown convincingly that multi-component prevention interventions for delirium should be considered as a cost-effective health-care strategy for medically ill people admitted to hospital. It is an attractive intervention for health-care planners as they strive to reconfigure their services to better meet the needs of an ageing population.

Introduction

Delirium is a common syndrome that particularly affects older people admitted to hospital. It is characterised by disturbed consciousness and changes in cognitive function and/or perception that develop over a short period of time [1]. In medical-in-patients, the occurrence rate per admission range from 11 to 42% [2]. Delirium is important because it is an unpleasant condition to experience, is poorly recognised and is associated with poor outcomes. Previous studies have reported that the adverse consequences of delirium developing in hospitalised patients include increased risk of hospital-acquired complications [3], new dementia [4], new admission to institution [5], extended in-hospital stay [6–8] and increased mortality [3, 6, 9]. Unsurprisingly, these adverse outcomes increase the expenditure of health resources, estimated as $6.9 billion per year (2004 USD) for Medicare in the USA [10] and are likely to reduce the patient's health-related quality of life (HRQoL).

In the UK, the evidence for the diagnosis, prevention and treatment of delirium has been recently reviewed by the National Institute for Health and Clinical Excellence (NICE) [1]. It was apparent from the evidence that, whereas enhanced systems of care to treat delirium did not improve outcomes, enhanced care systems based on multi-component interventions were associated with the potential to prevent incident delirium in hospitals. The multi-component prevention interventions involve assessment of at risk patients to identify and then modify risk factors associated with delirium. The research literature suggested that about a third of incident delirium in hospitals could be prevented by this approach. However, the prevention of one episode of delirium would require interventions in many people at risk of delirium and therefore the widespread introduction of multi-component systems of care to prevent delirium would incur additional costs to the UK National Health Service (NHS). It was therefore important to examine the balance of cost (intervention and non-intervention) and health benefit of such interventions. This article aims to summarise the detailed cost-effectiveness analysis carried out during the development of the delirium guideline by NICE. It reports the cost-effectiveness of using multi-component interventions in delirium prevention in people admitted for urgent care to a general medicine service. The findings are particularly timely for the NHS and other health services as health-care planners internationally strive to reconfigure their services to better meet the needs of an ageing population.

Method

Full details of the methods are described in the NICE guideline [1]. A novel health economics model was developed for a multi-component intervention targeted at people at risk of delirium admitted for urgent care to a general medicine hospital ward. The model required inputs to provide estimates for the incremental costs, and the incremental quality-adjusted life years (QALYs) gained that were associated with the intervention. These were used to calculate the incremental cost-effectiveness ratio (ICER) and the incremental net monetary benefit (INMB). A systematic review to identify the multi-component delirium prevention studies was conducted as part of the NICE guideline. Of the eight studies investigating medical patients, the study assessed as the highest quality was a non-randomised study that recruited 852 patients at intermediate and high risk of delirium, and aged 70 years or older [11]. The multi-component intervention addressed six delirium risk factors: cognitive impairment; sleep deprivation; immobility; visual and hearing impairments and dehydration. Patients in the control group received usual care on a medical unit. The relative risk of delirium in the intervention group was 0.66 (95% confidence limits 0.46–0.95).

Model structure

The model used a decision tree analysis (Supplementary data are available in Age and Ageing online, Figure S1), constructed using the NICE reference case for economic evaluations [12] with health outcomes measured in QALYs; costs calculated from a UK NHS and Personal Social Services (PSS) perspective and future costs and QALYs discounted at 3.5%. The development of the model and decisions on model parameters was conducted by the NICE guideline development group which comprised clinicians, patient representatives, health economists and systematic reviewers. The decision tree included outcomes that would impact negatively on patients' HRQoL, or on health-care resources. These included falls in hospital, pressure ulcer, new dementia, new admission to institution, extended in-hospital stay and death. We compared the multi-component delirium prevention intervention and usual care. The decision tree was applied to each strategy separately and was used to estimate the impact of each strategy on the expected number of delirium cases, costs and QALYs associated with the adverse outcomes. The mean starting age for the model was 79 years.

