Drug use by demented and non-demented elderly people

Transcription

1 Age and Ageing 1997; 26: Drug use by demented and non-demented elderly people PHILUPA WILLS', CECILIA B. CLAESSON 2, LAURA FRATIGUONI' I3, JOHAN FASTBOM 1, MATS THORSLUND 4, BENGT WINBLAD 1 'Stockholm Gerontology Research Centre and Department of Geriatric Medicine, Karolinska Institute, S-I7I 77 Stockholm, Sweden department of Social Medicine, University Hospital, Uppsala, Sweden and National Corporation of Swedish Pharmacies 3 Neurologic Clinic, Department of Neurologic and Psychiatric Sciences, University of Florence, Florence, Italy 4 Department of Social Work, Stockholm University and Stockholm Gerontology Research Centre, Stockholm, Sweden Address correspondence to: J. Fastbom. Fax: (+46) johan.fastbom@kfcmail.hs.sll.se Abstract Aim: to determine the use of drugs by demented and non-demented elderly people in a population, by dementia status and type, age, sex and accommodation type. Method: data were obtained from the Kungsholmen project, a longitudinal community study of people over 75 in Stockholm, Sweden. Results: 85% used at least one medicinal drug, and of these 12% were demented. Mean numbers of drugs used were 2.8 for demented and 3-2 for non-demented people. 45% of demented people and 38% of non-demented people used psychotropic agents. Psychotropic use was higher in women and increased with institutionalization. Antipsychotic agents were used more by demented (22%) than by non-demented (35%) people: this was largely explained by differences in accommodation type. The odds ratio (OR) for use of antipsychotics by those in institutions compared with those living in their own homes was Opioids were commonly prescribed for demented people. The proportions taking opioids in those using analgesics were 42% in demented and 23% in nondemented people (OR 2.07). Laxatives were used by 18% of the demented people in institutions compared with 39% of non-demented people in institutions. Conclusion: being in an institution had a stronger association with the use of certain drugs (e.g. psychotropics) than did dementia status. Demented people, especially those in institutions, used a large number of antipsychotics and opioids, but fewer laxatives and minor analgesics. Prescribers and institutional staff should be aware of these factors so they can optimize patient treatment. Keywords: Alzheimer's disease, dementia, drug utilization, elderly, institutional care, nursing home, opioids, psychotropics Introduction Dementia accounts for much morbidity [1-3] and mortality [4] in old age, and will become increasingly common as the population ages [1, 5, 6]. It has a financial cost, medication being a significant part of this [7]. Some drugs (e.g. antidepressants and some anxiolytics) may be helpful, but the use of drugs in these people may be difficult [8-10], with patients experiencing problems remembering and reporting their medical histories, taking medication and reporting adverse effects [3, 11]. Access to doctors and drugs may be hindered by communication problems and the personality changes that may occur with dementia [3,10]. Pharmacokinetic and pharmacodynamic changes due to age and illness can make treatment more difficult and potentially hazardous [5, 9, 12]. Polypharmacy can increase the incidence of adverse drug effects [13, 141. There are few studies on medicine use in demented elderly people [2, 3, 8, 15-19] and none is comprehensive. Some are limited to those attending hospital clinics [8, 15, 19] or living in nursing homes [16, 17]; others make no comparison between demented and non-demented people [8] or limit their analysis to drug groups, such as psychotropics [2, 3, 18] or cardiovascular drugs [16]. We have used data from the Kungsholmen project, a population study of people 75 years and over in a district of Stockholm, which focuses on the medical, 383

