Psychiatric status, somatisation, and health care utilization of frequent attenders at the emergency department: A comparison with routine attenders

Journal of Psychosomatic Research 50 (2001) 161 ± 167

Psychiatric status, somatisation, and health care utilization of frequent attenders at the emergency department A comparison with routine attenders Edwina R.L. Williamsa, Elspeth Guthriea,*, Kevin Mackway-Jonesa, Marilyn Jamesb, Barbara Tomensona, Joe Easthamb, Deborah McNallya a

Department of Psychiatry, Manchester Royal Infirmary, University of Manchester, Rawnsley Building, Oxford Road, Manchester M13 9BX, UK b The Centre for Health Economics, University of Keele Received 11 June 1999; accepted 5 April 2000

Abstract Seventy-seven frequent attenders at an emergency department (ED) in an inner-city hospital in the UK (defined as seven or more visits in the previous 12 months) were compared with 182 patients who were attending the same department on a routine basis. Patients completed the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) and the Short Form (SF)-36. Information was obtained on 64% of the frequent attenders and 45% underwent a detailed psychiatric assessment. Of the frequent attenders, 45% had psychiatric disorder and 49% had some form of an alcoholrelated disorder. Compared with routine attenders, frequent attenders reported lower health status, had more psychiatric

disorder (odds ratio: OR = 8.2, 95% confidence interval: CI = 3.8 ± 18.1), had more general hospital admissions (OR = 19.9, 95% CI = 8.3 ± 47.8), more psychiatric admissions (OR = 167.5, 95% CI = 9.5 ± 2959.0), and more GP visits (95% CI for difference = ÿ 10.2 to ÿ 5.7). There was no evidence that frequent attenders had more somatisation than routine attenders. Specific treatment and management strategies need to be developed for this group of patients, although a substantial proportion may be difficult to engage in the treatment process. D 2001 Elsevier Science Inc. All rights reserved.

Keywords: Frequent attenders; Emergency department; Psychiatric status; Resource utilisation

Introduction Since 1981, emergency departments (EDs) have experienced a progressive rise in the overall number of new attendances by approximately 2% per year [1]. Several descriptive reports have suggested that a relatively small group of patients account for a large proportion of the overall total of ED attendances [2 ± 4], as they make repeated unscheduled visits. These patients have been characterised as often being vagrants [5], or having high rates of alcohol problems or social difficulties [3,4]. It has also been suggested that they may use EDs in preference to seeking care from the primary sector. A recent study from Ireland [6] has shown that increasing frequency of atten* Corresponding author. Tel.: +44-161-276-5383; fax: +44-161-2732135. E-mail address: [email protected] (E. Guthrie).

dance at an ED was significantly associated with increasing age, being male, and not being married. There have been no studies, however, which have examined the psychiatric profiles and health status of frequent attenders at EDs. The phenomenon of frequent attendance has been studied most in the primary care setting. Several studies from around the world [6 ±14] have shown that frequent attenders in primary care have high rates of psychiatric disorder in comparison with normal attenders, in particular, they have high rates of anxiety, depression, and somatisation. A recent study of frequent attendance at a medical out-patient clinic [15] has shown that frequent attenders in the secondary care sector are also high users of primary care services. The main goal of this study was to determine the health status and prevalence of psychiatric disorder, including somatisation, of patients who attend the ED on a frequent basis in comparison with routine attenders. A secondary goal was to establish whether ED frequent attenders use

0022-3999/01/$ ± see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 0 2 2 - 3 9 9 9 ( 0 0 ) 0 0 2 2 8 - 2