Risk estimates

The baseline and relative risks for the adverse consequences included in the model are summarised in Table 1. The values were obtained from the most relevant studies identified in the systematic reviews supporting the NICE guideline. The selected studies reported odds ratios but these were converted to relative risk estimates. The additional hospital length of stay attributable to delirium was estimated as 16.8 days. This estimate was obtained by applying a Weibull function to previously published Kaplan–Meier plots of lengths of stay of patients with and without delirium [6].

Table 1.

Parameters used in the economic model

Input parameter Point estimate Source Type of distribution for PSA Distribution parameters for PSA 
Baseline risk of delirium 15.0% [11Beta α = 64, β = 362 
Baseline risk estimates for adverse consequences 
New dementia 5.60% [4Beta α = 7, β = 117 
New admission to institution 17.40% [5α = 40, β = 190 
Pressure ulcer 4.00% [22α = 360, β = 8575 
Falls 6.90% [3α = 9, β = 122 
Mortality 5.00% [3α = 7, β = 124 
Estimated relative risk of adverse consequences (95% CI) 
New dementia 4.67 (1.43–15.29) [4Lognormal Log (mean) = 1.54, SE = 0.60 
New admission to institution 2.05 (0.65–6.57) [5Log (mean) = 0.72, SE = 0.59 
Pressure ulcer/Falls 2.18 (1.61–4.73) [3Log (mean) = 0.78, SE = 0.27 
Mortality 2.41 (0.65, 5.74) [3Log (mean) = 0.88, SE = 0.56 
Efficacy of MCI (relative risk of delirium) (95% CI) 0.66 (0.46–0.95) [11Lognormal Log (mean) = − 0.42, SE = 0.19 
Estimated utility values 
New dementia 0.29 [13Beta α = 730, β = 1094 
New admission to institution 0.18 [13α = 293, β = 880 
Falls 0.69 [16α = 249, β = 102 
Cost 
New dementia (per year) £16,302 [18Gamma Mean = £16,302, SEa = £2,079 
Stay in long-term care (per week) £656 [15Mean = £656, SEa = £84 
Hospital stay (per day) £152 [23Mean = £152, SEa = £19 
Pressure ulcer £1,364 (£1,228–£1,500)a [24Mean = £1,364, SE = £69 
Falls £1,875 [25Mean = £1,875, SEa = £239 
Multi-component intervention £377 [11Mean = £377, SEa = £48 
Duration 
Stay in long-term care (months) 18.9 [26— — 
Extended hospital stay (days) 16.83 (9.36–25.34)b [6Gamma Mean = 16.83, SE = 4.08 
Life with dementia (years) 1.2 [14— — 
Input parameter Point estimate Source Type of distribution for PSA Distribution parameters for PSA 
Baseline risk of delirium 15.0% [11Beta α = 64, β = 362 
Baseline risk estimates for adverse consequences 
New dementia 5.60% [4Beta α = 7, β = 117 
New admission to institution 17.40% [5α = 40, β = 190 
Pressure ulcer 4.00% [22α = 360, β = 8575 
Falls 6.90% [3α = 9, β = 122 
Mortality 5.00% [3α = 7, β = 124 
Estimated relative risk of adverse consequences (95% CI) 
New dementia 4.67 (1.43–15.29) [4Lognormal Log (mean) = 1.54, SE = 0.60 
New admission to institution 2.05 (0.65–6.57) [5Log (mean) = 0.72, SE = 0.59 
Pressure ulcer/Falls 2.18 (1.61–4.73) [3Log (mean) = 0.78, SE = 0.27 
Mortality 2.41 (0.65, 5.74) [3Log (mean) = 0.88, SE = 0.56 
Efficacy of MCI (relative risk of delirium) (95% CI) 0.66 (0.46–0.95) [11Lognormal Log (mean) = − 0.42, SE = 0.19 
Estimated utility values 
New dementia 0.29 [13Beta α = 730, β = 1094 
New admission to institution 0.18 [13α = 293, β = 880 
Falls 0.69 [16α = 249, β = 102 
Cost 
New dementia (per year) £16,302 [18Gamma Mean = £16,302, SEa = £2,079 
Stay in long-term care (per week) £656 [15Mean = £656, SEa = £84 
Hospital stay (per day) £152 [23Mean = £152, SEa = £19 
Pressure ulcer £1,364 (£1,228–£1,500)a [24Mean = £1,364, SE = £69 
Falls £1,875 [25Mean = £1,875, SEa = £239 
Multi-component intervention £377 [11Mean = £377, SEa = £48 
Duration 
Stay in long-term care (months) 18.9 [26— — 
Extended hospital stay (days) 16.83 (9.36–25.34)b [6Gamma Mean = 16.83, SE = 4.08 
Life with dementia (years) 1.2 [14— — 

