Outcome of Cognitive Behavior Therapy for Patients With Severe Health Anxiety Treated in Group Only. A Randomised Controlled Trial. The CHAG-Trial. Categorical and Dimensional Characteristics of Personality Pathology and Predictors of Outcome
Outcome of Cognitive Behavior Therapy for Patients With Severe Health Anxiety Treated in Group Only. A Randomised Controlled Trial. The CHAG-Trial. Categorical and Dimensional Characteristics of Personality Pathology and Predictors of Outcome
Background:
The prevalence of severe health anxiety is reported to be 1-2% in Western communities. This functional disorder is difficult for medical doctors to treat, the course of the disorder is often chronic, and that is costly for the social and health care systems as well as for the patients. A Cochrane metaanalysis from 2009 finds evidence for effectiveness of individual cognitive behavior therapy (CBT) for patients with hypochondriasis. But no randomised controlled trials (RCT) of the effectiveness of classical CBT delivered only in groups for patients with severe health anxiety (hypochondriasis/illness anxiety disorder) has yet been conducted.
Aims:
1) to examine the effectiveness of group-CBT for patients with severe health anxiety compared to a wait-list group receiving usual care, 2) to perform a categorical and dimensional assessment of personality, 3) to examine predictors of outcome especially comorbid personality disorders, 4) to examine the relation between personality, illness perception and treatment outcome, 5) to compare the cost-effectiveness of these two treatments, 6) at a 2 years follow up to examine the course and long-term effectiveness of group-CBT for patients with severe health anxiety and some also followed by psychological treatment for comorbid personality disorders.
Main hypothesis:
Patients with severe HA who have received group CBT will at 6-month follow-up compared to a wait-list group receiving usual care show a significantly reduction in health anxiety.
Methods:
84 patients referred from medical doctors during 2014-15 to the Clinic of Liaison Psychiatry in Koege, Region Zealand, Denmark, will be included and block randomised per 14 patients to either weekly group-CBT with 7 patients and 2 therapists for 3 hours a week in 12 weeks or wait-list with usual care for 9 months.
Inclusion: Severe health anxiety (dominant mental disorder), score on WI-7>21,4, age 18-65 years, Danish speaking, informed consent.
Exclusion: Another severe treatment demanding mental disorder, risk of suicide or psychosis, a serious somatic disease, pregnancy, dependency of drugs, alcohol or medication.
Diagnostic assessment:
The patients are included using research criteria for severe health anxiety (for ICD-11) and semi-structured interviews developed for DSM-IV, SCAN (general psychopathology) and SCID-II (personality disorders). Criteria for hypochondriasis from ICD-10 and illness anxiety disorder/somatic symptom disorder from DSM-5 are used for subcategorising. Dimensions and traits of personality are assessed by the questionnaire PID-5 included in DSM-5, section III.
Outcome measures:
The primary outcome measure is the questionnaire for health anxiety, Whiteley Index 7 (WI-7), with a cut-off for remission on 21,4 or a blinded diagnostic assessment of no severe health anxiety present 6 months after end of treatment.
The secondary outcome measures are questionnaires for health anxiety (HAI), general psychopathology (SCL-90-R), level of personality disorders (PID-5), level of functioning (SF-36), quality of life (WHO-5, EQ-5D), Illness perception (IPQ), alcohol consumption (CAGE) and register data for number of sick days and use of social and health care and a blinded global assessment of functioning (F-GAF).
Time frame:
Data wil be analysed, and results wil be disseminated from 2016.
BACKGROUND
Classification and comorbidity
Hypochondriasis is broadly defined in the international classification systems for mental disorders ICD-10 (WHO) and DSM-IV (APA) under somatoform disorders as a preoccupation with fears and the belief of having a serious disease with the condition lasting for at least 6 months. It is non-delusional, and the symptoms are not better explained by another mental disorder, it causes distress or impairment of functioning and persists despite appropriate medical reassurance. Somatoform disorders are defined as syndromes with physical symptoms, which are not sufficiently explained by physical causes alone, defined somatic illnesses or comorbid mental disorders.
