This chapter assesses the evidence on interventions for treatment for symptoms associated with, or that define, chronic multisymptom illness (CMI). The 38 studies and nine systematic reviews (described in 51 manuscripts) reviewed were found in the systematic search detailed in Chapter 3 for treatments of multiple physical symptoms associated with CMI. A narrative synthesis of the evidence is presented here by type of treatment: pharmacologic treatments and other biologic interventions, psychotherapies, mind–body approaches (including biofeedback, cognitive rehabilitation therapy, and complementary and alternative therapies), and exercise interventions. The committee looked for evidence of efficacy or effectiveness1 of interventions in alleviating symptoms and improving quality of life in all populations affected by multiple symptoms or syndromes similar to CMI. This chapter also includes the committee’s evaluation of the body of evidence on each treatment on the basis of the Agency for Healthcare Research and Quality (AHRQ) strength-of-evidence grading (Owens et al., 2010) and conclusions about the effectiveness of the evaluated treatments.
The committee did not limit its investigation to particular types of treatments but rather chose to evaluate all treatments on which there was evidence. The treatments considered varied widely and included nontraditional medicine, such as complementary and alternative medicine,
1Efficacy is benefit from an intervention under the best possible conditions, such as in a randomized controlled trial in which a selected sample of a specific population is carefully monitored by physicians, whereas effectiveness is benefit in real life as shown by many types of research studies (Agency for Healthcare Research and Quality, 2011).
psychologic therapies, and stress-management techniques. The committee also considered various outcomes, including symptoms, functioning, quality of life, health care use, and harms. Although the symptoms of CMI are physical, the committee embraced the need for a “whole-person” approach because of the complexity of CMI and its potential comorbidities. The nontraditional treatments were included in an effort to identify potentially effective pathways for treatment of the whole person instead of focusing on each specific symptom.
Only three studies that were identified were conducted in samples of veterans. Each reported the effects of a different intervention: cognitive rehabilitation therapy (Jakcsy, 2002), doxycycline (Donta et al., 2004), and cognitive behavioral therapy (CBT) and exercise (Donta et al., 2003; Guarino et al., 2001; Mori et al., 2006). The veteran populations were generally male (85% in Donta et al., 2003, and Mori et al., 2006; 86% in Donta et al., 2004, and 50% in Jakcsy, 2002), and the average age ranged from 37.5 years (Jakcsy, 2002) to 40.7 years (Donta et al., 2004). The committee considers those studies with others of similar interventions but different populations below.
The committee believed it necessary to consider additional evidence so that it could offer recommendations about the best treatment and management approaches for veterans who have CMI. Thus, the recommendations presented in Chapter 8 result from careful consideration of the evidence presented in the present chapter, evidence on the best treatments for comorbid and related conditions in Chapter 5, and issues surrounding patient care and communication in Chapters 6 and 7.
People who have many of the conditions described in this report are treated with pharmacologic agents that are also used to treat for other conditions. Often, the mechanisms of action of the pharmacologic agents are unknown. For example, patients who have fibromyalgia may benefit from duloxetine, which works independently of depression (which duloxetine is often prescribed for). Such agents include selective serotonin reuptake inhibitors, serotonin–norepinephrine reuptake inhibitors (for example, duloxetine), tricyclic medications (for example, amitriptyline), monoamine oxidase inhibitors (for example, phenelzine), dopaminergic blockers (for example, haloperidol), anxiolytics (for example, benzodiazepines), and medications that potentiate gaba-ergic transmission (for example, gabapentin), potentiate binding of voltage-gated calcium channels (for example, pregabalin), and potentiate voltage-dependent sodium channels (for example, topiramate). Analgesic medications include nonsteroidal anti-inflammatory analgesics, acetaminophen, opioid analgesics, and tramadol,
which is serotonergic and partially binds to the mu-opiate receptor. Because of the variety of effects of these pharmacologic agents and their varied biologic targets, the committee considered the evidence on specific agents rather than broad classes.
The committee reviewed 11 studies: nine clinical studies and two Systematic reviews. Of the clinical trials, five were randomized (Donta et al., 2004; Han et al., 2008a; Kroenke et al., 2006; Muller et al., 2008; Volz et al., 2000), six were blinded (Altamura et al., 2003; Donta et al., 2004; Han et al., 2008a; Kroenke et al., 2006; Muller et al., 2008; Volz et al., 2000), and three were pre–post2 studies (Garcia-Campayo and Sanz-Carrillo, 2002; Han et al., 2008b; Menza et al., 2001). Six of the clinical trials included agents typically known as antidepressants (mirtazapine, venlafaxine, nefazodone, escitalopram, paroxetine, and opipramol), one an antipsychotic (sulpiride), one an anticonvulsant (topiramate), and one an antibiotic (doxycycline).
Most of the pharmacologic studies were limited by a high number of dropouts in both treatment and placebo arms. Only one study involved the veteran population (Donta et al., 2004); it was the only study that used a nonpsychopharmacologic intervention (doxycycline), and it did not demonstrate efficacy. The other eight studies enrolled people in the general population, most of them female, who had somatoform disorder variously described as multisomatoform disorder, undifferentiated somatoform disorder, and somatization disorder. Two studies excluded patients who had comorbid major depressive disorder, social anxiety disorder, or generalized anxiety disorder but used anxiolytics, alprazolam, or lorazepam in over 50% of the subjects in addition to the medications being assessed (Han et al., 2008a,b). The generalizability of studies of nonveteran populations to the veteran population is unclear.
In all the studies except the doxycycline study (Donta et al., 2004), the authors concluded that there may be benefit from specific drug interventions but that future research should include larger cohorts of patients and fixed doses of medications. The evidence is limited by small sample sizes, high dropout rates, and potentially confounding effects of depression and anxiety. The influence of pharmaceutical manufacturers, who funded or employed the authors of several of the clinical trials, may have affected the reported results (Altamura et al., 2003; Donta et al., 2004; Han et al., 2008a; Kroenke et al., 2006; Menza et al., 2001). Publication bias may also be an issue if it limited the evidence available for consideration by the committee.
Regarding the risk of bias in each study, the committee judged four of the studies (Donta et al., 2004; Kroenke et al., 2006; Muller et al., 2008;
2Pre–post is used to describe clinical study designs where pre-intervention measures are compared to post-intervention measures.
Volz et al., 2000) to be at low risk for bias and five to be at high risk (Altamura et al., 2003; Garcia-Campayo and Sanz-Carrillo, 2002; Han et al., 2008a,b; Menza et al., 2001).
The two systematic reviews were of relatively good quality. A review by Sumathipala et al. (2007) was particularly relevant in that the studies reviewed focused on adults who had medically unexplained symptoms and excluded symptom syndromes such as chronic fatigue syndrome (CFS), fibromyalgia, and irritable bowel syndrome (IBS). The studies that were reviewed were limited by high dropout rates, focus on short-term outcomes, and use of varied methods but did indicate support for antidepressants. A review by Kroenke (2007) noted that pharmacologic studies were limited by short follow-up and included drugs that are not available in the United States.
On the basis of guidance provided by AHRQ (Owens et al., 2010), the committee rated the strength of evidence as shown in Table 4-1 and concluded that
• There is insufficient strength of evidence to determine the effectiveness of paroxetine, nefazodone, and opipramol in people who have somatoform disorders.
• There is low strength of evidence that venlafaxine, escitalopram, levosulpiride, and topiramate improved symptoms in people who had somatoform disorders.
• There is high strength of evidence that doxycycline is not effective in improving symptoms and functioning in veterans who have CMI (referred to as Gulf War illnesses in Donta et al., 2004).
Each of the studies on pharmacologic interventions is summarized in Table 4-2. Clinical trials are presented first, followed by systematic reviews.
Fontani et al. (2011) studied the use of noninvasive radioelectric asymmetric conveyer brain stimulation (REAC-BS) to treat for pain and physical problems related to stress. REAC-BS is a series of painless electric pulses applied to specific reflex auricular points. The authors found that patients treated with REAC-BS reported fewer pain and physical symptoms after treatment after a 4-week treatment cycle than those treated with a placebo. The study was limited because there was a lack of follow-up and because the dose was not standardized but varied on the basis of patient response. Patients may have been able to sense the stimulations and deduce their allocation to the treatment or placebo group. The clinical significance of pre–post score change from 122 to 96 on the psychologic stress measure questionnaire is unclear (Fontani et al., 2011).
|RCT: 2 (207)a||High||Consistent||Indirect||Imprecise||Low|
|Pre-post: 1 (22)b||High||N/A||Indirect||Imprecise||Insufficient|
|Pre-post: 1 (15)c||High||N/A||Indirect||Imprecise||Insufficient|
|RCT: 1 (31)d||Low||N/A||Indirect||Imprecise||Low|
|RCT: 1 (208)e||Unclear||N/A||Indirect||Imprecise||Insufficient|
|RCT: 1 (74)f||High||N/A||Indirect||Imprecise||Low|
|Pre-post: 1 (35)g||High||N/A||Direct||Imprecise||Low|
|RCT: 1 (491)h||Low||N/A||Direct||Precise||Highi|
NOTES: RCT = randomized controlled trial; RoB = risk of bias.
a Han et al., 2008a; roenke et al., 2006.
b Han et al., 2008b.
c Menza et al., 2001.
d Muller et al., 2008.
e Volz et al., 2000
f Altamura et al., 2003.
g Garcia-Campayo and Sanz-Carrillo, 2002.
h Donta et al., 2004
i No beneficial effect was found in the study.
On the basis of guidance provided by AHRQ (Owens et al., 2010), the committee rated the strength of evidence as shown in Table 4-3 and found that there was insufficient evidence to draw conclusions about the effectiveness of REAC-BS in treating for CMI. Details of Fontani et al. (2011) are summarized in Table 4-4.
Cognitive Behavioral Interventions
The committee reviewed five individual CBT studies and five group CBT studies that assessed CBT for patients who presented with somatic symptoms. Somatic-focused CBT, reviewed below, is not general CBT but rather targets somatic symptoms directly and includes relaxation and activity components.
|Altamura et al., 2003
|RCT, crossover||Levosulpiride 50 mg/day vs placebo for 4 weeks.||64.9% female, mean age 38.1 years
N levosulpiride = 37
N placebo = 37
|Donta et al., 2004
Veterans with Gulf War veterans illness and positive for mycoplasma DNA
US VA and DOD medical centers
|RCT||Doxycycline 200 mg/day vs identically matched placebo capsules for 12 months.||14% female, mean age 41 years
N screened = 2,712
N doxycycline = 245
N placebo = 246
|Garcia-Campayo and Sanz-Carrillo, 2002||Pre-post||Topiramate starting at 50 mg, increased by 50 mg every 4 days to a maximum of 300–400 mg (150–200 mg bi-daily) or until adverse event occurred; trial lasted 6 months; dose averaged 365.7 mg (range 300–400 mg).||68.6% female, mean age 41.8 years|
|Multisomatoform disorder||N screened = 84
N enrolled = 35
|Han et al., 2008a
Undifferentiated somatoform disorder
|RCT||Venlafaxine vs mirtazapine.
Venlafaxine starting at 37.5 mg/day, increased each week by 37.5–75 mg/day to maximum of 225 mg/day. Mirtazapine starting at 15 mg/day, increased by 15 mg/day to maximum of 60 mg/day. Both were adjusted for tolerability and clinical response. Conducted over 12 weeks.
|61% female, mean age 45 years
N venlafaxine = 45 N mirtazapine = 50
|Outcomes and Results|
|After treatment, levosulpiride patients reported fewer symptoms (measured by CISSD-SDS) compared with baseline (1,709 vs 1,205 symptoms, p = 0.007) and compared with placebo at 4 weeks (1,205 vs 1,597 symptoms, p < 0.001).
No difference between treatments in anticholinergic or neuroendocrine side effects reported; extrapyramidal system involvement was reported more frequently by levosulpiride-treated patients (9 vs 2 patients, p < 0.03).
|After treatment, treatment and placebo groups did not differ significantly in any outcome measured. In primary outcome, physical health functioning (SF-36), treatment and placebo groups’ average scores were 30.2 and 30.1 (physical) and 37.4 and 36.2 (mental), respectively, at baseline; at 12 months, treatment and placebo groups’ average scores were 32.0 and 30.9 (physical) and 37.6 and 36.7 (mental). No differences in pain (MPQ), fatigue (MFI), or cognition (Cognitive Failures Questionnaire). After 12 months, 154 of 200 doxycycline and 159 of 211 placebo patients were negative for mycoplasma DNA. Increased adverse events considered related to study drug were myalgia in placebo group (4.5% vs 1.2%, p = 0.05), nausea in doxycycline group (37.1% vs 10.2%, p < 0.001), and photosensitivity in doxycycline group (14.7% vs 6.1%, p = 0.002); all other reported events were in the two groups.
Blood doxycycline concentrations were undetectable in 38.9% of treatment-group patients and 98% of placebo-group patients at 12 months (adherence).
|From baseline to 6-month follow-up, improvements in severity (CGI, 3.8 to 3.3, p < 0.001) and functioning (GAF, 52.4 to 58.5, p < 0.001) were significant; no significant changes were seen in pain (pain VAS, 68.8 to 68.7; MPQ, 36.6 to 36.2) or anxiety and depression (HADS, 9.5 to 9.6).|
|Most frequently reported side effects were somnolence (25.7%), fatigue (20%), paresthesia (14.2%), nervousness (8.5%), and nausea (5.7%).|
|After treatment, mean somatic symptom severity scores (PHQ-15) decreased significantly for mirtazapine (–8.4, p < 0.0001) and venlafaxine (–6.1, p < 0.0001) over 12 weeks; betweengroup difference was significant (p = 0.046) in favor of mirtazapine.
For mirtazapine, scores for psychologic distress (GHQ-12) and depression (BDI) decreased significantly between baseline and end of study (–4.9, p < 0.0001 and –13.5, p < 0.0001, respectively); for venlafaxine, psychologic distress and depression scores also decreased significantly (–4.3, p = 0.001 and –9.02, p < 0.0001) with no significant difference between groups.
Adverse effects reported by participants were dry mouth in both groups; somnolence, yawning, and dizziness in mirtazapine group; and nausea in venlafaxine group.
|Han et al., 2008b
Undifferentiated somatoform disorder
|Pre–post||Paroxetine Immediate Release (IR) administered starting at 10 mg/day and increased to a maximum of 40 mg/day on basis of patient’s response over 8 weeks; average dose was 19.5 mg/day (range 10’40 mg/day).||59% female, mean age 37.4 years
N screened = 43
N enrolled = 22
|Kroenke et al., 2006
|RCT||Venlafaxine ER vs placebo. Capsules given once a day in following doses: week 1 = 75 mg; week 2 = 150 mg; weeks 3–12 = 225 mg.
Dose was decreased to tolerable level for participants who experienced intolerance.