PSA, probabilistic sensitivity analysis; SE, standard error.

The type of distribution used for the PSA of the difference in mortality between delirious and non-delirious patients was lognormal and the distribution parameters are Log (mean) = 0.54, SE = 0.27. The type of distribution used for the PSA of the population utility was multinomial and the distribution parameters are age-utility intercept: 1.06, age-utility gradient: −0.00 (data were taken from Ward et al. [27]).

aReported as mean (±10%).

b95% confidence interval.

SEa we assumed that upper and lower confidence intervals would be 125 and 75% of mean estimate, respectively.

Life expectancy

There is evidence that delirium is associated with increased mortality [2]. One study presented a Kaplan–Meier survival curve suggesting a lower survival rate in the first 3 years post-discharge and a calculated adjusted hazard ratio of occurrence of death of 1.71 [4]. We applied these estimates to capture the different survival expectations in the 3 years after discharge for people with and without delirium. We then applied the same general population rates to both groups (Interim Life Tables for England and Wales, 2005–07) up to age 100. We estimated a life expectancy of 3.6 and 5.4 years for patients with and without delirium, respectively, using a simple Markov survival model from the Life Table.

Quality of life (QALY estimates)

QALYs were estimated for the adverse outcomes of new dementia, falls, pressure ulcers and new admission to long-term care by combining utility scores describing health states with the expected life years or duration spent in the health state. The utility values used in the model are listed in Table 1 and were obtained from the systematic reviews conducted for the guideline. The utility score of 0.4 was reported for moderate dementia [13] and we have applied this as a utility multiplier in the model. We could not identify useful utility score for long-term care and therefore used the value for severe dementia of 0.25 [13] on the basis that severe dementia is a common reason for admission to long-term care after an episode of delirium. This was also applied in the model as a utility multiplier. In order to estimate QALY gains, we have used a life expectancy of 1.2 years for new dementia (weighted mean of estimates in [14]) and, 18.9 months as the survival time in the long-term care [15]. For falls, the mean utility at 12 months was reported as 0.71 [16]. This was applied in the model as a utility multiplier in the first year. The QALY gains for the rest of the patient's life expectancy were estimated from a Markov model from the UK Life Tables. We could not identify estimates for utility value for the adverse event of pressure ulcers. A conservative assumption was therefore made that the life-time expected QALY gain for pressure ulcer was equal to the QALY gain of a person without any adverse consequences of delirium. Deaths were ascribed a utility value of zero.

Costs

The cost estimates used in the model include the costs of the adverse consequences and cost of the intervention. The costs of adverse consequences are based on published estimates of the unit cost and the duration of time associated with each adverse consequence (Table 1). The costing of the intervention used the published study protocol [11, 17] but imputed UK relevant staffing costs obtained from the Personal and Social Services Research Unit [18].

Cost-effectiveness criteria

We estimated cost-effectiveness using the INMB which is the monetary value of an intervention compared with an alternative for a specific cost-effectiveness threshold (λ). It is estimated as: 

formula

The λ was £20,000 per QALY. The multi-component intervention is cost-effective when compared with usual care if it has an INMB that is greater than zero.