A less stigmatizing name: health anxiety has been suggested, and new research has reported more valid diagnostic criteria with continuous ruminations of fears and belief of having a serious disease lasting for at least 2 weeks and the presence of minimum one of five other symptoms: preoccupation with illness or body, preoccupation with medical information, suggestibility or autosuggestibility, fear of infection or contamination and fear of medication. A degree for severity has also been introduced here with either mild or severe health anxiety split by the presence of minimum one severely distressing symptom or a significant impairment of functioning [1]. Hypochondriasis and Health Anxiety (HA) are from now on used synonymously here.
HA like other somatoform disorders show high degree of mental comorbidity with depressive disorders between 15,3-45,2%, anxiety disorders between 22-52,4%, somatoform disorder between 8-21,4% and mental comorbidity in total between 54-78,6% [11,34,46]. HA also show a high degree of mental comorbidity with personality disorders (PD) between 63-76% [34,35] when using questionnaire for assessment, but they seem to overestimate the prevalence [36] and between 40-74% when using SCID-II, the semi-structured interview developed for DSM, that is the present golden standard for diagnosing PD [39,40].
A high degree of mental comorbidity has risen questions of classification, if HA is a primary mental disorder and therefore its own disease entity, or the symptoms are only secondary to another mental disorder like a melancholic depression, or if HA is better classified as a personality disorder [37] or an anxiety disorder [38]. In DSM-V and ICD-11 the disorder entity is suggested preserved and kept categorized under somatoform disorders, but the name is suggested changed to illness anxiety disorder in DSM-V and health anxiety in ICD-11.
Epidemiology
Studies have reported a prevalence for HA in communities between 0,2-4,5% [2,28] and in primary care between 0,8-7.0% [3-7], and using the new and more valid criteria on 9,5% (12.1% incl. mild degree) [1] and in secondary care on 10-20% [8-9]. Severe HA has an early onset with mean age of onset = 25 years [1], an equal sex distribution, it is fluctuating but persistent with a median duration before treatment on 2- 20 years [10,11,31,45], and it is distressing and disabling and therefore also costly for our health and social care system [12,13].
Personality characteristics
Only 2 former studies have assessed patients with HA using SCID-II. One had a small sample size of 21 psychiatric outpatients, and 74% of these patients were found to have comorbid PD compared to 52% for other psychiatric outpatients in Greece. Here the most prevalent PD for HA were histrionic and obsessive-compulsive personality disorders [39]. The other study included 42 patients with HA and no comorbid OCD or Social Phobia recruited from a community in USA. This study found, that 40% with HA had comorbid PD, and the most prominent types were paranoid, obsessive-compulsive and avoidant PD. Here the group of HA were compared to groups of patients with OCD and Social Phobia, and they found no significant difference between the prevalence of PD with HA compared to the prevalence of PD with either of these 2 anxiety disorders [40].
Few studies of personality dimensions for HA have been made. One study used the questionnaire Personality Assessment Schedule (PAS) and found for hypochondriacal PD a high degree of the personality traits: hypochondriasis, anxiousness, consciousness and dependence [37]. Other personality traits for HA that have been reported are high degree of neuroticism, narcissism and avoidance and histrionic and borderline traits [37,43,44].
Treatment
Patients suffering from HA are not often diagnosed and treated well by doctors [5]. It causes frustration for both doctors and patients. Doctors feel powerless towards the patients continuing suffering and complaints, patients feel they are not taken seriously by the doctor and do not get sufficient help, and the extensive somatic focus and medical examinations contribute to the chronic development of the disorder [48,49,59].
On Zealand with app. 2.5 million inhabitants only about 40 patients with HA were referred from medical doctors to specialized treatment in the existing 2 clinics of Liaison Psychiatry in 2011.
