Homeopathic Treatments in Psychiatry: a Systematic Review of Randomized Placebo-Controlled Studies

Homeopathic Treatments in Psychiatry

Homeopathic Treatments in Psychiatry: A Systematic Review of Randomized Placebo-Controlled Studies

Jonathan R. T. Davidson, MD; Cindy Crawford, BA; John A. Ives, PhD; and Wayne B. Jonas, MD

se of complementary and alternative medicine (CAM) to treat psychiatric problems is widespread, and the Objective: To systematically review placebo- U controlled randomized trials of homeopathy for need has been identified for more high-quality controlled psychiatric conditions. trials. A task force of the American Psychiatric Association Data Sources: Eligible studies were identified using concluded that several CAM treatments, including omega-3 the following databases from database inception to fatty acids, St John’s wort, folate, acupuncture, and others, April 2010: PubMed, CINAHL, PsycINFO, Hom-Inform, show promise for depression, but that more rigorous and Cochrane CENTRAL, National Center for Complementa- 1 ry and Alternative Medicine grantee publications database, larger studies were needed. A meta-analysis by Freeman 2 and ClinicalTrials.gov. Gray literature was also searched et al reported comparable efficacy but greater safety for a using Google, Google Scholar, the European Commit- number of herbal and dietary supplements than for standard tee for Homeopathy, inquiries with homeopathic experts antidepressants. Among the many forms of CAM, home- and manufacturers, and the bibliographic lists of included opathy is one of the most widely used on a global basis.3 published studies and reviews. Search terms were as follows: (homeopath* or homoeopath*) and (placebo or sham) Meta-analyses and systematic reviews have drawn mixed and (anxiety or panic or phobia or post-traumatic stress or conclusions as to whether homeopathy is more effective than PTSD or obsessive-compulsive disorder or fear or depress* or placebo in general medicine.4–10 In assessing these studies, dysthym* or attention deficit hyperactivity or premenstrual Lewith7 has pointed out that where reports are few and based syndrome or premenstrual disorder or premenstrual dys- on small samples, results of systematic reviews depend on phoric disorder or traumatic brain injury or fibromyalgia or chronic fatigue syndrome or myalgic encephalitis or which studies are included and which are excluded. Thus, any insomnia or sleep disturbance). Searches included only fair assessment needs to be systematic and comprehensive English-language literature that reported randomized and use established quality and scoring approaches on all controlled trials in humans. studies. No comprehensive review of research on homeopa- Study Selection: Trials were included if they met 7 cri- thy for psychiatric conditions has been conducted. Our aim teria and were assessed for possible bias using the Scottish Intercollegiate Guidelines Network (SIGN) 50 guidelines. in this article was to undertake such a systematic review. Overall assessments were made using the Grading of Rec- Although widely used in many parts of the world, home- ommendations Assessment, Development and Evaluation opathy remains controversial within the Western medical procedure. Identified studies were grouped into anxiety or paradigm. This is due principally to discordance between the stress, sleep or circadian rhythm complaints, premenstrual principles of homeopathy and those of accepted biomedical problems, attention-deficit/hyperactivity disorder, mild traumatic brain injury, and functional somatic syndromes. theory. The system of homeopathy rests on 2 fundamen- Results: Twenty-five eligible studies were identified tal principles: (1) similarity, whereby the indicated remedy from an initial pool of 1,431. Study quality according to for particular symptoms is that which elicits similar symp- SIGN 50 criteria varied, with 6 assessed as good, 9 as fair, toms when given to a healthy person, and (2) the power of and 10 as poor. Outcome was unrelated to SIGN quality. the minimum dose, whereby a substance that is repeatedly Effect size could be calculated in 16 studies, and number needed to treat, in 10 studies. Efficacy was found for the diluted and agitated (“succussed”) is believed to preserve its 4 functional somatic syndromes group (fibromyalgia and effect even into “ultramolecular” solutions. chronic fatigue syndrome), but not for anxiety or stress. In all major reviews of homeopathy, there is an absence For other disorders, homeopathy produced mixed effects. of comprehensive reviews of studies relevant to psychiatry, No placebo-controlled studies of depression were identi- even though there are some encouraging findings. For ex- fied. Meaningful safety data were lacking in the reports, but the superficial findings suggested good tolerability of ample, in one review homeopathy was superior to placebo homeopathy. A funnel plot in 13 studies did not support on at least 1 clinically meaningful measure in 6 of 7 trials of 2 publication bias (χ 1 = 1.923, P = .166). fibromyalgia, anxiety, agitation, traumatic brain injury (TBI), Conclusions: The database on studies of homeopathy and premenstrual syndrome (PMS).9 On the other hand, a and placebo in psychiatry is very limited, but results do not Cochrane review of attention-deficit/hyperactivity disorder preclude the possibility of some benefit. J Clin Psychiatry 2011;72(6):795–805 (ADHD) showed no overall benefit for homeopathy over pla- 11 © Copyright 2011 Physicians Postgraduate Press, Inc. cebo in 3 randomized clinical trials. Two systematic reviews in depression12 and anxiety13 found insufficient good quality Submitted: September 15, 2010; accepted December 15, 2010 data to judge the efficacy of homeopathy for these conditions. (doi:10.4088/JCP.10r06580). A Cochrane review of homeopathy for dementia found no Corresponding author: Jonathan R. T. Davidson, MD, 3068 Baywood Dr, placebo-controlled studies of adequate quality.14 Evidence in Seabrook Island, SC 29455 ([email protected]). support of homeopathy for fibromyalgia is more encouraging,

© JC ClinOPYRIGHT Psychiatry 2011 72:6, P HYSICIANSJune 2011 POSTGRADUATE PRESS, INC. © COPYRIGHT 2011 PHYSICIANS POSTGRADUATE PRESS, INC795. Davidson et al Clinical Points however.15,16 A review of randomized controlled trials homeopathy for insomnia ■■Randomized placebo-controlled studies suggest that (RCTs) in human subjects. called for more research.17 homeopathy is without benefit for anxiety, that it may All searches were performed Because patients with psy- be useful for functional somatic syndromes, and that across titles/abstracts where chiatric problems are well for other conditions such as ADHD, premenstrual and possible. Where some of represented in homeopathic these restrictions were not practice,18,19 it is important sleep-related problems, its benefit is undetermined. For possible, we screened the ti- to examine whether home- other common psychiatric conditions such as depression, tles and abstracts manually. opathy is beneficial in the posttraumatic stress disorder, and dementia, there are no more commonly seen psy- informative data. Inclusion and chiatric conditions, defined Exclusion Criteria Although homeopathic medicines are well tolerated and here as anxiety, depression, ■■ Four investigators (C.C., sleep problems, ADHD, believed to carry few side effects, there has to date been J.A.I., W.B.J., and J.R.T.D.) PMS, mild TBI, and somatic no adequate demonstration of their safety. independently screened spectrum disorders. ■■It is unknown whether a single individually chosen titles and abstracts for rele- Our objective was to medicine is more effective than a fixed-dose combination vance based on the inclusion conduct a comprehensive, criteria for this systematic systematic literature re- formula. review. Any disagreements view of placebo-controlled, about including a study were randomized clinical trials of homeopathy for psychiatric resolved through discussion and consensus. Articles were conditions, to assess the quality and risk of bias in each included in this systematic review if they met the following study’s design and execution, to report on outcome when criteria: (1) randomized controlled trial (RCT) design was possible by means of effect size (ES) or number needed used; (2) a placebo control was used; (3) between-treatment to treat (NNT) statistics, to review safety, and to grade the comparisons were made of homeopathic treatment versus overall evidence for each condition according to interna- placebo; (4) treatment was given in a double-blind fashion; tionally standardized methods. Because of the heterogeneity (5) the report assessed a psychiatric condition as specified of studies in each psychiatric category, we did not under- in the keyword list above; (6) the report was presented in take meta-analysis of the data but did check for likelihood of English; and (7) the study involved treatment-seeking hu- publication bias in a subset of the data. man subjects; that is, we did not review any animal model studies, studies in healthy volunteers, or studies in patient METHOD groups in which the focus was on mechanism of action or prediction of treatment effect. Data Sources and Search Strategy A systematic search was conducted for literature that de- Quality Rating of Individual Studies scribed homeopathic treatment of the following 7 groups Methodological quality of the included studies was as- of psychiatric conditions: depression, anxiety, sleep and sessed independently by the 4 reviewers for the individual circadian rhythm problems, ADHD, PMS, mild TBI, and studies and then by 2 reviewers (W.B.J. and J.R.T.D.) on the functional somatic syndromes (FSS), specifically fibromy- quality of the overall literature pool with regard to the mini- algia and chronic fatigue syndrome. The following databases mization of bias. The individual studies were all RCTs and were examined for studies reported from database inception were evaluated for study quality and bias using the Scottish to April 2010: PubMed, CINAHL, PsycINFO, Hom-Inform, Intercollegiate Guidelines Network (SIGN) 50 checklist for Cochrane CENTRAL, National Center for Complementary RCTs.20 SIGN is an internationally developed and accepted and Alternative Medicine grantee publications database, assessment approach widely used for both conventional and and ClinicalTrials.gov. Gray literature was also searched complementary medicine research. Once the quality assess- using Google, Google Scholar, the European Committee ment of the individual studies was completed, 2 reviewers for Homeopathy, inquiries with homeopathic experts and conducted a quality assessment of the overall literature pool manufacturers, and the bibliographic lists of included studies for each condition using the Grading of Recommenda- and published reviews. Search terms used were as follows: tions Assessment, Development and Evaluation (GRADE), (homeopath* or homoeopath*) and (placebo or sham) and looking at the (1) confidence in the estimate of the ES, (2) (anxiety or panic or phobia or post-traumatic stress or PTSD or magnitude of the effect, (3) safety grade, and (4) strength of obsessive-compulsive disorder or fear or depress* or dysthym* the recommendation.21 GRADE is also an internationally or attention deficit hyperactivity or premenstrual syndrome accepted approach for quality assessment of literature sets. or premenstrual disorder or premenstrual dysphoric disorder All reviewers were trained in the quality assessment of or traumatic brain injury or fibromyalgia or chronic fatigue individual studies (SIGN) and the quality assessment of syndrome or myalgic encephalitis or insomnia or sleep distur- overall literature pool (GRADE) by 1 of the authors (C.C.), bance). The following limits were placed on searches: only and each article was assessed by 2 reviewers. For any discrep- literature presented in the English language that reported ancies, discussion occurred between reviewers in order to

