Morita therapy for anxiety disorders in adults

link管理

链接快照平台

输入网页链接，自动生成快照
标签化管理网页链接

Summary of findings 2

Morita therapy plus pharmacological therapy compared with pharmacological therapy alone for anxiety disorders

Morita therapy plus pharmacological therapy compared with pharmacological therapy alone for anxiety disorders
Patient or population: people with anxiety disorders Settings: inpatient Intervention: Morita therapy plus pharmacological therapy Comparison: pharmacological therapy
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Pharmacological therapy alone	Morita therapy plus pharmacological therapy
Global state responders	See comment	See comment	Not estimable	‐	See comment	One OCD study reported global state, but data were incomplete and unusable
Acceptability drop‐outs for any reason	33 per 1000	58 per 1000 (15 to 220)	RR 1.76 (0.47 to 6.67)	172 (2 studies)	⊕⊝⊝⊝ very low ^1,2	2 OCD studies. Short term (up to 12 weeks post‐treatment) Insufficient information on drop‐outs for adverse effects
Measures of anxiety overall mean endpoint score of HAMA (high = poor)	‐	The mean measure of anxiety in the intervention groups was 2.47 lower (4.78 to 0.16 lower)	‐	103 (1 study)	⊕⊝⊝⊝ very low ^2,3	Short term (up to 12 weeks post‐treatment)
Measures of depression overall mean endpoint score of HAMD (high = poor)	‐	The mean measure of depression in the intervention groups was 2.3 lower (4.19 to 0.41 lower)	‐	56 (1 study)	⊕⊝⊝⊝ very low ^1,2	Short term (up to 12 weeks post‐treatment)
Measures of obsession and compulsion overall mean endpoint score of Y‐BOCS (high = poor)	‐	The mean measure of obsession and compulsion in the intervention groups was 0.47 standard deviations lower (0.78 to 0.15 lower)	‐	159 (2 studies)	⊕⊝⊝⊝ very low ^1,2	Short term (up to 12 weeks post‐treatment) SMD ‐0.47 (95% CI ‐0.78 to ‐0.15)
Quality of life measures ‐ endpoint data or change data	See comment	See comment	Not estimable	‐	See comment	No information
Measures of functioning ‐ endpoint data or change data	See comment	See comment	Not estimable	‐	See comment	No information
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; HAMA: Hamilton Rating Scale for Anxiety; HAMD: Hamilton Rating Scale for Depression; OCD: obsessive‐compulsive disorder; RR: risk ratio; SMD: standardised mean difference; Y‐BOCS: Yale‐Brown Obsessive Compulsive Scale.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

Open in a separate window

¹ High risk of selective reporting bias. Unclear risk of selection bias, detection bias and attrition bias. Lack of information on therapists' qualifications and treatment fidelity.
² Total sample size was lower than the calculated optimal information size (OIS). Total number of events was fewer than 300. Wide 95% confidence interval.
³ Unclear risk of selection bias, detection bias, selective reporting bias and attrition bias. Lack of information on therapists' qualifications and treatment fidelity.

Background

Description of the condition

Anxiety disorders are one of the most common classes of mental disorders, with prevalence in the range of 2.4% to 18.2% (inter‐quartile range percentiles, 5.8% to 8.8%) across countries ( Demyttenaere 2004 ). Prevalence of anxiety disorders is 5.3% in Japan ( Demyttenaere 2004 ), and 5.6% in China ( Phillips 2009 ). According to the Diagnostic and Statistical Manual of Mental Disorders ‐ fourth edition (DSM‐IV), anxiety disorders include panic disorder (PD), obsessive‐compulsive disorder (OCD), post‐traumatic stress disorder (PTSD), generalised anxiety disorder (GAD), social phobia, agoraphobia, specific phobia and anxiety disorder not otherwise specified ( APA 2000 ). The International Classification of Disease (ICD‐10) categorises phobic anxiety disorders (including agoraphobia, social phobia, etc.), other anxiety disorders (including PD, GAD, anxiety disorder unspecified, etc.), OCD, reaction to severe stress and adjustment disorders (including PTSD, etc.) as neurotic, stress‐related and somatoform disorders ( WHO 1992 ). The Chinese Classification of Mental Disorder (CCMD‐3) is widely used in China and references DSM‐IV and ICD‐10. The classification of CCMD‐3 is similar to ICD‐10, although it does isolate stress‐related disorders (including PTSD) from neurotic disorders ( CSP 2001 ).

In DSM‐IV, social phobia is characterised as an excessive or unreasonable fear of social situations; specific phobia as an excessive or unreasonable fear of a specific object or situation; agoraphobia as anxiety about having an unexpected panic attack in places or situations from which escape might be difficult; PD as recurrent unexpected panic attacks and persistent concern about having additional attacks; GAD as excessive anxiety and worry about a number of events or activities; OCD as recurrent and persistent thoughts and repetitive behaviours that cause marked anxiety or distress; PTSD as a delayed or protracted response to a traumatic experience including persistent re‐experience, avoidance of stimuli and increased arousal. Description of characteristics and diagnosis of these anxiety disorders are similar in ICD‐10 and CCMD‐3.