Earlier on HA was considered untreatable, but psychotherapy with cognitive behavioural therapy (CBT) for HA has been developed over the past 2 decades and has shown effective. There is also evidence from 2 studies reporting moderate treatment effectiveness of SSRI medication for HA [22,23], but prior to treatment patients with HA favour psychological treatment over pharmacological treatment, although pharmacological treatment is equally tolerated [22-24]. A meta-analysis from 2007 includes 6 randomized controlled trials (RCT) and finds evidence for effectiveness of treatment of HA with short term individual CBT. The effective factors of the CBT on health anxiety symptoms as primary outcome measure were found to be cognitive restructuring (cognitive therapy), response-prevention, exposure, and behavioural stress management, and with resource use as secondary outcome measure also psychoeducation [14].
CBT for personality disorders has so far only shown effective in long term treatment [65].
Group CBT might have some advantages over individual CBT in cost-effectiveness, because of less therapist time spend per patient, individual CBT is more vulnerable to session cancellation from the patient and possible effective group dynamics. No studies so far have compared effectiveness of group CBT to individual CBT for HA, and generally cost-effectiveness and group dynamics have little empirical evidence and need more research [16-21].
5 studies on effectiveness of group CBT for HA have formerly been made. They showed treatment effective and acceptable [25,26,29,30], and one study also showed cost-effective for society in saved health care [27]. But the sample sizes of 3 of the studies were small, and their design were not controlled, and for the two RCT studies one began therapy with 8 individual sessions of CBT [30], and the intervention of the second study was mostly psychoeducational [29].
Prior to treatment patients generally prefer individual CBT over group CBT, but attrition and satisfaction with both treatments have shown equal at the end of treatment with dropout rates between 10-47% [20,21,47].
Predictors of outcome
Only a few studies so far have examined predictors of outcome of CBT for HA and often with conflicting results. 2 studies found no predictive value for sociodemographic variables [33,62], but one study found a positive predictive value for marital status [32] and one a negative predictive value for high age [66].
3 studies have found a negative predictive value for a higher degree of health anxiety before treatment [32,33,66], but the results for pretreatment anxiety, general psychopathology and chronicity are inconsistent [32,33,62,66]. One study also found negative predictive value for more dysfunctional cognitions related to bodily functioning, disability at work and utilization of health care before treatment but no predictive value for comorbid anxiety or depressive disorders [33].
The influence of comorbid PD in HA for treatment outcome has to our knowledge only been investigated in one former study. The design included no control group, and the included 36 patients had been non-respondant to previously given pharmachological treatment, and afterwards they were treated with different duration of individual insight psychotherapy and anti-anxious medication. Here comorbid PD had a negative predictive value for outcome [62].
Studies of patients with depressive and anxiety disorders and comorbid PD have generally found worse treatment outcome for both pharmacological and psychological treatments [41,42,60].
RATIONALE
No RCT on outcome and effectiveness of classical group CBT for severe health anxiety and hypochondriasis has yet been conducted. Group CBT compared to individual CBT (and usual care) for patients with health anxiety might have some therapeutic advantages due to manualised and structured treatment, favourable group dynamics, improving social skills and creating a support group, group-CBT might be less costly and therefore have better cost-effectiveness, but attrition might be worse, so it needs further investigation.
Initial assessment prior to CBT should also include categorical and dimensional assessment of personality because of possible relevance for prognosis, treatment preference and possibilities, creation of homogenous groups for group therapy improving acceptability and outcome, and individual designed CBT as an element of group CBT. But better general knowledge of personality vulnerability for patients with HA here by using best diagnostics and a larger sample size than before and new and better assessment questionnaires for personality dimensions and traits is also relevant for designing more effective treatment and protocols for individual and group CBT in the future. As more general practitioners become educated in CBT and programs of Shared Care are introduced, in time this treatment for HA might also be possible to integrate in primary care more close to the patients.