2 P. Wills et al. psychological and social problems of ageing, with the emphasis on dementia [20]. The aim of this paper is to describe the use of drugs in an elderly population with respect to dementia status while considering age, sex and housing type. Methods The study population consisted of all the inhabitants of Stockholm's Kungsholmen district on 1 October 1987, who were born in or before 1912 (n = 2368). In phase I, most of these residents had a general health interview. Exceptions were those who dropped out due to death (181), moved out of district (69) or refused to participate (308), so that 1810 people participated in the first phase. Those who died before they could be contacted by the project staff were older and there was a greater proportion of men in that group than in the participant group. The other withdrawals did not differ significantly from the study participants with respect to age or sex. The age range was from 75 to 101 years, with a mean of 81.7 years. Those who did participate were interviewed by a nurse following a structured protocol which included health, medicinal drug use, social and general welfare questions. The nurse also performed a brief cognitive test [21] and took blood for analysis. In phase n of the study, those who had had a Mini- Mental State Examination (MMSE) score <23 (n = 385) in phase I were clinically examined by a geriatrician and reviewed by a neurologist, to see if they had dementia according to modified Diagnostic and Statistical Manual (DSM) m-r criteria [1]. Withdrawals due to death, moving away or refusal reduced the number examined to 314. Of those, 202 cases were diagnosed as having dementia and 10 questionable dementia. A similar number of age- and sex-matched subjects (n = 354) were taken from the group of people with an MMSE ^24 and were also fully examined for dementia. Of this group, nine were discovered to be demented and are thus included in the demented group. In addition, four people were diagnosed as having questionable dementia. The phase n withdrawals were more frequently men and were younger than the examined group. These withdrawals brought the total number included to The average time interval between the two data collection phases was 2.3 ± 3 months. The category of questionable dementia was added for those people who satisfied all but one of the DSMm-R criteria (n = 14). Dementia severity was also assessed using the Clinical Dementia Rating scale [22]. Those who were diagnosed as having confirmed dementia are included in the demented group. All other participants, except those with questionable dementia, are in the non-demented group, whether they were examined in phase II or not. Because of this, there may be a small number of false negatives in this group (i.e. labelled non-demented but actually demented). However, this is likely to be too small a number to significantly affect the results. In the present report 'dementia status' is used to describe those who either have confirmed dementia or are not demented: data from those with questionable dementia have been excluded from these analyses. For completeness, they are included in the discussions on dementia type. Where 'dementia type' is discussed, we have grouped the dementias into Alzheimer's disease, vascular dementia and 'other' dementias, which include alcoholic dementia (13 cases), mixed dementia (three cases), dementia in Parkinson's disease (seven cases), dementia associated with a subdural haematoma (one case) and unspecified types of dementia (14 cases). In discussions of drug use we have excluded the nine people who did not have their drug information recorded. The information on drug use was gathered from the phase I data where participants were asked about both prescription and non-prescription drug use and drug containers were inspected to confirm the information. If a person could not remember their drugs or was suspected of having memory problems, a close relative or friend was asked. This person may not have always been the primary caregiver. If medication was administered by health care personnel they were interviewed and medical records consulted. If a drug name or dose could not be identified this was noted. A drug was considered to be 'used' if it was being taken at the time. An 'as required' drug (i.e. taken intermittently) was considered to be used if it had been used in the previous 2 weeks. The drug data were classified according to the Anatomical Therapeutic Classification system as recommended by the World Health Organisation [23]. This is a five level coding system based on anatomical, therapeutic, pharmacological and chemical groups. The classification, coding and entering of data into a computer was done by an experienced pharmacist. Drug tables in this paper only include drugs used by 50 people or more. Because the use of pain relief medications may be related to dementia status [24], we looked at the use of opioids relative to the use of all pain relieving drugs and called this the 'opioid proportion'. It was calculated as: No. using opioids/[no. using opioids, minor analgesics, non-steroidal anti-inflammatory drugs (NSAIDs) or muscle relaxants] x 100 There was further analysis of psychotropic drugs (neuroleptics, hypnotics, sedatives and antidepressants). Housing was grouped into three types: own homes (either owned or rented), sheltered accommodation (individual apartments with access to communal 384

