Table 1 Characteristics of included reviews and summary of findings - older population with dementia or cognitive impairment
From: Systematic reviews on behavioural and psychological symptoms in the older or demented population
First author | Search date | BPS | Popu-lation | N reviewed | Summary of results | Meta-analysis | Recommendations future research | Reported limitations | Quality |
|---|---|---|---|---|---|---|---|---|---|
Prevalence and co-occurrence | |||||||||
Monastero [21] | Aug 2008 | BPS | MCI | 27 | Prevalence: 35 to 85%. Most common: dep, anx and irr. Hospital-based studies reported higher prevalence than population based studies | - | - Large cohort studies - Using standardised MCI criteria - Standardised behavioural instruments - Prognostic role of BPS in MCI - Not exclude those with dep at baseline Previously proposed: - Prevalence - Correlates - Different MCI subtypes, - Genetic/biological markers | Original studies - Differences in methodology, including setting, age and sex distribution, inclusion and exclusion criteria and differential sensitivity of BPS instruments Review - English language only - Heterogeneity affected ability to compare data | 4 |
Apostolova [22] | Dec 2006 | BPS | MCI | 21 | Prevalence: 35 to 75%. Most common BPS: dep, apa, anx and irr. Least common: ela, hal, dis and wan | - | - Large, prospective longitudinal studies - Standard MCI diagnostic criteria not excluding those with depression - Neuropsychiatric instruments designed for the cognitively impaired | Original studies - Different sampling methods - Exclusion of subjects with depression - Heterogeneity of MCI diagnostic criteria - Heterogeneity of BPS instruments Review - Only English literature - Heterogeneity limits comparison | 3 |
Seitz [23] See 1B | Mar 2009 | BPS Dep, Anx | Care home | 35 | Prevalence BPS in dementia: 78% (median) | - | See Table 2 | See Table 2 | 3 |
Zuidema [24] | Aug 2005 | BPS | MMSE < 24, care home | 25 | Prevalences ranged considerably, from 3 to 54% for del, 1 to 39% for hal, 8 to 74% for dep, 7 to 69% for anx, 17 to 84% for apa, 48 to 82% for agg or agi, and 11 to 44% for psychical agg. | - | See risk factors | See risk factors | 2 |
Shub [25] | NR | Psy Agg | Dem | 54 | Of seven cross and two long studies directly examining correlation agg and psy, most showed a positive association. | - | - Prospectively designed studies - Temporal relationship | Original studies - Limited data of sufficient methodological rigor - Majority cross-sectional studies | 3 |
Wragg [26] | NR | Dep Psy | Dem | 30 | Dep and Psy occurred in 30 to 40% of AD patients. Isolated symptoms were two to three times as frequent as diagnosable affective or psychotic disorders. Paranoid del were the most common psy symptoms. | - | - Specify an a priory hypothesis - Clearly report characteristics study sample - Prospective and longitudinal design - Use standardised case definitions - Reliable and valid BPS instruments - Use appropriate statistical analysis - Clearly report conclusions - Both theoretical and clinical focus. | Original studies - Methodological limitations including small samples, misclassification bias Review - Relatively small number of studies spanning several decades - Publication bias | 2 |
Ropacki [27] | 2003 | Psy | Dem | 55 | Prevalence psy: 41% (del 36%, hal: 18%) | - | - Longitudinal designs - Incidence and persistence of psychosis - Develop or utilise diagnostic criteria and rating scales for psychosis - Take into account medication use - Mechanisms underlying psychosis in AD | Original studies - Severely cognitively impaired subjects not included - Potential effects of medication on cognition - Small standard deviations age at onset and illness duration - Assessing psy: diagnostic criteria nonspecific to dementia and inconsistencies interpreting criteria | 1 |
Course and progression | |||||||||
Monastero [21] | Aug 2008 | BPS | MCI | 27 | Prospective studies showed that BPSD, particularly depression, may represent risk factors for MCI or predictors for the conversion of MCI to AD. | - | See prevalence | See prevalence | 4 |
Verkaik [34] | Mar 2006 | Dep | Dem | 24 | 1/4 (continous) and 0/3 (categorical) high quality studies found a significant association between severity of AD and prevalence of dep. | - | - Longitudinal study - Using a standardised dementia definitions - Assessment severity of AD - Assessing dep with dem specific instrument - Control for confounders | Review - Only English language, studies that did not have depression or depressive disorder as a keyword were not identified | 5 |
Ropacki [27] | 2003 | Psy | Dem | 55 | Incidence increased progressively the first three years, after that plateau. Duration several months but less prominent after one year. Associated with more rapid cognitive decline | - | See prevalence | See prevalence | 1 |