Co‐morbidities are common and include depressive disorders or substance abuse. Approximately 75% of people with an anxiety disorder have their first episode by 21.5 years of age ( Kessler 2005a ; Kessler 2005b ). Anxiety disorders are commonly associated with impaired quality of life compared with non‐clinical controls ( Olatunji 2007 ). Both pharmacotherapy and psychotherapy are used in clinical practice for the treatment of anxiety disorders.

Description of the intervention

Both psychological and pharmacological treatments are used for anxiety disorders. Psychological therapies include cognitive behavioural therapy (CBT), applied relaxation, psychoanalysis and interpersonal psychotherapy while pharmacological treatments include anxiolytics (such as benzodiazepines, 5‐HT _1A receptor agonist, etc.) and antidepressants (such as selective serotonin reuptake inhibitors (SSRIs), serotonin and noradrenergic reuptake inhibitors (SNRIs), noradrenergic and specific serotonergic antidepressants (NaSSAs), tricyclic antidepressants (TCAs), monoamine oxidase inhibitors (MAOIs), etc.) ( Jiang 2011 ; Stein 2009 ). National Institute for Health and Care Excellence (NICE) guidelines recommend CBT or SSRIs as first‐line treatment for GAD, PD, agoraphobia and OCD; trauma‐focused CBT as priority for PTSD and medications not recommended for a routine first‐line treatment; and individual CBT preferred before drug treatments such as SSRIs for social phobias (www.nice.org.uk).

An alternative treatment of anxiety disorders is Morita therapy, founded in 1919 by Shoma Morita (1874 to 1938), which is a systematic psychological therapy based on Eastern philosophy ( Kitanishi 1995 ; Miura 1970 ; Takeda 1964 ). The aim of the therapy is to help people with anxiety to live productive lives and complete life goals by encouraging them to accept anxiety as a natural feeling, while at the same time engaging them in constructive behaviours ( Lu 1995 ; Takeda 1964 ). Anxiety feelings and neurotic symptoms are likely to fade when people learn to accept that anxiety is natural and capitalise their characteristics and potentials to pursue life goals and make achievements. Morita therapy is used mainly in Japan and China. In China, there are 66 Morita therapy institutes or units for inpatient/outpatient/self help groups or for research ( Wen 2001 ). However, the therapy has also spread to western countries including US, Canada, Australia and England. It is provided mainly by doctors and psychologists in both inpatient and outpatient settings.

The theory and treatment methods of Morita therapy have been modified since its founding. Its original application in anxiety disorders has been expanded for use in other mental disorders and the treatment duration is now shortened to six weeks, although the longer four phases are still used (see How the intervention might work , which explains the four phases) ( Lu 1995 ). Several studies have reported the efficacy of Morita therapy for people with anxiety disorders with varying foci of treatment outcomes ( Kitanishi 1995 ; Nakamura 2004 ; Nakamura 2006 ; Reynolds 1980 ; Zhang 1994 ; Zhang 1995 ).

How the intervention might work

Morita therapy attempts to lead people away from preoccupation with anxiety and neurotic symptoms and help them to live productive lives and achieve life goals through the acceptance that anxiety is a natural feeling while engaging the person in constructive behaviours through the following four phases ( He 2007 ; Takeda 1964 ).

Phase 1: bed rest in isolation for seven days, during which people remain in their room all day. Activity during this phase is restricted to meals, a morning wash and going to the toilet, with no access to reading, radio, television or any other form of entertainment. The aim of this phase is to minimise conflicting thoughts and focus intrapsychically. At the end of the stage, people should ideally start working voluntarily.
Phase 2: light work for four to seven days. During this phase, the person: a) initiates, under therapist's guidance, graded activity and work needed in daily living at the hospital; b) writes a daily diary documenting what they have done. The therapist then reads and comments on the diary and return it to the person (diary guidance); c) has interviews two or three times a week with their therapist focusing on the daily activities rather than the symptoms of the disorder (strategic inattention to symptoms, contingency management). The theory behind this phase is that increased physical work enables people to break down conflicting thoughts and feelings and minimise intrapsychic‐focusing by observation and attentiveness to nature and outside reality.
Phase 3: work for one to two months with gradual engagement in activities such as gardening, caring for animals and cooking, eventually doing these together with other people. This phase is a continuation and enhancement of the second phase with the theory being that it prepares people for normal life by coping with necessary tasks and socialising with others.
Phase 4: preparation for normal daily living over one to four weeks, which may include commuting to work or school from the ward ( Takeda 1964 ). All restrictions are removed. People may be aware of their change and find themselves functioning better in spite of anxiety and neurotic symptoms.

Classic Morita therapy is conducted as described above in Morita therapy institutes and inpatient settings. Over the years it was modified to suit the changing economy and society. Modified Morita therapies are more practical in medical settings as well as more flexible for people who wish to receive therapy while attending school and going to work. The therapy can also be carried out in outpatient clinics. In such cases phase one is modified to learning theories of Morita therapy, instead of bed rest; and phases two, three and four are conducted by clinical visits once or twice a week depending on the need and individual settings ( Nie 2008 ; Zhang 2003 ). Diary keeping is common in outpatient treatments. People are not supervised everyday by care providers but report to them regularly and get feedback on clinical visits.

