The research evidence base for homeopathy

[Guest guest]

http://www.sciencedirect.com/science?_ob=ArticleURL & _udi=B6WXX-48HXDX3-8 & _user=1\

0 & _handle=W-WA-A-A-AE-MsSAYWW-UUA-AUCCZZEUAU-AUUWCUBZB-AE-U & _fmt=full & _coverDate\

=04%2F30%2F2003 & _rdoc=7 & _orig=browse & _srch=%23toc%237170%232003%23999079997%2342\

6583! & _cdi=7170 & view=c & _acct=C000050221 & _version=1 & _urlVersion=0 & _userid=10 & md5=\

0becc21b96c314d96dae8f1e3bb67256

 

Introduction

To sceptics, homeopathy is an archaic and ineffective method of treatment that

proclaims an implausible mode of action. In contrast, the individual and

collective experience of homeopathic practitioners paints a convincing picture

of its clinical effectiveness. Homeopathy's possible mechanisms of action remain

intangible theories, and it will be important ultimately to substantiate these.

The principal issues for the present and immediate future, however, are whether

there is compelling research evidence that homeopathy actually does exert a

remedial effect, and whether such an effect is remedy-specific.

What comprises `compelling evidence'? Many homeopaths are cheerfully sustained

by their own successful clinical experience: `We know it works, so why do

research to prove it?' However, if homeopathy truly enables people attain better

health, then it is vitally important that it achieves much wider respect within

medicine at large. To do this, homeopathy must convince sceptics `by rigorous

research evidence' of its clinical effectiveness. It must also face up to the

challenge of demonstrating the specific efficacy of at least some of its

medicines. These challenges are best met by data obtained from randomised

controlled clinical trials (RCTs), where homeopathy is compared to another

intervention or with placebo. Some homeopaths throw up their hands in horror at

such a proposition, but new research programmes can grow from observational

(cohort) studies,[1] for example, which can inform later RCTs.

A key to success of any study design is that clinical outcome measures must

properly reflect the whole-person approach to healthcare that typifies

homeopathy. Quality of life assessment and other patient-centred measures, for

example, may be at least as important as the measurement of biochemical markers

or other physical determinants of health status. After all, in an holistic

therapy like homeopathy, the individual symptom- (and thus remedy-) picture

often does not equate to a named disease, and so it is inappropriate to measure

only disease-specific outcomes in such studies.

Another vital concern is that any statistically significant difference/s between

treatment groups in trials should not automatically be equated to clinical

importance. It is only the latter that truly matters, and this must be properly

accommodated in the power calculations (and thus statistical conclusions)

connected with clinical trials. The existing homeopathic research literature has

not reflected this issue well––emphasis has been placed on statistical

significance with incomplete regard to the clinical value of any inter-group

differences observed.

What is the current evidence from which new research in homeopathy can develop?

Given the limitations of past research in homeopathy, the optimum question that

can be asked of the available evidence base is limited to: What is the weight of

the original evidence from published RCTs that homeopathy has an effect that is

statistically significantly different from that in a comparative group?

Only since 1975 have rigorous research methods been applied to homeopathy. Prior

literature comprised only four minor trials or brief communications.[2, 3, 4 and

5] The work reported up to 1997 is captured in the comprehensive meta-analysis

of placebo-controlled trials of homeopathy published by Linde et al in The

Lancet.[6] Nearly half the trials cited in that paper showed a homeopathic

treatment effect statistically significantly greater than that of placebo; none

of the trials found placebo more effective than homeopathy. The authors

concluded that homeopathy's clinical effects are not attributable solely to

placebo, though they could not single out any medical condition for which

homeopathy seemed clearly efficacious. [6]

Since 1997, some 50 new clinical trials or meta-analyses in homeopathy have been

published. Many of these newer trials have essentially the same Null hypothesis:

`Homeopathy has an effect which is not statistically significantly different

from that of placebo'; in effect, the Null hypothesis that was rejected by

Linde's meta-analysis. In 2002, the NHS Centre for Reviews and Dissemination

based at the University of York, UK, published an overview of all the published

systematic reviews and meta-analyses in homeopathic research.[7] The authors

conclude: `There are currently insufficient data ... to recommend homeopathy as

a treatment for any specific condition'. This statement is hardly

distinguishable from one of Linde's conclusions 5 years earlier.