162

E.R.L. Williams et al. / Journal of Psychosomatic Research 50 (2001) 161±167

emergency services instead of primary care, or as an addition to primary care. Patients and method The study was conducted at Manchester Royal Infirmary. The ED serves a deprived inner-city area and treated approximately 67,500 new patients in 1997. It has a computerised register that records every attendance and enables frequent attenders to be easily identified. Frequent attendance at the ED has been defined as four or more visits within a 12-month period [3]. For the purposes of this study, we chose a more conservative figure of seven or more visits as we wished to focus upon patients who had a very high rate of attendance. During the study period (August 1995 ±January 1996), all consecutive patients who had attended the ED seven or more times within the previous year were approached to take part in the project. Seventy-seven patients were identified as frequent attenders. For the control group, every tenth consecutive, routine, attender at the department, during specific periods of time, was approached and asked to complete a detailed psychiatric assessment and health economics questionnaire. The day and night was divided into six time periods; 8 a.m. ±12 noon, 12 noon ±4 p.m., 4 ± 8 p.m., 8 p.m. ± midnight, midnight ±4 a.m., and 4 ±8 a.m. Over the 3-month period, sampling occurred during different time periods, so that a representative sample of routine attenders was recruited for comparison with the frequent attenders. Approximately two 4-hour sessions were completed per day. Controls had to fulfil the following criteria: (a) aged between 16 and 65 years, (b) able to speak and write English, (c) be physically well enough to be interviewed, (d) not be a frequent attender. From previous research, the estimated psychiatric morbidity in routine attenders at the ED is approximately 20% [16]. We estimated that approximately half of the frequent attenders might have psychiatric disorder. We calculated that a sample size of 70 per group would allow us to test the main hypothesis at the .02 level of significance with a power of 90%. We determined to recruit a larger number of patients to the control group (approximately 180 ± 200), so that a subanalysis comparing patients with psychiatric disorder, in both routine and frequent attenders groups (estimated at approximately 30 ± 40 per group), could be undertaken. Main assessments Health status was measured using the Short Form (SF)36 [17]. The SF-36 is a shortened version of a battery of 149

health status questions developed and tested on a population of over 22,000 patients as part of the medical outcome study [18]. It is a profile measure that can be used to measure health status in a wide variety of different medical conditions. It has eight separate domains: physical function, role limitation due to physical illness, role limitation due to emotional problems, social functioning, general mental health, energy, bodily pain, and general health perception. A score of 100 on each scale indicates maximum health status for that particular domain. Low scores indicate poor health status. Psychiatric assessment was made using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) [19]. The SCAN is a detailed semistructured psychiatric interview that has to be administered by an experienced clinician who has been trained in its use at a designated centre. Diagnoses of mental disorders are made by computer algorithm according to criteria described in the International Classification of Diseases [20]. Patients were only considered to have psychiatric disorder if they reached an Index of Definition (ID) of 5 or above. They were grouped into the following categories; F06 = organic disorders; F20 = schizophrenia; F30± F39 = depressive/affective disorders; F40 ± F49 anxiety disorders/other neurotic disorders. Alcohol-related diagnoses do not reach an ID level of 5 on the SCAN. These were, therefore, recorded separately. Patients were also asked to complete a brief questionnaire concerning their physical status. This included details of any symptom complaints and current diseases or conditions. Health care utilisation was assessed using a detailed interview. Patients were asked about the frequency of their contact with (a) local emergency services in Manchester, (b) primary care services, (c) psychiatric services, and (d) general hospital services. One of the researchers (DM) was employed to specifically check and verify all data concerning GP and hospital attendance. The study received ethical approval from Central Manchester Ethics Committee. All patients who participated in the study provided written informed consent. Statistical analysis Statistical analysis was carried out using the Statistical Package for Social Sciences (SPSS) [21]. Data are expressed as mean values and standard error of the mean (S.E.M.). Normally distributed data was compared using the Student's t test. Results are given in the form of 95% confidence intervals (CIs) for difference. Categorical data was compared using the chi-square test and results are presented as odds ratios (ORs) and CIs. The distribution of the resource use data was severely skewed so nonparametric techniques were applied. The Mann ±Whitney U test was used for continuous data and results are expressed in the form of median and interquartile range (IQR). The demographic details of frequent attenders who

E.R.L. Williams et al. / Journal of Psychosomatic Research 50 (2001) 161±167

declined to take part in the study, or who could not be traced, were compared with those who agreed to participate, to determine whether the participants were representative of the whole group.