Why it is important to do this review

Morita therapy is mainly used as an alternative therapy for people with anxiety disorders in China and other Asian countries. Today, it is also used for other mental disorders and has been introduced to the West. The history of Morita therapy for anxiety disorders in particular is a long one ( He 2007 ; Kitanishi 1995 ; Li 2008 ; Reynolds 1980 ; Takeda 1964 ); however, varying foci of treatment outcomes have been reported and to date there has been no systematic review to investigate the strength of evidence for Morita therapy in anxiety disorders.

Objectives

To assess the effects of Morita therapy compared with pharmacological therapy, other psychological therapy, no intervention or wait list for anxiety disorders in adults.

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCTs). We placed no restrictions on language or publication status.

We excluded quasi‐randomised studies, such as those allocating using alternate days of the week.

We identified no cross‐over trials for this version of the review. Should we identify any in future updates of this review, we will only include the first randomisation period.

We planned to include cluster RCTs in this review; however, we identified none.

Types of participants

Adults (18 years old or over) diagnosed with anxiety disorders (GAD, PD, OCD, PTSD, social phobia, agoraphobia or specific phobia), irrespective of gender, race or nationality, were eligible. Diagnostic systems included the DSM‐IV ( APA 1994 ), ICD‐10 ( WHO 1992 ), CCMD‐3 and other validated diagnostic instruments or earlier version of those. We placed no restrictions on setting.

We excluded studies in which anxiety was a secondary symptom of a different disorder (e.g. depression or other psychiatric diagnoses) ( Li 2010 ). We also excluded co‐morbid disorders (physical and psychological).

Types of interventions

Experimental intervention

For this review, we defined Morita therapy as any care practice defined as Morita therapy by the carers and involving at least two of the four phases described in How the intervention might work . There is debate among the international Morita therapy community as to how many phases are essential for Morita therapy. We set what we believe to be the broad inclusion criteria of two or more phases of Morita therapy instead of all four phases in order to retrieve a maximal number of trials.

In situations where both two‐ and four‐phase Morita therapy were compared with the same control intervention, we pooled the different types of Morita therapy ( He 2007 ; Li 2010 ).

Morita therapy can be delivered as monotherapy or as an adjunct to any of the pharmacological or psychological therapies listed in comparator interventions.

Comparator intervention

Comparators may be one or a combination of:

pharmacological therapy (e.g. anxiolytics, antidepressants, adrenergic blocking agents, etc.);
other psychological therapy (e.g. CBT, insight‐oriented psychotherapy, psychoanalysis, counselling, etc.). Because the range of potential comparators is large, it may not be reasonable to assume homogeneity. If all or most studies in this group compared Morita therapy to specific psychological therapy, then we described results as a comparison of Morita therapy to specific psychological therapy (not psychological therapy in general);
no intervention or wait list.

Types of outcome measures

Primary outcomes

1. Global state ‐ clinical response. Criteria of response was scored using the Clinical Global Impression (CGI) scale ( Guy 1976 ), a widely used global outcome measure, or defined by individual studies.

2. Acceptability: drop‐outs for any reason as proxy measure of total acceptability.

Secondary outcomes

3. Acceptability: drop‐outs for adverse effects

Continuous measures

4. Any measure of anxiety, such as Hamilton Anxiety Scale (HAMA) ( Hamilton 1959 ), State‐Trait Anxiety Inventory (STAI) ( Spielberger 1977 ), Beck Anxiety Inventory (BAI) ( Beck 1988 ), Zung Self‐rating Anxiety Scale (SAS) ( Zung 1971 ), Hospital Anxiety and Depression Scale (HADS) ( Zigmond 1983 ), and Symptom Checklist 90 (SCL‐90) ( Derogatis 1975 ). Some measures are designed for specific type of anxiety disorders, such as Liebowitz social anxiety scale (LSAS) ( Liebowitz 1987 ) for social phobia, Generalized Anxiety Disorder‐7 (GAD‐7) ( Spitzer 2006 ), and Generalized Anxiety Disorder‐2 (GAD‐2) for GAD ( In 2007 ).

5. Any measure of depression, such as Hamilton Depression Scale (HAMD) ( Hamilton 1960 ), Zung Self‐rating Depression Scale ( Zung 1965 ), Hospital Anxiety and Depression Scale (HADS) ( Zigmond 1983 ), and SCL‐90.

6. Any measure of obsession and compulsion, such as Yale‐Brown Obsessive Compulsive Scale (Y‐BOCS) ( Goodman 1989 ).

7. Quality of life measures, such as Activity of Daily Living Scale (ADL) ( Lawton 1969 ).

8. Any measure of functioning, such as the Sheehan Disability Scale, which includes subscales to assess work‐, social‐ and family‐related impairment ( Sheehan 1996 ).

Timing of outcome assessment

We grouped outcomes as short term (up to 12 weeks), medium term (13 to 52 weeks) and long term (over 52 weeks) to observe possible efficacy changes after treatment ( Li 2010 ). We classed the short term as our primary time point and used this time point in the 'Summary of findings' table.

Search methods for identification of studies

We used a number of sources to identify studies for possible inclusion in this review (see below). We did not limit studies to any particular language.

