Psychotherapy patients : are they " the worried well " ?

for many years to describe patients who make frequent and inappropriate use of medical and psychiatric services. In mental illness settings these patients are felt to occupy inappropriately the time of psychiatrists, psychologists, social workers, and community psychiatric nurses, particularly those who work in specialist psy chotherapy departments. The implication of the use of the term in this context is that psychother apy departments see patients who are not 'really ill' psychiatrically but who enjoy chatting about

The term 'the worried well' has been used loosely for many years to describe patients who make frequent and inappropriate use of medical and psychiatric services.In mental illness settings these patients are felt to occupy inappropriately the time of psychiatrists, psychologists, social workers, and community psychiatric nurses, particularly those who work in specialist psy chotherapy departments.The implication of the use of the term in this context is that psychother apy departments see patients who are not 'really ill' psychiatrically but who enjoy chatting about themselves.The implication is also that psy chotherapists are happy to indulge them in this, wasting valuable resources which could be more appropriately applied to other patient groups.
The relevance of psychotherapy services in the National Health Service (NHS) is topical (Holmes & Marks, 1994), particularly in view of the current NHS review of psychotherapy services.
The aim of this study is to describe the symptom and personality disorder characteristics of patients waiting for treatment at a district psychotherapy service and to assess the severity of psychiatric disorder in this population.

The study
The psychotherapy department for the Swindon district serves a population of 250 000.Patients referred to the department, and who had con firmed their wish to attend, were surveyed by letter.The survey included three groups of patients.The first group included all patients on the waiting list at the end of July 1994.The second group included those patients referred after July 1994 who were surveyed at the time of referral for a four month period.These two groups together constitute a sample of 75 consecutive patients awaiting treatment.The third group came from an earlier survey of all patients on the waiting list at the end of December 1993 and was constituted of 39 consecutive patients await ing treatment, making a total of 114 patients in the study.
A covering letter described the survey and explained that cooperation was optional and would not affect the patient's place on the waiting list.Eighty-one patients (71%) of 114 patients responded.Average age was 34 and 55% were female.Fifty-five per cent were referred by psychiatrists and 45% by general practitioners.At the time they completed the questionnaires, patients had been waiting between two weeks and six months.
The patients completed the following symptom measures: the Symptom Rating Test (SRT) (Kellner & Sheffield, 1973); the Beck Depression Inventory (BDI) (Beck et cd, 1961); the Hope lessness Scale (Beck & Weissman, 1974); the Beck Anxiety Inventory (Beck et al 1988); and the modified PDQ-R (the screening test for co-morbid personality disorder; Dowson, 1992) which is a shortened form of the Personality Diagnostic Questionnaire-Revised (Hyler & Reider, 1987), and predicts the likelihood that the patient would receive one or more personality disorder diag noses, if that patient completed the full PDQ-R.It also provides specific information on the most frequently diagnosed personality disorders, viz histrionic, dependent, avoidant and borderline personality disorders.
The referral letters were perused to take note of features particularly mentioned by the referrer, and to see if there were obvious differences between those who returned questionnaires and those who did not.The results returned by the two waiting list groups were very similar, so these results are presented together.

Findings
The most important results on the symptom measures are presented in Tables 1 and 2. The mean on the Symptom Rating Test (SRT) was 69.17 which is much higher (P< 0.001) than the mean of 50 established by Kellner & Sheffield (1973) on two samples of 'mainly neurotic' psychiatric out-patients.In fact, this difference Psychiatric Bulletin (1996), 20, 153-156 is even greater since the 30 item version has been used in the present study, whereas Kellner & Sheffield's figure was established using the original 38 item version.(If the present scores were scaled accordingly, the figure for the present study would be approximately 88.) Twenty-eight (35%) patients were severely distressed by symp toms (>75 on the SRT).The mean on the Beck Depression Inventory was 24.3, significantly higher than the mean for the patients with major depression (P<0.01), and for patients with mixed psychiatric diagnoses (P<0.05)originally studied by Beck (Beck et al, 1961).It was similar to the mean (24.50) for 117 patients in a large treatment study of major depression (Shapiro et ai. 1994).Overall, 26 patients (32%) of patients in the present study were severely depressed (BDI>30).On the Beck Anxiety Inventory, 15 (19%) had severe anxiety symptoms (Table 1).
The results for personality disorder are sum marised in Tables 2 and 3.The average number of personality disorder diagnoses per patient was 5.36.This is higher than a mixed sample of out patients with symptom disorders and/or person ality disorders, and not much lower than the mean (6.0) for patients admitted to an in-patient unit for severe personality disorder (the Hender son Hospital; Dolan et al 1995).In Table 3, two comparisons for the frequency of the four commonest personality disorders are shown.
The data relevant to the assessment of suicide risk are summarised in Table 4.The modified PDQ-R contains an item which refers to the occurrence of at least two past suicide attempts.Previous attempts are the most important histor ical predictor of suicide risk.Overall, 25% of patients responded positively to this item ("More than once I have tried to end my life").In addition, 33% of patients reported a history of other episodes of self-harm.The scores on the Hope lessness Scale were striking.Sixty (74%) of the patients scored above 9, considered by Beck to be the cut-off point for the group at risk of completing suicide (Beck et cd, 1990).Thirtythree (41%) scored above 15, defining a group at high risk.Overall, the mean was 12.48, not significantly less than the mean (13.27) for 14 patients who completed suicide, in a ten year follow-up study of 207 patients who had been hospitalised with suicidal ideation.It was sig nificantly higher (P< 0.001) than the group who did not complete suicide (Beck et cd, 1985).