Conducted over 12 weeks, with up to 2 weeks of taper.
|80% female, mean age 47 years
N screened = 231
N venlafaxine ER = 55
N placebo = 57
|Menza et al., 2001
|Pre-post||Nefazodone administered starting at 50 mg bi-daily, increased to 100 mg bi-daily after first week, increased to 150 mg bi-daily in weeks 2–8.||86% female, mean age 48.6 years
N screened = 46
N enrolled = 15
|Outcomes and Results|
|After treatment, average symptom severity (PHQ-15) score dropped significantly, by 75%, from 17.2 to 4.3 (p = 0.001) between baseline and 8 weeks. Average depression (BDI) scores also decreased significantly, 50.8% (p < 0.001), and average psychologic distress (GHQ-12) scores decreased by 13%, but difference was not significant.
Most common adverse events reported were nausea, dry mouth, and somnolence.
|After treatment, both treatment and placebo groups had significant (p < 0.0001) decrease in somatic symptoms (PHQ-15 change of .8.3 and .6.6 respectively) but no significant difference between groups (p = 0.097). From baseline to 12 weeks, greater improvement was seen in Venlafaxine ER participants than in those who received placebo for pain (PHQ-15 pain subscale, .3.3 and .2.6, respectively, p = 0.03), psychic anxiety (HAM-A, .8.6 and .5.9, respectively, p = 0.02), clinical improvement (CGI-Improvement, .1.8 and .1.4, respectively, p = 0.009), physical symptoms (MQOL Physical symptoms, .11.7 and .6.0, respectively, p = 0.01), mental health (SF-36, 28.6 and 16.8, respectively, p = 0.03), bodily pain (SF-36, 26.1 and 14.5, respectively, p = 0.03), and concentration (MOS-CS, 30.1 and 15.3, respectively, p = 0.007).
Significant (p . 0.001) improvement from baseline to week 12 was noted for several measures without difference between groups (depression, HAM-D; HAM-A total score and HAM-A somatic anxiety score; CGI-Severity score; and SF-36 physical health score).
Percentage of patients reporting bothersome headache and stomach pain symptoms decreased significantly more in venlafaxine ER than in placebo-treated patients (both p = 0.03).
Venlafaxine ER.treated participants reported more adverse effects than placebo; more than 10% of venlafaxine ER patients reported nausea, headache, fatigue, dizziness, constipation, and tremor.
|11 of 15 patients who completed study showed nonsignificant improvement by 24% in symptom severity on VAS after 8 weeks of treatment compared with baseline. From baseline to week 8, 73% of participants showed improvement on CGI (significance not noted), and 73% showed significant improvement in functioning (SF-36, p < 0.025). Significant improvement in depression (p < 0.005) and cognitive and somatic subscales (both p < 0.025) of HAM-D and HAM-A were reported, but there was no correlation with VAS, CGI, or SF-36 score.
Most common adverse effects were sedation, GI discomfort, anxiety, and dry mouth.
|Muller et al., 2008
Cape Town, South Africa
|RCT||Escitalopram vs placebo. Administered at 10 mg/day and could be increased to 20 mg/day after week 4 (dose decreased to 10 mg/day for intolerance). Dosage was constant for weeks 8–12; after 12 weeks, dosage was tapered and stopped over 1 week.
Entire study took place over 18 months.
|90% female, mean age 40 years
N screened = 60
N escitalopram = 25
N placebo = 26
|Volz et al., 2000
Somatoform disorders (ICD-10 codes F45.0, F45.1, F45.3)
|RCT||Opipramol vs placebo. Started at 50 mg/day, increased by 50–200 mg/day on day 4.
Number of placebo capsules was identical with number of opipramol capsules. Treatment lasted 6 weeks.
|64% female, mean age 46 years
N opipramol = 104
N placebo = 104
|Systematic review||Multiple||N RCTs = 34|
|Systematic review||Multiple||N abstracts screened = 368
N systematic reviews = 6
N RCTs = 14
|Outcomes and Results|
|After treatment, symptom severity (PHQ-15) scores in treatment group decreased significantly more than in placebo group (14.6 vs 17.3 at baseline and 5.6 vs 12.5 at 12 weeks, respectively, p < 0.0001). Clinical improvement (CGI-Improvement, treatment: 3.0 to 1.6 vs placebo: 3.4 to 3.2, at baseline and 12 weeks, respectively) and severity (CGI-Severity, treatment: 4.5 to 2.6 vs placebo: 4.7 to 4.1, at baseline and 12 weeks, respectively) also showed significant improvement in treatment subjects (p < 0.05).
Likewise, escitalopram showed improvement on somatic anxiety (HAM-A somatic subscore, p < 0.0001, and HAM-A psychic subscore, p = 0.0002), depression (MADRS, p = 0.017), and disability (SDS, p = 0.015) scores after 12 weeks compared with placebo. No difference between treatment and placebo in illness behavior (SAIB) was detected.
After 12 weeks, 80% of treatment and 26.9% of placebo subjects met criteria for being responsive to treatment, and 52% of treatment and 3.8% of placebo subjects met criteria for remission.
Most common adverse events were headache, nausea, and abdominal discomfort in first 2 weeks and headache, nasopharyngitis, and diarrhea in weeks 2–12, with no significant differences between treatment and placebo groups in reported adverse events.
Average dose was 14.4 mg of escitalopram per day and 18.1 mg of placebo per day.
|After treatment, anxiety scores decreased significantly more in opipramol group than in placebo group for somatic anxiety (HAM-A somatic subscore, 16.8 to 7.3 vs 16.9 to 9.1, p = 0.013) and total anxiety (HAM-A, 25.1 to 11.8 vs 25.3 to 14.5, p = 0.013), and decrease in psychic anxiety score was nearly significant (HAM-A psychic subscore, 8.3 to 4.5 vs 8.4 to 5.4, p = 0.052). Scores followed same trend with greater decrease in opipramol group and in placebo group for depression (HAM-D, 14.6 to 8.0 vs 14.3 to 9.5, p = 0.006) and for symptoms (SCLR-90-R, 0.95 to 0.57 vs 0.77 to 0.55, p = 0.041), including somatic symptoms (1.61 to 0.93 vs 1.40 to 0.98, p = 0.043) and anxiety symptoms (1.15 to 0.63 vs 0.94 to 0.64, p = 0.049).
Adverse events were reported by 37% of opipramol patients and 38% of placebo patients—gastrointestinal complaints in 7.1% of opipramol patients and 20.8% of placebo patients and musculoskeletal complaints in 14% of opipramol patients and 7% of placebo patients. Tiredness occurred more frequently in opipramol group than in placebo group (9% vs 2%).
|Positive results for at least one outcome (symptoms, functional, or psychologic) were noted in 11 of 13 CBT studies and 4 of 5 antidepressants studies, and 8 of 16 studies of other interventions reported improvement—consultation letter to primary care physician (4 studies), primary care physician training (2 studies), and non-CBT psychotherapies (2 studies), hypnosis (2 studies), and one study each investigating multicomponent nurse-care management, aerobic exercise, writing disclosure, paradoxic intention, and explanatory therapy.
CBT is most effective for variety of somatoform disorders and outcomes.
|Studies assessed interventions with antidepressants (1 review and 3 RCTs), CBT (5 reviews and 2 RCTs), psychiatric consultation (3 RCTs), other forms of psychotherapy (3 RCTs), exercise (2 RCTs), and collaborative care (1 RCT).
Studies have many limitations and varied methods, so interventions cannot be compared.
Authors note support for antidepressants and CBT and limited evidence on other interventions.
NOTES: BDI = Beck Depression Inventory; CGI = Clinical Global Impressions Scale; CISSD-SDS = Conceptual Issues in Somatoform and Similar Disorders, somatoform disorders schedule; GAF = Global Assessment of Functioning; GHQ-12 = 12-item General Health Questionnaire; HADS = Hospital Anxiety and Depression Scale; HAM-A = Hamilton Anxiety Rating Scale; HAM-D = Hamilton Depression Rating Scale; MADRS = Montgomery-Åsberg Depression Rating Scale; MFI = Multidimensional Fatigue Inventory; MPQ = McGill Pain Questionnaire; MQOL = McGill Quality of Life Questionnaire; MOS-CS = Medical Outcomes Study Concentration Scale; PHQ-15 = Patient Health Questionnaire 15-Item Somatic Symptom Severity Scale; RCT = randomized controlled trial; SAIB = Scale for the Assessment of Illness Behaviour; SCL-90-R = Symptom Checklist-90-Revised; SDS = Sheehan Disability Scale; SF-36 = Medical Outcomes Study 36-Item Short Form Health Survey; VAS = Visual Analogue Scale.
|RoB||Consistency||Directness||Precision||Strength of Evidence|
|RCT: 1 (888)a||Low||N/A||Indirect||Precise||Insufficient|
NOTES: RCT = randomized controlled trial; REAC-BS = radioelectric asymmetric conveyer brain stimulation; RoB = risk of bias.
a Fontani et al., 2011.
|Fontani et al., 2011
Pain and physical problems and medically unexplained symptoms
|RCT||Noninvasive REAC-BS or inactive REAC: 18 sessions of 7 pulses on alternating days over 4 weeks.||77% female, mean age 43 years
N REAC = 688
N inactive REAC = 200
NOTE: PSM = Psychological Stress Measure; RCT = randomized controlled trial; REAC-BS = radioelectric asymmetric conveyer brain stimulation; SD = standard deviation.
In general, the goals of somatic-focused CBT are to modify dysfunctional beliefs about symptoms, reduce psychophysiologic arousal through relaxation, enhance activity regulation through increased exercise and pleasurable activities, and increase awareness and enhance communication of thoughts and emotions. Therapy is usually brief, ranging from 5 to 12 weekly therapy sessions.
Individual Cognitive Behavioral Interventions
Of the six clinical trials of individual sessions, all but one (Arnold et al., 2009) were RCTs (Allen et al., 2006; Escobar et al., 2007; Sharpe et al., 2011; Sumathipala et al., 2000, 2008). CBT in those trials ranged from 5 to 12 individual sessions, typically administered weekly for about 50–90 minutes in a mental health specialty clinic or a primary care (PC) clinic, most often by a mental health professional but some by a PC physician. Components of the CBT included various behavioral and cognitive techniques, including time-contingent activity planning, pleasant-activity scheduling, sleep hygiene, cognitive restructuring of negative thinking, affect management, and problem-solving skills. Most trials used standard medical care, routine clinical care, psychiatric consultation, or treatment as usual as the control condition. In two instances, a consultation letter to the PC physician was added to the control condition (Allen et al., 2006; Escobar et al., 2007). Most of the studies targeted somatoform conditions, including mainly somatization disorder, abridged somatization, and high levels of unexplained physical symptoms. Moreover, substantial comorbidity, particularly anxiety and depressive syndromes, was documented in the studies.
All the RCTs reported at least some benefit from the intervention (Allen et al., 2006; Escobar et al., 2007; Sharpe et al., 2011; Sumathipala et al., 2000, 2008). Most of the studies documented a consistent, beneficial effect
|Outcomes and Results|
|At baseline, average total pain and physical problems (PSM) score was 122.53 (SD ± 6.75) in treatment group and 122.96 (SD ± 7.04) in placebo group. After treatment, average total scores for treatment and placebo groups were 96.01 (SD ± 8.5) and 122.11 (SD ± 7.5), respectively, with significant decrease in PSM score in treated subjects (p < 0.005) but no change in placebo group. In number-needed-to-treat analysis, there were 403 of 688 subjects positive for stress-related pain and physical problems in treatment group and 159 of 200 in placebo group; after treatment, 196 of 403 (48.6%, asymptotic significance = 0.000) in treatment group and 135 of 158 (84.9%) in placebo group remained positive for stress-related pain and physical problems. RR = 0.42; number needed to treat = 2.56.|
of CBT relative to the control condition, most notably immediately after treatment, with gains largely maintained through long-term follow-up, which was up to 15 months (Allen et al., 2006). The pattern of improvement was largely consistent in all assessed outcomes, including assessment of physical symptoms by a blind interviewer and self-reported physical symptoms, such psychiatric symptoms as depression and anxiety, patient satisfaction, and measures of functioning. Retention rates were generally similar in the CBT and control groups (about 10–30%) and appear to compare favorably with rates for other interventions (such as medications) and with rates in the broader CBT literature for such mental disorders as depression and posttraumatic stress disorder. Generally, sample sizes in the RCTs were appropriate for detecting large, clinically significant effects.
Four of the individual studies (Allen et al., 2006; Escobar et al., 2007; Sharpe et al., 2011; Sumathipala et al., 2000, 2008) can be viewed as predominantly low risk of bias and one as high risk (Arnold et al., 2009).
Group Cognitive Behavioral Interventions
Five studies examined group CBT (Bleichhardt et al., 2004; Donta et al., 2003; Lidbeck, 2003; Martin et al., 2007; Mori et al., 2006; Rief et al., 2002; Zaby et al., 2008). The Lidbeck (2003) study included patients who met criteria for functional somatic symptoms, including hypochondriasis and irritable colon syndrome. The study by Bleichhardt et al. (2004) and Rief et al. (2002) and that by Zaby et al. (2008) examined multiple somatoform symptoms. Finally, the study by Donta et al. (2003) and Martin et al. (2007) examined medically unexplained symptoms. Interventions were conducted in a one- to eight-session group format with groups of up to 10. Interventions were typically carried out in a PC setting or a specialized setting with a trained psychotherapist. Only one of the studies was conducted in military or veteran samples (Donta et al., 2003; Mori et al., 2006).
Of the cognitive behavioral group interventions, four (Bleichhardt et al., 2004; Donta et al., 2003; Martin et al., 2007; Mori et al., 2006; Rief et al., 2002; Zaby et al., 2008) were original studies, and one was a follow-up of an earlier clinical trial (Lidbeck, 2003). The largest study, in which 1,092 Gulf War veterans who had multisymptom illness were enrolled, compared CBT plus exercise, CBT alone, exercise alone, and usual care (Donta et al., 2003; Guarino et al., 2001; Mori et al., 2006). For the main outcome of physical functioning, the group receiving CBT had an odds ratio (OR) of 1.73 (95% confidence interval [CI] 1.21–2.41) compared with no CBT, and exercise had an OR of 1.00 (95% CI 0.76–1.50) compared with no exercise. Those results suggest that group CBT rather than exercise conferred the main therapeutic benefit with respect to physical symptoms. Adherence (attendance to 8 out of 12 CBT or exercise sessions) was not
related to changes in physical function in the CBT group; however, adherent exercise participants were twice as likely than nonadherent participants to improve (p = 0.02). In a secondary data analysis, however, exercise alone or in combination with group CBT showed improvement in fatigue (general, physical, mental, activity level, and motivation). The observed sizes of treatment effects were lower than reported in other trials, and this might reflect the group format of CBT, the low number of sessions attended (6 of 12 exercise sessions or 5 of 12 CBT sessions attended on average), the range of therapist experience, and variability in the administration of CBT among sites.
Bleichhardt et al. (2004) and Rief et al. (2002) examined standard treatment plus either a specific somatic intervention or relaxation (another CBT intervention), and a waitlist control (N = 191). In the analyses, only the randomized arms for somatic-specific CBT intervention and relaxation were examined, and they showed small to large treatment effects through 1-year follow-up, with the largest effects on reduction in number of somatoform symptoms. There were no differences between the somatic CBT and relaxation groups except the lower number of doctor visits in the somatic CBT group. In the waitlist control group, there were no changes over the 4-month waiting period.