Electronic searches

The Cochrane, Depression, Anxiety and Neurosis Review Group's Specialised Register (CCDANCTR)

We searched the CCDANCTR‐Studies Register using the following terms:
Diagnosis = anxiety or anxious or *phobi* or panic or obsess* or compulsi* or OCD or post‐trauma* or "post trauma*" or posttrauma* or PTSD and Intervention = morita
We searched the CCDANCTR‐References Register to identify additional untagged/uncoded references using the following terms:
Keyword = anxiety or anxious or *phobi* or panic or obsess* or compulsi* or OCD or post‐trauma* or "post trauma*" or posttrauma* or PTSD and Free‐text = morita

Appendix 1 shows a description of the CCDANCTR. We ran searches to 7 December 2014, with no date or language restrictions.

Chinese biomedical databases

We searched the following Chinese biomedical databases to 7 December 2014 (search strategy in Figure 1 ).

An external file that holds a picture, illustration, etc. Object name is nCD008619-AFig-FIG01.jpg

Open in a separate window

Search phrases in Chinese.

Chongqing VIP Database (VIP).
Wanfang Database.
China Hospital Knowledge Database (CHKD).
Chinese Biology Medicine Database (CBM‐disc).

Searching other resources

Reference searching

We inspected the reference lists of all identified studies (see 'Methods' below) for additional trials. We searched studies of some experts in the field such as Prof. Richard Spates, Prof. Ishu Ishiyama and Prof. David Richards for potential eligible trials.

Personal contact

We contacted first authors of potential studies (studies that we had performed a full‐text inspection of) and some experts in the field for information regarding other published and unpublished trials.

Data collection and analysis

Selection of studies

Two review authors (CL and HW) independently inspected citations identified from the search. We identified potentially relevant reports by reviewing titles and abstracts and ordered full papers for assessment. If there were doubts about a certain report, we re‐assessed the full article and discussed whether it should be included. If we could not reach a consensus, we consulted a third review author to resolve the disagreement. When some reports were lacking crucial information, we added them to the reports awaiting assessment and contacted the authors of the reports for clarification ( Li 2010 ).

Data extraction and management

Two review authors (CL and HW) extracted the following data independently:

design (allocation method, blindness, duration of the intervention);
participant characteristics (diagnosis, number of participants, age, gender, duration of illness, inpatient/outpatient setting);
intervention details (phases of Morita therapy, co‐intervention if any, comparator, number of participants per arm, therapist qualifications, treatment fidelity, sponsorship);
outcome measures of interest (e.g. clinical response, drop‐outs for any reason).

Extraction

Two review authors (CL and HW) independently registered included studies on Characteristics of included studies tables and extracted data onto standard data extraction sheets. We resolved any disagreement by discussion, documented and reported decisions, and contacted authors of studies for clarification ( Li 2010 ).

Management

We extracted data onto standard, simple forms ( Li 2010 ).

Scale‐derived data

We included continuous data from rating scales only if the measuring instrument had been described in a peer‐reviewed journal ( Marshall 2000 ), and the instrument was either self report or completed by an independent rater or relative (not the therapist) ( Li 2010 ).

Planned treatment comparisons

Morita therapy versus pharmacological therapy.
Morita therapy versus other psychological therapy.
Morita therapy versus no intervention or wait list.
Morita therapy plus pharmacological therapy versus pharmacological therapy alone.
Morita therapy plus other psychological therapy versus other psychological therapy alone.

Due to lack of available data, we were only able to perform comparisons of Morita therapy versus pharmacological therapy and Morita therapy plus pharmacological therapy versus pharmacological therapy alone from the list above.

Assessment of risk of bias in included studies

Two review authors (CL and HW) independently assessed risk of bias using the 'Risk of bias' tool described in the Cochrane Handbook for Systematic Reviews of Interventions ( Higgins 2011 ). This tool encourages consideration of how the sequence was generated, how allocation was concealed, the integrity of blinding at outcome, the completeness of outcome data, selective reporting and other biases.

For this type of psychotherapy learnt by apprenticeship, therapist qualifications and treatment fidelity were respectively assessed in the evaluation of risk of bias.

Descriptions of what was reported to have happened were provided in each study in the 'Risk of bias' table. We made decisions of risk of bias (low, unclear, high) on each domain within and across studies.

We evaluated

studies with low risk of bias for all key domains as low risk of bias; studies with unclear risk of bias for one or more key domains as unclear risk of bias and studies with high risk of bias for one or more key domains as high risk of bias. When judging across studies, low risk of bias requires that most information is from studies at low risk of bias; unclear risk of bias requires that most information is from studies at low or unclear risk of bias and high risk of bias requires that the proportion of information from studies at high risk of bias is sufficient to affect the interpretation of results ( Higgins 2011 ).

Measures of treatment effect

Binary data

For binary outcomes, we calculated a standard estimation of the fixed‐effect risk ratio (RR) and 95% confidence interval (CI). For statistically significant results, we calculated the number needed to treat for an additional beneficial outcome (NNTB) or additional harmful outcome (NNTH), and 95% CI using Visual Rx (www.nntonline.net/), taking account of the event rate in the control group ( Li 2010 ).

Continuous data

Summary statistic

For continuous outcomes, anticipating different measures across studies, we pooled data by calculating the fixed‐effect standardised mean difference (SMD) and 95% CIs ( Li 2010 ).

Endpoint versus change data

We preferred to use scale endpoint data, which typically cannot have negative values and are easier to interpret from a clinical point of view. Change data can be problematic if baseline and final measurements are reported for different numbers of participants due to missed visits and study withdrawals ( Higgins 2011 ). Only if endpoint data were unavailable, would we have used change data ( Li 2010 ).