Sensitivity analyses

The robustness of the model results was explored by considering the impact of the model assumptions and the uncertainties in the model input parameters. This was done using deterministic and probabilistic sensitivity analyses (PSA). We carried out the following deterministic sensitivity analyses (DSA):

  • We assumed that only one of the six adverse consequences was the adverse outcome associated with delirium.

  • We included nursing home admission and mortality as a composite outcome.

  • We reduced the baseline risk of delirium to 12.5%, the lower range of incidence estimate for general medical patients [1].

  • We assumed that the survival chances for patients with or without delirium were the same.

  • We increased the life expectancy of patients with dementia with and without delirium to 3.6 and 5.4 years, respectively.

  • For patients in long-term care, we increased the life expectancy of patients with dementia with and without delirium to 3.6 and 5.4 years, respectively.

  • The annual cost of dementia was reduced to £5,859 to avoid a potential double counting of the cost of accommodation of a dementia patient.

  • The cost of stay in long-term care was reduced by 30% to reflect the assumption that the full cost of long-term care is not borne by the public.

  • Cost of pressure ulcer was increased to reflect a more severe ulcer.

  • Costs of the intervention were increased by 7% to reflect a higher payment to hospital staff.

The PSA explored the uncertainties in the model parameters by randomly sampling 5,000 times from each parameter distribution. We calculated the cost, QALYs, ICERs and INMB from this sample. This allowed both mean values to be calculated and for an estimate of the proportion of the simulations for which the intervention was cost-effective. Finally, cost-effectiveness acceptability curves (CEAC) were constructed to explore the uncertainty surrounding the optimal choice over different threshold values for λ. This is particularly important for planners whose willingness-to-pay for additional improvement in HRQoL is low.

Results

The multi-component delirium intervention was associated with a lower mean total cost of £12,690 (compared to £13,200 for usual care) and a higher QALY gain of 2.22 (compared to 2.140 for usual care) (Table 2). The intervention was therefore the dominant strategy because it reduced cost and increased QALY gains when compared to the usual care. It was associated with an INMB of £2,200 when estimated across 5,000 PSA simulations. Further, the PSA demonstrated a higher INMB estimate and therefore intervention cost-effectiveness in 96.8% of the simulations (Figure 1). These results are supported by the result of the deterministic analysis which suggests that the intervention is cost-effective with an INMB of £2,130.

Table 2.

The costs, QALYs and cost-effectiveness of the multi-component delirium prevention intervention compared with usual care

  Usual care Multi-component intervention 
Probabilistic analysisa Mean cost £13,200 £12,690 
Mean QALYs 2.140 2.220 
Incremental cost N/A −£520 
Incremental QALYs 0.084 
Incremental cost/QALY (ICER) Multi-component intervention dominates usual care 
Incremental net monetary benefit (INMB) £2,200 
% of simulations where strategy was most cost-effective 3% 97% 
Deterministic analysis Incremental net monetary benefit (INMB) N/A £2,130 
  Usual care Multi-component intervention 
Probabilistic analysisa Mean cost £13,200 £12,690 
Mean QALYs 2.140 2.220 
Incremental cost N/A −£520 
Incremental QALYs 0.084 
Incremental cost/QALY (ICER) Multi-component intervention dominates usual care 
Incremental net monetary benefit (INMB) £2,200 
% of simulations where strategy was most cost-effective 3% 97% 
Deterministic analysis Incremental net monetary benefit (INMB) N/A £2,130 

aCosts and QALYs are mean total costs and QALYs across 5,000 probabilistic sensitivity analysis simulations.

Figure 1.

Cost-effectiveness plane for multi-component targeted intervention compared with usual care.

Figure 1.

Cost-effectiveness plane for multi-component targeted intervention compared with usual care.

The CEAC show the probability that the intervention was cost-effective over a range of λ values (£0 to £60,000 per QALY) (Supplementary data are available in Age and Ageing online, Figure S1). The probability of being more cost-effective at λ of £30,000 per QALY was 98% and, at this threshold, the INMB was £3,040.