AIMS
In a randomised controlled trial to examine the treatment outcome of specialized classical cognitive behaviour therapy (CBT) delivered in a group setting for patients with severe health anxiety (HA) including the evidence based effective factors of individual CBT for HA [14]. More specifically:
Hypotheses
Main hypothesis:
Patients with severe HA who have received group CBT will at 6-month follow-up compared to a wait-list group receiving usual care show a significantly reduction in health anxiety.
Additional hypothesis:
Patients with severe HA who have received group CBT will at 6-month follow-up compared to a wait-list group receiving usual care have significantly: a) better mental and physical health and improved global and social functioning, b) lower use of health care services and c) less sick days.
30-60% of patients with severe HA have a comorbid personality disorder. Both patients with severe HA with or without comorbid PD show reduction in health anxiety from short term group CBT, but patients with comorbid personality disorder show a significantly lower cure-rate, than patients without comorbid personality disorder. Patients with severe health anxiety and comorbid PD (or other comorbid mental disorders) afterwards therefore need long term psychotherapeutic treatment to be cured. This hypotheses will be examined at 24 month follow-up, where a part of the patients after group CBT also have received long term psychotherapy for their PD.
Group CBT is significantly more cost-effective than a wait-list group receiving usual care and a standard duration of individual CBT for patients with severe HA.
Group CBT is a feasible and acceptable treatment for the patients (high satisfaction and low drop-out).
Patients with severe HA who have received group CBT will at 6-month follow-up compared to a wait-list group receiving usual care show a significantly more sound illness perception.
Patients with severe HA and a comorbid personality disorder have before group CBT a less sound illness perception than patients without comorbid personality disorder.
METHODS
84 patients with severe health anxiety consecutively referred from family physicians, medical specialists or hospital wards on Zealand to The Clinic of Liaison Psychiatry in Koege, Region Zealand.
The project is communicated through relevant homepages, regionsjaelland.dk and sundhed.dk, and through arranged visits with medical doctors.
The patients are block randomized per 14 patients into 2 groups: group CBT or a wait-list receiving usual care and after 9 month offered group CBT. Expected drop-out rate on 20% in group CBT and on wait-list.
DATA AND PROCEDURES
Diagnostic assessment
The patients referred are initially screened for inclusion by a telephone interview of a specialist in psychiatry (MS). The telephone interview consist of direct questions for the inclusion and exclusion criteria, WI-7, the diagnostic criteria for severe health anxiety and seeking verbal consent to send the first online questionnaire. Each patient fulfilling the inclusion is assessed by use of the SCAN interview (Schedules for Clinical Assessment in Neuropsychiatry), which is a thorough semi-structured diagnostic interview recommended by the WHO and is applied in the diagnostics of general mental disorders as defined by ICD-10 and DSM-IV criteria [50], and by use of the SCID II [51], which is a semi-structured diagnostic interview for personality disorders developed for DSM-IV. Just before the semi-structured interviews are performed a non-structured clinical interview and assessment is performed and ended by psychoeducation and the joint creation of the individual illness model for cognitive behaviour therapy. Global assessment of functioning (GAF-F) and diagnostic assessment using DSM-V, ICD-10 and the criteria for severe health anxiety will also be performed in this initial clinical assessment. The clinical and diagnostic interviews are performed by a specialist in psychiatry (MS), certified at the SCAN training centre in Copenhagen and trained in SCID-II assessment before project start. If needed a second opinion on the patients from another medical doctor at the clinic in Koege is possible (JS). For inter-rater-reliability assessment 5% of the patients are similarly clinically assessed by a blinded second rater and specialist in psychiatry. After the patients have answered the 5. online questionnaire (6 month after end of treatment), they are diagnostically assessed for severe health anxiety with direct questions for the diagnostic criteria for severe health anxiety in a telephone interview performed by a blinded second rater and specialist in psychiatry. Right after this telephone interview MS will see the patients in the clinic as a conclusion of the treatment and a possibility for the patient to be referred to further psychotherapeutic treatment, if it is indicated by a comorbid mental disorder and wanted by the patient.
