Residential care for elderly people: the prevalence of cognitive impairment and behavioural problems

Age and Ageing, Nov, 1997 by Carol Jagger, James Lindesay

Introduction

The recent moves towards the care of older people within the community are likely to further increase the growing levels of dependency, both mental and physical, found in residential care [1-4]. Moreover it is the care of the most severely dependent residents which incurs the greatest burden on staff tune and resources [1]. Although a high proportion of elderly people with dementia are cared for within their own homes, the onset of behavioural problems as the disease progresses can create crises which result in subsequent institutionalization [5]. Such problems do not necessarily cease once the patient is institutionalized. Aggressive behaviour in residents may cause distress to both staff and other residents [6]. It can be a serious management problem which requires higher than average staff: resident ratios [7].

There is much research into the nature and type of behavioural disturbance in demented elderly people, both in the community and institutional care. Although studies of the latter have investigated in detail the frequency, cause and type of specific behavioural abnormality with and without the presence of dementia [8-11], they have tended to cover only one type of residential care facility. Local differences in the types of care available and the criteria for admission can give false impressions of the prevalence of disorders and their relationships to severity of dementia, and there is a need for work in this area over a wider range of provision of care.

The aim of this paper is to address this deficit and to describe the prevalence of behavioural disturbance and cognitive impairment in the elderly in residential care from a 1-day census of all types of institutional care within the Leicestershire District Health Authority. We also explore the relationships between behavioural problems and cognitive impairment and the differences in prevalence and patterns between the different types of care provision.

Methods

The study has been described fully elsewhere [4]. The study population comprised all those aged 65 years or over on 27 November 1990 and resident at midnight on that date in any type of ward, hospital, home or hostel provided by the National Health Service (NHS), social services, private or voluntary agencies. Only homes with four or more residents were included as smaller homes did not require registration with the health or local authority.

Care staff at each establishment received a doublesided 21 x 30 cm card of questions which they were to complete for each elderly person resident on the census night. These were delivered to homes in the week preceding the census night and the mode of completion explained to staff. The information collected fell under two broad headings: basic demographic data (date of birth, sex, marital status, date and place of admission) and functioning during the previous week (both mental and physical).

Since the assessment was completed by care staff rather than by personal interviewing of residents, the confusion sub-scale of the modified Crichton Royal Behavioural Rating Scale (CRBRS) was chosen to assess cognitive impairment [12]. The sub-scale consists of three items covering memory, orientation and communication with a total score of 11. Scores of 0 or 1 classify the patient as lucid, 2 and 3 as intermediate, 4-6 as moderately confused and 7-11 as severely confused. Scores of four or more have been shown to be closely associated with a diagnosis of dementia [13]. Behaviour was assessed by a single question with three categories: observes accepted social standards; bouts of unconventional behaviour that do not offend others or fellow residents; and behaviour causes others or fellow residents actual distress or discomfort.

Statistical methods

To investigate age and sex differences in cognitive impairment, logistic regression models were fitted with the dependent variable being cognitive impairment classified by four ordered categories (CRBRS score 0 or 1, 2 or 3, 4-6, 7-11). The analysis was repeated with behaviour (on the three ordered categories stated previously) as the dependent variable.

Similar logistic regression models were fitted to assess whether differences in behavioural problems between types of care were accounted for by levels of cognitive impairment at the patient level. In these models the covariates were age group, sex, cognitive impairment (four categories) and type of care (seven categories). To explore the effect of home size on behavioural problems a further logistic regression model was fitted at the home level with the proportion of patients aged 85 years and over, the proportion who were male, the proportion with moderate or severe confusion (CRBRS score [is greater than or equal to] 24), home size (number of residents) and the type of care as covariates. As size of home was only meaningful in non-NHS settings, the analysis was confined to four care types: private residential, local authority, private nursing, and other.

Results

On the night of the census, 6079 people aged 65 years and over were enumerated in 241 establishments within the Leicestershire District Health Authority, including 21 NHS hospitals, 160 homes for elderly people (46 local authority, 105 private residential and nine run by voluntary agencies), 32 private nursing homes and four homes or hospitals for the mentally or physically handicapped. Only 12 homes, with a maximum of 281 residents, refused to take part in the census, giving response rates of 95.3% for establishments and 95.6% for residents.