Figure 1. Study Flowchart

1,431 Articles identified and screened for inclusion

Other sources and 10 Systematic reviews Online databases reference mining identifying 526 69 836 potential studies

59 Excluded (11 duplicates, 8 secondary analyses, 501 Excluded for not consisting of the target population, 3 psychiatric but not RCTs, 24 not psychiatric 827 Excluded or not having the correct study design, not being or homeopathic, 11 review/methods papers, unobtainable placebo controlled, or not being in English 2 insufficient data)

19 Duplicate articles from various sources excluded

25 Articles included in quality assessment

++ Level of evidencea + Level of evidencea – Level of evidencea

6 RCTs 9 RCTs 10 RCTs

aLevel of evidence with respect to minimizing bias was rated as “good” (++), “fair” (+), or “poor” (–) according to Scottish Intercollegiate Guidelines Network (SIGN) quality analysis. Abbreviation: RCT = randomized controlled trial.

achieve consensus. Final judgment was reserved for the first and on the combination, assigning the higher ratings (if they author (J.R.T.D.). differed) if one report gave more complete information than the other; the overall evaluation was based on information Data Analysis from both. For every study that provided a mean score and SD, SE, t, or F statistic, we calculated the ES between treatments us- RESULTS ing the Hedges unbiased g,22 which mathematically adjusts for small samples. The ES was recorded as positive if it fa- Study Selection and Quality vored homeopathy and negative if placebo was more effective. The search strategy led to the identification of 69 reports Consistent with the GRADE conventions, an ES that ranges from online databases (Hom-Inform, n = 29; ClinicalTrials. from 0.20 to 0.49 is considered to be small, 0.50 to 0.79 is gov, n = 1; MEDLINE [PubMed], n = 14; PsycINFO, n = 6; medium, and 0.8 or greater is large. For studies reporting Cochrane CENTRAL, n = 19; CINAHL, n = 0), 836 from rates of response, the NNT was calculated.23 For a subset of 13 other sources (n = 834 from the following sources: European studies that gave sufficient information to derive ES and the Committee for Homeopathy List of Dissertations and Theses 95% CIs, we calculated an estimate for possible publication in Homeopathy, n = 644; theses and dissertations from Dur- bias by graphing the 1/var to g and running the nonparamet- ban University of Technology [Health Sciences], n = 66; and ric selection model applied to the 13 studies.3 The outcome theses and dissertations from University of Johannesburg, measures chosen for calculating ES were those identified in n = 124; plus reference mining, n = 2), and 526 from 10 sys- the respective publications as primary. When more than 1 tematic reviews (Figure 1). As shown in Table 1, 25 studies primary measure was identified and results were conflicting, fulfilled the specified criteria.24–51 According to SIGN quality ES were calculated separately for the most and least favorable analysis, 6 studies were rated as “good” (++) with respect to toward homeopathy. In studies in which primary outcomes minimizing bias, 9 as “fair” (+), and 10 as “poor” (–). These uniformly failed to show a statistically significant difference, are shown individually in eAppendix 1. a single scale was chosen at random. Of the 25 studies, 6 were conducted in populations In some cases, as shown in the tables, a study appeared suffering from anxiety or stress; 5, in subjects with sleep or cir- more than once, usually because it was published as a thesis at cadian rhythm disturbances; 4, in subjects with premenstrual a university Web site, and elsewhere as a peer-reviewed publi- problems; 3, in subjects with ADHD; 1, in subjects with mild cation. SIGN ratings were conducted on each communication TBI; and 6, in subjects with functional somatic syndromes.

© JC ClinOPYRIGHT Psychiatry 2011 72:6, P HYSICIANSJune 2011 POSTGRADUATE PRESS, INC. © COPYRIGHT 2011 PHYSICIANS POSTGRADUATE PRESS, INC797. Davidson et al a – + + + + SIGN SIGN Score (continued) treatments; unlikely that that unlikely treatments; been have this would a larger with different sample treatments better was homeopathy placebo than homeopathy No difference between difference No between difference No evidence that No benefitSome for homeopathy benefit for No – Mixed results Mixed overturn the results overturn failed to therapy, cognitive be can regarded study work; rather study a “failed” as for study a negative than words, other In homeopathy. informative not is it been inadequate placebo (weak ES). favored test stress-provoking Lack of evidence No a limitation. the specific of presence that a made the remedy for profile the outcome to difference Weak the primary measures. on homeopathy evidence for scale items scale depersonalization on MBI of significant improvement on on improvement significant to unlikely anxiety Trait sleep. any term with in short change Effect sizes based treatment. anxiety state sleep and on trait and pulse measures; either as considered anxiety to useful likely not or (pulse) term (trait in the short change anxiety) Length of treatment may have have may treatment Length of Larger sample unlikely to to unlikely sample Larger homeopathy for result Positive effective Homeopathy + Rates of of Rates Response Reviewer Comments Conclusions Main Authors’ 40% 80% 42% 46% = = = = Not givenNot of scores effect the total on No H P givenNot GAD, for treatment A proven givenNot Results powered. Adequately H P Not givenNot but anxiety state benefit on No .05) < ( P value) .001) < (NS) (NS) ( P Primary Outcomes Primary Outcomes Thought interference interference Thought BAI (NS) BAI (NS) PPQ (NS) STAI(S) ( P Sleep Pulse (NS) Trial Trial Duration 5 d8 wk anxiety of Feelings 10 d subscales MBI (NS) given Not score daily Snoring placebo than worse Homeopathy 10 wk (NS) HARS 4 wk (NS) HARS 4 d (NS) RTA 15 d (NS) STAI(T) 14, 44, 18 21, 14, = = = = = 23, P2 14/10 18 16 46 = Entered/ = = = = 22/20 14/11 38/35 Completed 22/19 13/10 39/37 (2 placebo groups)/H P1 enrolled/H P P enrolled/H P = = = No. of Subjects Subjects of No. = = = H H total 70 Enrolled H 47 Total given/H Not 101 Total P CBT P P homeopathy homeopathy product product homeopathy product Argentum nitricum Argentum 9-remedy Combined Individualized 9-remedy Combined anxiety burnout snoring Condition Treatment GAD Individualized GAD Individualized anxiety Test trait High anxietyTest 3-remedy Combined Job-related Severe 29,c 27,c 30,c 24,b 26,d 25,c 28,c Ngobese (2006) Ngobese Baker et al (2003) (1996) McCutcheon (2005) Vaithilingam et al (1999) Lipman Traub (2000) Traub Table 1. Placebo-Controlled, Randomized Clinical Trials of Homeopathy for Common Psychiatric Conditions: Quality Individual Studies 1. Placebo-Controlled, Randomized Clinical Trials Table Reference Bonne et al (2003)