The intervention remained cost-effective for the majority of the DSA conducted (Supplementary data are available in Age and Ageing online, Table S1). Exceptions were when we assumed that pressure ulcer, falls, in-hospital mortality and extended hospital length of stay were the only adverse outcomes associated with delirium. It remained cost-effective when we excluded the survival difference between delirious and non-delirious cases, removed the cost of dementia due to stay in long-term care, and increased the cost of pressure ulcer. It was cost-effective when the life expectancy of dementia was increased from 1.2 years to 3.6 and 5.4 years for dementia with and without delirium, respectively. When we used the composite outcome of new admission to institution and mortality, and assumed that the UK NHS and PSS would pay only 70% of the cost of stay in long-term care, the intervention remained cost-effective.

Discussion

We have estimated the cost-effectiveness of a multi-component targeted intervention in older patients at intermediate or high risk of delirium admitted to a general medical service. We found this prevention intervention to be cost-effective as it was associated with an INMB of £2,200. This finding was robust to the majority of sensitivity analyses.

There are several factors that might have affected our findings. We assumed that the adverse outcomes on the decision tree branches were mutually exclusive. For example, a patient with delirium who subsequently develops dementia might also be admitted to a nursing home and the total cost and QALY gain for this patient might be different from the modelled estimate as the two outcomes are occurring in the same patient rather than in separate individuals. We tested the impact of this assumption by considering that each of the six adverse outcomes was the only outcome to be associated with delirium therefore removing the risk of double counting. The intervention remained cost-effective when some of the outcomes were the only one to be associated with delirium.

We assumed that the relative risk of falls and pressure ulcer was the same on the basis of the best available study. The base-case cost estimate for pressure ulcer was based on the assumption that it would be a less severe ulcer. We made an alternative assumption for a more severe one, and the intervention remained cost-effective. In the base-case, we assumed that all the cost of long-term care would be paid by the UK NHS and PSS. We made an alternative assumption that only 70% of this cost will be paid by the public. The cost of dementia used in the model included the cost of stay in long-term care and it could be argued that the cost of long-term care was already accounted for as a different model outcome and that we had double counted. We therefore made an alternative assumption and removed the proportion of cost of dementia attributable to long-term care. The life expectancy for dementia used in the base-case was 1.2 years on the basis of a weighted estimate. Arguably, this may be too short and could even be different between patients with and without delirium. We therefore increased the life expectancy to 3.6 and 5.4 years for dementia patients with and without delirium, respectively. In all these various alternative situations, the intervention remained cost-effective.

There are some uncertainties in our model input parameters. They are published estimates with a range of plausible values. We explored the impact of such uncertainties in the PSA. The results of this analysis suggest that the intervention remains cost-effective over the plausible range of parameter values. Further, the simulations conducted in the PSA for different cost-effectiveness thresholds support this finding and showed the high level of certainty with the results presented here.

An initial systematic review conducted for the NICE guideline failed to identify any relevant economic evaluations for multi-component delirium prevention interventions applicable to the guideline population. This is therefore the first evaluation of this type of complex intervention which estimates cost-effectiveness using cost and QALYs as primary outcomes. Previous research suggests that the multi-component intervention is cost-effective when directed towards elderly patients at intermediate, but not high baseline risk, for delirium [19]. Our findings here suggest that the intervention should target elderly patients at both intermediate and high risk for delirium. The strength and uniqueness of our analysis lies in the methods used. Our model was developed by a multi-disciplinary group comprising of clinical experts, patient representatives and a technical team who guided the selection of model structures and input parameters. We have included HRQoL in our analysis and have also included outcomes that would reflect more lasting benefits of delirium prevention. These are outcomes which were not included in previous research but were acknowledged to be useful in order to show the benefit of the multi-component intervention [19].