3 Drug use in dementia facilities and a professional, but not medically skilled, caregiver) and institutional care (nursing homes, geriatric institutions and two individuals who were in temporary' hospital care for an extended period of time). Statistical analysis was performed using advanced computer programs [25, 26]. The \ 2 test was used for comparing proportions in different groups. When proportions were small, Fisher's exact test was used. Logistic regression was used for examining the effect of dementia status, age, sex and accommodation type on the number of drugs used, where each variable was analysed taking the others into account. Logistic regression was also used for looking at drug use by dementia status, taking age, sex and accommodation type into the model. This analysis was then repeated using dementia type rather than dementia status. The odds ratios (ORs) for age indicate the average OR of using a drug with an increase in age of 1 year, compared with those aged 75- The ORs for accommodation types are compared with those living in their own home. Logistic regression was also used for selected drugs to see if severity of dementia was significantly associated -with drug use. This model included only those with dementia and took into account dementia severity (from the Clinical Dementia Rating scale), age, sex and accommodation type. Finally, logistic regression was used (selecting just those who used the analgesic drugs listed above) to analyse die association of opioid use with dementia status (then dementia type), age, sex and accommodation type. Results The distribution of study participants by dementia status and type and age, sex and housing type are given in Table 1. Number of drugs used by dementia status and type It appears that 2.4% of demented people (or their proxies) did not know whether they were using drugs or what types they used, compared widi 0.2% of nondemented people (P < 0.001). Eighty-five percent of demented and non-demented people used at least one drug, while 34% of demented people and 40% of nondemented people used four or more drugs. Table 2 gives the mean number of drugs by dementia status and Table 3 the results of the logistic regression. The model shows the effect on the number of drugs used of having dementia, age, sex and accommodation. For each variable the analysis controls for the other variables in the model. Age indicates the average change with increase in age of 1 year, and accommodation types are compared with those living at home. The analysis showed that demented people used fewer drugs on average than those who were not demented. Average drug use increased with increasing age and was higher in women. Drugs used by dementia status and type Use of drugs by dementia status is given in Table 4, with the ORs for use and levels of significance in Table 5. Statistical data are controlled for age, gender and accommodation type. A smaller proportion of demented people used hypnotics and sedatives, nitrates and potassium. Psychotropic drugs and vitamin B12 and folate were used by a larger proportion of demented than nondemented people (Table 4); this was associated with accommodation type, age and sex, rather than dementia (Table 5). The individual variables in this study are closely connected (e.g. people in institutions are more likely to be demented). This is reflected in the population demographics (Table 1). Because of this, multivariate analyses have been used to study the effect of these variables (Tables 3 and 5). This revealed that accommodation type played a large part in the use of many drugs in demented and non-demented elderly people (as could be seen in the case of antipsychotics and all psychotropics). The OR for antipsychotic and opioid use in sheltered accommodation "was three-fold. Antipsychotic use in institutions had the greatest OR (932). Likewise, hypnotics and sedatives, anxiolytics and all psychotropic