There has been a recent increase in the number of research papers that carry the

more original Null hypothesis: `Homeopathy has an effect which is not

statistically significantly different from that of another active treatment'.

Such comparative trials of clinical effectiveness are included in the focus of

the present review.

The present paper reports the total currently available statistical evidence

from the published research literature for and against the specific efficacy or

clinical effectiveness of homeopathy, based on the frequency with which the

above two Null hypotheses have been rejected or not rejected in clinical trials.

This is neither a meta-analysis nor a formal systematic review––the paucity and

heterogeneity of published research in homeopathy limits the value of such

formalised approaches. The originality of the present approach lies mainly in

the following: (a) it considers the number of individual trials in homeopathy

that report outcome statistics, and by type of medical condition studied; (b) it

reports only full, published, research papers; © it includes comparative as

well as placebo-controlled trials. The coverage reflects the entire range of

medical conditions in which the effects of homeopathy have been the subject of

research and, like Linde et al,[6] it only includes trials

that are randomised and/or double-blinded. Based on my review, I conclude by

recommending some opportunities for future research development in homeopathy.

Literature search methods

This review examines all published clinical trials on human subjects that

appeared in the literature from 1975 until December 2002, and which compared

homeopathic treatment with placebo or with another medication, where a

randomised and/or double-blind study design has been used. The analysis includes

only full papers reporting original research, and excludes conference

proceedings, brief communications, book chapters and theses. Reports of case

series, clinical outcomes (cohort) studies and other non-controlled

investigations are also excluded. All forms of homeopathic intervention are

included, from classical to single-remedy and isopathy. Principal information

sources were: The National Library of Medicine (Medline); The Cochrane Library;

The Centralised Information Service for Complementary Medicine (CISCOM); The

British Homoeopathic Library (Hom-Inform); cross-referencing between published

papers.

Two principal outcomes are identified: `Null hypothesis rejected' and `Null

hypothesis not rejected'. For papers in the category Null hypothesis rejected,

the findings are classed as either `positive' (ie for homeopathy) or `negative'

(ie against homeopathy), based on a two-tailed test. A positive trial is one

where at least one outcome measure was statistically significantly improved by

homeopathy compared with placebo or alternative treatment (P0.05). A negative

study is one where homeopathic treatment was statistically significantly

inferior to the compared treatment (placebo, in the two actual cases cited

below) in at least one outcome measure (P0.05). Null hypothesis not rejected is

the conclusion from a study where no significant inter-group difference in

outcome/s is evident (P>0.05). In most cases, these conclusions have been based

on the authors' own reported hypothesis testing; for results obtained in a few

less accessible or non-English language papers, a probability of

P0.05 was equated with a reported odds ratio (±95% confidence interval) 1.[6]

A balance of evidence in favour of one of the above three categories or classes

for a given medical condition is concluded if, using simple arithmetic, it

contains at least two more papers than the sum of the papers in the other two

categories or classes. This approach has been adopted separately for

placebo-controlled and comparative trials. Results of literature search and

analysis

The database scrutinised and reported here comprises a total of 93 original

articles in homeopathic research. Of this total, 79 trials were

placebo-controlled, while the remaining 14 compared homeopathy with a

conventional medical treatment (controlled comparative trials). Table 1 lists

all those medical conditions (35 in total), in 11 broad types, for which there

exists at least one published clinical research trial in homeopathy that

satisfies the inclusion criteria for this analysis.

 

Table 1. Medical conditions for which there is at least one published

randomised/double-blind clinical research trial in homeopathy

(33K)

References cited by superscript beside the number of articles published. Key to

symbols: * Two studies set up formally as equivalence trials.[44 and 101] +

positive trial; – negative trial; ± inconclusive trial; includes common cold,

cough, sinusitis and pharyngitis.

