Intended for healthcare professionals

Education And Debate

North of England evidence based guidelines development project: guideline for the primary care management of dementia

BMJ 1998; 317 doi: (Published 19 September 1998) Cite this as: BMJ 1998;317:802

This article has a correction. Please see:

  1. Martin Eccles (Martin.Eccles{at}, professora,
  2. Julie Clarke, general practitionerb,
  3. Moira Livingston, consultantc,
  4. Nick Freemantle, senior research fellowd,
  5. James Mason, research fellowd,
  6. for the North of England Evidence Based Dementia Guideline Development Group
  1. aCentre for Health Services Research, University of Newcastle upon Tyne, Newcastle upon Tyne NE2 4AA
  2. bDepartment of Primary Health Care, School of Health Sciences, Medical School, Newcastle upon Tyne NE2 4HH
  3. cPsychiatry of Old Age Service, North Tyneside General Hospital, North Shields NE29 8NH
  4. dCentre for Health Economics, University of York, York YO1 5DD
  1. Correspondence to: Professor Eccles
  • Accepted 23 April 1998

Other members of the guideline development and project groups are listed at the end of the paper.

This guideline aims to provide recommendations to assist general practitioners manage people with all forms of dementia and help their carers. This is a summary version of the full guideline.1 The areas covered were developed in conjunction with the guideline development group. They were felt to reflect areas that were important in daily clinical practice. The guideline is for the management of patients with dementia; although it covers the area of screening instruments, it is not meant to cover the area of differential diagnosis. All recommendations are for general practitioners and apply to patients attending general practice with dementia. The development group assumes that doctors will use their knowledge and judgment in applying the principles and recommendations given below in managing individual patients, since the recommendations may not be appropriate for all circumstances. Doctors must decide to adopt any particular recommendation in the light of available resources and the circumstances of individual patients. Throughout this guideline categories of evidence (cited as I, II, or III) and the strength of recommendations (A, B, C, or D) are as described in previous method papers and the full version of the guideline. 1 2 A summary of categories of evidence and strength of recommendations is given in the box.

Summary points

In general practice, dementia has an incidence of 1.6, a prevalence of 3.6, and a workload of 7.4 consultations per year

General practitioners should use formal cognitive testing as well as clinical judgment in diagnosing dementia

The impact of caring depends on factors such as behaviour and affect more than on the severity of the cognitive impairment

Search strategy and synthesis

The search strategy was carried out using Medline (covering 1966-96), BIDS-EMBASE(1980-96), and PsycLIT (1974-96). Searches were limited to English language studies. We conducted MeSH heading and free text searches in the area of dementia using the terms: meta-analysis, randomised controlled trials, reviews, cohort studies, or case-control studies. The search was backed up by the expert knowledge and experience of group members. The quality of relevant studies retrieved was assessed and the information from relevant papers was synthesised using narrative summary.

General practice workload

The incidence, prevalence, and workload associated with dementia can be estimated from the morbidity statistics from general practice.3 The statistics classify patients with dementia within the group “Senile and pre-senile organic psychotic conditions” (ICD-9 290). Although this includes presenile conditions, these form only a small part of the total. Given an assumed list size of 2000 patients, dementia is associated with an incidence of 1.6 new patients per general practitioner per year, a prevalence of 3.6 patients consulting per general practitioner per year, and a workload of 7.4 consultations per general practitioner per year.

The morbidity statistics also allow us to estimate where the consultations occur. For all conditions, 14% of contacts with patients aged 65-74 years occur as home visits; for patients with dementia the equivalent proportion is 51%. For patients aged 75 years and over, the proportions seen at home are 40% for all conditions and 71% for patients with dementia.