drugs were used more in institutions. For these drugs, a second regression analysis was done for just the demented cases with dementia severity included in the model (along with age, sex and accommodation type). Dementia severity was not significantly associated with the use of these drugs and so is not included in the Table. Certain cardiac drugs cardiac glycosides, agents acting on arteriolar smooth muscle and thiazides were used significandy less often in institutions (even after controlling for age, sex and dementia status). Gender and age were also important in determining the use of drugs. For many drug classes the proportion of users was larger in women than in men. For some cardiovascular drugs, laxatives, vitamin B12 and folate and antiglaucoma drugs, use increased -with increasing age. There was decreased use with age for /3-blockers and antipsychotics. In institutions, fewer demented than non-demented people used laxatives (18% versus 39%, logistic regression controlling for age and sex P = 0.030), hypnotics and sedatives (27% versus 54%, P 0.017) and minor analgesics (14% versus 36%, P = 0.006). However opioid use was much die same (14% versus 12%). People widi vascular dementia were significantly more likely to use cardiac glycosides than nondemented people (OR 2.49, P < 0.01). People with 385

4 P. Wills et al. Table I. Study participant demographics Group Gender Male Female Age group (years) Accommodation type Own home Sheltered accommodation Institutional care Percentage of group, by dementia status Demented Total («= 211) Questionable dementia (n 14) Non-demented (n = 1475) Alzheimer's disease used fewer cardiac vasodilators (OR 0.41 P < 0.05) than did non-demented people. Antipsychotics were used by significantly more people with Alzheimer's disease than by nondemented people (OR 3.68, P < 0.0O1), as were opioids (OR 2.09, P = 0.018) and vitamin B12 and folate (OR 2.07, P = 0.044). Antidepressants were used by 1.1% of the nondemented population and 3-3% of the demented group. Opioid use the 'opioid proportion* Many of the people using opioids were using dextropropoxyphene, either alone or in combination with a minor analgesic such as paracetamol. The data show that demented people had a higher opioid proportion than non-demented people (41.9% versus 23.2%, P = 0.001). Logistic regression analysis using dementia status showed that those with dementia had a higher opioid proportion than those who were non-demented (OR 2.07, P ). When broken down by dementia type, this was true only for those with Alzheimer's disease (OR 3-55, P 0.002). There was also a large difference by accommodation type, with those in sheltered accommodation having the greatest risk of a high opioid proportion (OR 352, P 0.001). Discussion This paper is based on a large population-based study of very old people, with well-defined dementia diagnoses and systematically categorized drug data. It is the first study to look at the use of both prescription and non-prescription drugs in elderly people with respect to dementia presence and type. We found important differences between demented and nondemented people in the number and type of drugs used. Accommodation type played a greater role than expected in influencing associations with drug use. While more demented than non-demented people did not know if they were using drugs or what types they used, the percentages were small. If the demented person was living at home, a close relative or friend (usually the primary caregiver) was asked instead. As this person may not have known what drugs the demented person was taking, there may be an underestimation in the number of drugs used by demented people living in their own home. Underestimation in numbers of people with dementia may have occurred from use of the MMSE as the screening Table 2. Mean number of drugs used (± standard deviation) Gender Male Female Age group (years) Accommodation type Own home Sheltered accommodation Institutional care Total Demented 2.8 ± ± ± ± ± ± ± ± ± ±2.3 Non-demented 2.6 ± ± ± ± ± ± ± ± ± ±