The INMB estimate of £2,200 implies a saving for each patient receiving the intervention over the remaining life-span. In many ways, this is a remarkable finding given that most novel health-care systems require increased costs to achieve improved outcomes. These more usual health economic situations are presented in terms of affordability, typically a particular threshold of a funder's willingness to pay in return for an additional QALY gained. The positive value for the INMB should make multi-component delirium prevention an attractive intervention to the NHS. However, the savings for the intervention are spread unevenly between the NHS and social care providers. The savings to the NHS may be modest and largely accrue through lower treatment costs resulting from reduced length of hospital stay, where as the savings to social services are likely to be more considerable resulting from an enduring and diminished burden of dependency and dementia, particularly reduced need for expensive institutional care. Indeed, we estimated an additional cost to implement the intervention of £377 per patient. This was based on the description of the intervention that required additional staff for delivery in the context of the underpinning explanatory, ‘proof of concept’ trial [11]. Thus NHS acute providers may need to invest to implement the intervention and to accrue savings to the wider public sector. The current NHS hospital funding system does not incentivise this type of investment. This is potentially a major structural barrier to a widespread uptake of delirium prevention systems of care in the UK. However, it is possible that the model we have developed provides an important under-estimate of cost-effectiveness. This is because it might be possible to implement the intervention within existing resources. Arguably, the intervention is designed to addresses risk factors for delirium by delivering the sort of person-centred routine care that patients might expect to receive [20]. For example, attention to hydration, nutrition, medication reviews, mobilisation etc. Such an approach appears feasible in routine care [21] though new research is needed to demonstrate reliably a reduction in incident delirium. However, our health economic analysis has demonstrated convincingly that delirium prevention systems of care can be considered as a cost-effective health-care strategy for medically ill people admitted to hospital.

Key points

  • The health-care strategy to implement a multi-component delirium prevention intervention in older people admitted to medical services is cost-effective.

  • The economic evidence is convincing and investment to implement this intervention will accrue savings to the NHS and social care services.

  • Delirium is a common and serious condition in acutely ill older people and multi-component prevention interventions should be encouraged by health care.

Supplementary data

Supplementary data mentioned in the text is available to subscribers in Age and Ageing online.

Acknowledgements

The method and results presented here draw completely from the full version of the NICE clinical guideline 103: the diagnosis, prevention and management of delirium (see reference [1]). The guideline was developed by the National Clinical Guideline Centre which is funded by NICE. The important input of the Guideline Development Group is gratefully acknowledged. In the guideline, NICE recommends that a tailored multi-component delirium prevention intervention should be provided in hospitals and in residential care. The economic evidence presented here was used to support this recommendation. The authors of this article are employed by the organisations for which they have provided contact details. They did not receive additional funding for writing this article. The views expressed in this article are those of the authors and not necessarily those of NICE or the organisations they work for. Anayo Akunne and Lakshmi Murthy were working at the National Clinical Guideline Centre at the time the guideline was developed. We are particularly grateful to Sarah Davis, whose contribution to the economic model was significant. We are also particularly grateful to Maggie Westby, whose work on the review that underpins the model was important. At the time the guideline was developed, Sarah Davis was a Senior Health Economist at the National Clinical Guideline Centre while Maggie Westby was a Clinical Effectiveness Lead. We are also grateful to the 2 independent peer reviewers of the article.

References

1
National Institute for Health and Clinical Excellence
Delirium: Diagnosis, Prevention and Management
 , 
2010
London: National Institute for Health and Clinical Excellence
 