Results Of the 77 patients who were identified as frequent attenders, 23 (29.9%) could not be traced (i.e. were of no fixed abode or gave a false address), 13 (16.9%) refused to be interviewed, 3 (3.9%) had died, 2 (2.6%) had learning difficulties, and 1 man was deaf ± mute. Several patients on whom assessments were completed required up to four home visits before face-to-face contact could be established. A total of 35 (45.5%) frequent attenders were finally interviewed and data from psychiatric case records was also available on a further 14 patients, who could not be interviewed. Thus, data was available on 49 (63.6%) of the frequent attenders. For the control group, 528 patients presented during the study sampling period. A total of 111 patients were below the age of 16 and 34 patients were above 65 years of age. A total of 88 patients were too physically ill to be interviewed, 48 left the ED department before assessment, 10 could not speak English, and 2 had learning difficulties. Out of a possible 235 patients who met the study criteria, 22 (9.4%) could not be interviewed as they were too intoxicated with alcohol, 2 (0.9%) were considered too dangerous, and 29 (12.3%) patients refused to take part in the study. A total of 182 patients were assessed accounting for approximately 77.4% of those patients eligible for the study. The main analyses involve a comparison of the 35 frequent attenders who were interviewed compared with the 182 controls. Some questionnaire responses were incomplete so the denominators for each group vary slightly. A small amount of supplementary data (e.g., demographic

163

data) was available on all the 77 frequent attenders and this is also presented in the results. The frequent attenders who could not be interviewed were similar to those interviewed in terms of age ( P = .34) and gender (interviewed = 17 males (48.6%) vs. noninterviewed = 26 males (61.9%) (x2 = 0.89, df = 1, P = .34). It was more difficult to assess the prevalence of psychiatric disorder in the noninterviewed group. A total of 21 (60%) of those interviewed had psychiatric disorder and at least 14 (33.3%) of those not interviewed had psychiatric notes (x2 = 4.45, df = 1, P = .035). Demographic data There was a predominance of males in both groups. A total of 123 (67.6%) of the routine attenders and 43 out of 77 (55.8%) of the frequent attenders were male (x2 = 2.8, df = 1, P = .09). The frequent attenders (mean age 36.6 + 2.3 years) were significantly older than the routine attenders (mean age 30.4 + 0.8 years) (t = 2.5, P = .02). Because of the difference in gender and age between the two groups, all comparisons were also analysed using an analysis of covariance with gender and age as covariates. Both groups had a high proportion of patients who were single (routine attenders, 58.2%; frequent attenders, 42.9%) and only 35% of the routine attenders and 17.1% of the frequent attenders were in a stable relationship. Health status Table 1 shows the mean scores on the SF-36 subdomains for each group. A small number of patients failed to complete the SF-36 in full, data are given for 178 routine attenders and 32 frequent attenders. The scores on the SF36 are expressed as percentages with a high score indicating good function and a low score indicating poor function or poor health status. The frequent attenders had extremely low scores on all of the subscales of the SF-36, particularly,

Table 1 Health status (SF-36) of routine attenders vs. frequent attenders Routine attenders (n = 178)

Frequent attenders (n = 32)

ANCOVAR with age and gender as covariates ( P)

Subscales of SF-36

Mean

S.E.M.

Mean

S.E.M.

Student's t test 95% CI for difference

Physical function Role limitation Ð physical illness Role limitation Ð emotional Social functioning General mental health Energy Bodily pain Health perceptions

91.1 84.7

1.3 2.4

57.0 33.6

6.4 6.5

20.8 ± 47.4a 37.1 ± 65.1a

< .001 < .001

83.7

2.4

28.1

6.2

43.5 ± 67.7

< .001

87.9 75.4 67.8 78.0 75.7

1.6 1.3 1.6 2.0 1.7

42.0 38.3 27.7 47.2 38.8

6.1 4.7 4.3 6.2 4.7

33.1 ± 58.7a 27.3 ± 47.0a 32.0 ± 48.3 17.5 ± 44.0a 28.1 ± 45.8

< .001 < .001 < .001 < .001 < .001

a

Equal variances not assumed.

164

E.R.L. Williams et al. / Journal of Psychosomatic Research 50 (2001) 161±167

Table 2 Psychiatry diagnoses generated by SCAN: routine attenders vs. frequent attenders Routine attenders, n = 182

Frequent attenders, n = 35

c2 test

Psychiatric diagnoses according to SCAN

n

%

n

%

OR

Psychiatric diagnosis (excluding alcohol) Alcohol-related diagnosis Schizophrenia/psychosis Depression/affective disorders Anxiety/neurotic disorders Somatoform disorders

28 63 2 10 26 7

15.4 34.6 1.1 5.5 14.3 3.8

21 17 7 13 13 1

60.0 48.6 20.0 37.1 37.1 2.9

8.2 3.8 ± 18.1 1.8 0.9 ± 3.7 15.0 2.8 ± 80.9 10.2 4.0 ± 25.9 4.0 1.8 ± 8.9 NA as numbers very small

95% CI

OR = odds ratio; CI = confidence interval. NA = not applicable. SCAN = Schedules for Clinical Assessment in Neuropsychiatry.