Mild dementia is difficult to diagnose, therefore it is difficult to assess its prevalence (II). The older the patient the more likely the diagnosis of a dementing illness is to be senile dementia of the Alzheimer type (II). The prevalence of dementia increases with age and is estimated to be about 20% at 80 years of age (II). The annual incidence of vascular dementia is 1.2/100 overall person years at risk, and is the same in all age groups (II). The annual incidence of senile dementia of the Alzheimer type is 34.3/100 person years at risk in the 90 year age group; the incidence is higher in women than in men (II). In a third of cases, dementia is associated with other psychiatric symptoms (depressive disorder, adjustment disorder, generalised anxiety disorder, alcohol related problems) (II). A complaint of subjective memory impairment is not a good indicator of dementia; altered functioning is a more important symptom (II).47

Strength of recommendation

  • A—Directly based on category I evidence

  • B—Directly based on category II evidence or extrapolated

  • C—Directly based on category III evidence or extrapolated recommendation from category I or II evidence

  • D—Based on the group's clinical opinion

Categories of evidence

  • I—Based on well designed randomised controlled trials, meta-analyses, or systematic reviews

  • II—Based on well designed cohort or case control studies

  • III—Based on uncontrolled studies or external consensus

Recommendations and important points

  • The incidence and prevalence of dementia increase with increasing age (B).

  • Complaints of subjective memory impairment are not a good indicator of dementia; a history of loss of function is more indicative (B).

  • Population screening for dementia in the over 65s is not recommended; a case finding approach is recommended (D).

Identifying people with dementia

Patient's or carer's history


As the dementing process progresses, awareness of memory problems decreases, leading to less reliable histories from patients (II). People with dementia cannot be relied on to complain of memory difficulties (II). The short mental questionnaire is a screening tool that is sensitive to mild dementia. It can be completed by the carer and may have a useful place in identifying people with dementia (II). Memory complaints by patients correlate with depression. Carers' complaints about the memory of their relatives correlate with dementia (II).812

Recommendations and important points

  • Insight diminishes as dementia progresses making the patient's history less reliable (B).

  • In assessing a person with cognitive impairment, a history of memory problems should be sought from the carer as well as the patient (B).

  • Dementia and other psychiatric symptoms (delusions or hallucinations, or both, usually persecutory in nature and simple in type) may coexist (B).

The general practitioner


The general practitioner's clinical judgment alone compares unfavourably with the use of formal cognitive testing in the diagnosis of dementia (II).1315


General practitioners should consider using formal cognitive testing to enhance their clinical judgment (B).

Short screening tests for cognitive impairment


The mini-mental state examination can be shortened for use in primary care with only a small reduction in specificity (II). The mini-mental state examination may be influenced by verbal fluency, age, education, and social grouping (II). Four items of the mini-mental state examination are predictors of dementia: orientation to day, spell WORLD backwards, recall three words, write a sentence (II). Reducing the mini-mental state examination to two items—recall and orientation for place—reduces the specificity only slightly (II). In the clock drawing test, the accuracy of the fourth quadrant of the clock face shows the greatest sensitivity (87.5%) and specificity (82.3%) for dementia (II). Deterioration in four domains of instrumental activities of daily living are significantly associated with cognitive impairment. These domains are: managing medication, using the telephone, coping with a budget, and using transportation (II). The shortened mental test score has proved statistically significant for discriminating between the normal and the abnormal, those with delirium and those with mixed dementia and delirium, those with dementia and those with mixed dementia and delirium, those who are not demented, and those with delirium (III).1634


  • Healthcare professionals should consider using the following instruments to identify cognitive impairment (B):

  • The mini-mental state examination

  • The clock drawing test

  • An instrument for assessing activities of daily living, for example, CAPE (Clifton assessment procedures for the elderly)

  • Abbreviated mental test score.

Physical screening in dementia patients


A small proportion of people with dementia have an underlying abnormality, and when this is treated cognitive function improves. The exact number of people thus affected is uncertain because of problems of study populations (II). People with Alzheimer type dementia do not complain of common physical symptoms, but experience them to the same degree as the general population (II).3548

Recommendations and important points

  • Healthcare professionals should be aware of the existence of reversible causes of dementia (B).

  • People with dementia experience physical morbidity to the same degree as the general population, but are likely to underreport their symptoms (B).

  • General practitioners should ensure that the following routine tests are performed: haematology (including erythrocyte sedimentation rate), biochemistry, serum calcium and phosphate, thyroid function, and simple urine analysis (B).