798© C OPYRIGHT 2011 PHYSICIANS POSTGRADUATE PRESS, INC. © COPYRIGHT 2011 PHYSICIANSJ Clin P OSTGRADUATEPsychiatry 72:6, P JuneRESS 2011, INC. Homeopathic Treatments in Psychiatry a – – – – + + ++ SIGN SIGN Score (continued) homeopathy and placebo and homeopathy response and subjects subjects and response for therapy received trauma abuse-related symptoms homeopathy, benefit for limited acknowledge but size sample not effective in treating effective in treating not PMS homeopathy when compared against against when compared placebo suggests additional additional suggests targeted more and in future questionnaires jet lag of studies No benefit. High placebo benefit. High No greater of Suggestive Homeopathic simillimum simillimum Homeopathic effect an for Suggests No benefit for homeopathy homeopathy benefit for No Benefit for homeopathyBenefit for + Results favor homeopathy; homeopathy; favor Results screened and 36 entered study. study. 36 entered screened and randomized 10 were Only beto post hoc choices for publication, thesis, and and thesis, publication, for assessment combined unpublished reports were were reports unpublished + as rated was assessed, SIGN for in both cases. ES provided other for only;SII unavailable scales on POMS-Vigor. Ambiguous, Ambiguous, POMS-Vigor. on further warrant results but study Impossible to interpret. Over 200 interpret. to Impossible seem secondarySome outcomes Negative study Negative the same remains rating SIGN Rates of of Rates Response Reviewer Comments Conclusions Main Authors’ 40% 90% 33% 60% 38% 50% ======Not givenNot homeopathy benefit for No to Equal response H P H P given Not given Not H P Not givenNot and the published When Not givenNot P values reported Inconsistently .05)) < .0001) ( P value) (NS) .05) e < < .05 on 5/28 items .05 on pattern (NS) pattern ( P Primary Outcomes Primary Outcomes < CAVT (NS) CAVT (NS) IIQ MDQ (NS) PAF P SII ( P SII DBAS Sleep satisfaction (NS) satisfaction Sleep in sleep Change POMS-Fatigue POMS-Fatigue (NS) POMS-Vigor Trial Trial 7 d for each for d 7 treatment of remedy remedy of given) was design 24 h each treatment Duration Crossover Crossover 4 mo Global (NS) 3 mo 1 dose(Only 3 mo6 mo MDQ (NS) MDQ on Each item 4 wk diary Sleep ( P 30 d sleep (NS) of Hours Crossover Crossover 16, = 15 Entered/ = 5/5 13/11 18/13 16/14 15/15 Completed 5/5 10/8 21/14 17/16 15/14 completed completed crossover Entered/H P completed completed crossover = = = = = No. of Subjects Subjects of No. = = = = = 34 Entered/28 34 Entered/28 H H H P P P H P P 23 Entered/19 23 Entered/19 product homeopathy homeopathy homeopathy homeopathy remedy product remedy Combined 5-remedy 5-remedy Combined 15CFolliculinum 35 34 or Individualized Individualized Coffea cruda 200CCoffea H night shift night workers insomnia than 1 y in than duration Condition Treatment Shift lag in Shift lag PMS Individualized PMS Individualized PMSPMS Individualized Primary Insomnia less less Insomnia Jet lagJet multiple Combined ; 34,c 31,c 38,c

40,c 36 33,c 39,c and and composite 32 35,c 37,c publication publication (2008) (1994) and Elizabeth and HD, Solomon, e-mail DO, ND, communications 12, April JRTD; to 14, July 2010, and 2010 thesis (2010) thesis and Andrew Andrew and Criglington, B Comm, e-mail communications 16, March JRTD, to August 2010, and 9, 2010 David Naudé, M Naudé, David e-mail (Hom), Tech to communication 14, 2010; July JRTD, thesis Maharaj and (2005) Kolia-Adam combined combined Kolia-Adam Yakir et al (2001) Yakir (2008) Laister Table 1 (continued). Placebo-Controlled, Randomized Clinical Trials of Homeopathy for Common Psychiatric Conditions: Quality Individual Studies 1 (continued). Placebo-Controlled, Randomized Clinical Trials Table Reference et al (2010) Naudé (1994) Kirtland Chapman et al Chapman Kumar (2010) Kumar La et al (2006) Pine

© JC ClinOPYRIGHT Psychiatry 2011 72:6, P HYSICIANSJune 2011 POSTGRADUATE PRESS, INC. © COPYRIGHT 2011 PHYSICIANS POSTGRADUATE PRESS, INC799. Davidson et al a – – – + ++ ++ ++ ++ SIGN SIGN Score (continued) of homeopathy, homeopathy, of particularly in cognitive and behavioral functions targeted the method of in choice remedy of Analysis fibromyalgia. outcomes predetermined significant gave pain on differences indicated sleep for and remedy evidence favoring homeopathy favoring homeopathy. homeopathy. favoring furtherCall studies for advantages for for advantages homeopathy improved more on on more improved than homeopathy placebo homeopathy, especially homeopathy, points tender on No benefit, no trend benefit, no No effectiveness suggests Trial of feasibility supports Study equivocal but Weak Significant improvement improvement Significant Guarded in stating in stating Guarded Overall hyperactivity Positive results for for results Positive not significantly so significantly not in adequately outcome study powered expected Missing direction. a problem. is information (Peter method (Minitab) FRCP, FF Hom, Fisher, communication, electronic 23, 2010) May supports measure rigorous homeopathy intent-to-treat analysis analysis intent-to-treat effects significant for showed all scales on homeopathy many measures, but statistical statistical but measures, many Some carried out. not analysis numbers the published of in subscales add up do not and disability, (fatigue, myalgia). review in Cochrane analysis review in Cochrane analysis significance on show failed to measure any cautious interpretation of of interpretation cautious be seems to a positive what study Information on randomization randomization on Information Mixed results, but the most the most but results, Mixed Mostly positive study. An An study. positive Mostly Advantages seem evident on seem on evident Advantages Lack of detail a cause for for detail a cause Lack of ) 55% 74% 5[p376] = = Rates of of Rates Response Reviewer Comments Conclusions Main Authors’ 43% 37% 26% 85%, P 74%, P 13% 9% 4% outcome significant improvement all primaryon outcomes) on SRS on DAS on ======H H Not givenNot givenNot H, but be than to better P tended primary on result Positive givenNot in the results but Underpowered, P (Global P clinically (For H H Not givenNot best. Subsequent effect at Weak P .01) < .05) .05) .05) < < < ( P value) .05) < ( P (NS) (NS) (NS) analyzed) General fatigue fatigue Physical (NS) fatigue Mental Reduced activity Reduced motivation Index-Parent Index-Parent ( P Index-Parent (NS) Index-Parent ( P FA effect”) Primary Outcomes Primary Outcomes Conners Global Conners given) (not Pain given) (not Sleep ( P points Tender Global (NS) VAS sleep (NS) VAS (NS) spots tender No use (“No Analgesic CCT (NS) Trial Trial crossover crossover each treatment 4 wk of each on treatment Duration 12 mo (not Global response 6 wk Crossover 18 wk Global Conners 4 mo for Overall MANOVA 3 mo (NS) pain VAS 6 mo 5 MFI scales H 8 wk CPSQ ( P 12, 10, = = 31/ = 12 10 Entered/ = = 22/22 31, P 33/27 32/30 53/43 Completed 21/21 28/23 50/43 32/31 all completed all completed crossover completed crossover P P = = = = = No. of Subjects Subjects of No. = = = = H H H Unknown/H Unknown/30 H H P P P Unknown/H P homeopathy homeopathy homeopathy or alba, Bryonia montana Arnica 6C homeopathy homeopathy homeopathy Rhus toxicodendron, toxicodendron, Rhus toxicodendron Rhus Condition Treatment ADHDADHD Individualized Individualized TBIMild Individualized Fibromyalgia Fibromyalgia CFS Individualized CFS Individualized ADHD Individualized ; 44 41,c 47,c

42,d 43,c 48,c 46,c 45,d 49,d (1999) (2004) composite and Cole (1998) and Fisher et al (1989) Fisher Table 1 (continued). Placebo-Controlled, Randomized Clinical Trials of Homeopathy for Common Psychiatric Conditions: Quality Individual Studies 1 (continued). Placebo-Controlled, Randomized Clinical Trials Table Reference et al (2005) Jacobs et al (2005) Frei (2000) Strauss et al Chapman (1986) Fisher (1996) Awdry et al Weatherley-Jones

a Relevant details of these studies are shown in Tables ++ SIGN SIGN Score 1 and 2. No placebo-controlled studies of depression were identified. Table 1 presents details of each study, and Table 2 presents the overall GRADE assessment. Taking a statistically significant P value as a crude indicator of

Hamilton Anxiety Hamilton possible efficacy, the following assessment for each = Profile of Mood Mood of Profile