5 Drug use in dementia Table 3. Odds ratios for using many (^3) drugs compared with few (0-2) drugs Having dementia Older age Being female living in sheltered accommodation Living in institutional care Odds ratio P-value* * Representing level of significance by logistic regression, for each variable, controlling for the others in the model. test for dementia: this resulted in an estimated 39 false negative cases being included in the non-demented group [21]. Demented people in all age groups used a smaller mean number of drugs than those who were not demented. A similar result was found in the US by Wolf-Klein et al., who concluded that people with Alzheimer's disease were healthier than age-matched controls [15]. Unfortunately, we do not have data on the prevalence of diseases other than dementia in our population, so we cannot correlate drug use with disease prevalence. However, a second American study noted that even when demented and non-demented groups had the same prevalence of cardiovascular disease, those with dementia received fewer cardiovascular drugs [16]. Another reason for the lower use of drugs by Table 4. Drug use as a percentage by dementia status; cases with questionable dementia (n = 14) are not included Drug group Hypnotics and sedatives 21.8 Cardiac glycosides 18.4 Loop diuretics 13.6 Minor analgesics and antipyretics 15.5 Potassium 9.2 Anxiolytics 15.0 Laxatives 17.5 Cardiac vasodilators (nitrates) 6.3 Agents acting on arteriolar smooth muscle 4.9 Thiazides 4.9 Opioids 12.6 Non-steroidal anti-inflammatory drugs 2.9 j3-blocking agents 1.5 Antipsychotics 22.3 Diuretics and potassium-sparing agents in combination 53 Potassium-sparing agents 39 Thyroid preparations 58 Vitamin B12 and folic acid 12.1 Antiglaucoma preparations and miotics 39 Muscle relaxants, centrally acting agents 1.5 Expectorants, excluding combination with antitussives 34 Antithrombotic agents 1.0 Other systemic anti-asthmatics 4.9 Systemic antihistamines 1.0 Iron preparations 4.4 Systemic corticosteroids 1.0 Oral antidiabetics 1.9 Multivitamins, combinations 0.5 All psychotropic drugs 451 Percentage of total group Demented (n = 206) Non-demented (n = 1472)

6 P. Wills et al. Table 5. Odds ratios for drug use Accommodation type* Drug group Dementia Older age Female sex Sheltered Institutional Hypnotics and sedatives Cardiac glycosides Loop diuretics Minor analgesics and antipyretics Potassium Anxiolytics Laxatives Cardiac vasodilators (nitrates) Agents acting on arteriolar smooth muscle Thiazides Opioids Non-steroidal anti-inflammatory drugs /S-blocking agents Antipsychotics Diuretics and K + sparers in combination K* sparing agents Thyroid preparations Vitamin B12 and folic acid Antiglaucoma preparations and miotics Muscle relaxants, centrally acting Expectorants, excluding combination with antitussives Antithrombotic agents Other systemic anti-asthmatics Systemic antihistamines Iron preparations Systemic corticosteroids Oral antidiabetic agents Multivitamins, combinations Psychotropic drugs 0.43"* " * " *" 1.08*" 1.03* 1.08*" *" " 0.94* *" 1.05* *" * 1.43* 1.90*" " " *" * " 1.87"* *" * * * 2.33" 0.32*" * 1.83* * 0.32* *" * *" * Compared with own home. *P < 0.05; "P < 0.01; ""P < demented people could be difficulties in remembering or communicating their symptoms to a caregiver or doctor [8, 11, 24]. Individuals with dementia may also be unwilling to admit symptoms or to have investigations [8]. They, their caregivers or indeed their doctors may also misinterpret symptoms. Moreover, there could be a preference by doctors, patients or relatives to treat some symptoms/diseases non-pharmacologically in demented people [8]. For both demented and non-demented groups, those in institutions used more drugs than those in their own homes. When dementia status, accommodation type, age and gender were taken into account in the regression analysis, only the last three factors were statistically significantly and independently associated with an increased use of drugs. The association between accommodation type and drug use has been described in a previous paper from the Kungsholmen project [27]. A probable explanation is easier access to medical care for those in sheltered accommodation or institutions, as well as a possibly healthier status of those living at home. In sheltered accommodation, the supervising caregiver may initiate medical contact for the patient if they feel it is necessary. In institutional care, there are trained nurses supervising patient care and doctors visit the institution. The regression analysis showed that advanced age and female gender were also associated with an increased use of drugs, independent of dementia status and living place. The high use of drugs in women has been the subject of discussion. Possible reasons include higher attendance rate at doctors, cultural acceptance of the sick role and different prescribing patterns to women [28, 29]. 388