(Clinical Guideline 103), Available from: http://guidance.nice.org.uk/CG103/Guidance.
2
Siddiqi
N
Horne
AO
House
AO
Holmes
JD
Occurrence and outcome of delirium in medical in-patients; a systematic literature review
Age Ageing
 , 
2006
, vol. 
35
 (pg. 
350
-
64
)
3
O'Keeffe
S
Lavan
J
The prognostic significance of delirium in older hospital patients
J Am Geriatr Soc
 , 
1997
, vol. 
45
 (pg. 
174
-
8
)
4
Rockwood
K
Cosway
S
Carver
D
Jarrett
P
Stadnyk
K
Fisk
J
The risk of dementia and death after delirium
Age Ageing
 , 
1999
, vol. 
28
 (pg. 
551
-
6
)
5
Bourdel-Marchasson
I
Vincent
S
Germain
C
, et al.  . 
Delirium symptoms and low dietary intake in older inpatients are independent predictors of institutionalization: a 1-year prospective population-based study
J Gerontol A Biol Sci Med Sci
 , 
2004
, vol. 
59
 (pg. 
350
-
4
)
6
Holmes
J
House
A
Psychiatric illness predicts poor outcome after surgery for hip fracture: a prospective cohort study
Psychol Med
 , 
2000
, vol. 
30
 (pg. 
921
-
9
)
7
Thomason
JW
Shintani
A
Peterson
JF
Pun
BT
Jackson
JC
Ely
EW
Intensive care unit delirium is an independent predictor of longer hospital stay: a prospective analysis of 261non-ventilated patients
Critical Care
 , 
2005
, vol. 
9
 (pg. 
R375
-
81
)
8
Ely
EW
Shintani
A
Truman
B
, et al.  . 
Delirium as a predictor of mortality in mechanically ventilated patients in the intensive care unit
J Am Med Assoc
 , 
2004
, vol. 
291
 (pg. 
1753
-
62
)
9
Pitkala
KH
Prognostic significance of delirium in frail older people
Dement Geriatr Cogn Disord
 , 
2005
, vol. 
19
 (pg. 
158
-
63
)
10
United State Department of Health and Human Services
CMS Statistics
 , 
2004
Washington, DC
Centers for Medicare and Medicaid Services
 
34 (CMS Pub. No. 03445)
11
Inouye
SK
Bogardus
ST
Jr
Charpentier
PA
, et al.  . 
Amulticomponent intervention to prevent delirium in hospitalized older patients
N Eng J Med
 , 
1999
, vol. 
340
 (pg. 
669
-
76
)
12
National Institute for Health and Clinical Excellence
Guide to the Methods of Technology Appraisal
 
London
 
13
Ekman
M
Berg
J
Wimo
A
Jonsson
L
McBurney
C
Health utilities in mild cognitive impairment and dementia: a population study in Sweden
Int J Geriatr Psychiatry
 , 
2007
, vol. 
22
 (pg. 
649
-
55
)
14
McNamee
P
Bond
J
Buck
D
Resource Implications Study of the Medical Research Council Cognitive Function and Ageing Study. Costs of dementia in England and Wales in the 21st century
Br J Psychiatry
 , 
2001
, vol. 
179
 (pg. 
261
-
6
)
15
Netten
A
Bebbington
A
Darton
R
Forder
J
Miles
K
1996 Survey of care homes for elderly people
1998
 
Final Report, PSSRU Discussion Paper 1423/2, University of Canterbury, Kent: Personal Social Services Research Unit
16
Hendriks
MR
Evers
SM
Bleijlevens
MH
van Haastregt
JC
Crebolder
HF
van Eijk
JT
Cost-effectiveness of a multidisciplinary fall prevention program in community-dwelling elderly people: a randomized controlled trial
Int J Tech Assess Health Care
 , 
2008
, vol. 
24
 (pg. 
193
-
202
)
17
Inouye
SK
Bogardus
ST
Jr
Baker
DI
Leo-Summers
L
Cooney
LM
Jr
The hospital elder life program: a model of care to prevent cognitive and functional decline in older hospitalized patients
J Am Geriatr Soc
 , 
2000
, vol. 
48
 (pg. 
1697
-
706
)
18
Personal Social Services Research Unit
 