= condition was found: measure used to power used power to measure many on and the study, others Menstrual Distress Questionnaire, Questionnaire, Distress Menstrual Positive outcome on main main on outcome Positive homeopathy benefit for No – cognitive-behavioral therapy, therapy, cognitive-behavioral = Daily Activities Scale, Activities Daily • There is no support for the efficacy of home- = =

.1 opathy in anxiety- or stress-related conditions. < State Trait Anxiety Inventory (State), (State), Anxiety Inventory Trait State In only 1 study,27 on a sleep measure, did the = HARS homeopathy,

= difference reach significance. • For sleep- and circadian rhythm–related problems, the evidence is mixed. Two studies30,31 yielded predominantly positive POMS syndrome, premenstrual

= results, and these were the studies that scored higher on GRADE evaluation (Table 2). accepted significance at P significance at accepted 7 of Three trends. for assess to longer no would comparisons if significant be statistically been had the .05 criterion used palpation on pain point tender measure Rates of response based the on response of Rates Overall positive trial. Author Overall trial. positive Author Because each study addressed a different prob- Maslach Burnout Inventory, MDQ Inventory, Burnout Maslach

= lem, however, we do not think the cumulative 15%

CBT Test, Vigilance Assisted Computer evidence for any one condition warrants either = generalized anxiety disorder, H disorder, anxiety generalized =

= a positive or a negative overall recommenda-

Rates of of Rates tion for this group. Response Reviewer Comments Conclusions Main Authors’ DAS Questionnaire, Symptom Parents Conners 50%, P on 25% on improvement in tender on pain point palpation = = • For premenstrual problems, there was little H givenNot trial Negative STAI(S) Injury, Brain Traumatic of Symptoms

= evidence of efficacy, other than 1 suggestive study,38 which was limited by a small sample .01) < Premenstrual Assessment Form, PMS Form, Assessment Premenstrual

MBI Affective Pain, McGill size. = =

.01) • Of 3 ADHD studies, 1 relapse prevention <

.05) 42 ( P value) .05) design was positive, and 2 acute symptom re- <

< 41,43,44 CAVT Beck Anxiety Inventory, duction trials were negative, although the ( P palpation ( P palpation Primary Outcomes Primary Outcomes = 43 Tender point count count point Tender MAP ( P (NS) MSP AF ( P (NS) F-VAS Fatigue Visual Analog Scale, Analog GAD Visual Fatigue PAF placebo, report by Strauss indicated statistical signifi- = = cance on 1 measure. Two41,42 of the 3 ADHD studies scored strongly on SIGN evaluation. SRS Index, Insomnia Severity of

Trial Trial 45

= • For mild TBI, the 1 available study scored Duration 3 mo CFS-Q (NS) 3 mo on pain point Tender favorably on attempts to reduce bias and produced weakly positive results in favor of nonsignificant, P nonsignificant, 15, = = Chronic Fatigue Syndrome Questionnaire, CPSQ Questionnaire, Syndrome Fatigue Chronic homeopathy. = 46–51 multivariate analysis of variance, MAP variance, of analysis multivariate • Of 6 FSS studies, all except 1 yielded posi- = BAI Fibromyalgia, of Appraisal 15

Entered/ tive evidence that homeopathy was superior to = = 30/26 Completed 32/27 P F-VAS Assessment, Functional 51 = No. of Subjects Subjects of No. = placebo, and the negative study was one of = H 37 Entered/H P the smallest and methodologically the weakest. Fisher’s first study46 failed to show positive ef- Revised Test Anxiety Scale, SII Revised Test traumatic brain injury. brain traumatic =

= fect for homeopathy on the all-comers sample McGill SensoryNS McGill Pain, = effect FA size, but was positive on 2 key predefined measures = when prospective matching of remedy to

homeopathy homeopathy clinical picture was taken into account. His chronic fatigue syndrome, CFS-Q syndrome, fatigue chronic 47 = second study was positive on 1 measure, but

32 impossible to interpret on 2 of the 4 primary outcomes.52,53 Three positive FSS trials46–48

Condition Treatment were given low ratings according to the SIGN Fibromyalgia Individualized CFS Individualized attention-deficit/hyperactivity disorder, AF disorder, attention-deficit/hyperactivity and GRADE assessments, but the 2 method- =

MANOVA Questionnaire, Intervention of Impact 49,50

= ologically strongest studies were positive 49 50,d

RTA Questionnaire, Perception Patient for homeopathy. In one of these, although =

TBI (Trait), Anxiety Inventory Trait State several outcomes failed to show a difference, 51,c = Dysfunctional Beliefs About Sleep, ES Sleep, Dysfunctional Beliefs About Children’s Checking Test, CFS Test, Checking Children’s MSP Inventory, Fatigue Multidimensional = the most rigorous measure of clinically signifi- = = cant improvement in all primary scales was STAI(T) PPQ States, MFI Rating Scale,Rating IIQ DBAS CCT Power calculation was done and adequate sample was achieved. was sample adequate and done was calculation Power Power calculation was done but sample was insufficient. was sample but done was calculation Power Level of evidence with respect to minimizing bias was rated as “good” (++), “fair” (+), or “poor” (–) according to Scottish Intercollegiate Guidelines Network (SIGN) quality analysis. quality (SIGN) Network Guidelines Intercollegiate Scottish to (–) according “poor” (+), or (++), “fair” “good” as rated was bias evidence minimizing respect Level with to of Only presented in Maharaj thesis. in Maharaj presented Only No power calculation. power No positive. Saul (2005) Saul Table 1 (continued). Placebo-Controlled, Randomized Clinical Trials of Homeopathy for Common Psychiatric Conditions: Quality Individual Studies 1 (continued). Placebo-Controlled, Randomized Clinical Trials Table Reference Bell et al (2004) a b c d e ADHD Abbreviations:

Effect Size and Number Needed to Treat b It was possible to calculate ESs in 16 of the 25 studies. Of the 12 studies in which a single outcome was used to determine ES, GRADE GRADE results favored homeopathy in 8 and placebo against use against against use against for use for Recommendation in 4 cases; The magnitude of effect in favor Weak recommendation recommendation Weak No recommendation No No recommendation No Weak recommendation recommendation Weak No recommendation No Weak recommendation recommendation Weak of homeopathy was large in 2, medium in 1,

b small in 2, and negligible in 3. ES in favor of placebo was small in 1 study and below 0.2 a in 3 studies. For the 4 studies with multiple primary outcomes that yielded discrepant Safety GRADE Safety provided reported provided reported reported reported results, the most favorable ES for homeopa- appears safe with infrequent adverse events events adverse infrequent with safe appears +2 Based on 2 studies +2 Based 2 studies on No information information No +2 Based on 2 studies +2 Based 2 studies on +2 Based on 1 study +2 Based 1 study on +2 Based on 2 studies +2 Based 2 studies on

= thy was medium in 1 study and small in 3 studies. For those outcomes least favorable appears to have safety concerns that include include that concerns safety have to appears =

b to homeopathy, the ES was small in 2 cases and negligible (ie, below 0.2) in 2 cases. Across the 13 studies in which it was possible to obtain confidence intervals for the ES, the upper and lower bounds of 95% confidence

of Effectof GRADE intervals crossed zero in all except 3 instanc- Magnitude of Estimate Estimate of Magnitude 0.78, 2.40) (−0.12, 0.34, 0.17) 0.27–0.49 on subscales 0.27–0.49 on measure (−0.07, −0.07, and (−0.07, −0.07, and measure 0.31–0.40) es, thus indicating substantial imprecision in

b the estimates of treatment effect, which are shown in Table 3. B effect (−0.43, −0.07, 0.50) No information No B small to None B and Small; 0.14 overall B to small, according or None C Small (0.03, 0.24, 0.17–0.24, C Small (−0.17, 0.94) In 10 studies, it was possible to obtain response rates and derive the NNTs, which are Confidence in Estimate of of in Estimate Effect GRADE given by category. NNT results were obtained

safety not well understood or conflicting, –1 or understood well not safety 47–50

= for 4 of the 6 FSS studies, which when pooled (N = 260) yield an NNT of 3.67.