Martin et al. (2007) examined patients who had medically unexplained symptoms (N = 140), comparing one session of 3–4 hours of group CBT with standard medical care. Through a 6-month follow-up, patients in both groups showed improvements; there was greater reduction in doctor visits and somatization severity after CBT than after standard care.
In a study similar to that of Bleichhardt et al. (2004), Zaby et al. (2008), in a smaller sample (N = 77), showed a moderate decrease in somatic symptoms after both a specific somatic CBT therapy and relaxation and little improvement in a separate waitlist condition. The specific CBT somatic therapy also reduced anxiety and improved mental health compared with relaxation. Relaxation in both studies consisted of progressive muscle relaxation. Lidbeck et al. (2003), in a smaller follow-up (N = 31) of an RCT (Lidbeck, 1997), showed sustained improvement at 6-month follow-up after CBT somatic intervention.
The committee found that the studies of group CBT were of low risk of bias.
Brief Psychodynamic Psychotherapy Interventions
The committee reviewed four individual clinical studies and one study of group intervention published after 2000 that have examined brief psychodynamic psychologic interventions (typically fewer than 12 sessions).
Individual Psychodynamic Psychotherapies
Five of the studies examined brief individual dynamic or brief psychodynamic therapies in people who had medically unexplained symptoms (Abbass et al., 2009a; Aiarzaguena et al., 2002), multisomatoform disorder (Sattel et al., 2012), functional neurologic symptoms (Reuber et al., 2007), and alexithymia with somatization (Posse, 2004). Reuber et al. (2007) and Posse (2004) included too little information for determining how similar or dissimilar the diagnosis was to CMI. There were problems related to standardization of therapy—such as a lack of a detailed therapy manual (Abbass et al., 2009a; Aiarzaguena et al., 2002; Posse, 2004; Reuber et al., 2007), the use of only one therapist (Posse, 2004; Reuber et al., 2007), failure to assess therapeutic adherence (Aiarzaguena et al., 2002; Posse, 2004; Reuber et al., 2007), or lack of acceptable therapeutic adherence (Sattel et al., 2012). Both Reuber et al. (2007) and Sattel et al. (2012) noted that components of cognitive behavioral approaches were allowed to be included in the intervention. No studies examined military or veteran samples.
Of the individual studies, two (Posse, 2004; Sattel et al., 2012) were RCTs, and three were pre–post designs (Abbass et al., 2009a; Aiarzaguena et al., 2002; Reuber et al., 2007). The Sattel et al. (2012) trial was the largest of the psychodynamic intervention trials, with 211 participants. It compared the results of 12 weekly 90-minute sessions of psychodynamic interpersonal therapy (PIT), including some cognitive behavioral components, with the results of three 30-minute sessions of enhanced medical care (EMC) separated by about 6 weeks. After treatment, PIT was not superior to EMC on any of the somatic or mental health outcomes. At 9-month follow-up, PIT emerged as superior to EMC only for somatic outcomes, with a small to moderate effect size (d = 0.42), but not for mental health or health care use outcomes.
Posse (2004), published in Spanish, reported that five patients who were receiving Jungian therapy improved significantly more than five controls on measures of anxiety, fatigue, and impulsivity. However, the study lacked description of methods—including selection and randomization of participants, the intervention, and the control condition—and therefore carries a high risk of bias.
The three brief uncontrolled trials (Abbass et al., 2009a; Aiarzaguena et al., 2002; Reuber et al., 2007) found moderate effects of therapy on outcomes. Abbass et al. (2009a), using open-ended dynamic intervention of one to 25 sessions with 50 participants, showed a 69% reduction (p < 0.001) in emergency department visits per person and improved somatic symptoms. Aiarzaguena et al. (2002), in a small sample of 12 participants with a pre–post structured psychotherapy approach handled by PC physicians, reported improvement in functioning and quality of life. Reuber et al. (2007)
used up to 19 sessions of psychodynamic intervention and failed to show an improvement in intent-to-treat analyses of somatic symptoms in 91 participants but found improvement in general clinical outcome and quality of life.
Of the individual psychodynamic psychotherapy clinical studies, one study (Sattel et al., 2012) was rated as of low risk of bias and the remainder (Abbass et al., 2009a; Aiarzaguena et al., 2002; Reuber et al., 2007) as of high risk of bias.
Group Psychodynamic Psychotherapies
One study used a group intervention based on psychodynamic principles in a sample of people who met criteria for somatoform disorder (Rembold, 2011). This small study (N = 40) reported improvements in number of somatic symptoms, impairment due to symptoms, and anxiety relative to treatment as usual under PC physician control. Although it was promising, lack of reporting of randomization procedures and inferential statistics limits any conclusions that can be drawn from the study. It was rated as of high risk of bias.
Psychotherapy Intervention Reviews
The committee looked at five systematic reviews published after 2000 that examined trials of CBT for patients who presented with somatic symptoms (Kroenke, 2007; Kroenke and Swindle, 2000; Looper and Kirmayer, 2002; Nezu et al., 2001; Sumathipala, 2007), and two meta-analyses (Abbass et al., 2009b; Kleinstäuber et al., 2011) and three systematic reviews (Allen et al., 2002; Kleinstäuber et al., 2011; van Rood and de Roos, 2009) that focused on psychotherapies.
Three review papers focused on RCTs of CBT or behavioral medicine for somatoform disorders, medically unexplained symptoms or complaints, and somatization and somatic syndromes (Kroenke and Swindle, 2000; Looper and Kirmayer, 2002; Nezu et al., 2001). The reviews were conducted before 2007 and therefore did not include several of the individual studies described above. Two other systematic reviews looked at multiple treatments for somatoform disorders and medically unexplained symptoms (Kroenke, 2007; Sumathipala, 2007). Both concluded that the evidence supported the use of CBT more than other interventions, including pharmacologic agents, consultation letters, and exercise. All five reviews fulfill most of the criteria for Assessment of Multiple Systematic Reviews (AMSTAR) (Shea et al., 2007). Results of the five reviews all support the effectiveness of CBT and other behavioral interventions for somatic syndromes.
Abbass et al. (2009b) specifically examined psychodynamic psychotherapy for a wide variety of medically explained and unexplained disorders
(for example, IBS, Crohn’s disease, atopic dermatitis, and heart disease). Using AMSTAR guidance, rating categories for the quality of this are mixed. The meta-analysis conducted by Abbass et al. (2009b) showed moderate improvements for psychiatric symptoms, anxiety, and somatic symptoms. In 91.3% of the studies reviewed, short-term psychodynamic psychotherapy resulted in at least some benefit (improvement in more than one characteristic) in patients who had somatoform disorders.
A meta-analysis by Kleinstäuber et al. (2011) assessed the value of psychotherapy and showed small but consistent between-group differences in effect size and small to large within-group differences in effect size for physical symptoms and cognitive, emotional, and behavioral symptoms of somatoform disorder through the posttreatment period and follow-up. Moderator analysis for effects on physical symptoms showed that a higher number of therapy sessions was associated with better outcome. Similarly, use of therapists who are mental health professionals was associated with better outcome.
Three systematic reviews of psychotherapy interventions included such conditions as CFS, fibromyalgia, IBS, chronic pain, phantom-limb pain, psychogenic nonepileptic seizures, body dysmorphic disorder, olfactory reference syndrome, myoclonic movement, and stress-related dermatologic disorders (Allen et al., 2002; Kleinstäuber et al., 2011; van Rood and de Roos, 2009). Such aggregation of conditions makes extrapolation to CMI difficult. A review conducted by Allen et al. (2002) included only one study published since 2000 (Sumathipala et al., 2000) that was included in the present review. The van Rood and de Roos (2009) review focused on eye-movement desensitization and reprocessing (EMDR) and consisted largely of case studies. Only one study included by Russell (2008) examined EMDR for combat-related medically unexplained physical symptom; it used a single-subject case design and showed the benefit of 5 sessions of EMDR maintained through 6-month follow-up. The Russell (2008) study is not included in the present review because it does not include case studies. Accordingly, neither the van Rood and de Roos nor the Russell reviews altered the committee’s conclusions.
Summary of Psychotherapy Interventions
The size of the observed effect of CBT was consistent among studies and generally in the small to moderate range. Indicators of response to interventions varied in definition among studies, and rates of response ranged from 18.5% to 60%; this suggests that although substantial gains occur in most patients, full response or remission is rare. That is similar to what happens in people who have other chronic physical and mental disorders. Review of the studies described above yields a perception that
the dose of CBT may be relevant: Studies that used fewer sessions (6 or fewer) or in which CBT was delivered by a PC or a family physician rather than a mental health professional yielded less positive results (Arnold et al., 2009; Martin et al., 2007; Sumathipala et al., 2008). Similarly, the effect of the intervention may be attenuated if CBT is delivered in a group rather than individual format (Donta et al., 2003). Those factors have not been systematically evaluated. In general, attendance at treatment sessions was far from perfect but not atypical in such patient populations. Moreover, active components of the CBT interventions used in the trials have not been systematically evaluated, nor have active treatment comparisons that control for amount of patient contact been used.
The systematic reviews varied in quality and in terms of applicability to the CMI samples. There is a consensus among the committee members that somatic-focused CBT confers a consistent small to moderate effect in treatment for CMI and related somatization conditions.
On the basis of guidance provided by AHRQ (Owens et al., 2010), the committee rated the strength of evidence as shown in Table 4-5 and concluded that
• There is high strength of evidence of the efficacy of somatic-focused CBT for somatic-symptom syndromes. Somatic-focused individual
|RoB||Consistency||Directness||Precision||Strength of Evidence|
|RCT: 5 (601)a||Low||Inconsistent||Indirect||Precise||High|
|Pre-post: 1 (65)b||High|
|RCT: 5 (1,531)c||Low||Consistent||Indirect||Precise||High|
|RCT: 2 (222)d||Moderate||Inconsistent||Indirect||Imprecise||Insufficient|
|Pre-post: 3 (153)e||High|
|Pre-post: 1 (40)f||High||N/A||Indirect||Imprecise||Insufficient|
NOTES: CBT = cognitive behavioral therapy; RCT = randomized controlled trial; RoB = risk of bias.
aAllen et al., 2006; Escobar et al., 2007; Sharpe et al., 2011; Sumathipala et al., 2000, 2008.
bArnold et al., 2009.
cBleichhardt et al., 2004/Rief et al., 2002; Donta et al., 2003, and Mori et al., 2006; Lidbeck 2003; Martin et al., 2007; Zaby, 2008.
dPosse 2004; Sattel et al., 2012.
eAbbass et al., 2009a; Aiarzaguena et al., 2002; Reuber et al., 2007.
or group CBT may be a useful intervention for veterans who have CMI.
• There is insufficient evidence that brief psychodynamic psychotherapies are beneficial in treating people who have somatic syndromes. The committee draws this conclusion largely because there is not a substantial body of high-quality randomized studies, in particular studies of psychodynamic interventions alone, that is, without CBT components.
Each of the studies of psychotherapy interventions is briefly summarized in Table 4-6. Clinical trials of CBT interventions are presented first, then psychodynamic interventions, and finally systematic reviews.
Psychophysiologic treatments comprise a set of self-regulation techniques targeted at particular symptoms or body systems. Biofeedback, the
|Cognitive Behavioral Therapies|
|Allen et al., 2006
Piscataway, New Jersey
|RCT||CBT + PCI vs PCI alone; 10 sessions over 3 months.||89% female, mean age 46.6 years, 82% white, 55% employed
N screened = 142
N treatment = 43
N control = 41
quintessential psychophysiologic technique, involves one or more physiologic measures (such as heart rate, muscle tension, and sweating) that target the likely pathophysiology of the patient’s symptoms and teaching the patient to control his or her physiology voluntarily by watching the physiologic measures change in real time on a computer screen and developing a personal strategy for controlling them. Other components of the therapy include muscle relaxation and a self-hypnotic method called autogenic training. Psychophysiologic treatments have been found to be generally successful as treatment for various stress-related illnesses (Barrios-Choplin et al., 1997; Chappell and Chasee, 2006; McCraty et al., 1998, 2003; Reineke, 2008; Strack et al., 2004; Tanis, 2008).
The committee reviewed three studies of biofeedback in patients who had complex somatic symptoms. Two of the studies were RCTs in the United States and Germany (Katsamanis et al., 2011; Nanke and Rief, 2003), and the third one was an open trial in India (Saldanha et al., 2007).
Katsamanis et al. (2011) recruited persons who were seeking medical care for unexplained physical symptoms. People who had common disorders (such as hypertension and diabetes) were included if their personal treating physicians did not think that the physical symptoms were results of
|Outcomes and Results|
|After 15 months, treatment group reported greater improvement in each outcome, and improvement increased with time (significant treatment x time interaction effects) for somatization severity, physical functioning, somatic-symptom diary, and severity of somatic symptoms. No differences at baseline.
At 15 months, severity of somatization rating (CGI for somatization disorder): CBT + PCI = 3.91 (95% CI 3.55–4.28), PCI = 4.68 (95% CI 4.32–5.05), p < 0.001; treatment x time interaction p < 0.001.
Physical functioning (SF-36): CBT + PCI = 68.57 (95% CI 60.00–77.14), PCI = 59.17 (95% CI 50.55–67.78), p = 0.01; treatment x time interaction p = 0.02.
Symptom diary: CBT + PCI = 2.59 (95% CI 2.23–2.95), PCI = 3.24 (95% CI 2.89–3.60), p = 0.002; treatment x time interaction p < 0.001.
Severity of symptoms (Severity of Somatic Symptoms Scale): CBT + PCI = 35.02 (95% CI 27.28–42.76), PCI = 45.39 (95% CI 37.66–53.12), p = 0.005; treatment x time interaction p = 0.01.
Total health care cost significantly decreased after treatment compared with control group.
Median total health care costs before treatment: CBT + PCI = $1,944, PCI = $1,888; after treatment: CBT + PCI = $1,205, PCI = $1,645; p = 0.01.
No significant differences between groups in number of physician visits or diagnostic procedures.
|Arnold et al., 2009
|Pre-post, controlled||CBT vs control; 5 45-min sessions over 3 months (first 4 sessions scheduled every 2 weeks, fifth planned 3 months later).||88% female, mean age 47 years, 7% unemployed or sick leave
N screened = 6,409
N treatment = 31
N control = 34
|Bleichhardt et al., 2004
Rief et al., 2002
|CBT group therapy vs PMR group therapy vs waiting-list controls.
8 sessions of group therapy lasting 1.5 hours, 2-3 times per week over 3-4 weeks. Average treatment duration was 52 days. Individual therapy and additional group therapy for disorders (social anxiety, depression, etc.) were also provided.
|73% female, mean age 44 years
N PMR = 84
N soma = 107
|Only 45% (14) of participants in treatment group completed intervention; 29% (9) completed 2-4 sessions of CBT, and 26% (8) withdrew before start of treatment.
According to patients’ judgment of improvement, 35% of control group and 42% of treatment group had improved at 12-month follow-up (not significant).