Dementia UK: The Full Report, London: The Alzheimer's Society. 2007. Available from, http://www.alzheimers.org.uk/site/scripts/download.php?fileID=2
19
Rizzo
JA
Bogardus
ST
Jr
Leo-Summers
L
Williams
CS
Acampora
D
Inouye
SK
Multicomponent targeted intervention to prevent delirium in hospitalized older patients: what is the economic value?
Medical Care
 , 
2001
, vol. 
39
 (pg. 
740
-
52
)
20
Rockwood
K
Need we do so badly in managing delirium in elderly patients?
Age Ageing
 , 
2003
, vol. 
32
 (pg. 
473
-
4
)
21
Vidan
M
Sanchez
E
Alonso
M
Montero
B
Ortiz
J
Serra
J
An intervention integrated into daily clinical practices reduces the incidence of delirium during hospitalisation in elderly patients
JAGS
 , 
2009
, vol. 
57
 (pg. 
2029
-
36
)
22
Clark
M
Watts
S
The incidence of pressure sores within a National Health Service Trust Hospital During 1991
J Adv Nursing
 , 
1994
, vol. 
20
 (pg. 
33
-
6
)
23
The NHS Information Centre
 
Hospital Episode Statistics for England, 2007–08
24
Bennett
G
Dealey
C
Posnett
J
The cost of pressure ulcers in the UK
Age Ageing
 , 
2004
, vol. 
33
 (pg. 
230
-
5
)
25
Iglesias
CP
Manca
A
Torgerson
DJ
The health-related quality of life and cost implications of falls in elderly women
Osteoporos Int
 , 
2009
, vol. 
20
 (pg. 
869
-
78
)
26
Netten
A
Darton
R
Bebbington
A
Forder
J
Brown
P
Mummery
K
Residential and nursing home care of elderly people with cognitive impairment: prevalence, mortality and costs
Aging Men Health
 , 
2001
, vol. 
5
 (pg. 
14
-
22
)
27
Ward
S
Lloyd
JM
Pandor
A
, et al.  . 
A systematic review and economic evaluation of statins for the prevention of coronary events
Health Tech Assess
 , 
2007
, vol. 
11
 (pg. 
1
-
160
)

Supplementary data

Comments

1 Comment
NICE, but not that nice?
31 May 2012
Robert H Skelly

Akunne et al [1] may have over-estimated the cost-effectiveness of multi-component interventions for the prevention of delirium. Their review aims to summarise the detailed cost-effectiveness analysis carried out during the development of the delirium guideline by NICE (the National Institute for Health and Clinical Excellence). The authors note that delirium is associated with adverse outcomes such as dementia, prolonged hospital stay, falls and pressure sores. Likewise, they cite Sharon Inouye's 1999 study [2] that demonstrated a reduced incidence of delirium following a multi-component intervention (relative risk of delirium in intervention group 0.66 [95% confidence limits 0.46-0.95]. Is it logical to conclude that long term adverse outcomes will be prevented if delirium is prevented? Akunne et al appear to have made this assumption in their cost-effectiveness model but the assumption is flawed. In fact, when the patients in the 1999 Inouye study were followed up 6 months after discharge the investigators could find "no lasting benefit of the intervention" [3]. There was no difference between intervention group and usual care group in the outcomes such as cognitive status, functional status, depression, or nursing home placement. The likely reason for this disappointing longer term outcome is the complex relationship between dementia and delirium: acute illness inflammatory processes may damage the brain in delirium and lead on to dementia; drugs used to treat delirium may have long term toxic effects; or delirium may be a marker of underlying brain disease [4].

1. Akunne A, Murthy L, Young J. Cost-effectiveness of multi-component interventions to prevent delirium in older people admitted to medical wards. Age Ageing 2012; 41: 285-291.

2. Inouye SK, Bogardus ST Jr, Charpentier PA et al. A multicomponent intervention to prevent delirium in hospitalised older patients. N Eng J Med 1999; 340: 669-676.

3. Bogdardus ST Jr, Desai MM, Williams CS et al. The effects of a targeted multicomponent delirium intervention on post discharge outcomes for hospitalized older adults. Am J Med 2003; 114: 383-390.

4. MacLullich AM, Beaglehole A, Hall RJ, Meagher DJ. Delirium and long-term cognitive impairment. Int Rev Psychiatry 2009; 21: 30-42.

Conflict of Interest:

None declared

Submitted on 31/05/2012 8:00 PM GMT