0.8) effect. We also report the number of studies from which we were not able to calculate an effect an calculate size. to able not were which we from studies of also the number report 0.8) effect. We The chance of obtaining a positive result 50 (1) 269 (6) 243 (5) 125 (3) 314 (6) favoring homeopathy was unrelated to study Completed Completed

(no. of studies) of (no. quality. Quality of the 25 studies was vari- No. of Participants Participants of No. able with respect to minimization of bias, but there was no suggestion that the more favorable outcomes for homeopathy were associated with lower quality or weaker methodology, in that a higher proportion (66%) of the 6 best-quality reports could be taken as supportive of homeopathy to

has serious safety concerns that include frequent and serious adverse events and/or interactions. and/or events serious adverse and frequent include that concerns serious safety has varying degrees, while only 4 of 10 (40%) = (range across studies) across (range in the weakest group provided positive studies studies Response Rates on Homeopathy Homeopathy on Rates Response evidence. This lack of association between 40%, but reported in only 1 of the 1 of in only reported 40%, but 80%, reported in only 1 of the 1 of in only 80%, reported 40%–90%, reported across 2 studies across 40%–90%, reported 87 (4) Not provided Not 74%–85% 26%–50% quality and outcome is possibly due to the low number of studies for which ES could be calculated (16), the rather crude nature of quality rating schemes in general, and the small sample sizes, which have a large impact on estimates of precision. Publication bias is another possible reason, with lower- (range across studies) across (range 1 of the studies 1 of the studies 2 studies quality studies simply not being published Response Rates on Placebo on Rates Response in only reported 42%, but 46%, reported in only 1 of 1 of in only 46%, reported Not provided Not 4%–15% or reported. We doubt that publication bias Grading of Recommendations Assessment, Development and Evaluation. and Development Assessment, Recommendations of Grading appears relatively safe but with frequent but not serious adverse events and interactions, 0 interactions, and events serious adverse not but frequent with but safe relatively appears

= was a significant factor, however, because of = our extensive search strategy, the inclusion of “gray” literature, and the generally low level of funding for research in this field. In addition, we conducted a funnel plot using 13 studies with sufficient information for disturbances hyperactivity disorder hyperactivity chronic fatigue syndrome fatigue chronic Formulating the safety grade is dependent on the frequency and severity of adverse effects and interactions. The criteria developed are as follows: +2 follows: as are developed criteria The effectsinteractions. adverse and severity of the frequency and on grade dependent is the safety GRADE. Formulating Safety +1 interactions, and –2 and interactions, and/or events serious adverse but infrequent This algorithm follows the GRADE working group approach. There are 4 possible levels, A–D, as follows. (A) High: Further research is very unlikely to change our confidence confidence our very is change to unlikely research Further (A) High: follows. as A–D, levels, 4 possible are There approach. group working the GRADE follows algorithm This effect. the of estimate the in Confidence or impact important an have to likely is research Further (B) Moderate: RCT. multicenter high-quality in special or cases: 1 large, results consistent with effect; of several RCTs high-quality in the estimate in confidence or impact important an veryhave is to (C) Low: likely research Further limitations. some with several or RCTs RCT 1 high-quality the estimate: change effectmay of and in the estimate confidence more 1 or evidence or effect very of is research direct no uncertain: expert estimate low: opinion, Any (D) Very limitations. severe with RCTs more 1 or the estimate: change to likely effectis of and the estimate very with limitations. severe RCTs effect if an not criteria size was inclusion meeting the studies effect calculate to size for attempted We large). and small, moderate, (none, 4 levels into element this data categorize We size. effect the of Magnitude (> large a small(0.5–0.8), and reported effect (0.2–0.5), moderate studies many how report We provided. GRADE has defined the levels as follows: strong recommendation in favor, weak recommendation in favor, no recommendation, weak recommendation against, and strong strong and against, weak recommendation recommendation, no in favor, weak recommendation in favor, recommendation strong defined follows: has as GRADE the levels recommendation. the of Strength against. recommendation Four major domains comprise the core of the evidence-basedmethodology: of evaluation the core comprise domains major Four See Table 1 for outcomes assessed. outcomes 1 for See Table this procedure (data not shown). Analysis of Table 2. Placebo-Controlled, Randomized Clinical Trials (RCTs) of Homeopathy for Common Psychiatric Conditions: Quality the Overall Literature Pool (RCTs) 2. Placebo-Controlled, Randomized Clinical Trials Table Condition Psychiatric stress Anxiety or a b GRADE Abbreviation: Sleep or circadian rhythm rhythm circadian or Sleep Premenstrual syndromePremenstrual across 38%–60%, reported Attention-deficit/ Mild traumatic brain injury brain traumatic Mild 55%–74% Fibromyalgia/ Fibromyalgia/

Table 3. Effect Size (ES) and Number Needed to Treat (NNT) in Studies of Homeopathy Versus Placeboa Reference ES (95% CI) Rating Used NNT Rating Used Anxiety or stress Bonne et al (2003)24 −0.07 (−0.70 to 0.55) HARS −47.5 50% Reduction in HARS score Baker et al (2003)26 −0.43 (−1.02 to 0.17) Revised Test Anxiety Scale McCutcheon (1996)27 0.50 (0.03 to 0.97) Sleep loss 0.22 (−0.25 to 0.68) State anxiety Sleep or circadian rhythm disturbances Lipman et al (1999)30 0.78 (0.35 to 1.22) Snoring diary 2.95 Global rating Kolia-Adam et al (2008)35 0.24 (−0.53 to 1.02) Hours asleep −5.99 Satisfaction with sleep La Pine et al (2006)34 0.03 (−0.49 to 0.56) Fatigue Naudé et al (2010)31 2.40 (1.46 to 3.34) Sleep Improvement index Kumar (2010)33 0.24 POMS-Fatigue 0.17 POMS-Vigor Premenstrual syndrome Yakir et al (2001)38 0.94 (−0.02 to 1.90) MDQ 1.87 Global improvement Laister (2010)39 −0.17 (−0.93 to 0.58) MDQ-Pain Chapman et al (1994)37 −5.00 Global improvement Attention-deficit/hyperactivity disorder Jacobs et al (2005)41 −0.12 (−0.72 to 0.48) Conners Parent Global Scale Frei et al (2005)42 0.34 Conners Parent Global Scale Strauss et al (2000)43 0.17 (−0.71 to 1.05) Conners Parent Symptom Questionnaire Mild traumatic brain injury Chapman et al (1999)45 0.14 Three-part Functional Assessment Scale 3.26 Global situations 621.00 Global activities Functional somatic syndromes Weatherley-Jones et al (2004)49 0.40 (−0.03 to 0.83) Multidimensional Fatigue Inventory-Fatigue 6.14 Clinically significant improvement on all primary scales −0.08 (−0.34 to 0.50) Multidimensional Fatigue Inventory-Reduced Motivation Bell et al (2004)50 0.31 (−0.23 to 0.86) Tender point palpation pain 2.84 25% Improvement on tender point −0.07 (−0.61 to 0.47) McGill sensory pain palpation pain Fisher et al (1990)53 4.28 Global improvement Awdry (1996)48 2.49 Global response: unchanged or slight improvement vs other categories aNegative values indicate that placebo was more effective than homeopathy. Abbreviations: HARS = Hamilton Anxiety Rating Scale, MDQ = Menstrual Distress Questionnaire, POMS = Profile of Mood States.

the funnel plot also did not support evidence of publication in sleep-related disorders is warranted; a recent polysomnog- 2 54 bias (χ 1 = 1.923, P = .166). raphy study by Bell et al offers some basis for believing in the activity of homeopathic remedies on sleep mechanisms. The DISCUSSION efficacy of homeopathy for FSS looks promising, but larger well-designed studies are needed. Principal findings of this systematic review are as fol- Functional somatic syndromes, which account for 25% to lows: Homeopathy had no effect over placebo in the studies 50% of all outpatient visits in the United States,55 are chronic, of anxiety and stress reaction. There are currently no stud- disabling conditions that are unlikely to show spontaneous ies meeting our selection criteria for depression. There was improvement. They are also among the more frequently stud- reasonable evidence for the efficacy of homeopathy in func- ied psychiatric disorders with respect to homeopathy. In this tional somatic syndromes. Findings for other conditions were review, 5 of the 6 studies provided some evidence for efficacy mixed and inconclusive. Sample sizes were generally small, in either fibromyalgia or chronic fatigue syndrome. The low and overall confidence in the results was graded as moderate placebo response (4%–15%) and modestly consistent rates of or low, suggesting that further research could well change response to homeopathy (26%–50%) in these disorders and the estimate of effect. Mainly because of the limited num- the larger sample size of over 200 patients may have yielded ber of studies in any single category and heterogeneity of the more precise estimates than in the other categories. Taking data set, we decided that meta-analysis was not meaningful. the best-case outcomes, ESs of 0.31 (pain) and 0.40 (fatigue) Disorders were grouped to provide some level of diagnos- are comparable to the ES ranges that have been reported for tic homogeneity, although this clearly worked better for selective serotonin reuptake inhibitor antidepressants of 0.39 some disorders (eg, ADHD, PMS) than for others (eg, sleep/ and 0.17.56 Other widely used psychotropic drugs for fibro- circadian rhythm and anxiety disorders). Possible reasons for myalgia have small ESs for pain and fatigue, in the range of the lack of effect in stress and anxiety include a high placebo 0.2 for pain and 0.1 to 0.3 for fatigue.57 Consistent with the response or spontaneous recovery for the conditions studied, efficacy of homeopathy in fibromyalgia is a pragmatic RCT clinical variability of the included syndromes, methodological that showed benefit for homeopathy over usual treatment in problems, or some other factor. Further study of homeopathy primary care.58 Relevant collateral support in this context