 

 

 

Given the above criteria, 52 of the total 93 published papers are in the

category `Null hypothesis rejected' (Table 1); the remaining 41 papers are thus

in the category `Null hypothesis not rejected' (ie no conclusive difference

between homeopathy and a comparison group). Of the 52 articles, 50 are classed

as `positive' and two are `negative'. Within the 50 `positive' reports, 47

observed a homeopathic effect superior to placebo. The three others found

homeopathy to be superior to another treatment for the given condition––otitis

media, [8] osteoarthritis [9] and back pain. [10] As regards the two `negative'

articles (rheumatoid arthritis, [11] tissue healing after dental surgery [12]),

the placebo group had a significantly better clinical response in at least one

outcome measure than patients treated with a homeopathic remedy. Balance of

evidence: placebo-controlled trials

Examining the relative number of papers published for each of the three

categories or classes of evidence, the present weight of evidence favours

homeopathic treatment effectiveness in eight conditions:

 

childhood diarrhoea;

 

 

fibrositis (fibromyalgia);

 

 

hayfever/allergic rhinitis;

 

 

influenza;

 

 

pain (of various origins);

 

 

side-effects of radio-/chemotherapy;

 

 

sprains;

 

 

upper respiratory tract infection.

 

A weight of evidence suggesting homeopathy has no effect above placebo is

apparent in three medical conditions:

 

headache;

 

 

stroke;

 

 

warts.

 

For the remaining 20 conditions in which studies have been carried out, there is

insufficient weight of evidence either to favour or to find no support for

homeopathy (Table 1). Some of these trials are examples of the `double positive

paradox', [13] where a homeopathy group and a placebo group have

indistinguishable results but both manifest some clinical improvement. Balance

of evidence: comparative trials

There is insufficient evidence either to favour or to find no support for

homeopathy in nine of the 10 medical conditions in which studies have been

carried out (Table 1). Within the group of conditions `upper respiratory tract

infection', however, a weight of evidence suggests that homeopathy and

conventional medicine (aspirin in both papers concerned) are equally effective

in treating the common cold. [14 and 15] DiscussionNumber of published trials in

homeopathy

The total number of original full research papers over a 27 year time-span (n =

93) is very modest, for research in homeopathy is still in an early stage of

development. The largest total number of articles for any particular type of

condition is 10 (upper respiratory tract infection). For 12 of the 35 included

medical conditions, a single published trial is all that exists. The paucity of

research literature is a serious drawback for meta-analyses and formal

systematic reviews in homeopathy. In the last 5 years, however, such syntheses

of (usually very few) published trials have provided support for homeopathic

treatment in hayfever,[16] post-operative ileus [17] and rheumatoid arthritis.

[18] Other meta-analyses or systematic reviews have concluded there is

insufficient evidence for or against homeopathy in asthma, [19] influenza, [20]

muscle soreness, [21] osteoarthritis, [22] prophylaxis for migraine or headache

[23] and in tissue trauma treated with Arnica.[24]

Publication bias towards the reporting of `positive' homeopathic treatment

effects has been a well-founded concern, and is properly recognised in

meta-analyses,[6, 25 and 26] though no single trial is individually suspected of

biased reporting. [27] There is thus very likely to be overall exaggeration of

reported positive treatment effects in the homeopathic research literature cited

in the present review. Quality of published trials in homeopathy

This review deliberately does not categorise published trials in homeopathy by

their intrinsic scientific quality, for information on this issue is already

available. Only 29% of placebo-controlled trials in homeopathy published before

1997 were judged to possess `high' methodological quality.[6] A statistical

re-evaluation of Linde's meta-analysis noted that treatment effects were larger

in smaller studies and in those with inadequate blinding of outcome assessment.

[27] Exaggeration of treatment effect can also be expected in the 14 comparative

trials highlighted in the present review: the sample sizes of these 14 studies

(mean, 97 patients; range, 10–184) are broadly equivalent to those for

placebo-controlled trials reported by Linde et al (118; 5–1270).[6]