Groups did not differ significantly at baseline regarding severity of main physical symptom (measured by self-reported VAS). Recovery at 12-month follow-up (30% decrease in severity score on VAS or VAS < 5 was similar between control and treatment groups: 29% and 32%, respectively.
VAS baseline to 12-month follow-up: control 7.6 (95% CI 7.2-7.9) to 6.0 (95% CI 5.2-6.8); treatment 7.6 (95% CI 7.1-8.0) to 6.0 (95% CI 5.1-6.8).
Self-reported symptoms (PSC-51), anxiety and depression (HADS), functional impairment (SF-36), and illness behavior (IAS) and health care use measurements were comparable except greater mean number of physical symptoms at baseline were reported in treatment group (12.6 [95% CI 10-15]) than in control group (9.4 [95% CI 8-11]), p = 0.02. At 12-month follow-up, physical symptoms were 10.9 (95% CI 8-14) and 8.5 (95% CI 6-10) for treatment and control groups, respectively (p > 0.05).
|At 1-year follow-up, significant improvements (p < 0.001) were made in all outcomes (number of symptoms, DSM-IV; general psychopathology, SCLR-90; anxiety and depression HADS; subjective health status, EQ-5D; and life satisfaction, FL2M), but there was no significant group or time x group effect. The effect of time was significant, p < 0.001, for all outcomes except no. of doctor visits in the previous year, p < 0.05. No change over 4-month waiting period was seen in controls.
Mean ES for PMR = 0.57 and mean ES for the soma group = 0.67.
In pre-post assessment of soma group only, at posttreatment assessments, 3 of 8 items on hourly journal evaluation had significant time effect from session 1 through session 8 (“I got suggestions for managing my symptoms,” “I felt the group was helpful today,” and “group atmosphere” all p < 0.05). Overall average score was about 4.0, indicating that treatment was well accepted.
Motivation for psychotherapy was not a good predictor of hourly evaluations, but items for “general therapeutic experience” and “experience with respect to psychotherapy” correlated with hourly evaluations (both p < 0.01). Only controlling conviction related to hourly evaluations was “internality” (p < 0.01).
Results indicate that the more positive expectations, positive previous experience, and higher internal control convictions, the more positive the hourly evaluation.
After 1-year follow-up, there was significant improvement in somatic symptoms.
|Donta et al., 2003; Mori et al., 2006
Gulf War Veterans’ Illnesses (GWVI)
US VA and DOD medical centers
|RCT||CBT vs CBT + exercise vs exercise vs usual care.
CBT was 60–90 minutes in duration; groups met weekly for 12 weeks.
Aerobic exercise one 1-hour session per week with exercise therapist for 12 weeks. 2–3 independent sessions per week for 12 weeks.
|15% female, mean age 41 years
N screened = 2,793
N CBT + exercise = 266
N CBT only = 286
N exercise only = 269
N usual care = 271
|Escobar et al., 2007
|RCT||CBT vs control, 10 sessions; each 45–60 minutes long except that first session was about 90 minutes long.
Sessions were delivered over 10–20 weeks with average of 3 months.
|88% female, mean age 40 years, and 66% Hispanic.
N screened = 416
N treatment = 87
N control = 85
|Lidbeck, 1997, 2003
Functional somatic symptoms
8 sessions lasting for 3 hours, weekly. Administered over 2 months.
|85% female, mean age 43.8 years
N treatment = 33
N follow-up = 31
|At 12 months, pairwise comparisons of physical functioning (SF-36): CBT vs usual care OR 1.72 (95% CI 0.91–3.23; p value raw 0.03, corrected 0.13); CBT + exercise vs usual care OR 1.84 (95% CI 0.95–3.55; p value raw 0.02, corrected 0.09); marginal comparisons: CBT vs no CBT OR 1.71 (95% CI 1.21–2.41; p value raw 0.002, corrected 0.005).
Improvement in those who attended at least 8 treatment sessions vs nonadherent patients, CBT + exercise: OR 1.71 (95% CI 0.92–3.20); CBT: OR 1.47 (95% CI 0.80–2.69; p value 0.27); exercise: OR 2.67 (95% CI 1.20–2.92).
Adjusted mean change from baseline and significance compared with usual care—mental health functioning (SF-36): CBT 0.97 (p < 0.025), CBT + exercise 2.30 (p < 0.01), exercise 2.33 (p < 0.01); pain (McGill Short Form): affective pain, CBT –0.43 (p < 0.01), CBT + exercise 0.50 (p < 0.01); fatigue (MFI): general fatigue, CBT + exercise –0.97 (p < 0.01), exercise –0.87 (p < 0.01), physical fatigue, CBT + exercise –0.70 (p < 0.025), exercise –0.73 (p < 0.01); reduced activity: CBT + exercise –0.68 (p < 0.01), exercise –0.69 (p < 0.01); reduced motivation, CBT + exercise –0.35 (p < 0.025), exercise –0.38 (p < 0.01); mental fatigue, CBT + exercise –1.08 (p < 0.01), exercise –0.84 (p < 0.025); cognitive difficulties: cognition failures questionnaire, CBT 2.66 (p < 0.01), CBT + exercise 3.38 (p < 0.01), exercise 2.98 (p < 0.01); V/SF-36 mental health index, CBT + exercise 2.95 (p < 0.01), exercise 3.27 (p < 0.01); other changes for CBT, CBT + exercise, and exercise not significant.
|After 6 months of follow-up, both intervention and control groups improved for all outcomes, but no significant differences between groups or group x time interactions were noted for physical functioning (SF-36 physical-functioning subscale), depression (HAM-D), anxiety (HAM-A), or medically unexplained symptoms (by visual analogue scale) except that somatic complaints (PHQ-15) were reduced in intervention group at 6 months compared with controls (p = 0.03, no interaction effect).
Mean scores in treatment and control groups, respectively, from baseline to 6-month follow-up—physical functioning: 63.28–73.22 vs 61.41–69.41; somatic complaints: 14.17– 9.11 vs 13.98–10.91, at 6 months p = 0.03; depression: 18.25–12.88 vs 17.41–14.29; anxiety: 20.46–14.85 vs 20.99–17.58; medically unexplained symptoms: 42.34–23.72 vs 39.62–25.25.
|After 1.5 years of follow-up, significant improvement was noted for social problems (SPQ, p = 0.004), illness behavior (IBQ, p < 0.001), hypochondriasis (WI, p < 0.001), depression (HADS, p < 0.005), and anxiety (HADS, p = 0.006). Sleep quality (SDI) and medication use were most improved after treatment and then gradually worsened (both p < 0.05).|
|Martin et al., 2007
|RCT||CBT vs control.
CBT was administered in single session lasting 3-4 hours.
|75% female, mean age 45.7 years in treatment group, 51.7 in control group
N screened = 146
N treatment = 70
N control = 70
|Sharpe et al., 2011
Functional (psychogenic or somatoform) symptoms
|RCT||CBT-based GSH vs control.
Up to 4 30-min sessions over 3 months.
|71% female, mean age 43 years, 54% working
N screened = 3,057
N treatment = 64
N control = 63
|Sumathipala et al., 2000
Medically unexplained complaints
|RCT||CBT vs structured care, 6 30-min sessions over 3 months.||78% female, mean age 35 years
N screened = 110
N treatment = 34
N control = 34
|Sumathipala et al., 2008
|RCT||CBT vs control, 3 mandatory, 30-min sessions over 3 weeks; additional 3 optional sessions were offered.||79% female, mean age 34 years
N screened = 504
N treatment = 75
N control = 75
|Zaby et al., 2008
|RCT||CBT group therapy vs PMR group therapy vs waiting-list controls.
Group interventions for CBT and PMR were 90-min weekly sessions over 8 weeks. Groups had average of 8 participants.
|78% female, mean age 46.3 years
N CBT = 42
N PMR = 35
|After 6 months of follow-up, both groups improved, but CBT-treated patients had greater reduction in doctor visits, medication use, and somatization (group x time interaction effect p < 0.05).
Outcomes at baseline to 6-month follow-up in treatment and control groups, respectively:
Health care use: mean number of doctor visits 13.4 to 8.5 vs 11.5 to 10.2, p < 0.05; mean days of medication use 501.9 to 398.1 vs 503.0 to 492.3, p < 0.05.
Somatization severity (mean BSI-somatization scale): 0.79 to 0.59 vs 0.60 to 0.61, p < 0.05.
Number of symptoms in last 7 days (mean SOMS-7): 9.5 to 7.8 vs 7.2 to 6.4, p > 0.05.
Health anxiety (mean WI): 6.5 to 5.4 vs 5.7 to 5.5, p > 0.05.
Global severity index (mean BSI-global severity index): 0.78 to 0.66 vs 0.61 to 0.59, p > 0.05.
Depression (mean BDI): 15.9 to 14.2 vs 13.2 to 11.9, p > 0.05.
Health-related internal control (mean KKG-I): 23.1 to 23.8 vs 22.7 to 21.8, p < 0.10.
Sick-leave days last month (mean): 3.8 to 1.6 vs 2.8 to 3.7, p < 0.10.
|Treatment effects at 6-month follow-up:
Change in overall health (CGI): OR 1.5 (95% CI 0.7-2.8).
Change in presenting symptoms (CPS): OR 2.3 (95% CI 1.2-4.5).
Symptom burden (PHQ-15): mean difference -0.7 (95% CI -1.5 to 0.1).
Depression (HADS): mean difference -1.0 (95% CI -2.2 to -0.2).
Anxiety (HADS): mean difference -1.4 (95% CI -2.7 to -0.2), p = 0.028.
Physical functioning (SF-12): mean difference 11 (95% CI 3-19), p = 0.008.
Belief that symptoms are permanent: mean difference -0.5 (95% CI -0.9 to -0.1), p = 0.024.
Overall quality-of-care satisfaction: OR 4.0 (95% CI 1.9-8.5)
|At 3-month follow-up: both groups showed improvement, but CBT-treated patients showed greater improvement.
Mean differences for psychiatric morbidity (GHQ-30) 4.1 (95% CI 0.5-7.6), p = 0.04; complaints 2.3 (95% CI 0.8-0.7), p = 0.001; number of visits 4.8 (95% CI 1.3-8), p = 0.004; symptoms (BSI) 2.3 (95% CI 0.5-5.2), p = 0.01.
|At 12-month follow-up, no significant differences between CBT and structured-care interventions were reported.
Mean differences for psychiatric morbidity (GHQ-30) -0.1 (95% CI -3.3 to 3.1), p = 0.9; complaints 0.8 (95% CI -0.94 to 1.13), p = 0.8; symptoms (BSI) 0.2 (95% CI -3.1 to 3.5), p = 0.9; hospital visits 0.2 (95% CI -0.8 to 1.1), p = 0.7.
|After treatment, there was significant time x group interaction over all outcomes (number and severity of somatic symptoms, SOMS-7; depression and anxiety, HADS; and healthbased quality of life, SF-12 for physical and mental health). Compared with waiting controls, CBT group had fewer (p < 0.05) and less severe (p < 0.01) symptoms, improved anxiety (p < 0.05), and better subjective mental health (p < 0.05) whereas PMR group only had fewer and less severe symptoms (both p < 0.1). Compared with PMR, CBT showed improvement in anxiety (p < 0.05) and subjective mental health (p < 0.1).|
|Abbass et al., 2009a
Halifax, Nova Scotia
|Pre–post||Dynamic psychotherapy 1–25 sessions (average 3.8). Sessions were not time limited. Physician and patient collectively made decision to terminate treatment.||70% female, mean age 36.9 years
N referred = 77
N enrolled = 50
|Aiarzaguena et al., 2002
Basque region, Spain
|Pre–post||DEPENAS (a biopsychosocial intervention).
Five 30-min visits conducted by 3rd-year family medicine doctors.
|Age 18–65 years, mean 48 years
N = 12: 9 female, 2 male, 1 uknown
|RCT||Jungian therapy vs “controls,” 1 session weekly for 6 months.||N Jungian therapy = 5
N control = 5
|Pre–post, controlled||Psychosocial group therapy vs usual care.
8 1.5-hour weekly sessions, up to 10 participants per group.
|75% were female.
N group therapy = 20
N control = 20
|Reuber et al., 2007
Functional neurologic symptoms
Barnsley, United Kingdom
|Pre–post||Psychotherapy. Initial “semistructured” assessment interview was 2 hours. Up to 19 sessions, each 50 min long.||81% female, mean age 44.2 years; 66.7% were “economically inactive”
N referred = 94
N enrolled = 91
|After 1-year follow-up, ED visits decreased from 232 visits among all 50 participants (mean 4.6 visits/participant) 1 year before treatment to 72 (mean 1.4/patient) in year after intervention (69% reduction per participant, p < 0.001). Difference in ED visits correlated, nonsignificantly, with number of treatment sessions (Spearman r = 0.22, p = 0.13). By patient subgroups, revisit rates were reduced by 16% for all ED patients and 4.3% for patients with same symptoms as those included in study, and there was no change in ED visit rates for patients referred to study but ultimately not included in it.
Self-reported symptom outcomes: 26 patients that attended two or more sessions (mean of 5.1 treatment sessions) reported significant symptom improvement (BSI 1.21 before treatment to 0.86 after treatment, p < 0.01) and somatization subscale (1.61 before treatment to 1.04 after treatment, p = 0.02). 13 patients who completed satisfaction surveys rated treatment 7.4 of 10 on average (between “satisfied” and “very satisfied”).
|Significant SF-36 pre-post improvements:
Physical role functioning –25.00 to 45.45, p = 0.03.
General health perceptions –46.18 to 58.73, p = 0.003.
Emotional role functioning –54.64 to 75.73, p = 0.17.
Moderate SF-36 improvements for bodily pain –40.91 to 49.00, p = 0.03; vitality 44.09 to 51.36, p = 0.14.
Small SF-36 improvement for mental health 58.55 to 61.09, p = 0.28.
|Treatment group improved more than controls in anxiety and fatigue measures (p < 0.001 or less).|
|Significant improvement was seen in all outcomes—number of symptoms, strength experienced and intensity of symptoms, impairment experienced because of symptoms and disease anxieties—after therapy. Frequency of PC physician visits and demand for medical tests were also reduced in treatment subjects after therapy.|
|63 patients completed median of 6 sessions (range, 1–24); 42.6% were considered by therapist to have completed treatment. Results of intent-to-treat analysis of 91 patients:
After treatment, improvement (baseline to after therapy and at 6 months) was reported for all three measures of well-being (CORE-OM; 49.9 to 42.5 to 41.3, p = 0.002), physical and mental functioning (SF-36; 77.1 to 85.0 to 85.7, p < 0.001), and symptoms (PHQ-15; 13.8 to 13.0 to 12.7, p > 0.05) that were maintained after 6 months of follow-up.
Number needed to treat to improve scored by 1 SD in at least one outcome measure = 2.0; to improve 2 or more outcome measures = 3.9; and to see improvement in all 3 outcome measures = 7.0.
|Sattel et al., 2012
|RCT||PIT vs EMC. PIT: first session was up to 90 min long; all others 45 min; 12 weekly sessions.