7 Drug use in dementia When age, sex and accommodation were taken into account, demented people used potassium and cardiac vasodilators (nitrates) less frequently than non-demented people. The low rate of use of some cardiovascular agents in demented people has been noted in other studies [15,16] and may be due to many factors, including survivor-selection and difficulties in the diagnosis and treatment of people with dementia. The number of both demented and non-demented people using psychotropics was very high, especially in institutions. The frequent use of these drugs in elderly subjects is of concern as these drugs can cause or exacerbate confusion [30]. Those with dementia are more susceptible to these effects. The use of antipsychotics in those with Alzheimer's disease was higher in our study than in a similar US study where only 15.9% of those with Alzheimer's took antipsychotics [8]. This may be due to the concerns at the time in the US about antipsychotic drug prescribing and the implementation of regulations to limit psychotropic drug use in nursing homes [31]. In Sweden there have been no such regulations, although a workshop on psychotropic use was conducted in 1989 [6, 32]. Our results support the concern about the high levels of psychotropic use in institutions. Our findings pose some important questions for those who work in institutions. Why are these drugs so commonly used in this setting? Many reasons have been proposed, including institutional characteristics (e.g. high patient to staff ratio, small facility size) [35], staff expectations (e.g. need to 'do something') [28], historical and socio-cultural reasons [28] and patient characteristics (e.g. psychiatric illness, sleep disturbance, behavioural symptoms) [2, 28, 35]. Of these, behavioural symptoms are the most commonly cited reasons for prescribing psychotropic drugs [2, 35]. However, psychotropic treatment of behavioural disturbances in dementia is often not effective and controlled trials of these drugs in demented people are lacking [6, 33]. There has been some success in reducing the use of these drugs by educating staff about the drugs and/or training on how to use behavioural interventions instead of drugs [34, 35]. To assess whether the use of these drugs was related to dementia severity rather than accommodation type we also took severity of dementia into account in the statistical analysis. However it did not significantly affect the results. Thus, the effect of accommodation type cannot be explained by the severity of dementia of their residents. Drugs with an impact on the quality of life include analgesics and laxatives. From drug statistics we know that in 1987 about 90% of muscle relaxants, 75% of NSAEDs and 95% of analgesics, for all age groups in Sweden were prescribed for pain [36]. As we do not have drug indications we cannot be sure that these were prescribed to our study participants for pain. People with dementia, in particular those with Alzheimer's disease, had a higher use of opioids than the non-demented, while the use of other analgesics, including minor analgesics, NSAIDs and muscle relaxants was low in the demented group. Demented people as a group and those with Alzheimer's dementia in particular, had a higher opioid proportion than non-demented people. However, logistic regression showed that accommodation type was also significantly related to opioid use. This may be related to demented people being unable to communicate their discomfort to staff, which may lead to inadequate care [37]. People in their own homes are more likely to be cared for by a family member or friend who can understand what they say and be more aware of deviations from normal behaviour. The high use of opioids is worrying, especially when one considers that opioids can both cause and exacerbate confusion. Their use in elderly people, especially the demented elderly, should be limited [38]. The high use of opioids in sheltered accommodation might also be explained by a higher incidence of painful disorders in elderly people living in this type of housing. However, in another study of the Kungsholmen project no significant differences were found in the prevalence of pain between different types of housing [39], which would argue against this hypothesis. Another unexpected finding was of a low use of laxatives by demented people in institutional care compared with non-demented people. This may be due to non-demented people in institutions being less active, or demented people using fewer cardiovascular drugs such as diuretics, which may give constipation. However, it may also be related to communication problems. The differences in use of laxatives and minor analgesics, i.e. 