© JC ClinOPYRIGHT Psychiatry 2011 72:6, P HYSICIANSJune 2011 POSTGRADUATE PRESS, INC. © COPYRIGHT 2011 PHYSICIANS POSTGRADUATE PRESS, INC803. Davidson et al comes from exploratory work by Bell et al,59,60 who found The same holds true for safety. For anxiety and stress-related links of clinical benefit to possible mechanisms of action problems, particularly generalized anxiety disorder, the data and predictors of response in their fibromyalgia sample. The are not encouraging, but most forms of anxiety remain un- overall NNT for homeopathy on global measures in the 4 studied. For fibromyalgia and chronic fatigue syndrome as a studies that provided source information compares favorably group, results suggest possible utility for homeopathy. For the with the NNTs of 5.0 to 9.2 (as determined for 30% relief of other disorders, the data are insufficient in quality or quan- pain) reported in 5 studies of gabapentin and pregabalin for tity to generate either positive or negative recommendations. fibromyalgia.57 All-cause dropout rates in 3 FSS studies were Overall, we believe the findings offer sufficient grounds to 11% for homeopathy and 10% for placebo, which compares warrant further clinical trials and are compatible with the use to the published dropout rate of 21% for adverse effects with of homeopathy to treat certain conditions. 57 15,16 pregabalin and gabapentin. As noted by others, studies Drug names: gabapentin (Neurontin and others), pregabalin (Lyrica). of homeopathy for fibromyalgia are currently neither suffi- Author affiliations: Department of Psychiatry and Behavioral Science, ciently rigorous nor sufficiently plentiful to warrant a definite Duke University Medical Center, Durham, North Carolina (Dr Davidson); and Samueli Institute, Alexandria, Virginia (Ms Crawford and Drs Ives and answer on its use, but the evidence is encouraging. Jonas). Full understanding of any treatment involves not only Potential conflicts of interest: In the last 12 months, Dr Davidson has evidence of efficacy, but also evidence of safety. Unfortu- received consulting fees from AstraZeneca and Euthymics Bioscience and royalties from the Davidson Trauma Scale, Social Phobia Inventory, nately, only 7 studies addressed this question, and even then Connor-Davidson Resilience Scale, Guilford Publications, and American the assessments were minimal, but all indicated there was Psychiatric Press. Ms Crawford and Drs Ives and Jonas report no potential no difference between homeopathy and placebo, which is conflict of interest. Funding/support: This project was partially supported by award number consistent with the general presumption about the safety W81XWH-08-1-0615-P00001 (United States Army Medical Research of homeopathy, where side effects and aggravations of the Acquisition Activity). underlying symptoms have not been found to occur more fre- Disclaimer: The views expressed in this article are those of the authors and do not necessarily represent the official policy or position of the US Army quently on homeopathy than on placebo in a major systematic Medical Command or the Department of Defense. analysis.61 In one study of ADHD,42 there were 3 dropouts Acknowledgments: The authors acknowledge the following individuals related to tics, depression, and disturbed behavior, which sug- for giving their permission to cite their unpublished results: Peter Fisher, FF Hom, FRCP, Royal London Homoeopathic Hospital and University gests that careful evaluations might indicate the existence of College London, United Kingdom; Elizabeth Solomon, HD, ND, DO, homeopathy-related adverse effects. What cannot be assessed Department of Homoeopathy, Faculty of Health Sciences, University here, however, is the “harm” caused by failing to offer an ef- of Johannesburg, South Africa; Andrew Criglington, B Comm, Miers Laboratories, Wellington, New Zealand; and David Naudé, M Tech (Hom), fective treatment to a condition that, if untreated, leads to Department of Homeopathy, Faculty of Health Sciences, Durban University disability or other morbidity. To the extent that the reports of Technology, Durban, South Africa. Mr Criglington is an employee of said little about safety, our GRADE-based recommendations and stock shareholder in Miers Laboratories; the other acknowledged individuals report no potential conflict of interest. have limitations, since safety evaluation should be taken into account when making such assessments. One surprising find- REFERENCES ing was the low rate of dropouts, which was 12% in 12 studies (range, 0%–21%). In that one of the more common reasons 1. Freeman MP, Fava M, Lake J, et al. Complementary and alternative medi- for early exit relates to side effects, a low dropout rate might cine in major depressive disorder: the American Psychiatric Association Task Force report. J Clin Psychiatry. 2010;71(6):669–681. doi:10.4088/JCP.10cs05959blu PubMed be seen as a favorable aspect of homeopathic treatment. On 2. Freeman MP, Mischoulon D, Tedeschini E, et al. Complementary and al- the other hand, an almost total lack of side effects is often ternative medicine for major depressive disorder: a meta-analysis of patient characteristics, placebo-response rates, and treatment outcomes relative to taken to imply lack of efficacy. This aspect of homeopathy standard antidepressants. J Clin Psychiatry. 2010;71(6): has consistently been neglected in the design and reporting 682–688. doi:10.4088/JCP.10r05976blu PubMed of clinical trials. 3. Linde K, Clausius N, Ramirez G, et al. Are the clinical effects of homeopathy placebo effects? a meta-analysis of placebo-controlled trials. Lancet. Limitations of this review include its inability to provide 1997;350(9081):834–843. doi:10.1016/S0140-6736(97)02293-9 PubMed information about major depression, which is such a large 4. Jonas WB, Kaptchuk TJ, Linde K. A critical overview of homeopathy. Ann Intern Med. 2003;138(5):393–399. PubMed health problem worldwide and for which there is quite an ex- 5. Linde K, Melchart D. Randomized controlled trials of individual- tensive literature on other CAM approaches. We also did not ized homeopathy: a state-of-the-art review. J Altern Complement Med. 1998;4(4):371–388. doi:10.1089/acm.1998.4.371 PubMed include the entire range of psychiatric problems in our review, 6. Ernst E. Homeopathy: what does the “best” evidence tell us? Med J Aust. such as dementia, alcohol and substance problems, eating dis- 2010;192(8):458–460. PubMed orders, or psychosis. Apart from an unrevealing Cochrane 7. Lewith G. The conundrum of homeopathy: a commentary on Rutten ALB & Stolper CF (2009). J Eval Clin Pract. 2009;15(6):1236–1237. doi:10.1111/j.1365-2753.2009.01261.x PubMed review of homeopathy for dementia, we are unaware of any 8. Cucherat M, Haugh MC, Gooch M, et al; Homeopathic Medicines systematic reviews, or even a double-blind placebo-controlled Research Advisory Group. Evidence of clinical efficacy of homeopathy: a meta-analysis of clinical trials. Eur J Clin Pharmacol. 2000;56(1):27–33. doi:10.1007/s002280050716 PubMed trial, of homeopathy in any of these disorders. Another con- 9. Shang A, Huwiler-Müntener K, Nartey L, et al. Are the clinical effects of sideration is that not all studies we reviewed presented their homoeopathy placebo effects? comparative study of placebo-controlled results as an intent-to-treat analysis. trials of homoeopathy and allopathy. Lancet. 2005;366(9487):726–732. doi:10.1016/S0140-6736(05)67177-2 PubMed 10. Lüdtke R, Rutten ALB. The conclusions on the effectiveness of home- In summary, our review demonstrates that well-designed opathy highly depend on the set of analyzed trials. J Clin Epidemiol. and comprehensively reported homeopathic studies in psy- 2008;61(12):1197–1204. doi:10.1016/j.jclinepi.2008.06.015 PubMed 11. Heirs M, Dean ME. Homeopathy for attention deficit/hyperactiv- chiatry are few and far between and preclude firm conclusions ity disorder or hyperkinetic disorder. Cochrane Database Syst Rev. about the efficacy of this treatment in any single disorder. 2007;4:CD005648. doi:1011002/14651858.CD005648.pub2.