EMC: 3 30-min sessions at 6-week intervals.
|66% female, mean age 48 years; 42% were employed
N screened = 662
N PIT = 107
N EMC = 104
|Abbass et al., 2009b
Somatic symptom disorders
Halifax, Nova Scotia
|Systematic review||Short-term psychodynamic psychotherapies; models included Malan, group-analytic, Hobson, Davanloo, Luborsky, relaxation, affect-focused, group, individual, Sifneos, Strupp, Binder, body work. Number of sessions 5–22.||57.8% female, mean age 41.3 years
N screened > 100
N reviewed = 23
N meta-analysis = 14
|Allen et al., 2002
Multiple unexplained physical symptoms
|Systematic review||Psychosocial treatments||N = 34 studies|
|Outcomes and Results|
|Average difference between PIT and EMC scores at baseline and 9-month follow-up.
Physical functioning (SF-36)–baseline: mean –0.6 (95% CI –2.43 to 1.21), p = 0.51; 9-month follow-up: mean 2.5 (0.16–5.09), p = 0.001.
Mental functioning (SF-36)–baseline: mean 0.6 (95% CI –2.85 to 3.94), 0.75; 9-month follow-up: mean 1.1 (–2.25 to 4.55), p = 0.73.
Somatization (PHQ-15)–baseline: mean 0.4 (95% CI (–0.94 to 1.77), p = 0.48; 9-month follow-up: mean –1.12 (–2.65 to 0.31), p = 0.01.
Depression (PHQ-15)–baseline: mean 0.5 (95% CI –1.03 to 2.06), p = 0.47; 9-month follow-up: mean –0.8 (95% CI –2.46 to 0.90), p = 0.08.
Health Anxiety (WI)–baseline: mean –0.2 (95% CI –0.79 to 0.34), p = 0.42; 9-month follow-up: mean –0.3 (95% CI –0.88 to 0.31), p = 0.62.
At 9-month follow-up, PIT vs EMC:
Mean number of primary care visits: 3.2 vs 3.9, p = 0.39.
Mean number of specialist consultations: 7.4 vs 7.5, p = 0.73.
% taking concurrent antidepressant medication: 36% vs 46%, p = 0.04.
% being treated with psychotherapy: 3% vs 12%, p = 0.05.
|21/23 studies reported significant or possible symptom benefits related to main physical condition; 11/12 observed significant or possible social-occupational function improvement; 16/21 observed significant or possible psychologic symptom benefits; 7 reported significant or possible reductions in health care use. Studies included 13 medical conditions.
Meta-analyses (14 studies included): short-term (> 3 months) outcomes fixed-effect model showed moderate improvements (ES = 0.58–0.78). There were significant (p < 0.05) differences of moderate magnitude in medium-term (3–9 months) outcomes and long-term (< 9 months) outcome for general psychiatric symptoms, depression, anxiety, and somatic symptoms with the fixed-effect model (all p < 0.001). There were significant differences (p < 0.05) of moderate magnitude in the long-term outcomes (> 9 months) with the random-effect model.
Fixed-effect models for short-term and long-term somatic symptoms were significant (both p < 0.0007).
Publication bias: for short-term effectiveness, N statistic range = 41–56, indicating little publication bias; for medium term, N statistic = 16–19, indicating that results were more subject to publication bias; for long term, N statistic = 42–44 for depression and anxiety, but N for general psychiatric conditions = 14, and N for somatic symptoms = 12, suggesting that results for psychiatric and somatic symptoms were more likely influenced by publication bias than results for depression and anxiety.
|No significant associations with treatment outcome were found for different diagnoses, treatment types, treatment formats, or control conditions. Only 2 of 34 studies looked at unexplained physical symptoms. Effect sizes were calculated for 11 studies and ranged from 0.20 comparing dynamic therapy with standard medical care in IBS patients to 4.01 comparing EMG biofeedback with false EMD in fibromyalgia patients. continued|
|Kleinstäuber et al., 2011
|Systematic review||Psychotherapies, number of sessions 1–72 (mean 11). Mean therapy duration 92 days (range, 1–365 days).||72% female, mean age 44.4 years
N screened = 171
N reviewed = 27
Somatization, somatoform disorders, persistent symptoms, or symptom syndromes
24 studies used fixed number of sessions with median of 8 sessions (range, 2–16). 7 studies tailored number of sessions to patient.
|N RCT = 29
N nonrandomized controlled trials = 2
|Systematic review||N RCTs = 34
N subjects = 3,922
|Outcomes and Results|
|Weighted mean effect sizes for outcomes after 1 year of follow-up:
Physical symptoms—between-group contrast (BGC): 0.40 (95% CI 0.1–0.7, p < 0.01); within-group contrast (WGC): 0.80 (95% CI 0.53–1.07, p < 0.001).
Cognitive emotional behavioral symptoms—BGC: 0.33 (95% CI 0.17–0.5, p < 0.001); WGC: 0.57 (95% CI 0.27–0.86, p < 0.001).
Depression—BGC: 0.16 (95% CI 0.02–0.30, p < 0.01); WGC: 0.40 (95% CI 0.18–0.63, p < 0.01).
General psychopathology—BGC: 0.06 (95% CI –0.12 to 0.24, p = 0.5); WGC: 0.63 (95% CI 0.42–0.83, p < 0.001).
Functional impairment—BGC: 0.20 (95% CI 0.05–0.34, p = 0.01); WGC: 0.42 (95% CI 0.32–0.52, p < 0.001).
Health care use—BGC: 0.33 (95% CI 0.04–0.61, p < 0.05); WGC: 0.44 (95% CI 0.22– 0.66, p < 0.001).
On basis of 18 controlled studies, effect of psychotherapy on physical symptoms after treatment was significant (ES 0.33, 95% CI 0.23–0.43, p < 0.001).
Moderator analyses showed largest effect on physical symptoms from reattribution training after follow-up (ES 0.76, 95% CI 0.46–1.06).
Weighted metaregression showed that efficacy increased with more sessions in models for physical symptoms, cognitive emotional and behavioral symptoms, and depression (all p < 0.05).
Authors concluded that it was unlikely that enough null-result studies were missed to decrease total weighted mean ES significantly from 0.22 to 0.05.
|Most studies showed definite (71% of studies) or possible (11% of studies) treatment effect for CBT.
Definite or possible treatment benefit for psychologic distress was established in 38% and 8% of 26 studies examining this outcome.
Benefits of CBT regarding functional status were intermediate, showing definite or possible treatment effects in 47% (n = 9) and 26% (n = 5) of 17 studies examining this outcome.
In 12 studies (3 on hypochondriasis, 2 on back pain, 2 on IBS, 1 each on CFS, chest pain, tinnitus, fibromyalgia, and unexplained somatic symptoms), patients who received CBT showed greater improvement than controls.
In 6 studies (2 on back pain, 1 each on chest pain, tinnitus, electric sensitivity, and unexplained somatic symptoms), patients who received CBT did not improve more with regard to cognitive-behavioral measures.
|Positive results for at least one outcome (symptoms, functional, or psychologic) were noted in 11 of 13 CBT studies and 4 of 5 antidepressants studies, whereas 8 of 16 studies of other interventions reported improvement (consultation letter to the primary care physician [4 studies], PC physician training [2 studies], and non-CBT psychotherapies [2 studies], hypnosis [2 studies], and one study each investigating multicomponent nurse-care management, aerobic exercise, writing disclosure, paradoxic intention, and explanatory therapy).
CBT is most effective for variety of somatoform disorders and outcomes.
Somatization disorder, MUS
|Systematic review||CBT; consultation letter; group therapy||N somatization disorder = 4 studies
N MUS = 5 studies
|Nezu et al., 2001
|Systematic review||CBT individual and group therapy.
Number of sessions 4–24, 40–180 minutes.
|N MUS = 9 studies|
|Systematic review||Miscellaneous.||N abstracts screened = 368
N systematic reviews = 6
N RCTs = 14
|van Rood and de Roos, 2009
Number of sessions 1–72 over 1 week–18 months.
|64% female; of 57 patients with available information, average age 43 years
N screened = 181 results screened
N included = 16 studies (102 patients)
NOTE: BDI = Beck Depression Inventory; BSI = Brief Symptom Inventory; CBT = cognitive behavioral therapy; CFS = chronic fatigue syndrome; CGI = Clinical Global Impressions Scale; CI = confidence interval; CORE-OM = Clinical Outcomes in Routine Evaluation Outcome Measure; DEPENAS = Detection-Explanation-Plan-Exploration-Normalization-Action-Follow-Up; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; ED = emergency department; EMC = enhanced medical care; EMDR = eye movement desensitization and reprocessing; EQ-5D = EuroQol health status questionnaire; ES = effect size; FLZM = Questions on Life Satisfaction; GHQ-30 = 30-Item General Health Questionnaire; GSH = guided self-help; HADS = Hospital Anxiety and Depression Scale; HAM-A = Hamilton Anxiety Rating Scale; HAM-D = Hamilton Depression Rating Scale; IAS = Illness Attitude Scale; IBQ = Illness Behaviors Questionnaire; IBS = irritable bowel syndrome; MFI = Multi-dimensional
|Outcomes and Results|
|Somatization disorder: 4 studies show consultation letters are effective in reducing excessive and expensive help-seeking associated with somatization disorder but do not improve psychologic distress of patients. Group therapy provided additional benefit not only in reducing health care costs but in improving psychologic well-being.
MUS: 3 RCTs demonstrate that individual CBT is effective in reducing somatic symptoms with effect sizes in moderate range.
|Of 9 studies, only 4 met criteria allowing for calculation of effect sizes regarding efficacy of CBT. Effect-size scores for CBT ranged from 0.23 to 0.67 (mean = 0.48) for physical symptoms and 0.1 to 0.72 (mean = 0.39) for psychologic distress; this provides quantitative support for notion that CBT is effective.|
|Resulting studies assessed interventions that used antidepressants (1 review and 3 RCTs), CBT (5 reviews, 2 RCTs), psychiatric consultation (3 RCTs), other forms of psychotherapy (3 RCTs), exercise (2 RCTs), and collaborative care (1 RCT).
Authors note level I (systematic review) evidence that antidepressants and CBT benefit patients who have MUS, with limited level II (RCT) evidence for other interventions. No studies compared different interventions for MUS.
|3 studies reported improvement in formal measures of chronic pain with decrease of 1.2–2 points (on a 10-point scale) (58 patients). Phantom-limb pain also improved (4 studies, 21 patients) by average of 4.7 points (on a 10-point scale). Another 9 studies reported clinically observed effects.
6 studies showed improvement in the form of decreased posttreatment Impact Event Scale scores compared with pretreatment score (scores decreased to nonclinical levels).
6 studies reported depression scores with BDI showing decreased score from pretreatment to posttreatment to nonclinical or mild levels.
Fatigue Inventory; MUPS = medically unexplained physical symptoms; MUS = medically unexplained symptoms; OR = odds ratio; PC = primary care; PCI = psychiatric consultation intervention; PHQ-15 = Patient Health Questionnaire 15-Item Somatic Symptom Severity Scale; PIT = psychodynamic interpersonal therapy; PMR = progressive muscle relaxation; PSC-51 = Physical Symptoms Checklist; RCT = randomized controlled trial; SCL-90-R = Symptom Checklist-90-Revised; SDI = Sleep Disturbance Inventory; SF-12 = Medical Outcomes Study 12-Item Short Form Health Survey; SF-36 = Medical Outcomes Study 36-Item Short Form Health Survey; SOMS-7 = Screening for Somatoform Symptoms-7; SPQ = Social Problems Questionnaire; V/SF-36 = Medical Outcomes Study 36-Item Short Form Health Survey for Veterans; VAS = Visual Analogue Scale; WI = Whitely Index.
underlying disorders. Participants met the abridged criteria for subthreshold somatization disorder (4 or more of a list of 42 possible somatic symptoms included in the Composite International Diagnostic Interview in men and 6 or more in women). Thirty-eight patients were randomized to the intervention or control (consultation letter) group and received an average of 10 sessions of manualized biofeedback tailored to a patient’s specific symptom profile, progressive muscle relaxation, and self-hypnosis. Results showed statistically significant improvement in somatic and related symptoms in the intervention group compared with controls despite the relatively low number of patients. There was no follow-up after the intervention, so it is not known whether gains were maintained after the intervention had stopped.
Nanke et al. (2003) included patients who had diagnoses of multiple somatoform symptoms. Patients met criteria for an abridged construct of somatization, which required eight or more unexplained somatic symptoms in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) list of somatic symptoms. The symptoms had to be chronic and disabling. Patients were part of a general treatment program in a hospital that consisted of CBT, physical therapy, and medical care. Fifty patients were randomly assigned to individual biofeedback or group-relaxation intervention. Outcomes focused primarily on cognitive elements (“catastrophizing”) rather than somatic symptoms. After treatment, patients who received biofeedback had statistically significant improvement for catastrophizing of somatic complaints compared with patients who received group relaxation. Changes in somatic symptoms themselves and maintenance of gains after the end of the intervention were not reported.
Finally, Saldanha et al. (2007) conducted a trial of biofeedback and relaxation alone (N = 30) or jointly with pharmacotherapy (N = 30) compared with pharmacotherapy alone (N = 30) for 12 weeks and followed for up to a year in 90 participants. It does not appear that somatic symptoms were a primary target of the study inasmuch as patients who entered the study had either anxiety or depressive disorders. Patients who received biofeedback plus pharmacotherapy attained better reduction in anxiety than those on drug treatment alone or biofeedback alone, but no statistical analyses were reported.
Only Katsamanis et al. (2011) appeared to present generally low or unclear risk of bias. Although Nanke et al. (2003) used random assignment, it is unclear whether subjects or interviewers were properly blinded, so the study was judged to have a high risk of bias. Saldanha et al. (2007) also was determined to have a high risk of bias.
There have been only two controlled studies, both of which used small samples, multiple other intervention components included with biofeedback, and no follow-up. Only one of the two (Katsamanis et al., 2011) reported somatic outcomes. No studies of biofeedback targeted veterans
who had CMI. On the basis of guidance provided by AHRQ (Owens et al., 2010), the committee rated the strength of evidence as shown in Table 4-7 and concluded that there is insufficient evidence to determine the effectiveness of biofeedback for somatic syndromes. That conclusion is largely because there is no substantial body of high-quality randomized studies in the subject. The studies of biofeedback interventions are briefly summarized in Table 4-8.
Cognitive Rehabilitation Therapy
Cognitive rehabilitation therapy (CRT) is a term used to describe a variety of interventions that have been developed to improve the cognitive function of people who have cognitive impairments. Interventions have been developed in a variety of domains—including attention, memory, processing speed, and executive function—and can be delivered individually or in a group format.
One study assessed the effect of CRT on symptoms of Gulf War syndrome in a small sample of veterans (N = 14) (Jakcsy, 2002). Slight improvements on neuropsychologic test scores were reported, but results were inconclusive because of methodologic limitations and the small sample.