Skewed data

Continuous data on clinical and social outcomes are often not normally distributed. To avoid the hazard of applying parametric tests to non‐parametric data, we applied the following standards to all data before inclusion ( Li 2010 ):

means and standard deviations (SD) were reported in the paper or obtainable from the authors;
when a scale started from the finite number zero, the SD, when multiplied by two, was less than the mean (as otherwise the mean was unlikely to be an appropriate measure of the centre of the distribution ( Altman 1996 ));
if a scale started from a positive value (such as the Positive and Negative Syndrome Scale (PANSS), which can have values from 30 to 210), we modified the calculation described in the bullet point above to take the scale starting point into account. In these cases, skew was present if 2 × SD > (S ‐ S _min ), where S was the mean score and S _min was the minimum score.

Endpoint scores on scales often have a finite start and endpoint, and these rules can be applied. Where continuous data are presented on a scale that includes a possibility of negative values (such as change data), it is difficult to determine whether data are skewed or not. We entered skewed data from studies of fewer than 200 participants in additional tables rather than into an analysis. Skewed data posed less of a problem when considering the means if the sample size was large and were entered into syntheses ( Higgins 2011 ).

Unit of analysis issues

Cluster‐randomised trials

Studies increasingly employ 'cluster randomisation' (such as randomisation by clinician or practice) but analysis and pooling of clustered data poses problems. First, authors often do not account for intraclass correlation in clustered studies, leading to a 'unit of analysis' error ( Divine 1992 ), whereby P values are spuriously low, CIs unduly narrow and statistical significance overestimated. This causes type I errors ( Bland 1997 ; Gulliford 1999 ).

There were no cluster‐randomised trials among the included studies. In future updates, if we find a cluster‐randomised trial and clustering is not accounted for in the primary study, we will present data in a table and indicate the presence of a probable unit of analysis error. We will seek to contact the first author of the study to obtain the intraclass correlation coefficient of their clustered data and to adjust for this by using accepted methods ( Gulliford 1999 ); if clustering is incorporated into the analysis of primary studies, we will present these data as if from a non‐cluster‐randomised study, but will adjust for the clustering effect.

If a cluster study has been appropriately analysed, taking into account the intraclass correlation coefficient and relevant data documented in the report, synthesis with other studies is possible using the generic inverse variance technique ( Li 2010 ).

Studies with multiple treatment groups

We found no studies with multiple treatment groups. In future updates, if we find a study that involves more than two treatment arms, if relevant, we will present the additional treatment arms in comparisons although we will use the control once in a particular meta‐analysis. Where the additional treatment arms are not relevant, we will not reproduce these data ( Li 2010 ).

Cross‐over trials

Cross‐over trials make comparisons on an individual level, which require relatively stable conditions of interest and short‐lived treatment effects ( Higgins 2011 ). A major concern of cross‐over trials is carryover effect. The effect of the intervention (e.g. pharmacological, psychological or physiological) in the first period may influence the effect of other intervention in the subsequent period ( Elbourne 2002 ).

In this review, Morita therapy, pharmacological therapy or other psychological therapy were target interventions. Initial conditions of participants would change after first period treatment with or without wash‐out. People with anxiety disorders were not ideal stable‐conditioned population. We found no cross‐over trials for this review. In future updates, if we identify cross‐over trials for inclusion, we will only use first‐period data ( Li 2010 ).

Dealing with missing data

Overall loss of credibility

In terms of loss of follow‐up, if more than 40% of data were unaccounted for by 12 weeks or longer we did not reproduce these data or use them in analyses ( Li 2010 ). If studies had more than 40% of missing data, the credibility will be very problematic ( Xia 2009 ).

Binary data

In the case where attrition for a binary outcome was between 0% and 40% and outcomes of these people were described, we included these data as reported. Where these data were not clearly described, we assumed the worst primary outcome, and rate of adverse effects similar to people who did continue to have their data recorded ( Li 2010 ).

Continuous data

In the case where attrition for a continuous outcome was between 0% and 40% and completer‐only data were reported, we reproduced these ( Li 2010 ).

Assessment of heterogeneity

Clinical heterogeneity

We inspected all included studies from clinical perspectives for heterogeneity, such as participant resources, intervention methods, etc. ( Li 2010 ).

Statistical heterogeneity

Visual inspection

We visually inspected overlap of CI values to investigate the possibility of statistical heterogeneity. In the next update, if forest plots are possible, we will also inspect the graphs to investigate statistical heterogeneity ( Li 2010 ).

Employing the Chi ² test and I ² statistic

We used a Chi ² test with a P value of 0.10 to assess heterogeneity. We also used the I ² statistic, which provides an estimate of the percentage of inconsistency thought to be due to chance, with 30% to 60% representing moderate heterogeneity, 50% to 90% substantial heterogeneity and 75% to 100% considerable heterogeneity ( Higgins 2011 ).

Assessment of reporting biases

Reporting biases arise when the dissemination of research findings is influenced by the nature and direction of results. These are described in Section 10.1 of the Cochrane Handbook for Systematic Reviews of Interventions ( Higgins 2011 ). We are aware that funnel plots may be useful in investigating reporting biases. We did not use funnel plots for outcomes because there were fewer than 11 studies and all studies were of similar sizes. In future updates, if funnel plots are possible, we will seek statistical advice in their interpretation ( Li 2010 ).