`role limitation Ð physical,' `role limitation Ð emotional,' and `general mental health.' Their scores were significantly lower, at the 0.1% level, on all of the subscales in comparison with the routine attenders, whether the Student's t test was used, or an analysis of covariance was carried out to adjust for age and gender differences between the two groups. Psychiatric and somatoform diagnoses Excluding alcohol-related diagnoses, 21 (60.0%) of the 35 frequent attenders had a psychiatric diagnosis on the SCAN (ID level 5 or above) compared with 28 (15.4%) of the routine attenders (Table 2). Frequent attenders were also more likely to have had psychiatric in-patient treatment than routine attenders (16/31 vs. 5/182: OR = 37.3, 95% CI = 12.0 ±116.1) and to have had a past history of psychiatric disorder (21/31 (67.7%) vs. 24/182 (13.2%): OR = 13.8, 95% CI = 5.8± 32.9). Many patients in the frequent attenders group had more than one diagnosis. All SCAN diagnoses above ID level 5 are shown in Table 2. In addition, the table also gives figures for the major alcohol-related diagnoses made by SCAN interview and somatoform or dissociative diagnoses (most of which did not reach ID level 5).

There was a high prevalence of alcohol-related disorders in both groups. A total of 17 (48.6%) of the frequent attenders and 63 (34.6%) of the routine attenders had at least one alcohol-related diagnosis. A total of 10 (28.6%) of the frequent attenders had comorbid alcohol-related problems and psychiatric disorder in comparison with 11 (6.0%) of the routine attenders ( P = .0003, OR = 6.2, 95% CI = 2.4 ± 16.1; Fisher's Exact Test). Only 1 (2.9%) of the frequent attenders received a somatoform diagnosis that was F45.3 (somatoform autonomic disorder), in comparison with 7 (3.8%) of the routine attenders; 4 had hypochondriacal disorder, 2 had neurasthenia (fatigue syndrome), and 1 had undifferentiated somatoform disorder. A further 14 patients in the frequent attenders group, who could not be interviewed (either refused or could not be traced), had a history of contact with the psychiatric services in Manchester. Four had been diagnosed as having schizophrenia, two were brain injured, two had anorexia nervosa, five were diagnosed as having personality disorder, and one had somatisation disorder. These diagnoses were taken from the notes and are clinically based. If data from these patients is added to those interviewed, the overall prevalence of psychiatric disorder in the frequent attenders group was 45.5%.

Table 3 History of chronic medical condition: routine attenders vs. frequent attenders Routine attenders (n = 182)

Frequent attenders (n = 33)

c2 test

History of chronic medical condition

n

%

n

%

OR

95% CI

Central nervous system, e.g., epilepsy Urogenital system Cardiology, e.g., ischaemic heart disease Gastrointestinal, e.g., peptic ulcer Respiratory Haematology Endocrine/pancreatic/diabetic Dermatology Joint/limb ENT/ophthalmology

9 4 3 11 22 1 3 11 20 8

4.9 2.2 1.6 6.0 12.1 0.5 1.6 6.0 11.0 4.4

14 6 5 8 9 2 3 4 6 4

42.4 18.2 15.2 24.2 27.3 6.1 9.1 12.1 18.2 12.1

12.8 9.2 9.9 4.6 2.5 10.9 5.6 2.0 1.7 2.8

4.9 ± 33.2 2.5 ± 34.6 2.4 ± 43.8 1.7 ± 12.5 1.0 ± 6.1 1.0 ± 124.5 1.1 ± 28.9 0.6 ± 6.7 0.6 ± 4.5 0.8 ± 9.9

ENT = ear, nose, and throat. CI = confidence interval.

E.R.L. Williams et al. / Journal of Psychosomatic Research 50 (2001) 161±167

165

Table 4 SF-36 subscale scores: routine attenders with psychiatric illness vs. frequent attenders with psychiatric illness Routine attenders with psychiatric illness (n = 24)

Frequent attenders with psychiatric illness (n = 20)

Subscale of SF-36

Mean

S.E.M.

Mean

S.E.M.