On the basis of guidance provided by AHRQ (Owens et al., 2010), the committee rated the strength of evidence as shown in Table 4-9 and concluded that the evidence is insufficient to determine the effectiveness of CRT for symptoms of CMI in veterans. Details of Jakcsy (2002) are summarized in Table 4-10.
Complementary and Alternative Therapies
Complementary and alternative medicine (CAM) is a broad field of health care that encompasses the use of techniques to promote well-being in conjunction with conventional medicine (complementary) or in place of conventional medicine (alternative). CAM techniques to promote physical,
|RoB||Consistency||Directness||Precision||Strength of Evidence|
|RCT: 2 (88)a||High||Inconsistent||Indirect||Imprecise||Insufficient|
|Clinical trial: 1 (90)b||High|
NOTES: RCT = randomized controlled trial; RoB = risk of bias.
aKatsamanis et al., 2011; Nanke et al., 2003.
bSaldanha et al., 2007.
|Katsamanis et al., 2011
Piscataway, New Jersey
|RCT||Psychophysiologic therapy (biofeedback) vs control.
10 sessions of psychophysiologic treatment over about 10 weeks.
|N screened = 57
N control = 20
N treatment = 18
|Nanke et al., 2003
Multiple somatoform syndrome
|RCT||Biofeedback vs relaxation therapy.
3 50-min sessions a week for 2 weeks.
|N biofeedback = 25
N relaxation therapy = 25
|Saldanha et al., 2007
Neuroses and psychosomatic disorders
|Nonrandomized uncontrolled clinical trial||Biofeedback vs psychoactive drug therapy vs biofeedback and drug therapy.
5 30-min sessions a week for first 4 weeks followed by 3 30-min sessions a week for 8 weeks.
|N drugs = 30
N biofeedback = 30
N drugs and biofeedback = 30
NOTE: BDI-II = Beck Depression Inventory Second Edition; CABAH = Cognitions About Body and Health Questionnaire; CGI = Clinical Global Impression Scale; MUPS = medically unexplained physical symptoms; SF-36 = Medical Outcomes Study 36-Item Short Form Health Survey; TMAS = Taylor Manifest Anxiety Scale.
|Outcomes and Results|
|75% of the treatment group attended all 10 sessions.
Estimated score change difference (p value) from baseline to final assessment (control group vs biofeedback group), corrected for age:
Somatization disorder severity (CGI): –0.77, p = 0.04, d = 0.8.
Depression score (BDI-II total): –5.7, p = 0.03, d = 0.81.
Depression, neurovegetative subscale (BDI-II): –2.21, p = 0.045, d = 0.70.
Functioning (SF-36) total: 21.57, p = 0.0011.
Physical functioning (SF-36): 6.83, p = 0.0037.
Mental functioning (SF-36): 8.27, p = 0.0027.
Normal (SF-36): 8.18, p = 0.0009.
Treatment group was significantly younger than the control group (p < 0.01).
|Results revealed significant group x treatment interaction effect for reduction of “catastrophizing interpretation of bodily complaints” (CABAH, p < 0.05) but no significant effect of group or treatment alone (pretreatment and posttreatment scores, 23.1 and 19.7 vs 15.2 and 16.7, for biofeedback and relaxation, respectively) with significant reduction from baseline to posttreatment in biofeedback group (p < 0.05); no significant improvement in “intolerance of bodily complaints,” “concept of bodily weakness,” or “sensitivity to autonomic sensations.”
From baseline to posttreatment, causal attribution scores for psychosocial attribution improved in biofeedback patients (2.7 to 3.0, p < 0.01) but not relaxation patients (2.9 to 2.9, p = 0.66); significant group x treatment interaction was also found (p = 0.011). Causal attribution of symptoms to organic, genetic, and environmental factors did not change from baseline to posttreatment in either group.
Sessions ratings for credibility and treatment success were significantly higher in biofeedback patients than relaxation patients (p < 0.001 and p < 0.05, respectively) and both groups' ratings improved throughout treatment (both p < 0.001). No differences in ratings of therapeutic rapport were noted.
|Numbers of patients with severe (> 36), moderate (31–35), and mild (26–30) anxiety scores (TMAS) from baseline to 1 year:
Drugs group: severe, 5 patients at baseline, none at follow-up; moderate, 15 patients at baseline and 5 at follow-up; mild, 10 at baseline and 17 at follow-up; no anxiety, none at baseline or follow-up (p < 0.05).
Biofeedback group: severe, 4 patients at baseline and none at follow-up; moderate, 16 patients at baseline and 10 at follow-up; mild, 6 at baseline and 9 at follow-up; no anxiety, none at baseline and 4 patients at follow-up (p < 0.05).
Drugs and biofeedback group: severe, 3 patients at baseline and none at follow-up;
moderate, 18 patients at baseline and 3 at follow-up; mild, 8 at baseline and 6 at follow-up;
no anxiety, 20 at baseline and 18 at follow-up (p < 0.02).
Depression scores (BDI) at baseline and follow-up were not reported.
|RoB||Consistency||Directness||Precision||Strength of Evidence|
|Pre-posta: 1 (14)||High||N/A||Direct||Imprecise||Insufficient|
NOTE: RoB = risk of bias.
Mild memory problems
Black Hills Health Care System, VA campus at Fort Meade
|Pre–post||Computer-assisted cognitive retraining vs control training, 6 4-hour sessions over 6 weeks.||Persian Gulf War veterans, 50% female, all white, average age 36 years in treatment group and 39 years in control group
N eligible = 80
N treatment = 10
N control = 10
mental, and emotional well-being vary widely and include the use of natural products, such as probiotics and herbal or botanic medicines, and mind– body practices, such as acupuncture, deep-breathing techniques, yoga, and Tai Chi (NIH, 2012). The literature search found several published studies of complementary and alternative therapies, including acupuncture, movement therapies, Kampo, and St. John’s wort supplements.
Acupuncture was investigated in one trial. Two publications reported on the same RCT that compared acupuncture with a control condition; Rugg et al. (2011) reported a longitudinal analysis of qualitative data, and Paterson et al. (2011) reported the quantitative results from the Classical Acupuncture Treatment for People with Unexplained Symptoms (CACTUS) study. Paterson et al. (2011) had a strong design for determining attribution, but it was not blinded. Results indicate that acupuncture improves well-being and health status. The qualitative information presented by Rugg et al. (2011) indicates that the acupuncturists in the study did a fair amount of counseling, and this leaves unclear what produced the effect. The risk of bias in the study was low.
Two of the studies assessed used body-movement-based approaches focused on a body–mind connection in psychotherapy. The study by Nickel
|Outcomes and Results|
|Outcomes assessed 1 week to 2 months after treatment. Comparing treatment and control groups on 19 subtests to assess left hemisphere function and verbal memory and right hemisphere function and motor skills showed no significant differences in mean changes in test scores (Weschler Adult Intelligence Scale–Revised or Third Edition; Weschler Memory Scale–Revised or Third Edition; Rey Auditory Verbal Learning Test; Wisconsin Card Scoring Test; Controlled Oral Word Association Test; Porteus Mazes; Halstead Reitan Battery; and Sensory Perceptual Examination) between treatment and control groups except for Sentence Repetition Test scores (mean change scores, treatment = 4.4 and control = –5.0 , p < 0.01).
Comparison of test scores on record for a cohort of normal veterans and a cohort of depressed veterans showed no significant differences in mean group scores.
14 veterans completed the study (7 in each group).
et al. (2006) included a reasonable sample size and RCT design. It was conducted on a population of 128 inpatient Turkish immigrants and found that bioenergetic exercises led to statistically significantly decreased somatization symptoms, social insecurity, depression, anxiety, hostility, anger, and tendency to direct anger inward. Payne and Stott (2010) examined the effects of dance-movement psychotherapy on 17 patients who had medically unexplained symptoms and found statistically significant increases in activity and well-being. The effect sizes were moderate to large (0.31–0.72), but the study design was poor and thus had a high risk of bias.
Beneficial effects of Kampo, an adaptation of traditional Chinese medicine, on quality of life in patients who had undifferentiated somatoform or conversion disorders were described in a poorly reported study (Yamada et al., 2005). After 3 months of Kampo, 120 subjects reported improvements in physical and psychologic health aspects of life. No information on the quality and standardization of the herbal formula was presented, and the study was judged to have a high risk of bias.
Two studies examined the potential effect of St. John’s wort on somatoform disorders (Muller et al., 2004; Volz et al., 2002). Both were conducted in Germany as RCTs. Muller et al. (2004) conducted a study at a single
site. Volz et al. (2002) recruited participants from 21 sites. Both studies conducted an intention-to-treat analysis with more than 150 participants. They both found consistent and fairly large effect sizes, reflecting significant improvements in anxiety and somatoform symptoms. Despite some weaknesses, both studies seem to have been rigorously conducted and to have had a low risk of bias.
On the basis of guidance provided by AHRQ (Owens et al., 2010), the committee rated the strength of evidence as shown in Table 4-11 and concluded that
• There is low strength of evidence that acupuncture improves well-being in people who have medically unexplained symptoms.
• There is insufficient evidence to support conclusions about the effectiveness of movement therapy in people who have medically unexplained symptoms or of Kampo in people who have undifferentiated somatoform disorder.
• There is moderate strength of evidence that St. John’s wort improved symptoms in people who had somatoform disorders.
Details of each of the studies of complementary and alternative therapies are briefly summarized in Table 4-12.
|RoB||Consistency||Directness||Precision||Strength of Evidence|
|RCT: 1 (80)a||Low||N/A||Direct||Imprecise||Low|
|RCT: 1 (128)b||High||N/A||Indirect||Imprecise||Insufficient|
|Pre–post: 1 (18)c||High|
|Pre-post: 1 (120)d||High||N/A||Indirect||Imprecise||Insufficient|
|St. John's Wort|
|RCT: 2 (324)e||Low||Consistent||Direct||Precise||Moderate|
NOTES: RCT = randomized controlled trial; RoB = risk of bias.
aPaterson et al., 2011; Rugg et al., 2011.
bNickel et al., 2006.
cPayne and Stott, 2010.
dYamada et al., 2005
eMuller et al., 2004; Volz et al., 2002.
Regular exercise promotes physical health, improves mood, and helps to protect from chronic diseases, such as heart disease. Exercise may also be helpful in managing a variety of illnesses (CDC, 2011).
Two studies conducted research on the effects of exercise on unexplained physical symptoms. One described the effects of CBT and exercise compared with each alone and usual care in veterans (Donta et al., 2003; Guarino et al., 2001; Mori et al., 2006). The authors showed that a combination 12-week program of group CBT and exercise led to a modest improvement in the primary outcome measure of physical function at 1 year. CBT accounted for nearly all the combined treatment effect. In contrast, almost no improvement in physical function was observed with exercise alone. The study was limited by serious lack of compliance. Only 87 (16%) attended all 12 exercise sessions, and 39 (7%) all 12 CBT sessions. How adherence with off-site exercise was monitored was not discussed. In the exercise group, participants who were adherent (attended at least eight sessions) were more than twice as likely to improve as nonadherent participants (OR 2.67, 95% CI 1.20–5.92). Even if patients were adherent, it is not clear whether the frequency and duration (1 hour per week for 12 weeks in the presence of a therapist and two or three times per week on their own) of exercise prescribed would have been adequate.
In the second study, a program of group aerobic exercise (1 hour twice a week for 10 weeks) augmented with presentation of explanatory models to link the content of training with potential amelioration of patients’ symptoms reduced primary care consultations and prescriptions compared with stretching alone (Peters et al., 2002). Although there were significant improvements in both groups after 6 months, there was no difference between the two groups in patient-reported outcomes. Patient effort was titrated to maintain heart rate at 60–65% of the age-adjusted maximum, and stretching was titrated to maintain heart rate below 50% of the age-adjusted maximum.
On the basis of guidance provided by AHRQ (Owens et al., 2010), the committee rated the strength of evidence as shown in Table 4-13 and found insufficient evidence of the benefit of exercise in veterans who had CMI. Details of each of the studies of exercise interventions are briefly summarized in Table 4-14.
The committee encountered several difficulties when trying to make determinations of the effectiveness of treatments for CMI that precluded the drawing of conclusions with a high level of confidence. The available
|Muller et al., 2004
Unexplained multiple somatic symptoms, somatization and undifferentiated somatoform disorders, and somatoform autonomic dysfunction
|RCT||SJW (Hypericum) vs placebo.
300 mg of SJW extract administered twice daily for 6 weeks.
|Treatment and control groups 60% and 76% female, respectively; mean age 47.6 years
N screened = 184
N treatment = 87
N placebo = 88
|Nickel et al., 2006
Chronic somatoform disorders
Simbach am Inn, Germany
|RCT||BE vs light gymnastics. Over 6 weeks, all patients received 2 individual psychotherapy sessions, 3 interactional group sessions, and 5 group sessions a week (all 60 mins). BE interventions were twice a week for 60 min. Control groups did light gymnastic exercise twice a week for 60 min.||About 15% of participants were Turkish immigrants; BE group mean age 48.3 years, spent 24.5 years in central Europe, and 76.6% were employed as laborers
Control group mean age 49.4 years, spent 23 years in central Europe, and 71.9% were laborers
N screened = 135
N treatment = 64
N control = 64
|Paterson et al., 2011||RCT||Acupuncture vs waitlist.
12 60-min sessions of acupuncture over 6-month period (approximately weekly, then fortnightly, and monthly).
|80% of participants female, mean age 51 years
N screened = 200
N treatment = 39
N control = 41
|Rugg et al., 2011
London, United Kingdom
|Longitudinal qualitative interview study nested in RCT trial by Paterson et al.||Acupuncture. Two interviews lasting 45–60 min.||80% female, mean age 56 years
N participants = 20
|Outcomes and Results|
|At the end of treatment, statistically significant improvement was noted for somatic symptoms as percent change in scores from baseline to week 6 between treatment and placebo groups as measured by SOMS-7: .46.5 vs .21.5, p . 0.0001; HAM-A somatization subscale: .54.7 vs .34.0, p . 0.0001; SCL-90-R somatization subscale: .44.1 vs .15.9, p . 0.0001. Clinical improvement was also shown (CGI improvement, CGI efficacy, and global judgment of efficacy by patient, all p . 0.0001). 45.4% of treatment and 20.9% of placebogroup patients were classified as responders (p = 0.0006).
Tolerability and safety of SJW treatment was comparable with those of placebo. 24.1% of treatment-group patients reported total of 27 adverse events, whereas 18.1% of placebogroup patients reported total of 24 adverse events.
|At end of treatment, symptom checklist (SCL-90-R) scores improved significantly more in BE than in control group for somatization (BE 75.4–64.1 vs control 74.4–69.3, p < 0.001), insecurity in social contexts (BE 63.5–58.1 vs control 62.4–58.7, p = 0.02), depression (BE 76.9–66.2 vs control 77.1–68.6, p = 0.03), anxiety (BE 65.3–58.1 vs control 66.1–61.1, p = 0.04), and hostility (BE 77.2–63.1 vs control 78.2–66.3, p = 0.01) but there was no significant difference in change for obsessiveness, phobic anxiety, paranoid thinking, or psychoticism. Differences in anger (measured with STAXI) were significantly greater in BE group than in control group for state anger (BE 31.9–27.3 vs control 32.2–30.7, p =0.01) and anger-in (BE 26.2–18.3 vs control 22.1–20.5, p < 0.001). Score improved more in control group than in BE group for anger-out (BE 20.9–22.0 vs control 23.0–22.0, p =0.022) with no significant difference for trait anger or anger control scores.|
|At 52 weeks, significant differences were noted between two groups. Adjusted mean differences for patient-reported symptoms (MYMOP) were 0.8 (95% CI 0.2–1.4), p =0.017; well-being (W-BQ12) 3.8 (95% CI 1.5–6.1), p = 0.022; and generic outcomes (EQ-5D) 0.13 (95% CI 0.02–0.24), p = 0.03.