'quality of life' drugs, should be of concern to health care workers. Vitamin B12 andfolatewere used more by Alzheimer's demented than non-demented people and this use increased with age. It has been debated whether people with Alzheimer's have lower B12 levels [40,41]. In recent work on the Kungsholmen population who were not on B12 treatment, those with Alzheimer's had the same or higher B12 levels than the non-demented population [40]. It has been suggested that B12 levels should be checked periodically and the decision to treat based on these results [40]. Conclusion Both the number and type of drugs used differed between demented and non-demented people. However, living in institutions or sheltered accommodation was more highly associated with the use of certain drugs e.g. psychotropics, than was dementia status. Demented people, especially those in institutions or sheltered accommodation, used a large number of antipsychotic drugs and a higher relative amount of opioids. The effect of accommodation type may reflect 389

8 P. Wills et al. increased access to medical care as well as changing health status. The relatively low use of laxatives and minor analgesics by demented people may be due to communication problems. Prescribers and institutional staff should be aware of these factors so that they can optimize patient treatment. Acknowledgements The authors thank all those involved with the Kungsholmen project for making this paper possible. Special thanks to Ingemar Kareholt for help with the statistics and Mike Lloyd for editorial assistance. This study was supported by grants from The National Corporation of Swedish Pharmacies' Fund for Research and Studies in Health Economics and Social Pharmaceutics, The Swedish Council for Social Research, The Einar Belven Foundation and The Municipal Pension Institute. Key points In people over 75 in a Swedish community, drug use was greater in women, the very old and those living in institutions. The use of psychotropic agents was high, especially in institutions. Demented subjects were prescribed fewer drugs overall, but took more opioids and fewer laxatives and minor analgesics than non-demented people. Cardiovascular drugs were used less often by people in institutional care. References 1. Fratlglionl L, Grut M, Forsell Y et al Prevalence of Alzheimer's disease and other dementias in an elderly urban population: relationship with age, sex and education. Neurology 1991; 41: Sandman P-O, Adolfsson R, Norbcrg A et al Long-term care of the elderly. A descriptive study of 3600 institutionalized patients in the county of Vasterbotten, Sweden. Compr Gerontol [A] 1988; 2: Skoog I, Nilsson L, Landahl S et al Mental disorders and the use of psychotropic drugs in an 85-year-old urban population. Int Psychogeriatr 1993; 5: Katzman R, Hill R, Yu ESH et al The malignancy of dementia. Predictors of mortality in clinically diagnosed dementia in a population survey of Shanghai, China. Arch Neurol 1994; 51: Ames D, Tuckwell V Psychopharmacology in the 1990s. What does the future hold for the aged patient' Drugs Aging 1991; 1: Olanders S. Neuroleptics in geriatrics. In: Strandberg K, Beermann B, Lonnerholm G, eds. Workshop: treatment with neuroleptics. Uppsala, Sweden: National Board of Health and Welfare Drug Information Committee, 1990; 1: Buckholtz NS. Alzheimer's disease drug development and testing at the National Institute on Aging. Psychopharmacol Bull 1994; 30: Semla TP Cohen D, Paveza G et al Drug use patterns of persons with Alzheimer's disease and related disorders living in the community. J Am Geriatr Soc 1993; 41: Landahl S. Drug treatment in year-old persons. A longitudinal study. Acta Mcd Scand 1987; 221: Peisah C, Brodaty H. Practical guidelines for the treatment of behavioural complications of dementia. Med J Aust 1994; 161: Frank EM. Effect of Alzheimer's disease on communication function. J SC Med Assoc 1994; 90: Aronow WS. Digoxin or angiotensin converting enzyme