Course
The patients will be followed by means of various self-reporting health measures reported online (Survey Xact); before clinical assessment (after telephone interview including verbal consent) (SCAN and SCID-II), 14 days before group start (baseline), at the end of treatment, and 3, 6 and 24 months after end of treatment (see attached Flowchart Figure 1).
Dimensions and traits of personality are assessed at baseline by the questionnaire Personality Inventory for DSM-V (PID-5) [52] developed for DSM-V, where it is being included in the Appendix.
6 months after end of treatment the patients are clinical assessed for severe HA by a telephone interview from a blinded second rater (Cure is defined here as either no diagnosis of severe HA or a WI-7 score<21,4 6 months after end of treatment).
Use of health care services and sick days from a period of 9 months preceding treatment and 6 months after end of treatment are drawn from relevant registers.
The 6. online questionnaire will be send to all patients at 2 years follow-up. Most patients have here received the short term group-CBT. Some of the patients have afterwards also received long term psychotherapeutic treatment for their comorbid personality disorder.
The patients in the control groups will be offered the group CBT immediately after answering the 6 month follow-up questionnaire of the RCT (no. 5), which then also functions as their baseline questionnaire. The patients in the control groups are then also given this questionnaire no. 5 again 9 month after (Their 6 month follow-up from the group CBT). The patients in the control group are then given one more questionnaire than the RCT treatment group then a total of 7 questionnaires for the 24 month follow-up prospective cohort study.
Randomisation
After the patients have given written informed consent, they are put on a wait-list for block randomization until there is a total of 14 patients. When 14 patients have answered the 2. online questionnaire, they are randomized to 2 groups; group therapy (n=7), starting shortly after randomization, or wait-list (n=7) with an offer to start group therapy after 9 months. Patients are randomized by means of a computer algorithm that use predefined concealed random numbers.
Treatment
Group CBT
The patients are treated in groups of 7 by 2 therapists, that are either medical doctors, psychologists, psychomotoric or nurses trained in cognitive behaviour therapy and experienced in treating patients with HA with CBT. 2 teams of therapists (4 therapists) deliver the treatment, and MS only take part as a therapist if necessary. The therapists are during the project supervised regularly from an external psychologist. The group therapy includes 12 weekly meetings of 3 hours excl. a 15 minutes break and a booster-meeting of 3 hours delivered 3 month after end of treatment. The treatment follows a written manual for each meeting, which is given to the patients at the first meeting. The treatment and manual generally consists of the CBT elements of goal setting, psychoeducation, symptom registration, relaxation training, stress management, cognitive restructuring including individual designed behavioural experiments, response prevention, exposure, relapse prevention and homework.
For assessment of therapist competencies and adherence to the manual all group-CBT is video-recorded, and app. 5% of the essential material is assessed by a blinded rater. The therapist satisfaction with the group-CBT is rated with a questionnaire.
Usual care group
After the diagnostic assessment and after randomization, patients, referring doctors, and family physicians are informed in writing about the diagnosis, in the same way as for the patients randomized for group CBT. But here the patient has been put in the group for usual care with the offer to receive group CBT after 9 months. This group CBT will not be part of the primary research project (RCT), but it still takes place at The Liaison Psychiatric Clinic in Koege. Patients in the usual care group receive a letter stating that they have been randomized to 9-month with usual care, and in this period they continue usual care by their other health care providers. The patients in the usual care group are followed-up by questionnaires in the same way as the patients randomized to group CBT (see attached flowchart Figure 1). For details regarding timing and characteristics of treatment elements delivered in the group CBT see attached Figure 3.
Drop-outs
The patients that drop-out (after inclusion) are followed-up by questionnaires in the same way as the patients randomized to either group CBT or wait-list. This is done for intention to treat analysis. Also these patients are telephoned to examine reasons for drop-out. Attending in less than 8 of the total 13 treatment meetings is also defined as drop-out.