At 52 weeks, improvements in acupuncture group were maintained. Control group had received 6 months of acupuncture by this point and appeared to show a catchup improvement in all outcome measures.
|“Almost all patients cited physical, psychological, and/or social changes that took place during or after their acupuncture. Some linked these directly to acupuncture, but others were unsure whether there was an association. Most of the cited changes were positive, although a few patients said that treatment had not tackled expected problems and/or didn't help at all” (Rugg et al., 2011, p. e310).|
|Payne and Stott, 2010; Payne, 2009
|Pre–post||BMA with DMP.
Four parts comprising 4 sessions lasting 2 hours each for 12 weeks.
|83% of participants female, mean age 48 years
N screened = 31
N enrolled = 18
|Volz et al., 2002
|RCT||SJW vs control.
Participants received placebo or SJW tablets twice daily for 6 weeks. SJW was administered as 300-mg tablets.
|72% of treatment group female, mean age 46.9 years; 62% of placebo group female, mean age 48.6 years
N treatment = 75
N control = 74
|Yamada et al., 2005
USD and CD
|Pre–post||Kampo (with supportive psychotherapy).
Duration of treatment 3 months.
|74% of participants female, mean age 47.1 years; 94 met criteria for USD, 26 criteria for CD
N = 120
NOTES: BE = bioenergetic exercise; BMA = body–mind approach; CD = conversion disorder; CGI = Clinical Global Impressions Scale; CORE-OM = Counseling Outcome Routine Evaluation–Outcome Measure; DMP = dance-movement psychotherapy; EQ-5D = EuroQol health status questionnaire; ES = effect size; HAM-A = Hamilton Anxiety Rating Scale; MUPS = medically unexplained physical symptoms; MUS = medically unexplained symptoms; MYMOP = Measure Yourself Medical Outcome Profile; SCL-90-R = Symptom Checklist-90-Revised; SJW = St. John's wort (Hypericum); SOMS-7 = Screening for Somatoform Symptoms-7; STAXI = State-Trait Anger Expression Inventory; USD = undifferentiated somatiform disorder; W-BQ12 = 12-Item Well-Being Questionnauire; WHOQoL-BREF = World Health Organization Quality of Life Scale Brief Version.
|Outcomes and Results|
|At 3-month follow-up, significant mean changes (CORE-OM, baseline to follow-up) were reported for well-being –0.76 (95% CI –1.15 to –0.38), p < 0.001, ES = 0.72; problems –0.70 (95% CI –1.13 to –0.26), p < 0.004, ES = 0.65; function –0.54 (95% CI –0.93 to –0.14), p < 0.011, ES = 0.58; risk –0.15 (95% CI –0.38 to 0.09), p < 0.204, ES = 0.31; all nonrisk –0.64 (95% CI –1.03 to –0.25), p < 0.003, ES = 0.66; all items –0.56 (–0.90 to –0.21), p < 0.003, ES = 0.66.
Significant improvements for specific symptoms were also shown (MYMOP, baseline to follow-up) for symptom one –1.7 (95% CI –2.6 to –0.8), p < 0.001, ES = 0.71; activity –2.1 (95% CI –3.4 to –0.9), p < 0.003, ES = 0.66; well-being –1.7 (95% CI –2.6 to –0.8), p < 0.001, ES = 0.71; symptom two –1.8 (95% CI –3.2 to –0.5), p < 0.012, ES = 0.67.
Facilitator and participant perceptions were compared and found to be largely congruent;
participants learned how their emotions are linked to physical well-being and how to avoid or cope with symptoms in different ways.
|At end of treatment, statistically significant differences were reported between treatment and control groups from day 0 to day 42 with regard to somatic anxiety (HAM-A somatization subscale SJW 15.39–6.64 vs placebo 15.55–11.97, p = 0.001), anxiety (HAM-A SJW 22.09–10.0 vs placebo 22.47–17.0, p = 0.0001), psychic anxiety (HAM-A psychic subscale SJW 6.71–3.36 vs placebo 6.92–5.03, p = 0.0001), depression (HAM-D SJW 10.59–5.43 vs placebo 10.80–8.08, p = 0.0001), psychic symptoms (SCL-90-R SJW 61.65–29.39 vs placebo 66.37–50.50, p = 0.0001), somatic symptoms (SCL-90-R somatization subscale SJW 15.57–6.84 vs placebo 15.95–12.50, p = 0.0001).
9 adverse events were observed in 8 participants in treatment group and 5 adverse events in 4 participants in control group.
|After treatment, 13 participants were “very much improved,” 37 “much improved,” 39 “minimally improved,” 9 “no change,” 1 “minimally worse,” and 1 “much worse.” Quality-of-life (WHOQoL-BREF) score improved from 3.08 at baseline to 3.21 after 3 months (p = 0.0001). Subscores also showed significant improvement in “physical health” (p = 0.0004) and “psychological health” (p = 0.0006), and nonsignificant changes were found in “social relationships” and “environment.”
Adverse events were reported in 6 patients.
|RoB||Consistency||Directness||Precision||Strength of Evidence|
|RCT: 2 (1,320)a||Intermediate||Inconsistent||Indirect||Imprecise||Insufficient|
NOTES: RCT = randomized controlled trial; RoB = risk of bias.
aDonta et al., 2003/Mori et al., 2006; Peters et al., 2002.
|Donta et al., 2003;
Mori et al., 2006
Gulf War Veterans' Illnesses
18 VA and 2 DOD medical centers
|RCT||CBT vs CBT + exercise.
CBT was 60–90 min long; groups met weekly for 12 weeks.
Aerobic exercise was one 1-hour session per week with an exercise therapist for 12 weeks and 2 or 3 independent sessions per week for 12 weeks.
|Gulf War Veterans deployed August 1990 to August 1991; 15% female, mean (SD) age 40.7 (8.7) years; 24% with disability claim
N screened = 2,793
N CBT + exercise = 266
N CBT only = 286
N exercise only = 269
N usual care = 271
|Peters et al., 2002
Liverpool, United Kingdom
|RCT||Aerobic exercise vs stretching, 1-hour sessions twice weekly for 10 weeks.
Homework: exercise or stretch for 20 min 3 times weekly.
|53% female, mean age 44 years (range 9–73), 46% employed
N screened = 323
N aerobic training = 114
N stretching training = 114
NOTES: CBT = cognitive behavioral therapy; GP = General Practitioner; HADS = Hospital Anxiety and Depression Scale; MSPQ = Modified Somatic Perception Questionnaire; PUPS = persistent unexplained physical symptoms; RCT = randomized controlled trial; SD = standard deviation; V/SF-36 = Medical Outcomes Study 36-Item Short Form Health Survey for Veterans.
|Outcomes and Results|
|At 12-month follow-up, physical functioning (V/SF-36): CBT + exercise adj OR = 1.84 (95% CI 0.95–3.55); CBT adj OR = 1.72 (95% CI 0.91–3.23); exercise adj OR = 1.07 (95% CI 0.63–1.82); CBT vs no CBT OR = 1.71 (95% CI 1.21–2.41); exercise vs no exercise adj OR = 1.07 (95% CI 0.76–1.50). Mean change from baseline: CBT 0.59, CBT + exercise 1.03, exercise 0.97.
Exercise alone or in combination with CBT significantly improved fatigue, cognitive symptoms, mental health functioning; CBT alone significantly improved cognitive symptoms and mental health functioning, p < 0.025. Neither had significant impact on pain.
In treatment phase, 44.9% in exercise group and 40.2% in CBT + exercise group complied (p = 0.28); in maintenance phase, 24.9% of exercise and 21.1% of CBT + exercise patients complied (p = 0.29).
|6 months after training, aerobic exercise and stretching resulted in improvement from baseline with fewer symptoms recorded by GP; fewer GP consultations; fewer prescriptions; fewer secondary-care contacts; and fewer new referrals to secondary care (all p < 0.01).
Anxiety and depression (HADS) and somatization (MSPQ) declined from start of program to 6-month follow-up (all p < 0.01). Reported life interference, energy, mental health, and social function improved from randomization to 6-month follow-up (all p < 0.001). No significant changes were noted in SF-36 scores. Most improvement occurred during training and did not continue in 6 months after.
No significant difference between effects of stretching and aerobic exercise on any variable.
Regression models show that treatment attendance significantly reduced visits to GPs (p < 0.001), number of prescriptions (p < 0.01), number and type of prescription refills (p < 0.001) from start of treatment to 6-month follow-up but was not related to measures of symptoms and somatization.
evidence had methodologic limitations and biases that may have masked true treatment effects, and there was great diversity in study populations, types of interventions conducted, controls used, and types of outcomes assessed and measures used.
Study quality was generally poor, and samples were small (from 10 to 1,092 participants, median 87, mean 150), and many studies had high dropout rates. Most of the studies reviewed were subject to multiple sources of bias or did not clearly describe methods, and this resulted in a high proportion of studies at high or unclear risk of bias. Confounding diagnoses, such as depression and anxiety, may affect the magnitude of treatment effect and were not examined. For example, using an antidepressant to treat patients who have CMI with minor depression may improve symptoms because of changes in mood rather than changes in CMI symptoms. Some of the studies attempted to avoid that problem by excluding people who met diagnostic criteria for some psychiatric disorders or medical illnesses, but subsyndromal or undiagnosed conditions may have remained in study samples.
The heterogeneous nature of the literature limited the committee’s ability to compare the efficacy of different interventions. With the exception of CBT, each intervention was the subject of only a few studies that evaluated its efficacy. The study populations were defined in different ways; some studies examined only a few symptoms not explained by medical illness, and others required diagnosis of a specific disorder for various durations and of various degrees of severity. Methods differed greatly as well: there were few blinded RCTs, and many clinical trials were influenced by varied sources of bias. The literature also used a variety of outcomes to define treatment success—such as number of symptoms, severity of symptoms, health care use, functional improvement, and changes in mood—and this made comparison difficult in that different studies used different outcomes. In addition, the numerous outcomes were measured in different ways with different tools, so comparability of studies was limited further.
Generalizability among the populations studied is questionable. Only three studies were conducted in military or veteran populations. Almost all the other study samples were dominated by middle-aged women. Thus, it is difficult to extrapolate the efficacy and acceptability of an intervention to the veteran population of interest, which is slightly younger on the average and mostly male.
Adverse effects of pharmaceuticals are typically reported, but other types of interventions may have harmful effects that are not identified or investigated. The method of adverse-event reporting may be important; use of a standardized list of adverse events may cause study participants who are suggestible (such as people who are seeking psychologic health
care) to report more events than they would have otherwise (Kroenke and Rosmalen, 2006).
Length of follow-up was variable, although generally short, ranging from 1 week to 18 months (median 3 months) between the start of a study and the last assessment. Many studies collected data only immediately after treatment, whereas others assessed the effect of treatment many months after treatment had ceased.
The committee’s recommendations—based on the weight of the evidence described above, treatments for comorbid and other related conditions (Chapter 5), and issues related to patient care and communication (Chapter 6 and 7)—are presented in Chapter 8.
Abbass, A., S. Campbell, K. Magee, and R. Tarzwell. 2009a. Intensive short-term dynamic psychotherapy to reduce rates of emergency department return visits for patients with medically unexplained symptoms: Preliminary evidence from a pre-post intervention study. Canadian Journal of Emergency Medical Care 11(6):529-534.
Abbass, A., S. Kisely, and K. Kroenke. 2009b. Short-term psychodynamic psychotherapy for somatic disorders: Systematic review and meta-analysis of clinical trials. Psychotherapy and Psychosomatics 78(5):265-274.
AHRQ (Agency for Healthcare Research and Quality). 2011. Effective Health Care Program. Glossary of Terms. http://effectivehealthcare.ahrq.gov/index.cfm/glossary-of-terms/ ( accessed July 1, 2012).
Aiarzaguena, J. M., G. Grandes, I. Alonso-Arbiol, J. L. del Campo Chavala, M. B. Oleaga Fernandez, and J. Marco De Juana. 2002. Bio-psychosocial treatment approach to somatizing patients in primary care: A pilot study. Atencion Primaria 29(9):558-561.
Allen, L. A., J. I. Escobar, P. M. Lehrer, M. A. Gara, and R. L. Woolfolk. 2002. Psychosocial treatments for multiple unexplained physical symptoms: A review of the literature. Psychosomatic Medicine 64(6):939-950.
Allen, L. A., R. L. Woolfolk, J. I. Escobar, M. A. Gara, and R. M. Hamer. 2006. Cognitive-behavioral therapy for somatization disorder: A randomized controlled trial. Archives of Internal Medicine 166(14):1512-1518.
Altamura, A. C., A. Di Rosa, A. Ermentini, G. P. Guaraldi, G. Invernizzi, N. Rudas, G. Tacchini, R. Pioli, M. T. Coppola, V. Tosini, E. Proto, E. Bolognesi, and B. Marchio. 2003. Levosulpiride in somatoform disorders: A double-blind, placebo-controlled cross-over study. International Journal of Psychiatry in Clinical Practice 7:155-159.
Arnold, I. A., M. W. de Waal, J. A. Eekhof, W. J. Assendelft, P. Spinhoven, and A. M. van Hemert. 2009. Medically unexplained physical symptoms in primary care: A controlled study on the effectiveness of cognitive-behavioral treatment by the family physician. Psychosomatics: Journal of Consultation Liaison Psychiatry 50(5):515-524.
Barrios-Choplin, B., R. McCraty, and B. Cryer. 1997. An inner quality approach to reducing stress and improving physical and emotional wellbeing at work. Stress Medicine 13(3):193-201.
Bleichhardt, G., B. Timmer, and W. Rief. 2004. Cognitive-behavioural therapy for patients with multiple somatoform symptoms: A randomised controlled trial in tertiary care. Journal of Psychosomatic Research 56:449-454.
CDC (Centers for Disease Control and Prevention). 2011. Physical Activity and Health. http://www.cdc.gov/physicalactivity/everyone/health/index.html (accessed September 27, 2012).
Chappell, H., and S. Chasee. 2006. The effect of biofeedback-assisted stress management training on migrant college students’ anxiety and personal growth initiative. Dissertation Abstracts International: Section B: The Sciences and Engineering 66(11-B):62-65.