Data synthesis

We employed a fixed‐effect model to calculate RR for dichotomous outcomes and SMD for continuous outcomes for analyses; if heterogeneity of intervention effects existed during clinical inspection or statistical calculation (I ² ≥ 50%) ( Higgins 2011 ), we would have used a random‐effects model. The random‐effects method incorporates an assumption that the different studies are estimating different, yet related, intervention effects. However, random‐effects put added weight onto the smaller of the studies (i.e. those trials that are most vulnerable to bias).

If data were considerably heterogeneous (I ² ≥ 75%) ( Higgins 2011 ), we checked that data had been correctly extracted and entered, and that we had made no unit of analysis errors. If the high levels of heterogeneity remained, particularly if there was inconsistency in the direction of effect and no clear reasons for heterogeneity, we did not undertake a meta‐analysis. We presented the final data in narrative form at that point.

We presented skewed data, incomplete data and non‐quantitative data descriptively in narrative forms and did not include them in meta‐analyses ( Li 2010 ).

Subgroup analysis and investigation of heterogeneity

Subgroup analyses may be done as a means of investigating heterogeneous results, or to answer specific questions about particular participant groups ( Higgins 2011 ). Had we identified a sufficient number of studies, we would have undertaken subgroup analyses to assess the degree to which methodological differences between trials might have systematically influenced differences observed in the primary treatment outcomes ( Li 2010 ).

We did not perform subgroup analyses because there were not enough studies. In consideration of the possibility of differential effect of Morita therapy for different conditions, we plan to perform three subgroup analyses in future updates if there are sufficient studies:

subtype of anxiety disorders (GAD, PD, OCD, PTSD, social phobia, agoraphobia or specific phobia);
source of participants (inpatient/outpatient);
type of Morita therapy (classic/modified, see How the intervention might work for details of classic and modified Morita therapy).

Sensitivity analysis

We planned to perform sensitivity analyses to investigate the effect of decisions made in the review process. However, there were insufficient data included in the review, so we did not conduct sensitivity analyses.

In future updates, if there are sufficient studies, we will perform two sensitivity analyses:

trials with unclear randomisation (we will remove trials from the analysis in a sensitivity analysis if they are described as 'double blind' but the methods of randomisation are unclear);
completer‐only data (we will remove completer‐only data for the primary outcome, but will report our data, assumption included, in the final analysis).

'Summary of findings' tables

We used 'Summary of findings' tables to summarise main outcomes (global state ‐ clinical response, acceptability, measure of symptoms, quality of life measures and measure of functioning). 'Summary of findings' tables present findings for patient‐important outcomes, including the amount of evidence, typical absolute risks for the different methods, estimates of relative effect, a depiction of the quality of the body of evidence, comments and footnotes ( Higgins 2011 ). The assessment of the quality of the body of evidence followed the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) system, which combines consideration of risk of bias, directness, heterogeneity, precision and publication bias ( GRADE Working Group 2004 ; Guyatt 2008a ; Guyatt 2008b ; Schünemann 2006 ).

Results

Description of studies

Results of the search

We initially identified 664 reports. After removing duplicates, 300 reports remained. Judging from titles and abstracts of these identified reports, we deemed 32 as potentially relevant and ordered full papers for assessment. We included seven studies in the systematic review, of which six were included in meta‐analyses.

For flowchart depicting the selection of studies from the search results, see Figure 2 .

An external file that holds a picture, illustration, etc. Object name is nCD008619-AFig-FIG02.jpg

Open in a separate window

Study flow diagram.

Included studies

We were able to include seven studies, with 449 participants. Attempts to contact the first authors of two included studies were unsuccessful ( Nie 2008 ; Shi 2011 ). See Characteristics of included studies table for full details. Only two comparisons were possible: Morita therapy versus pharmacological therapy; and Morita therapy plus pharmacological therapy versus pharmacological therapy.

Design

There were no cluster‐randomised trials, cross‐over trials or trials with multiple treatment groups. All trials were parallel‐group RCTs. However, one trial had double the participants in the experimental group than the control group (24 participants versus 12 participants) ( Zhang 2003 ), and another trial only recruited women ( Lu 2008a ).

Sample sizes

The mean sample size of included studies was 64, ranging from 36 to 108. All included studies had small sample sizes. In the comparison of Morita therapy versus pharmacological therapy alone (comparison 1), one study was well balanced with a sample size of 39 ( Nie 2008 ), and the other study had a sample size of 36, which was not well balanced between groups (experimental/control = 24/12) ( Zhang 2003 ). The mean sample sizes of experimental group (Morita therapy) was 43 and control group (pharmacotherapy alone) was 32. In the comparison of Morita therapy plus pharmacological therapy versus pharmacological therapy alone (comparison 2), all studies were well balanced in sample size. The mean sample size of the experimental group (Morita therapy plus pharmacotherapy) was 182 and control group (pharmacotherapy alone) was 192.

Setting

All studies were conducted in hospitals in the People's Republic of China between about 1994 and 2007. Two studies recruited outpatients while the remaining studies recruited inpatients.