Biological | |||||||||
Flirski [42] | ? | BPS | Dem | 73 | Behavioural genetics of BPS reviewed: genes coding for APOE E, serotonin receptors, serotonin transporter, COMT, MAO-A, tryptophan hydroxylase and dopamine receptors. A general conclusion is the striking inconsistency of the findings, unsurprising in the field of psychiatric genetics. | - | - Precisely define symptoms - Fusing multidisciplinary data | Original studies - Inconsistency of results - Recruitment solely based on clinical diagnosis - Variability in study design - BPS fluctuation, studies rely heavily on average disease state - Cross sectional studies - Variety of BPS instruments - Studying isolated symptoms or symptom clusters - Selection bias: ethnicity or genetic homogeneity, choice of setting - Insufficient number of study participants - No correction for multiple testing - Carrier status versus dose - Multifactorial aetiology | 0 |
Risk factors | |||||||||
Zuidema [24] | Aug 2005 | BPS | MMSE < 24, care home | 25 | BPSD predicted not only by dem type or stage, but also by the psychosocial environment and the amount of psychoactive medication and physical restraints used. | - | - Effects of manipulation the physical and social environments in nursing homes. | Original studies - Uncertainty defining dem, its type and severity - Few accurately diagnosed dementia - Different BPS instruments and definitions | 2 |
Wragg [26] | NR | Dep Psy | Dem | 30 | Â | - | See prevalence | See prevalence | 2 |
Ropacki [27] | 2003 | Psy | Dem | 55 | Associations: age, age at onset AD, illness duration. Weak/inconsistent: gender, education, family history dem or psychiatric illness | - | See prevalence | See prevalence | 1 |
Care | |||||||||
Gaugler [51] | 2006 | BPS | Dem | 80 | Behavioural symptoms one of most consistent predictors of nursing home admission in persons with dementia. | - | - Interventions should consider long-term efficacy and timing of nursing home admission in course of dem - Power for subgroup analyses - More complex models of institutionalisation. | Review - Research synthesis method: requires descriptive information from samples composed of subjects a similar age; studies providing only correlations and not means or standard deviations are excluded and some assumption made about shape of distribution. | 5 |
Black [52] | Dec 2001 | BPS | Dem | 55 | Pooled correlation coefficients for relationship BPS and caregiver burden (0.57, 95%CI 0.52 to 0.62), caregiver psychological stress (0.41, 0.32 to 0.49) and caregiver depression (0.30, 0.21 to 0.39). Multivariate data supported BPS are predictor of burden of care, psychological distress and dep. Limited long data. Caregiver variables may be more important in predicting institutionalisation than BPS. | See summary of results | - Concept of burden of care is too broad and more clinically relevant measures such as caregiver depression are preferred. - Cohort studies | Original studies - Relatively few studies - Majority clinic-based samples, few representative - Majority cross-sectional and correlational - Little about which care-recipient symptoms are most distressing or particular risk factors for subgroups - Concept of burden may be too broad - Other variables are likely to be important Review - Publication bias - Pooling data assumes homogeneity (questionable) | 3 |
Disease outcome | |||||||||
Lee [54] | NR | BPS | Dem | NR | There was no consensus regarding the association with dementia prognosis | - | - Guideline for dementia prognostication - Risk score to better estimate survival. | Original studies - Uncertainty of etiologic diagnosis of dementia | 1 |
Fischer [55] | NR | Psy | Dem | 6 | Three of six studies showed and association with real-world functioning | - | - Longitudinal studies - More detail about delusional severity, - Use cognitive and functional measures that are better at detecting executive impairment to clarify the association. | Original studies - Basic measures functional performance + cognition - Confounders not always taken into account - Considerable variation definition del - Psychoactive medication not taken into account - No longitudinal studies | 2 |
Quality of life | |||||||||
Banerjee [53] | Oct 2007 | BPS | Dem | NR | Strong suggestion dep is consistently associated with decreased health related quality of life in dem. Magnitude of associations is moderate and the proportion of variance explained is low. | - | - Quality of life in dementia: determinants, in dementia subtypes, self- versus proxy-report, in different settings, association with outcomes and interventions | NR | 0 |