inhibitors for congestive heart failure in geriatric patients. Which is the preferred treatment? Drugs Aging 1991; 1: Nolan L, O'MalleyK_ Prescribing for the elderly. Part 1: sensitivity of the elderly to adverse drug reactions. J Am Geriatr Soc 1988; 36: Colt HG, Shapiro AP. Drug-induced illness as a cause for admission to a community hospital. J Am Geriatr Soc 1989; 37: Wolf-Klein GP, Silverstone FA, Brod MS et al Are Alzheimer's patients healthier? J Am Geriatr Soc 1988; 36: Renner EA, Engle VF, Graney MJ. Prevalence and predictors of regularly scheduled prescription medications of newly admitted nursing home residents. J Am Geriatr Soc 1992; 40: Andersson M. Drugs prescribed for elderly patients in nursing homes or under medical home care. Compr Gerontol 1989; 3 (suppls A+B): Travis SS, Moore SR. Nursing and medical care of primary dementia patients in a community hospital setting. Appl Nurs Res 1991; 4: Lacro JP, Jeste DV. Physical comorbidity and potypharmacy in older psychiatric patients. Biol Psychiatry 1994; 36: Fratiglioni L, Viitanen M, Backman L et al. Occurrence of dementia in advanced age: the study design of the Kungsholmen project. Neuroepidemiology 1992; 11 (suppl. 1): Grut M, Fratiglioni L, Viitanen M et al Accuracy of the Mini-Mental Status Examination as a screening test for dementia in a Swedish elderly population. Acta Neurol Scand 1993; 87: Forsell Y, Fratiglioni L, Grut M et al Clinical staging of dementia in a population survey: comparison of DSM-m-R and the Washington University Clinical Dementia Rating Scale. Acta Psychiatr Scand 1992; 86: World Health Organisation. Guidelines for ATC (Anatomical Therapeutic Classification). Oslo: World Health 390

9 Drug use in dementia Organisation Collaborating Centre for Drug Statistics Methodology, Robinson D, Bucci J. Fenn H. Pain assessment in the Alzheimer's patient. J Am Geriatr Soc 1995; 43: SPSS for windows. Release 6.0. Chicago, IL: SPSS Inc., Stata Release 3.1. College Station, TX: Stata Corp., Thorslund Met al Drug use in an elderly population the role of the organizational environment. Submitted. 28. Hansen EH. How widely do women and men differ in their use of psychotropic drugs? J Soc Adm Pharm 1989; 6: van der Waals FW, Mohrs J, Foets M. Sex differences among recipients of benzodiazepines in Dutch general practice. BrMedJ 1993; 307: Fastbom J, Claesson CB, Cornelius C et al. The use of medicines with anticholinergic effects in older people: a population study in an urban area of Sweden. J Am Geriatr Soc 1995; 43: Semla TP, Palla K, Poddig B et al Effect of the Omnibus Reconciliation Act 1987 on antlpsychotic prescribing in nursing home residents. J Am Geriatr Soc 1994; 42: Sandman P-O, Adolfsson R, Eriksson S. Prescription of psychophannaceutical drugs to the elderly in institutions in Sweden. In: Strandberg K, Beermann B, Lonnerholm G, eds. Workshop: treatment with neuroleptics. Uppsala, Sweden: National Board of Health and Welfare Drug Information Committee, 1990; 1: Yeager BF, Farnett LE, Ruzicka SA. Management of the Received 28 June 1996 behavioral manifestations of dementia. Arch Intern Med 1995; 155: Ray WA, Taylor JA, Meador KG et al Reducing antipsychotic drug use in nursing homes. Arch Intern Med 1993; 153: Sloane PD, Mathew IJ, Scarborough Metal Physical and pharmacologic restraint of nursing home patients with dementia. Impact of specialized units. JAMA 1991; 265: Nordenstam I, Wennberg M, Kristoferson K. Svensk Lakemedels-Statistik Stockholm: Apoteksbolaget, 1994; Ekman S-L, Robins Wahlin T-B, VUtanen M et al Preconditions for communication in the care of bilingual demented persons, hit Psychogeriatr 1994; 6: Schor JD, Levkoff SE, Iipsitz LA et al Risk factors for delirium in hospitalized elderly- JAMA 1992; 267: Cornelius C, Fastbom J, Claesson CB et al Self-reported symptoms in the elderly. Prevalence and association with drug use. Clln Drug Inv 1997; 13: Basun H, Fratiglioni L, Winblad B. Cobalamin levels are not reduced in Alzheimer's disease: Results from a population study. J Am Geriatr Soc 1994; 42: Crystal HA, Ortof E, Frishman WH et al Serum vitamin B] 2 levels and incidence of dementia in a healthy elderly population: a report from the Bronx longitudinal aging study. J Am Geriatr Soc 1994; 42:

10 Mrs Ann Scarth in the doorway of her home, Rock Head Cottage, Glaisdale, near Whitby. Photograph: Frank Meadow Sutcliffe ( ). The Sutcliffe Gallery, Whitby. Tel: (+44)