Donta, S. T., D. J. Clauw, C. C. Engel, Jr., P. Guarino, P. Peduzzi, D. A. Williams, J. S. Skinner, A. Barkhuizen, T. Taylor, L. E. Kazis, S. Sogg, S. C. Hunt, C. M. Dougherty, R. D. Richardson, C. Kunkel, W. Rodriguez, E. Alicea, P. Chiliade, M. Ryan, G. C. Gray, L. Lutwick, D. Norwood, S. Smith, M. Everson, W. Blackburn, W. Martin, J. M. Griffiss, R. Cooper, E. Renner, J. Schmitt, C. McMurtry, M. Thakore, D. Mori, R. Kerns, M. Park, S. Pullman-Mooar, J. Bernstein, P. Hershberger, D. C. Salisbury, J. R. Feussner. 2003. Cognitive behavioral therapy and aerobic exercise for Gulf War veterans’ illnesses: A randomized controlled trial. Journal of the American Medical Association 289(11):1396-1404.
Donta, S. T., C. C. Engel, Jr., J. F. Collins, J. B. Baseman, L. L. Dever, T. Taylor, K. D. Boardman, L. E. Kazis, S. E. Martin, R. A. Horney, A. L. Wiseman, D. S. Kernodle, R. P. Smith, A. L. Baltch, C. Handanos, B. Catto, L. Montalvo, M. Everson, W. Blackburn, M. Thakore, S. T. Brown, L. Lutwick, D. Norwood, J. Bernstein, C. Bacheller, B. Ribner, L. W. P. Church, K. H. Wilson, P. Guduru, R. Cooper, J. Lentino, R. J. Hamill, A. B. Gorin, V. Gordan, D. Wagner, C. Robinson, P. DeJace, R. Greenfield, L. Beck, M. Bittner, H. R. Schumacher, F. Silverblatt, J. Schmitt, E. Wong, M. A. K. Ryan, J. Figueroa, C. Nice, J. R. Feussner. 2004. Benefits and harms of doxycycline treatment for Gulf War veterans’ illnesses: A randomized, double-blind, placebo-controlled trial. Annals of Internal Medicine 141(2):85-94.
Escobar, J. I., M. A. Gara, A. M. Diaz-Martinez, A. Interian, M. Warman, L. A. Allen, R. L. Woolfolk, E. Jahn, and D. Rodgers. 2007. Effectiveness of a time-limited cognitive behavior therapy type intervention among primary care patients with medically unexplained symptoms. Annals of Family Medicine 5(4):328-335.
Fontani, V., S. Rinaldi, L. Aravagli, P. Mannu, A. Castagna, and M. L. Margotti. 2011. Noninvasive radioelectric asymmetric brain stimulation in the treatment of stress-related pain and physical problems: Psychometric evaluation in a randomized, single-blind placebo-controlled, naturalistic study. International Journal of General Medicine 4:681-686.
Garcia-Campayo, J., and C. Sanz-Carrillo. 2002. Topiramate as a treatment for pain in multisomatoform disorder patients: An open trial. General Hospital Psychiatry 24(6):417-421.
Guarino, P., P. Peduzzi, S. T. Donta, C. C. Engel, D. J. Clauw, D. A. Williams, J. S. Skinner, A. Barkhuizen, L. E. Kazis, and J. R. Feussner. 2001. A multicenter two by two factorial trial of cognitive behavioral therapy and aerobic exercise for Gulf War veterans’ illnesses: Design of a Veterans Affairs cooperative study (CSP #470). Controlled Clinical Trials 22(3):310-332.
Han, C., C. U. Pae, B. H. Lee, Y. H. Ko, P. S. Masand, A. A. Patkar, S. H. Joe, and I. K. Jung. 2008a. Venlafaxine versus mirtazapine in the treatment of undifferentiated somatoform disorder—a 12-week prospective, open-label, randomized, parallel-group trial. Clinical Drug Investigation 28(4):251-261.
Han, C., D. M. Marks, C. U. Pae, B. H. Lee, Y. H. Ko, P. S. Masand, A. A. Patkar, and I. K. Jung. 2008b. Paroxetine for patients with undifferentiated somatoform disorder: A prospective, open-label, 8-week pilot study. Current Therapeutic Research—Clinical and Experimental 69(3):221-231.
Jakcsy, S. D. 2002. Remediation of Acquired Memory Problems (Post War) in Gulf War Veterans Using Computer Technology. University of Idaho.
Katsamanis, M., P. M. Lehrer, J. I. Escobar, M. A. Gara, A. Kotay, and R. Liu. 2011. Psychophysiologic treatment for patients with medically unexplained symptoms: A randomized controlled trial. Psychosomatics 52(3):218-229.
Kleinstäuber, M., M. Witthoft, and W. Hiller. 2011. Efficacy of short-term psychotherapy for multiple medically unexplained physical symptoms: A meta-analysis. Clinical Psychology Review 31(1):146-160.
Kroenke, K. 2007. Efficacy of treatment for somatoform disorders: A review of randomized controlled trials. Psychosomatic Medicine 69(9):881-888.
Kroenke, K., and J. G. M. Rosmalen. 2006. Symptoms, syndromes, and the value of psychiatric diagnostics in patients who have functional somatic disorders. Medical Clinics of North America 90(4):603-626.
Kroenke, K., and R. Swindle. 2000. Cognitive-behavioral therapy for somatization and symptom syndromes: A critical review of controlled clinical trials. Psychotherapy and Psychosomatics 69(4):205-215.
Kroenke, K., N. Messina, I. Benattia, J. Graepel, and J. Musgnung. 2006. Venlafaxine extended release in the short-term treatment of depressed and anxious primary care patients with multisomatoform disorder. Journal of Clinical Psychiatry 67(1):72-80.
Lidbeck, J. 1997. Group therapy for somatization disorders in general practice: Effectiveness of a short cognitive-behavioural treatment model. Acta Psychiatrica Scandinavica 96(1):14-24.
Lidbeck, J. 2003. Group therapy for somatization disorders in primary care: Maintenance of treatment goals of short cognitive-behavioural treatment one-and-a-half-year follow-up. Acta Psychiatrica Scandinavica 107(6):449-456.
Looper, K. J., and L. J. Kirmayer. 2002. Behavioral medicine approaches to somatoform disorders. Journal of Consulting & Clinical Psychology 70(3):810-827.
Martin, A., E. Rauh, M. Fichter, and W. Rief. 2007. A one-session treatment for patients suffering from medically unexplained symptoms in primary care: A randomized clinical trial. Psychosomatics 48(4):294-303.
McCraty, R., B. Barrios-Choplin, D. Rozman, P. Atkinson, and A. D. Watkins. 1998. The impact of a new emotional self-management program on stress, emotions, heart rate variability, DHEA and cortisol. Integrative Physiological and Behavioral Science 33(2):151-170.
McCraty, R., M. Atkinson, and D. Tomasino. 2003. Impact of a workplace stress reduction program on blood pressure and emotional health in hypertensive employees. Journal of Alternative and Complementary Medicine 9(3):355-369.
Menza, M., M. Lauritano, L. Allen, M. Warman, F. Ostella, R. M. Hamer, and J. Escobar. 2001. Treatment of somatization disorder with nefazodone: A prospective, open-label study. Annals of Clinical Psychiatry 13(3):153-158.
Mori, D. L., S. Sogg, P. Guarino, J. Skinner, D. Williams, A. Barkhuizen, C. Engel, D. Clauw, S. Donta, and P. Peduzzi. 2006. Predictors of exercise compliance in individuals with Gulf War veterans illnesses: Department of Veterans Affairs cooperative study 470. Military Medicine 171(9):917-923.
Muller, J. E., I. Wentzel, L. Koen, D. J. Niehaus, S. Seedat, and D. J. Stein. 2008. Escitalopram in the treatment of multisomatoform disorder: A double-blind, placebo-controlled trial. International Clinical Psychopharmacology 23(1):43-48.
Muller, T., M. Mannel, H. Murck, and V. W. Rahlfs. 2004. Treatment of somatoform disorders with St. John’s wort: A randomized, double-blind and placebo-controlled trial. Psychosomatic Medicine 66(4):538-547.
Nanke, A., and W. Rief. 2003. Biofeedback-based interventions in somatoform disorders: A randomized control trial. Acta Neuropsychiatrica 15(4):249-256.
Nezu, A. M., C. M. Nezu, and E. R. Lombardo. 2001. Cognitive-behavior therapy for medically unexplained symptoms: A critical review of the treatment literature. Behavior Therapy 32 (3):537-583.
Nickel, M., B. Cangoez, E. Bachler, M. Muehlbacher, N. Lojewski, N. Mueller-Rabe, F. O. Mitterlehner, P. Leiberich, N. Rother, W. Buschmann, C. Kettler, F. Pedrosa Gil, C. Lahmann, C. Egger, R. Fartacek, W. K. Rother, T. H. Loew, and C. Nickel. 2006. Bioenergetic exercises in inpatient treatment of Turkish immigrants with chronic somatoform disorders: A randomized, controlled study. Journal of Psychosomatic Research 61(4):507-513.
NIH (National Institutes of Health). 2012. What Is Complementary and Alternative Medicine? http://nccam.nih.gov/health/whatiscam (accessed September 27, 2012).
Owens, D. K., K. N. Lohr, D. Atkins, J. R. Treadwell, J. T. Reston, E. B. Bass, S. Chang, and M. Helfand. 2010. AHRQ Series Paper 5: Grading the strength of a body of evidence when comparing medical interventions: Agency for Healthcare Research and Quality and the Effective Health-Care Program. Journal of Clinical Epidemiology 63(5):513-523.
Paterson, C., R. S. Taylor, P. Griffiths, N. Britten, S. Rugg, J. Bridges, B. McCallum, and G. Kite. 2011. Acupuncture for “frequent attenders” with medically unexplained symptoms: A randomised controlled trial (CACTUS study). British Journal of General Practice 61(587):e295-305.
Payne, H., and D. Stott. 2010. Change in the moving bodymind: Quantitative results from a pilot study on the use of the bodymind approach (BMA) to psychotherapeutic group work with patients with medically unexplained symptoms (MUS). Counselling & Psychotherapy Research 10(4):295-306.
Peters, S., I. Stanley, M. Rose, S. Kaney, and P. Salmon. 2002. A randomized controlled trial of group aerobic exercise in primary care patients with persistent, unexplained physical symptoms. Family Practice 19(6):665-674.
Posse, M. 2004. Jungian psychotherapy for somatization. European Journal of Psychiatry 18:23-29.
Reineke, A. 2008. The effects of heart rate variability biofeedback in reducing blood pressure for the treatment of essential hypertension. Dissertation Abstracts International: Section B: The Sciences and Engineering 68(7-B):4880.
Rembold, S. M. E. 2011. Somatoform disorders in primary care settings: Construction and evaluation of a psychosocial group programme. Gruppenpsychotherapie und Gruppendynamik 47(1):2-21.
Reuber, M., C. Burness, S. Howlett, J. Brazier, and R. Grunewald. 2007. Tailored psychotherapy for patients with functional neurological symptoms: A pilot study. Journal of Psychosomatic Research 63(6):625-632.
Rief, W., G. Bleichhardt, and B. Timmer. 2002. Group therapy for somatoform disorders: Treatment guidelines, acceptance, and process quality. Verhaltenstherapie 12(3):183-191.
Rugg, S., C. Paterson, N. Britten, J. Bridges, and P. Griffiths. 2011. Traditional acupuncture for people with medically unexplained symptoms: A longitudinal qualitative study of patients’ experiences. British Journal of General Practice 61(587):e306-e315.
Russell, M. C. 2008. War-related medically unexplained symptoms, prevalence, and treatment: Utilizing EMDR within the armed services. Journal of EMDR Practice and Research 2(3):212-225.
Saldanha, D., S. Chaudhury, A. A. Pawar, V. Ryali, and K. Srivastava. 2007. Reduction in drug prescription using biofeedback relaxation in neurotic and psychosomatic disorders. Medical Journal Armed Forces India 63(4):315-317.
Sattel, H., C. Lahmann, H. Gundel, E. Guthrie, J. Kruse, M. Noll-Hussong, C. Ohmann, J. Ronel, M. Sack, N. Sauer, G. Schneider, and P. Henningsen. 2012. Brief psychodynamic interpersonal psychotherapy for patients with multisomatoform disorder: Randomised controlled trial. British Journal of Psychiatry 200:60-67.
Sharpe, M., J. Walker, C. Williams, J. Stone, J. Cavanagh, G. Murray, I. Butcher, R. Duncan, S. Smith, and A. Carson. 2011. Guided self-help for functional (psychogenic) symptoms: A randomized controlled efficacy trial. Neurology 77(6):564-572.
Shea, B. J., J. M. Grimshaw, G. A. Wells, M. Boers, N. Andersson, C. Hamel, A. C. Porter, P. Tugwell, D. Moher, and L. M. Bouter. 2007. Development of AMSTAR: A measurement tool to assess the methodological quality of systematic reviews. BMC Medical Research Methodology 7(10).
Strack, B., R. Gevirtz, and W. Sime. 2004. Effect of heart rate variability (HRV) biofeedback on batting performance in baseball. Applied Psychophysiology and Biofeedback 29(4):299.
Sumathipala, A. 2007. What is the evidence for the efficacy of treatments for somatoform disorders? A critical review of previous intervention studies. Psychosomatic Medicine 69(9):889-900.
Sumathipala, A., S. Hewege, R. Hanwella, and A. H. Mann. 2000. Randomized controlled trial of cognitive behaviour therapy for repeated consultations for medically unexplained complaints: A feasibility study in Sri Lanka. Psychological Medicine 30(4):747-757.
Sumathipala, A., S. Siribaddana, M. Abeysingha, P. De Silva, M. Dewey, M. Prince, and A. Mann. 2008. Cognitive-behavioural therapy v. structured care for medically unexplained symptoms: Randomised controlled trial. British Journal of Psychiatry 193(1):51-59.
Tanis, C. 2008. The effects of heart rhythm variability biofeedback with emotional regulation on the athletic performance of women collegiate volleyball players. Dissertation Abstracts International: Section B: Sciences and Engineering 69(4-B):2666.
van Rood, Y. R., and C. de Roos. 2009. EMDR in the treatment of medically unexplained symptoms: A systematic review. Journal of EMDR Practice and Research 3(4):248-263.
Volz, H. P., H. J. Moller, I. Reimann, and K. D. Stoll. 2000. Opipramol for the treatment of somatoform disorders: Results from a placebo-controlled trial. European Neuropsychopharmacology 10(3):211-217.
Volz, H. P., H. Murck, S. Kasper, and H. J. Moller. 2002. St. John’s wort extract (Li 160) in somatoform disorders: Results of a placebo-controlled trial. Psychopharmacology (Berl) 164(3):294-300.
Yamada, K., R. Den, K. Ohnishi, and S. Kanba. 2005. Effectiveness of herbal medicine (Kampo) and changes of quality of life in patients with somatoform disorders. Journal of Clinical Psychopharmacology 25(2):199-201.
Zaby, A., J. Heider, and A. Schroder. 2008. Waiting, relaxation, or cognitive-hehavioral therapy: How effective is outpatient group therapy for somatoform symptoms? Zeitschrift für Klinische Psychologie und Psychotherapie 37(1):15-23.