95% CI for difference in means

Physical function Role limitation Ð physical Role limitation Ð emotional Social functioning General mental health Energy Bodily pain General health perception

84.5 66.7 56.9 74.1 56.8 49.4 65.7 61.2

4.6 7.9 8.9 5.4 4.6 5.4 6.8 4.6

56.8 32.5 16.7 38.9 33.0 24.0 43.9 38.5

8.0 8.5 6.2 7.3 5.6 5.4 7.5 6.4

9.2 ± 46.8 10.7 ± 57.7 18.5 ± 62.1 17.2 ± 53.2 9.3 ± 38.4 9.8 ± 40.9 1.5 ± 42.2 7.1 ± 38.3

Physical health Out of the 35 frequent attenders who were interviewed, 33 supplied information concerning their physical status. Of these 33 frequent attenders, 30 (90.9%) reported a history of chronic medical illness, compared with 68 (37.4%) of the routine attenders (OR = 16.8, 95% CI = 4.9 ±57.0). Many patients reported having several medical conditions. Table 3 shows the main kinds of reported medical conditions according to general bodily system. This data is based upon the patients' self-report and was not validated by an independent medical opinion. The most common conditions reported by frequent attenders were central nervous system problems, respiratory problems, and gastrointestinal problems.

ANCOVAR with age and gender as covariates ( P) .014 .016 < .001 .002 .002 .004 .093 .021

with 39/179 (21.8%) of the routine attenders (OR = 2.6, 95% CI = 1.2 ± 5.7). Ten frequent attenders had been admitted to a psychiatric bed within the previous 12 months compared to none of the routine attenders (OR = 167.5, 95% CI = 9.5 ± 2959.0). The mean number of GP visits by the frequent attenders in the previous 12 months was 11 (S.E.M. = 1.1) compared with 3 (S.E.M. = 0.4) for the routine attenders (t = 7.0, P .001, 95% CI for difference = ÿ 10.2 to ÿ 5.7, unequal variance version of the Student's t test). Very few of either group were in contact or used community services, e.g., visits from a practice nurse, community psychiatric nurse or any other kind of community health-related service.

Health status of patients with psychiatric disorder

Discussion

The health status of patients with psychiatric disorder (ID level 5 or above on the SCAN) in both the frequent attenders and the routine attenders group was compared. SF-36 data was complete for 20 out of the 21 frequent attenders with psychiatric disorder and 24 out of the 28 routine attenders with psychiatric disorder (Table 4). The frequent attenders with psychiatric problems had significantly poorer health status than the routine ED patients with psychiatric disorder. This was apparent on all subscales, with the exception of bodily pain.

This is the first study in the UK that has focused upon patients who make repeated visits to the ED. Its findings are strengthened by the use of a detailed interview as opposed to reliance upon a self-report questionnaire. Out of the 77 frequent attenders identified for inclusion in the project, however, only 35 could be interviewed face-to-face, although further data was obtained on 64% of the original sample. This limits the potential representativeness of the study, but it may also reflect certain characteristics of the population studied. A substantial proportion of subjects has no fixed abode or gave false names or addresses, making them impossible to trace. Another proportion refused to take part in the study. Of the patients who were interviewed, in many instances, several visits had to be arranged before contact with them was established. The patients who could not be interviewed were similar in terms of their demographic profile to those who were interviewed and at least 30% had had psychiatric illness. We cannot be certain, however, of the exact prevalence of psychiatric disorder in this group, or whether they differed from those interviewed on any other characteristics. The power of the study to detect differences between the frequent attenders group and controls was also reduced,

Health utilisation A substantial proportion of the routine attenders (n = 52, 28.6%) had attended the ED on a previous occasion during the 12-month study period. On all parameters of health service utilisation, however, the frequent attenders had used more health service resources than the routine attenders in the previous 12 months. Of the 34 frequent attenders, 23 (67.6%) had been admitted to the general hospital in the previous year compared with 17 (9.5%) out of 179 of the routine attenders (OR = 19.9, 95% CI = 8.3 ±47.8. Of 33, 14 (42.4%) had attended a medical out-patient clinic compared