Participants

All participants were diagnosed with anxiety disorders (social phobia in two studies, OCD in three studies, GAD in one study and anxiety neurosis in one study). Six studies applied CCMD‐3 as the diagnostic instrument, while one used the Chinese Classification of Mental Disorder, Second Edition, Revised (CCMD‐2‐R) ( Mei 2000 ). Two studies added severity criteria ( Lu 2008b ; Shi 2005 ). Only two studies specifically reported exclusion of other severe mental disorders, general medical conditions, drug/alcohol abuse and pregnancy/lactation ( Shi 2005 ; Shi 2011 ). Overall, participants ranged from 16 to 60 years old, and the mean age was 33 years (SD 8). Six of the seven trials included both men and women, with one study using an all women sample ( Lu 2008a ). One study did not report the gender ratio of the control group ( Zhang 2003 ). Duration of illness varied between studies (from three months to 20 years, mostly on a yearly basis). All participants were recruited by researchers when receiving treatments in the hospital. Unfortunately, none of the authors of included studies provided further information about the recruitment.

Interventions

Most inpatient trials completed the four‐phase Morita therapy, while the majority of outpatient trials modified Morita therapy by clinic interviews (mostly on a weekly basis) and home practice afterwards. For specific details for each study, see the Characteristics of included studies table. According to our categories of duration of treatment, there were three short‐term trials (up to 12 weeks), three medium‐term trials (13 to 52 weeks) and one long‐term trial (more than 52 weeks). Duration of Morita therapy varied from six weeks ( Lu 2008a ) to 12 months ( Shi 2011 ).

Comparisons

Comparison 1. Morita therapy versus pharmacological therapy

Two outpatient social phobia studies (75 participants) compared Morita therapy versus pharmacological therapy ( Nie 2008 ; Zhang 2003 ). They conducted modified Morita therapy using regular outpatient interviews (mostly once a week) and home practices afterwards. Phase 1 (bed rest) was modified as understanding principals of Morita therapy and accepting anxiety symptoms. One study continued phase 2 (light work), phase 3 (heavier work) and phase 4 (preparation for normal daily living) ( Zhang 2003 ), and the other study stopped the therapy at phase 3 ( Nie 2008 ). The two studies used alprazolam ( Zhang 2003 ) and clonazepam ( Nie 2008 ) as their control treatment. These anxiolytics were all prescribed at clinical dosages.

Comparison 2. Morita therapy plus pharmacological therapy versus pharmacological therapy alone

Five inpatient studies compared Morita therapy plus pharmacological therapy with pharmacological therapy alone ( Lu 2008a ; Lu 2008b ; Mei 2000 ; Shi 2005 ; Shi 2011 ). Three of the studies focused on OCD ( Mei 2000 ; Shi 2005 ; Shi 2011 ), one on GAD ( Lu 2008b ), and one on anxiety neurosis ( Lu 2008a ). In each of the studies, Morita therapy consisted of four phases. In one study, phase 1 was modified to learning principals of Morita therapy and accepting anxiety symptoms ( Shi 2005 ), while the other studies used the traditional four‐phase Morita therapy described in Types of interventions . The pharmacological therapies prescribed were citalopram, buspirone, clonazepam, fluoxetine, chlorimipramine, or paroxetine at clinical dosages. The experimental and control groups in each study were prescribed the same anxiolytics. In one study, two types of anxiolytics were prescribed to both the experimental and control groups ( Lu 2008b ). One study did not provide information regarding the type of anxiolytics used ( Lu 2008a ).

Outcomes

Two social phobia studies reported global state (primary outcome) ( Nie 2008 ; Zhang 2003 ). One used a four‐level evaluation method, dividing participants into cured, much improved, improved and no change ( Zhang 2003 ). The other study introduced a three‐level evaluation method, which included cured, improved and no change ( Nie 2008 ). Only two studies reported data for drop‐outs for any reason (primary outcome) ( Mei 2000 ; Shi 2011 ). Only two studies provided limited description of adverse effects ( Lu 2008b ; Shi 2005 ). Three studies, including two on OCD and one on GAD measured symptoms of anxiety using the HAMA ( Lu 2008b ; Mei 2000 ; Shi 2011 ). One social phobia study ( Nie 2008 ), and one anxiety neurosis (according to CCMD‐3, anxiety neurosis contains PD and GAD) study ( Lu 2008a ), used SCL‐90, among which, the social phobia study only reported four dimensions of SCL‐90, namely somatisation, depression, anxiety and phobic anxiety ( Nie 2008 ). One OCD study used the HAMD ( Shi 2005 ). Three OCD studies measured Y‐BOCS for obsession and compulsion ( Mei 2000 ; Shi 2005 ; Shi 2011 ). Two OCD studies used measures of functioning from Social Disability Screen Scale (SDSS; Shen 1985 ) ( Mei 2000 ; Shi 2011 ).

Unusable data

Two studies presented some findings by inexact P value, providing no other detailed statistical data ( Shi 2005 ; Shi 2011 ). One study reported scores of Y‐BOCS into two parts, which were obsessive thoughts and compulsive behaviours, rather than total score ( Mei 2000 ). Since data were skewed and incomplete, we reported them narratively. One of the studies reported medians of Clinical Global Impression Scale ‐ Severity of Illness (CGI‐SI) as one of the outcomes ( Shi 2005 ). Since means and SDs of CGI‐SI were not reported or obtainable from the authors, we did not use these but instead reported them narratively. One study did not provide direct or indirect description of drop‐outs, and the first author gave no further information regarding this problem after personal contact ( Lu 2008a ). We did not use data with no information of number of participants. Some data from studies were skewed and we also reported them narratively ( Lu 2008b ; Mei 2000 ; Shi 2011 ).