Dementia of the Lewy body type


The clinical course of dementia of the Lewy body type differs from that of Alzheimer's disease, showing clear fluctuations with the following clinical features: complex visual hallucinations (48%), auditory hallucinations (14%), paranoid delusions (57%), clouding of consciousness (81%), falls or collapses (38%), depression (38%), extrapyramidal features (9.5%) (II). There is high neuroleptic sensitivity (61.5%) and a high risk of increased morbidity and mortality if neuroleptic drugs are prescribed (II). 4950

Recommendations and important points

  • Differential diagnosis of dementia of the Lewy body type is important because of the high risk of increased morbidity and mortality with neuroleptic agents in these patients (B).

  • Doctors should be aware of the importance of avoiding neuroleptic drugs in people known to have dementia of the Lewy body type (B).

Depression in patients with dementia


Depressive illness is commoner in people with dementia than in those without (I). Mortality is increased in people with dementia and depression (II). The prevalence of depression in patients with dementia varies widely according to study population. In a general population, the prevalence varies from 10%-40% of patients with dementia (II). Depression is more commonly diagnosed or recognised in early dementia (I). Treatment is likely to be of value, with reported response rates of up to 85% (I). Depression commonly leads to difficulties in communication and independent activities of daily living and has a less common effect on cognitive function (I). Presenting symptoms relate to inner feelings (anxiety, mood, loss of interest, helplessness, hopelessness and worthlessness) and less to vegetative symptoms (I).5085

Recommendations and important points

  • Depression can be suspected in patients with dementia at any stage in the dementing process (B).

  • The history should be gathered from both the patients and their carers (B).

  • Relevant risk factors for depressive illness, such as personal or family history of depression, or recent adverse events, such as bereavement or relocation, should be considered (B).

  • Consider a trial of antidepressant medication evaluated against explicit criteria such as activities of daily living, level of functioning, behavioural disturbance, and biological features of recent onset (B).

Non-psychotic behavioural disorders


An association exists between acute underlying medical illness and outbursts of aggressive behaviour in people with dementia (II). A placebo response is seen in 67% of people treated with neuroleptic agents for the control of behavioural disorders in dementia; there is no difference between neuroleptic agents used and no identifiable differences between responders and non-responders (I). A high proportion of people with dementia of the Lewy body type are sensitive to neuroleptic agents, and an appreciable number of these experience a severe reaction (II). Delusions or misidentifications are associated with a high number of aggressive episodes (II). 49 6772

Recommendations and important points

  • Any underlying causes of behavioural disorder, for example, an acute physical illness, environmental distress, or physical discomfort, should be excluded (D).

  • Where underlying causes are identified they should be managed before prescribing drugs for the behavioural disorder (D).

  • General practitioners should, wherever possible, resist using tranquillisers routinely to control behaviour disorders in dementia (D).

  • In crisis situations, the short term use of neuroleptic drugs may be appropriate (D).

  • Patients with dementia of the Lewy body type should not be treated with neuroleptics (B).

  • There is a relation between delusions and aggressive behaviour; aggressive behaviour should be assessed with this in mind (D).

  • The care setting and the attitudes of carers (or care teams in an institutional setting) may influence the emergence of behavioural problems (D).



Fracture of the hip is the commonest fracture in falls associated with dementia (II). Medication increases the risk of falling in people with dementia (II). Falls are not associated with the severity of the dementia but are associated with wandering and reversible confusion (II). Those people who fall are more likely to fall again and falls are associated with their doing too much, for example, wandering, restlessness (II). Falls are increased in the more capable groups of people with dementia (II).7375

Recommendations and important points

  • Falls are increased in people with dementia (B).

  • Medication, wandering, and reversible confusion contribute to the risk of falls (B).

  • People with dementia who fall are more likely to fall again (B).

  • The risk of falls is not associated with the severity of the dementia but with the functional capability of the person, and is increased in more capable people (B).