166

E.R.L. Williams et al. / Journal of Psychosomatic Research 50 (2001) 161±167

but despite this, highly significant differences were found between the two groups. These findings, however, only correspond to the frequent attenders who were interviewed. The sampling procedure that was used ensured that the control group included patients who were representative of those attending the ED department on a routine basis. A total of 528 patients were identified during a 3-month period. As sampling took place for one-third of each day, the total number of patients attending the department during this 3-month period would have been approximately 528  10  3 = 15,840. This accords with the average annual attendance of 67,500. As with a previous study of frequent attenders at the ED [6], there was a predominance of males and frequent attenders were of an older age than routine attenders. Primary care studies [11,12], however, have found that women are more likely to frequently attend than men. This suggests that the two populations may be different. The most striking finding of this study was the extremely poor health status of the frequent attenders, whether or not they had psychiatric disorder. The prevalence of psychiatric disorder was also very high in the frequent attender group, the most common conditions being anxiety and depressive disorders, with many patients having a combination of both, plus, in addition, alcohol-related problems. Nearly half of the frequent attenders and one-third of the routine attenders received at least one alcohol-related diagnosis on the SCAN. Although there is a high prevalence of psychiatric disorder amongst frequent attenders in primary care [6 ± 14], alcohol abuse does not appear to feature as a major problem in this group. This again suggests that frequent attenders at the ED may differ from patients who are high utilisers of primary care services. Frequent attendance in primary care has been linked to somatisation [10]; however, the prevalence of somatisation and somatoform disorders in the ED frequent attenders was much lower than we expected. Only one of the frequent attenders and seven of the routine attenders received any kind of somatoform diagnosis. One of the advantages of the SCAN instrument is the inclusion of items related to somatisation and the diagnosis of somatoform disorders. Although, the criteria for awarding somatoform diagnoses are fairly stringent, there was no suggestion from our study that ED frequent attenders had higher rates of somatisation than routine attenders. However, the SCAN instrument may not be sensitive enough to detect subclinical processes of somatisation and subtle differences in health perception as have been found between primary care frequent and routine attenders [22], may have been missed. It is also possible that somatisation is much less common in ED frequent attenders than those in primary care. The ED frequent attenders reported higher rates of `chronic medical illness' in comparison with the routine attenders. Their reports of illness were not validated by an independent medical opinion, so it possible that some of their complaints were related to somatisation rather than

organic medical disease. Many of the frequent attenders, however, appeared to have frank organic disorder such as diabetes, epilepsy, and alcohol-related physical disease. It appeared to be the combination of chronic medical disorder plus psychiatric disorder that accounted for the frequent attendance at the ED. It is possible that the presence of anxiety or depression increases patients' concern about their physical problems and makes them less able to cope. In addition, many of the frequent attenders were socially isolated and therefore unlikely to receive adequate social support and help. The frequent attenders, who were interviewed, were not only high users of emergency services, but also visited their GP on a frequent basis, with a mean number of attendances in the previous year of 11.0 (S.E.M. = 1.1). Studies of frequent attenders in primary care have used attendance rates per annum of between 4 and 12 as a cut off to define a primary care frequent attender [7 ± 12]. Thus, frequent attenders at the ED, in this study, also satisfied criteria for frequent attendance in primary care. These findings are consistent with a study of medical out-patients [16], which found that 35% of frequent clinic attenders were also frequent attenders in primary care ( > 12 visits per annum). There are sufficient differences between the characteristics of ED frequent attenders in the present study and known characteristics of primary care frequent attenders to suggest that the two populations are not the same, although clearly, there is some overlap. Despite the heavy use of a wide variety of services by the ED frequent attenders, there was little evidence of a coordinated response. The psychiatric services were unaware that many of the patients were attending the ED frequently and there was no direct communication between the ED and primary care. Few of the patients were receiving structured community treatment, although one-third would be considered to have `severe mental illness' (e.g. schizophrenia, other psychosis, or bipolar affective disorder). None of the patients had a coordinated community care programme involving mental health, primary care, and the ED. As a result of the findings of this study, we recommend that EDs should keep a special register of frequent attenders (defined as those attending the department seven or more times within the previous 12 months), which should be updated monthly. For those frequent attenders who are known to the psychiatric services, joint planning meetings with the ED, the patient's GP, primary consultant physician, and psychiatric team should be arranged. As at least half of frequent attenders may not be traceable by conventional services, each district should, therefore, have a policy for the treatment of the mentally and physically ill who are of no fixed abode. Frequent attenders with psychiatric disorders should receive the same degree of priority for the care programme approach as patients with `severe mental illness'. Specific treatment and management strategies need to be developed for this group of patients, which focus upon

E.R.L. Williams et al. / Journal of Psychosomatic Research 50 (2001) 161±167

both their physical and psychosocial needs, although a substantial proportion may be difficult to engage in the treatment process. Acknowledgments This work was funded by the North West Regional Health Authority.