Missing outcomes

We found there were insufficient data on global state and drop‐outs for any reason (primary outcome). Data of drop‐outs for adverse effects and quality of life measures were missing. Measure of functioning and description of adverse effect were insufficient.

Excluded studies

We identified 32 candidate studies from 664 reports. Among these, we excluded six studies for not meeting inclusion criteria of intervention, five for incorrect population, three for improper comparison and 11 for incorrect study type. See Characteristics of excluded studies table for details.

Ongoing studies

We identified no ongoing trials.

Studies awaiting classification

There were no trials awaiting classification.

Risk of bias in included studies

The quality of reporting in studies was low. There was insufficient information on the methods of randomisation and blinding as well as attrition in included studies. Reporting of expected outcomes of interest was of low quality. Additionally, the studies provided no information of sponsorship, therapist qualifications or treatment fidelity.

Figure 3 and Figure 4 show a graphical summary of the risk of bias of individual studies. Four of the studies were at high risk of bias ( Lu 2008a ; Mei 2000 ; Nie 2008 ; Shi 2005 ), and we judged three studies at unclear risk of bias ( Lu 2008b ; Shi 2011 ; Zhang 2003 ). Judging across studies, both comparisons (Morita therapy versus pharmacological therapy and Morita therapy plus pharmacological therapy versus pharmacological therapy alone) were at high risk of bias. As a result, we rated these trials at high risk of bias. Therefore, any findings from these included studies should be interpreted carefully, since it is very likely to be biased to favour Morita therapy and overestimate the effect ( Juni 2001 ).

An external file that holds a picture, illustration, etc. Object name is nCD008619-AFig-FIG03.jpg

Open in a separate window

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

An external file that holds a picture, illustration, etc. Object name is nCD008619-AFig-FIG04.jpg

Open in a separate window

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Allocation

All trials were described as 'randomised' but few described the method used. By means of personal contact with first authors, we found out that two studies generated a series of random numbers by computer and assigned numbers to participants according to admission order ( Mei 2000 ; Shi 2005 ). Three authors could not recall which allocation method they had used ( Lu 2008a ; Lu 2008b ; Zhang 2003 ). One study applied 'minimisation' on allocation taking gender and residence of participants into account ( Shi 2011 ). There was insufficient information regarding allocation concealment in all studies.

Blinding

One study mentioned a single‐blind allocation in the study, yet provided insufficient information about the exact nature of the blinding and how this was conducted ( Zhang 2003 ). Another study described the study as open but gave no further details ( Lu 2008a ). Blinding of participants and personnel was assessed as being at high risk of bias for all studies, on the basis that Morita therapy is a psychological therapy intervention and so blinding is not possible. Two studies mentioned that outcomes were rated by trained, qualified physicians or psychotherapists, but neither specifically reported them as independent raters ( Lu 2008b ; Shi 2005 ). Five studies ( Lu 2008a ; Mei 2000 ; Nie 2008 ; Shi 2011 ; Zhang 2003 ) gave no description of whether or not independent raters were used. Blinding of outcome assessment was assessed as being at unclear risk of bias for all studies.

Incomplete outcome data

One study reported three drop‐outs at the time of follow‐up and did not include the data in the follow‐up analysis ( Mei 2000 ). We contacted the first author to ascertain whether the three drop‐outs were from the experimental group, they replied that this was the case. Another study reported two drop‐outs in the experimental group and three in the control group, and data of drop‐outs were not included in the statistical analysis ( Shi 2011 ). The other five studies provided no direct description of drop‐outs. Judging from 'n' values reported in 'Results' section, we became aware of the number of drop‐outs in four studies, but not in one study ( Lu 2008a ). After contacting the first author of the study, sufficient information was still not received ( Lu 2008a ).

Selective reporting

Studies were poorly reported and did not report on all of the outcomes of interest that we considered to be expected outcomes. No study protocols were available. Four studies did not report on all outcomes listed in the methods ( Lu 2008a ; Mei 2000 ; Nie 2008 ; Shi 2005 ), which was clear evidence of high risk of selective reporting bias.

Therapist qualification

The studies did not provide sufficient information on the credentials of the therapists. We rated the risk of bias for this domain as unclear for all the studies.

Treatment fidelity

Insufficient information was provided regarding whether treatment manuals were used. We rated the risk of bias for this domain as unclear for all the studies.

Other potential sources of bias

It was unclear from the study reports whether there were other potential areas for risk of bias.

Effects of interventions

See: Table 1 ; Table 2

Where data allowed, we reported outcomes at short term (up to 12 weeks), medium term (13 to 52 weeks) and long term (over 52 weeks). Table 1 and Table 2 summarise the main outcomes of each comparison.

Comparison 1. Morita therapy versus pharmacological therapy

We included two outpatient social phobia studies with 75 participants in this comparison ( Nie 2008 ; Zhang 2003 ). See Table 1 .