12. Pilkington K, Kirkwood G, Rampes H, et al. Homeopathy for depression: insomnia: a double-blind trial. Simillimum. 2008;21:91–99. a systematic review of the research evidence. Homeopathy. 2005;94(3): 36. Kolia-Adam N. The Efficacy of Coffea cruda 200cH on Insomnia [mini- 153–163. doi:10.1016/j.homp.2005.04.003 PubMed dissertation]. Johannesburg, South Africa: Technikon Witwatersrand; 2010. 13. Pilkington K, Kirkwood G, Rampes H, et al. Homeopathy for anxiety http://152.106.6.200:8080/dspace/bitstream/10210/3070/1/Kolia-Adam. and anxiety disorders: a systematic review of the research. Homeopathy. pdf. Accessed June 21, 2010. 2006;95(3):151–162. doi:10.1016/j.homp.2006.05.005 PubMed 37. Chapman EH, Angelica J, Spitalny G, et al. Results from a study of the 14. McCarney RW, Warner J, Fisher P, et al. Homeopathy for dementia. homeopathic treatment of PMS. J Am Inst Homeo. 1994;87:14–21. Cochrane Database Syst Rev. 2003;1:CD003803. doi:1002/14651858.CD003803. > 38. Yakir M, Kreitler S, Brzezinski A, et al. Effects of homeopathic treatment 15. Perry R, Terry R, Ernst E. A systematic review of homoeopathy for the in women with premenstrual syndrome: a pilot study. Br Homeopath J. treatment of fibromyalgia. Clin Rheumatol. 2010;29(5):457–464. doi:10.1007/s10067-009-1361-2 PubMed 2001;90(3):148–153. doi:10.1054/homp.1999.0491 PubMed 16. De Silva V, El-Metwally A, Ernst E, et al; Arthritis Research Campaign 39. Laister C-A. The efficacy of homoeopathic simillimum in the treat- Working Group on Complementary and Alternative Medicines. Evidence ment of premenstrual syndrome (PMS). http://ir.dut.ac.za/bitstream/ for the efficacy of complementary and alternative medicines in the handle/10321/383/Laister_2008.pdf?sequence=4, accessed June 21, 2010. management of fibromyalgia: a systematic review. Rheumatology (Oxf). 40. Kirtland KA. The Efficacy of Folliculinum in the Treatment of Premenstrual 2010;49(6):1063–1068. doi:10.1093/rheumatology/keq025 PubMed Tension [thesis]. Durban, South Africa: Technikon Natal; 1994. 17. Cooper KL, Relton C. Homeopathy for insomnia: a systematic review of 41. Jacobs J, Williams A-L, Girard C, et al. Homeopathy for attention-deficit/ research evidence. Sleep Med Rev. 2010;14(5):329–337. doi:10.1016/j.smrv.2009.11.005 PubMed hyperactivity disorder: a pilot randomized-controlled trial. J Altern 18. Jacobs J, Crothers D. Who sees homoeopaths? a study of patients’ Complement Med. 2005;11(5):799–806. doi:10.1089/acm.2005.11.799 PubMed characteristics in a homeopathic family practice. Br Homeopath J. 42. Frei H, Everts R, von Ammon K, et al. Homeopathic treatment of children 1991;80(1):57–58. doi:10.1016/S0007-0785(05)80428-4 with attention deficit hyperactivity disorder: a randomised, double blind, 19. Davidson JRT, Rampes H, Eisen M, et al. Psychiatric disorders in pri- placebo controlled crossover trial. Eur J Pediatr. 2005;164(12):758–767. doi:10.1007/s00431-005-1735-7 PubMed mary care patients receiving complementary medical treatments. Compr 43. Strauss LC. The efficacy of a homeopathic preparation in the man- Psychiatry. 1998;39(1):16–20. doi:10.1016/S0010-440X(98)90027-X PubMed agement of attention deficit hyperactivity disorder. Biomed Ther. 20. SIGN Network. A Guideline Developer’s Handbook. Edinburgh, Scotland: 2000;18(2):197–201. SIGN; 2001. http://www.sign.ac.uk/pdf/sign50.pdf. Accessed March 8, 44. Cole CL. The Efficacy of Selenium homaccord in the Management of 2011. Attention Deficit Hyperactivity Disorder [dissertation]. Johannesburg, South 21. Jacobs BP, Gundling K. Introduction. In: The ACP Evidence-Based Guide Africa: Technikon Witwatersrand, 1998; 1–73. to Complementary and Alternative Medicine. Philadelphia, PA: ACP Press; 45. Chapman EH, Weintraub RJ, Milburn MA, et al. Homeopathic treatment 2009. http://wdn.ipublishcentral.net/acp/viewinside/25931038387829. of mild traumatic brain injury: a randomized, double-blind, placebo-con- Accessed March 7, 2011. trolled clinical trial. J Head Trauma Rehabil. 1999;14(6):521–542. doi:10.1097/00001199-199912000-00002 PubMed 22. Hedges L, Olkin I. Statistical Methods for Meta-Analysis. London, England: 46. Fisher P. An experimental double-blind clinical trial method in homeopa- Harcourt Brace Jovanovich; 1985. thy: use of a limited range of remedies to treat fibrositis. Br Homeopath J. 23. NNT Calculator. Evidence Based Emergency Medicine Web site. 1986;75(3):142–147. doi:10.1016/S0007-0785(86)80009-6 http://www.ebem.org/nntcalculator.html. Accessed August 19, 2010. 47. Fisher P, Greenwood A, Huskisson EC, et al. Effect of homeopathic treat- 24. Bonne O, Shemer Y, Gorali Y, et al. A randomized, double-blind, placebo- ment on fibrositis (primary fibromyalgia). BMJ. 1989;299(6695):365–366. doi:10.1136/bmj.299.6695.365 PubMed controlled study of classical homeopathy in generalized anxiety disorder. 48. Awdry R. Homeopathy may help ME. Int J Altern Complement Med. J Clin Psychiatry. 2003;64(3):282–287. doi:10.4088/JCP.v64n0309 PubMed 1996;14:12–16. 25. Ngobese JC. The Relative Efficacy of Homeopathic Simillimum Treatment as 49. Weatherley-Jones E, Nicholl JP, Thomas KJ, et al. A randomised, controlled, Compared to Psychological Counseling (Cognitive Therapy and Behavioral triple-blind trial of the efficacy of homeopathic treatment for chronic fa- Therapy) in the Management of Generalized Anxiety Disorder [mini- tigue syndrome. J Psychosom Res. 2004;56(2):189–197. doi:10.1016/S0022-3999(03)00377-5 PubMed dissertation]. Durban, South Africa: Durban University of Technology; 50. Bell IR, Lewis DA 2nd, Brooks AJ, et al. Improved clinical status in fibro- 2006. http://ir.dut.ac.za/handle/10321/26. Accessed April 19, 2010. myalgia patients treated with individualized homeopathic remedies versus 26. Baker DG, Myers SP, Howden I, et al. The effects of homeopathic Argentum placebo. Rheumatology (Oxford). 2004;43(5):577–582. doi:10.1093/rheumatology/keh111 PubMed nitricum on test anxiety. Complement Ther Med. 2003;11(2):65–71. doi:10.1016/S0965-2299(03)00059-1 PubMed 51. Saul W. The Effectiveness of Homoeopathic Simillimum Treatment in 27. McCutcheon LE. Treatment of anxiety with a homeopathic remedy. J Appl Chronic Fatigue Syndrome (CFS) [M. Tech thesis]. Durban, Johannesburg, Nutr. 1996;48(1&2):2–6. http://homeoinst.org/node/358 South Africa: Technikon Natal; 2005. http://ir.dut.ac.za/bitstream/ 28. Traub G. The Influence of Homeopathic Medicines on Thought Interference, handle/10321/44/Saul_2005.pdf?sequence=16. Accessed June 21, 2010. Nervousness and Anxiety in University Students Under Examination 52. Colquhoun D. Re-analysis of clinical trial of homoeopathic treatment in Conditions [M. Tech thesis]. Johannesburg, South Africa: Technikon fibrositis. Lancet. 1990;336(8712):441–442. doi:10.1016/0140-6736(90)91989-N PubMed Witwatersrand; March 2000. http://ujdigispace.uj.ac.za:8080/dspace/ 53. Fisher P, Huskisson EC, Turner P, et al. Homoeopathic treatment of handle/10210/2580. Accessed Jun 21, 2010. fibrositis. Lancet. 1990;336(8720):954. doi:10.1016/0140-6736(90)92334-E PubMed 29. Vaithilingam H. The Effectiveness of Homoeopathic Simillimum in 54. Bell IR, Howerter A, Jackson N, et al. Effects of homeopathic medicines the Treatment of Job Burnout in the Human Services Field [mini- on polysomnographic sleep of young adults with histories of coffee-related dissertation]. Durban, South Africa: Durban University of Technology; insomnia [published online ahead of print July 27, 2010]. Sleep Med. doi:10.1016/j.sleep.2010.03.013 PubMed 2005. http://ir.dut.ac.za/bitstream/handle/10321/20/Vaithilingam_2005. 55. Aaron LA, Buchwald D. A review of the evidence for overlap among unex- pdf?sequence=7. Accessed June 21, 2010. plained clinical conditions. Ann Intern Med. 2001;134(9 Pt 2):868–881. PubMed 30. Lipman D, Sexton G, Schlesser J. A randomized double-blind 56. Häuser W, Bernardy K, Uçeyler N, et al. Treatment of fibromyalgia syn- placebo-controlled evaluation of the safety and efficacy of a natural over- drome with antidepressants: a meta-analysis. JAMA. 2009;301(2):198–209. doi:10.1001/jama.2008.944 PubMed the-counter (OTC) medication in the management of snoring. Sleep Breath. 57. Häuser W, Bernardy K, Uçeyler N, et al. Treatment of fibromyalgia syn- 1999;3(2):53–56. doi:10.1007/s11325-999-0053-0 PubMed drome with gabapentin and pregabalin—a meta-analysis of randomized 31. Naudé DF, Stephanie Couchman IM, Maharaj A. Chronic primary insom- controlled trials. Pain. 2009;145(1-2):69–81. doi:10.1016/j.pain.2009.05.014 PubMed nia: efficacy of homeopathic simillimum. Homeopathy. 2010;99(1):63–68. doi:10.1016/j.homp.2009.11.001 PubMed 58. Relton C, Smith C, Raw J, et al. Healthcare provided by a homeopath as an 32. Maharaj A. The Efficacy of Homoeopathic Simillimum in the Treatment adjunct to usual care for fibromyalgia (FMS): results of a pilot randomised of Chronic Primary Insomnia [M. Tech thesis]. Durban, South Africa: controlled trial. Homeopathy. 2009;98(2):77–82. doi:10.1016/j.homp.2008.12.004 PubMed Durban Institute of Technology; 2005. http://ir.dut.ac.za/bitstream/ 59. Bell IR, Lewis DA 2nd, Schwartz GE, et al. Electroencephalographic handle/10321/52/Maharaj_2005.pdf?sequence-5. Accessed June 21, cordance patterns distinguish exceptional clinical responders with fibro- 2010. myalgia to individualized homeopathic medicines. J Altern Complement 33. Kumar K. No Jet Lag Scientific Test. www.nojetlag.com/jetlag6.html. Med. 2004;10(2):285–299. doi:10.1089/107555304323062275 PubMed Accessed February 9, 2010. 60. Bell IR, Lewis DA 2nd, Lewis SE, et al. EEG alpha sensitization in in- 34. La Pine MP, Malcomson FN, Torrance JM, et al. Night shift: can dividualized homeopathic treatment of fibromyalgia. Int J Neurosci. a homeopathic remedy alleviate shift lag? Dimens Crit Care Nurs. 2004;114(9):1195–1220. doi:10.1080/00207450490475724 PubMed 2006;25(3):130–136. doi:10.1097/00003465-200605000-00015 PubMed 61. Grabia S, Ernst E. Homeopathic aggravations: a systematic review of ran- 35. Kolia-Adam N, Solomon E, Bond J, et al. The efficacy of Coffea cruda on domised, placebo-controlled clinical trials. Homeopathy. 2003;92(2):92–98. doi:10.1016/ S1475-4916(03)00007-9 PubMed