Leaving home


Factors in patients that lead to an increased risk of institutionalisation are physical dependence, irritability, nocturnal wandering, and incontinence (III). Stress in carers can lead to an increased risk of institutionalisation (III). Institutionalisation offers the best duration of survival for people with dementia followed by a formal care package at home. Survival in this context means time until death rather than quality of life (III). Day care for people with dementia can delay institutionalisation (III).7683

Recommendations and important points

  • The following factors increase the likelihood of patients having to leave their own homes: carer stress, physical dependence, irritability, nocturnal wandering, and incontinence (B).

  • A patient should not be assessed for optimal home care independently of the carer (B).

Non-drug therapies—residential care


>Skills training for people with dementia in residential care may lead to an improvement in personal care skills (I). Music therapy for people with dementia in residential care leads to an improvement in personal recollection, social disposition, enjoyment, and interaction during treatment (I). Activities and education for people with dementia in residential care lead to a decrease in behavioural disorder and an increase in the activity levels (I).8486


  • General practitioners should consider the use of structured programmes to encourage continuing independence for people with dementia resident in nursing homes (A).

Drug treatments


Aspirin in vascular dementia

Aspirin is of benefit in preventing vascular events or vascular death in patients with a history of prior transient ischaemic attack or stroke (I). Atrial fibrillation has been shown to have an association with cerebrovascular dementia (I).8791


Both reviews and studies show small improvements of variable sustainability. Those who will respond to hydergine cannot be predicted in advance (I).9296


There is no consistent evidence of clinical benefit from vasodilators in dementia (I).97


Tacrine has a moderate effect on cognitive function, but this effect does not seem to translate to differences in activities of daily living scores (I). Tacrine has potentially serious side effects, although in patients who received the drug in clinical trials these did not seem to lead to permanent damage (I). The side effects lead to large number of withdrawals from treatment, and these, coupled with the expense involved in regular hepatic monitoring, seem to mediate against the use of tacrine (I).

Velnacrine maleate is pharmacologically identical to the primary metabolite of tacrine. Similar in efficacy to tacrine, velnacrine also leads to substantial hepatotoxicity and seems, on current evidence, to have no advantage over tacrine.98111

Donepezil hydrochloride

Donepezil has shown a moderate effect on cognitive function in short term treatment trials of patients with mild to moderate Alzheimer's disease (I). These changes in cognitive function have not been accompanied by measured changes in quality of life, and evidence of the effects on activities of daily living is inadequate (I). Whether donepezil is a worthwhile treatment for Alzheimer's disease has not been established by current trials, and longer term randomised trials are required to evaluate its benefits and costs (I).112116

Recommendations and important points

  • Aspirin (75 mg) may reduce the risk of further vascular events for people with early dementia related to cerebral ischaemia, but the size of any effect on cognitive impairment is unclear (B).

  • Stroke is a significant risk factor in the development of vascular dementia (A).

  • Hydergine is not sufficiently effective to be routinely recommended as a treatment for dementia (A).

  • Vasodilators should not be prescribed as a treatment for dementia (A).

  • In the light of current knowledge tacrine should not be used by general practitioners for the treatment of dementia (A).

  • In the light of limited current knowledge, general practitioners should not initiate treatment with donepezil (Aricept) (A).

  • In the light of limited current knowledge, general practitioners should not continue hospital initiated treatment with donepezil (A).


Caring for carers


Carer support groups, though generally perceived as beneficial and helpful, do not reduce the burden or alter the stress of looking after someone with a dementing illness (I). Although information is seen as valuable by carers and best given in a standardised way, it does not alter outcome (I). Respite services offer satisfaction and relief to carers and may delay institutionalisation of the care recipient, but they do not seem to change the overall wellbeing of the carer (I). Formal training for carers may reduce their psychological morbidity and may delay institutionalisation of the dementing person (I). Day care delays institutionalisation by reducing the influence of exhaustion and stress on the carer (II). Day care does not influence cognitive function. The only activity of daily living influenced by day care is dressing skills (II). In one study, intensive community support with a counselling package for a person with dementia and their carer appeared to increase the likelihood of staying at home (II). Depression is common in carers, and is not particularly associated with the relationship of the carer to the patient or to any previous history of depression in the carer (II). Depression in carers is not eased by institutionalisation or death of the care recipient, nor by membership of a support group (II). Low income is associated with depression in carers (II). Depression becomes more likely as care recipients deteriorate, especially if behavioural problems are evident or there are greater care needs, or both (II). Carers experience dissatisfaction with the medical care of their demented relatives in the areas of information received (leaflets, education) and dealing with carer distress (III).117133

Recommendations and important points

  • Referral of carers to groups for the provision of support and information about dementia is valued by the carer, but is unlikely to reduce their burden (A).