[10] [11] [12]

[13]

References [1] Audit Commission. By accident or design: improving A&E services in England and Wales. London: HMSO, 1995. [2] Purdie RFJ, Honigman B, Rosen P. The chronic emergency department patient. Ann Emerg Med 1981;10:298. [3] Andren KG, Rosenqvist U. Heavy users of an emergency department: psycho-social and medical characteristics, other health care contacts and the effect of a hospital social worker intervention. Soc Sci Med 1985;21:761 ± 70. [4] Hansagi H, Edhag O, Allebeck A. High consumers of health care in emergency units: how to improve their quality of care. Qual Assur Health Care 1991;3:51 ± 62. [5] Ullman R, Block JA, Stratmann WC. An emergency room's patients: their characteristics and utilisation of hospital services. Med Care 1975;13:1011 ± 20. [6] Murphy AW, Leonard C, Plunkett PK, Brazier H, Conroy R, Lynam F, Bury G. Characteristics of attenders and their attendances at an urban accident and emergency department over a one year period. J Acid Emerg Med 1999;16:425 ± 6. [7] Westhead JN. Frequent attenders in general practice: medical, psychological and social characteristics. J R Coll Gen Pract 1985;35: 337 ± 40. [8] Katon W, Von Korff M, Lin E. Distressed high utilizers of medical care. DSM-III-R diagnoses and treatment needs. Gen Hosp Psychiatry 1990;12:355 ± 62. [9] Karlsson H, Lehtinen V, Joukamaa M. Psychiatric morbidity among

[14] [15] [16] [17] [18]

[19] [20] [21] [22]

167

frequent attender patients in primary care. Gen Hosp Psychiatry 1995;17:19 ± 25. Portegijis PJ, vander Horst FG, Proot IM, Kraan HF, Gunther NC, Knottnerus JA. Somatization in frequent attenders of general practice. Soc Psychiatry Psychiatr Epidemiol 1996;31:29 ± 37. Heywood PL, Blackie GC, Cameron IH, Dowell AC. An assessment of the attributes of frequent attenders to general practice. Fam Pract 1998;15:198 ± 204. Jyvasjarvi S, Keinanen-Kiukaanniemi S, Vaisanen E, Larivaara P, Kivela SL. Frequent attenders in a Finnish health centre: morbidity and reasons for encounter. Scand J Primary Health Care 1998;16: 141 ± 8. Vested P, Olesen F. Frequent attenders in out-of-hours general practice care: attendance prognosis. Fam Pract 1999;16:283 ± 8. Bellon JA, Delgado A, Luna JD, Lardelli P. Psychosocial and health belief variables associated with frequent attendance in primary care. Psychol Med 1999;29:1347 ± 57. Bass C, Bond A, Gill D, Sharpe M. Frequent attendance without organic disease in a gastroenterology clinic. Patient characteristics and health care use. Gen Hosp Psychiatry 1999;21:30 ± 8. Bell G. Screening for psychiatric morbidity in an accident and emergency department. Arch Emerg Med 1990;7:155 ± 62. Ware JE, Sherbourne CD. The MOS (medical outcome study) short form health survey (SF 36): conceptual framework and item selection. Med Care 1992;30:473 ± 83. Tarlov AR, Ware JE, Greenfield S, Nelson EC, Perrin E, Zubkoff M. The medical outcomes study: an application of methods for monitoring the results of medical care. JAMA, J Am Med Assoc 1989;262: 925 ± 30. World Health Organisation. Schedules for Clinical Assessment in Neuropsychiatry. Geneva: WHO, 1992. World Health Organisation. The ICD-10 classification of mental and behavioural disorders: clinical descriptions and diagnostic guidelines. Geneva: WHO, 1992. SPSS. SPSS for Windows version 6.1.3. 444 N. Michigan Ave., Chicago, IL: SPSS, 1995. Sensky T, MacLeod AK, Rigby MF. Causal attributions about common somatic sensations among frequent general practice attenders. Psychol Med 1996;26:641 ± 6.