Predictors of outcome (start outcome measures and diagnostic assessment)
+/- Comorbid personality disorder (PD)
The degree of comorbid PD (number of significant traits (SCID-II)
The degree of comorbid PD (total score on PID-5)
The degree of comorbid PD (number of PD)
Duration of comorbid PD (age-18)
Type(s) of comorbid PD
Confounders (predictors/mediators) of outcome (pretreatment if not defined)
Patient activity in group-CBT (attendance in meetings (objective), homework (subjective))
Patient group-destructiveness (devaluation, expressed emotions)
Therapists CBT-competencies and adherence to the treatment manual. (A blinded rater on essential 5% of the total videorecordings)
Comorbid mental disorders (somatic symptom disorder (BDS), anxiety disorders (Panic disorder, GAD or OCD) or depressive disorders (HA is primary and dominant)
Duration of comorbid mental disorders
Degree of Health anxiety (WI-7)
Duration of HA (chronicity>3 years)
Illness perception
Global functioning (objective) GAF-F
Global functioning (subjective) SF-36
General psychopathology (SCL-90-R)
Comorbid somatic diseases (serious diseases excluded)
Distressing life events during group-CBT
Other treatments during group-CBT or follow up
Sociodemographic variables (age, sex, network (also partner), education, employment, economy, home, material goods, concurring litigation, interests/activities)
Expectations and satisfaction with the treatment along the full course of the treatment in the clinic
Definition of sufficient treatment: Sufficient treatment demands attendance to group-CBT meetings in at least 8 meetings out of 13 meetings in total. If less attendance occurs, the patient is defined as drop-out.
Perspectives
Health anxiety is costly for the social and health care system and cause suffering and disability for the patients. This study will hopefully form the basis of an effective and evidence-based group therapy for patients with impairing health anxiety and thereby help increase the patients' quality of life and functional level and reduce social and health care costs. As more general practitioners are trained in cognitive behaviour therapy and Shared Care is introduced into primary care, elements of this developed group-CBT can be used for prevention and treatment of patients with lesser degree of HA in primary care.
Time frame
Recruitment of patients for the pilot-study will begin primo 2013. The 2 pilot-groups will be conducted in 2013. Patient inclusion to the RCT will begin primo 2014 and end in 2015 depending on referrals. The rate of referrals is expected to be 4 patients a month. Treatment groups are expected to run during a course of approx. 2 years. In the same period the PhD student will participate in teaching, training and courses. In 2016 the last patients will be followed up, and data will be analyzed and interpreted, and papers to relevant journals will be written. Please see the enclosed time schedule for the three-year PhD program (Figure 4).
The 2 years follow-up assessment (prospective cohort study) will not be a part of the PhD study, but MS is responsible for this study also. The responsibility of the RCT study of cost-effectiveness and illness perception lies with other researchers mentioned as collaborators and named specifically under the relevant planned articles.
DATA ANALYSIS
The results will be analysed and published according to the Consort statement [58].
The power calculation is based on change in Whiteley-7 scores (scale 0-100). An uncertainty of the change in Whiteley-7 scores of 25 is estimated from a pilot study and other randomised controlled trials [11,15,61]. A sample size of 70 was estimated to provide 84.9% power, at the 5% significance level, to find a difference in mean reductions in Whiteley-7 scores from baseline to 6 month follow-up of the size 20 and allowing for a drop-out rate of 20% in both groups.
Relevant statistical methods for group comparison will be applied, including multiple regression. In the data analysis, both "intention-to-treat" analyses and analyses of the clinic's existing data will be performed. There will be applied multiple imputations of missing outcomes. A statistician is involved in the study.
The study will also be registered in an international database for clinical trials (www.clinicaltrials.gov))
SUPERVISORS
COLLABORATORS
Inclusion Criteria:
Exclusion Criteria:
Koege, 4600, Denmark