Referred letters
The referral letters of the 1994 patients were examined to take note of the clinical features most often mentioned by referrers.In 60% unsatisfactory response of symptoms to a variety of psychotropic medications was mentioned; in 32% of cases the referrer noted the chronicity of the disorder as a main feature.In 21% a history of physical or sexual abuse was mentioned.

Comment
The response rate of 71% is slightly above average for this type of study.There is no way of knowing whether the responders are more or less severe than the non-responders.However, when the 'Missing data on six patients referral letters were examined, the non-responders could not be distinguished from the respon ders in terms of chronicity, of unsuccessful drug treatment, the frequency of past suicide attempts and other episodes of self-harm reported by the referrer, suggesting non-response depends on factors other than chronicity and severity.The patients showed high levels of distress and morbidity from a variety of symptom and person ality features.In fact, only one patient (1%) could be identified whose symptomatology was mild on all symptom measures, and who also had no personality disorder diagnosis.A further five (6%) had symptoms that were mild but these patients had at least one personality disorder diagnosis.Anxiety symptoms were generally not high but the likely explanation is that patients with anxiety disorders (including phobias and obsessions) are referred to behaviour therapists or the psychology department.
This kind of referral pattern is common in NHS settings.
Particularly striking was the high level of hopelessness, with 74% of patients scoring above the level, which distinguishes those at risk of suicide from those who are not, and 40% exhibiting very high levels of hopelessness.Estimation of suicide risk is an inexact science but the very high levels of hopelessness, coupled with the frequency of previous self-harm (33%) and actual suicide attempts (25%), support the conclusion that about half these patients are at significant risk of completing suicide.To date, one patient in the 1993 group has committed suicide.
The figures for personality disorder on the modified PDQ-R, 87% of patients with borderline personality disorder and an average five person ality disorder diagnoses per patient, may seem very high.There are three reasons for this: they are high, namely this group exhibits a large number of severe personality features; the use of specific diagnostic criteria means that multiple diagnoses can be made (Dolan et al 1994); and the lower threshold of self-report measures of personality disorder compared to interview diag nosis.
It is helpful in interpreting these data for personality disorder to compare the figures for other patient groups, as shown in Tables 2 and 3. Interview measures produce lower frequencies of personality disorder diagnoses than self-report measures as illustrated by Hyler et als (1990) comparative figures for patients assessed by selfreport measure and SCID-II interview, on admis sion to an in-patient unit for severe personality disorder (Table 3).Unsurprisingly, histrionic personality disorder produces the highest figures for self-report compared with SCID-II interview.Using Hyler et afs figures as a guide, it is probable that the following approximate frequen cies would be obtained by SCID-II interview for patients in the present study: borderline 67%; Psychotherapy patients: are they "the worried well"?dependent 39%; avoidant 54%; and histrionic 26%.
The National Institute of Mental Health Treat ment of Depression Collaborative Research Pro gram (Shea et al 1992) and the treatment study of Shapiro et al (1994) provide data on the amount of treatment needed by patients at each level of severity of depression.Using these studies as a guide, the department where this study was carried out is funded by the local purchaser to provide less than half of the minimum amount of treatment needed by these patients.

Comments
This study shows that patients referred to a general psychotherapy service could not be described as "well", and far from being simply "worried" they show a high incidence of severe symptoms and maladaptive personality features."Despair" rather than "worry" is the most notice able feature.Many are likely to be chronic attenders at general practices and psychiatric services and they respond poorly to pharmaco logical treatment of their co-morbid symptom disorders.
The evidence for the effectiveness of psychotherapeutic methods in a wide range of clinical settings has been summarised by Holmes & Marks (1994).Many of these patients can be helped (both in terms of response and relapse/ recurrence prevention) by modern brief psychotherapeutic techniques, especially cognitivebehavioural psychotherapy, interpersonal dy namic psychotherapy and cognitive-analytic therapy.

Table 1 .
Comparison of group norms

Table 2 .
Study ratings on the Personality

Table 3 .
Results for personality disorder

Table 4 .
Suicide risk factors in these data