  • Referring patients to respite services offers satisfaction and relief to carers, but does not seem to alter the overall wellbeing of the carer (A).

  • Referral of care recipients to respite and day care services may allow them to stay in their own homes for longer (B).

  • Depressive illness is common in carers of people with dementia and is influenced by behavioural problems and higher care needs in the care recipient (B).

  • Healthcare professionals can improve satisfaction for carers by acknowledging and dealing with their distress and providing more information on dementia (C).

Impact of caring on caregivers


The impact of caring for a person with dementia arises from a complex interplay of factors and is related to the risk of institutionalisation (II). Factors in carers that increase “the burden of care” are usually secondary to the caring role and include stress, vulnerability, deterioration in social networking, and economic issues (II). Male and female carers experience the impact of care equally, although men are less likely to discuss it (II). The impact of care continues for carers even after the care recipient is placed in a nursing home and may be greater than in carers continuing to offer care at home (II). Stressors perceived by the carer vary over time as the caring situation and the presentation of the dementia change. The burden of care does not always increase with time (II).134139

Recommendations and important points

  • Healthcare professionals should be aware of the impact of caring for a person with dementia and the effect this has on the caregiver (B).

  • Many factors contribute to the impact of caring, and male carers are less likely to complain spontaneously (B).

  • The impact of caring is dependent not on the severity of the cognitive impairment but on the presentation of the dementia, for example, on factors such as behaviour and affect (B).


We thank Dr Brian Ballinger, Dr Suzanne Hill, John Keady, Dr Iain McIntosh, Professor Ian McKeith, and Dr Sarah Marriott for reviewing the full version of the draft guideline, and Janette Boynton, Julie Glanville, and Susan Mottram for their contribution to the literature search.

Contributors: ME designed and conducted the guideline development and was responsible for joint running of the group and joint writing of the full guideline and summary. JC conducted the guideline development, and was responsible for joint running of the group, joint writing of the full guideline, and commenting on the summary. ML was responsible for synthesis and presentation of papers, joint writing of the full guideline, and commenting on the summary. NF and JM were responsible for data extraction and analysis for tacrine, velnacrine, and donepezil, presenting these to the group, and joint writing of the full guideline. The guideline development group members are given in the appendix. They considered evidence, generated recommendations, and commented on drafts.

The guideline project group comprised Professor Martin Eccles, Dr Julie Clarke, Dr Moira Livingstone. In addition, Mr Nick Freemantle and Dr James Mason participated in the section on antidementia drugs.

The guideline development group comprised the following members, in addition to the authors: Richard Curless, consultant physician, North Tyneside Health Care NHS Trust; Steve Iliffe, reader in primary care, Combined Department of Primary Care, UCLMS and Royal Free Hospital School of Medicine, London; Varun Kaura, general practitioner, Gateshead; Ruth Loughran, senior nurse and coordinator, Elderly Resource Team, Newcastle upon Tyne; Margaret Mace, former practice nurse responsible for elderly assessment; Liz Matthew, directorate manager, Tameside and Glossop Community and Priority Services NHS Trust; Alastair Rigg, general practitioner, Gretna; Brian Roycroft, chair, Alzheimer's Disease Society; John Wattis, consultant psychiatrist, Leeds; Neil Westhead, general practitioner, Whitehaven.

Funding: The development of the guideline was funded by the Health Services Research Panel, Research and Development Directorate, Northern and Yorkshire Regional Health Authority. JC was supported by the Royal College of General Practitioners Alzheimer's Disease Society Educational Fellowship.

Conflict of interest: None.


View Abstract