eAppendix 1 is available at

© JC ClinOPYRIGHT Psychiatry 2011 72:6, P HYSICIANSJune 2011 POSTGRADUATE PRESS, INC. © COPYRIGHT 2011 PHYSICIANS POSTGRADUATE PRESS, INC805. N N N N N N N N N N N N N N NNN NNN [1.10] (continued) Results Are Results Are for All Sites All for Is Multisite, Is Multisite, If the Study Comparable Comparable P P P P P P P P P P A A A W PPP PPP [1.9] Allocated Allocated All SubjectsAll Are Analyzed Analyzed Are (intention-to- in the Groups in the Groups treat analysis) treat to Which They They Which to Were Randomly Were P P P P P P P P A A A A A A [1.8] WPW Completed in Each Treatment Treatment in Each Arm Dropped Out Arm Dropped 6% placebo 6% placebo 20% placebo 14% placebo 9% placebo 28% placebo 5% placebo What % of SubjectsWhat 7% placebo Before the Study Was Was the Study Before 12% Homeopathy, PAA Homeopathy, 7%–20% 11% Total 11% 13% Homeopathy, 21% Homeopathy, 8% Homeopathy, 32% Total 32% Total 18% No information Total 11% Total 17% 0% 14% Homeopathy, 9% Homeopathy, P P A A W W W W W W W W W W PPP [1.7] WWW Valid, and Valid, Outcomes Outcomes Reliable Way Reliable in a Standard, in a Standard, Are Measured Measured Are P A A A A A A A W W W W W W PPP AAA [1.6] Between Investigation Groups Is the Groups Only Difference Only Difference Treatment Under Treatment P P P P P P P P P P A W W W PPP APA Trial [1.5] Treatment Treatment Start of the Groups Are Are Groups and Control and Control Similar at the Similar at P P A A A A A A A A W W W W AAA AAA [1.4] Allocation Subjects and Investigators Are Kept “Blind” “Blind” Kept Are About Treatment Treatment About P P P P P A A A A A A A W W PPP AAA [1.3] Adequate Adequate Concealment Concealment Method Is Used P P P P P A A A A A W W W W PPP AAA [1.2] Group Group Treatment Treatment Randomized Assignment Is Assignment P A A A A A W W W W W W W W [1.1] PWW AWW Study Study Clearly Focused Focused Question Addresses Addresses Appropriate, Appropriate,

35 31,a 24 29 30 34 41 24 36 32

26 38 a 25 40 39 33 28 37 3. Combined (1994) 2. Thesis (2010) Thesis 2. 1. Publication (2008) 1. Publication 3. Combined Bonne et al (2003) eAppendix 1. Results of SIGN Evaluations Kumar (2010) Kumar Laister (2008) Laister Chapman et al, PMS PMS Chapman et al, McCutcheon (1996) McCutcheon Ngobese (2006) Baker et al (2003) Traub (2000) Traub Vaithilingam (2005) Vaithilingam et al Kolia-Adam Lipman et al (1999) 1. Naudé et al (2010) et al (2006) La Pine Yakir et al (2001) Yakir 2. Maharaj (2005) Jacobs et al (2005) Jacobs Kirtland (1994) Reference

© COPYRIGHT 2011 PHYSICIANS POSTGRADUATE PRESSdoi:10.4088/JCP.10r06580, INC. © COPYRIGHT 2011 PHYSICIANS POSTGRADUATE PRESS, INC. N N N N N N N N N [1.10] Results Are Results Are for All Sites All for Is Multisite, Is Multisite, If the Study Comparable Comparable P P P P P W W W W [1.9] Allocated Allocated All SubjectsAll Are Analyzed Analyzed Are (intention-to- in the Groups in the Groups treat analysis) treat to Which They They Which to Were Randomly Were traumatic brain injury, injury, brain traumatic = P P P P A A A W W [1.8] Completed in Each Treatment Treatment in Each Arm Dropped Out Arm Dropped 3% placebo 16% placebo 3% placebo 19% placebo 16% placebo What % of SubjectsWhat Before the Study Was Was the Study Before 14% Homeopathy, 9% Homeopathy, given No information 20% Homeopathy, No information given No information No information given No information 6% Homeopathy, 19% Total 19% 13% Homeopathy, P A W W W W W W W [1.7] Valid, and Valid, Outcomes Outcomes Reliable Way Reliable in a Standard, in a Standard, Are Measured Measured Are P P P P P A A W W [1.6] Between TBI Network, Guidelines Intercollegiate Scottish Investigation Groups Is the Groups = Only Difference Only Difference Treatment Under Treatment P P P P A W W W W Trial [1.5] Treatment Treatment Start of the Groups Are Are Groups and Control and Control Similar at the Similar at P A A A A A W W W SIGN syndrome, premenstrual [1.4] = Allocation Subjects and Investigators Are Kept “Blind” “Blind” Kept Are About Treatment Treatment About P P P A A A W W W [1.3] Adequate Adequate poorly addressed, PMS addressed, poorly Concealment Concealment = Method Is Used P A A A A W W W W [1.2] Group Group not applicable, P applicable, not Treatment Treatment Randomized = Assignment Is Assignment P A A A W W W W W [1.1] Study Study Clearly Focused Focused Question Addresses Addresses Appropriate, Appropriate, adequately addressed, N addressed, adequately 47 =

50 42 43 48

46 51

45 49 well addressed. well = (2004) (1999) W Where ratings were done of published, unpublished, and composite of both, the sequence was published report, unpublished thesis, and composite. and thesis, unpublished report, both, published the was sequence of composite and unpublished, published, of done were ratings Where Abbreviations: A Abbreviations: Frei et al (2005) Frei (2000) Strauss (1986) Fisher et al Weatherley-Jones Bell et al (2004) Saul (2005) a eAppendix 1 (continued). Results of SIGN Evaluations Chapman et al, TBI Chapman et al, Awdry (1996) Awdry Fisher et al (1989) Fisher Reference