An example of some of the key issues can be taken from the two studies that have

examined the effectiveness of homeopathy in otitis media in children. In the

first of these papers, more patients on homeopathy than those on standard care

were found to have a normal tympanogram after a treatment period of 12

months.[8] The second article reported decreased symptom scores in acute otitis

media after 24 and 64 h in patients receiving homeopathy compared with those

given placebo.[28] These were the only statistically significant effects

observed in either investigation. Both studies were preliminary in nature and

had small sample sizes––33 and 75 patients, respectively. Both study designs

involved randomisation of patients, but of course the comparative trial was not

double-blinded. [8] The intrinsic quality of these studies is higher than

average in the homeopathic research literature. Balance of research evidence

The above caveats (low volume, publication bias, low quality) are important in

considering the research evidence base of homeopathy. Nevertheless, positive

effects of homeopathic treatment are apparent in 50 published trials (over half

of the included research literature). This represents a body of research where,

for at least one outcome measure per trial, the Null hypothesis has been

rejected in favour of homeopathy. Three of the papers (in otitis media,

osteoarthritis and back pain) provide research evidence that homeopathy can

actually be superior to conventional treatment.[8, 9 and 10] Only two trials

have found homeopathy to have less effect than placebo.

Based on the relative number of placebo-controlled studies with positive

results, the balance of research evidence currently favours homeopathy in

childhood diarrhoea, fibrositis, hayfever, influenza, pain, side-effects of

radio-/chemotherapy, sprains and upper respiratory tract infection. Based on

published research to date, it seems unlikely that homeopathy is efficacious for

headache, stroke or warts.

Insufficient research in 20 medical conditions prevents clear conclusions from

being drawn. There has been no research at all in homeopathy for many other

ailments. It is thus not surprising that comprehensive meta-analyses or

systematic reviews in homeopathy have not discerned any clear pattern of medical

conditions that appear especially promising for effective homeopathic

intervention.[29] Meanwhile, homeopathic practitioners continue to chronicle

their successful clinical cases. [30] Future opportunities and direction in

homeopathic research

Key issues that must be addressed in future research development in homeopathy

include: the research question, the associated trial design and the outcome

measures chosen. Fresh research should consider carefully whether

placebo-controlled trials and physical determinants of health are necessarily

the wisest approach. Such approaches may be particularly inappropriate in

researching homeopathy for chronic illness or where the treatment does not

address a named disease. Conditions of this nature are nearly always treated by

individualised homeopathy, where remedy selection is based on a person's

`constitutional' character, and so non-placebo-controlled designs using `quality

of life' measures, for example, might be the most relevant. The most appropriate

place for placebo-controlled trials in homeopathy might be in examining acute

(as opposed to chronic) medical conditions, where any homeopathic treatment

effect is likely to be swift acting, and thus better discernable. The fact that

prescribing is relatively simple (minimal range of indicated remedies) in some

acute conditions could offer useful advantages in study design. It may be

noteworthy that acute, rather than chronic, conditions feature prominently among

those for which homeopathy has the greatest weight of positive research evidence

in placebo-controlled trials.

Investigators might also consider testing the effects of homeopathy as an

adjunct to conventional medication, thus reflecting its complementary nature.

Equivalence trials offer another promising way forward. In these, a conclusion

of `similar clinical outcome' between homeopathy and an orthodox treatment would

be based on an ability to accept statistically equivalent confidence intervals

in the two groups of data.[31] This approach would be greatly preferable to

assuming equivalence based merely on failure to reject the Null hypothesis in a

typical superiority trial, as is the case in nearly all of the 41 examples

presented here. Formal equivalence trials could enable researchers to examine

more robustly the relative safety and cost-effectiveness of homeopathy compared

with a conventional medicine that was shown to possess similar treatment

effectiveness in a particular clinical situation.

It remains to be seen whether passage of time sees the publication of increased

numbers and quality of published research papers relevant to homeopathy that

report clinically significant findings in its favour. For its practitioners,

such robust research data would have an important impact in improving the

credibility as well as the intrinsic quality of the homeopathy that they provide

to their patients.

 

 

 

Acknowledgements

It is a pleasure to thank Conrad Harris, Peter Fisher and Bob Leckridge for

their helpful comments on the manuscript in the early stages of its preparation.

 

 

 

References

1. Mathie RT. Clinical outcomes research: contributions to the evidence base for

homeopathy. Homeopathy 2003; 92: 56–57.

2. Paterson J. Report on the mustard gas experiments (Glasgow and London). Br

Hom J 1943; 33: 1–12.

3. Ritter H. Ein homöotherapeutischer doppelter Blindversuch und seine

Problematik. Hippokrates 1966; 12: 472–476.

4. Kennedy CO. A controlled trial. Br Hom J 1971; 60: 120–127.

5. Ustianowski PA. A clinical trial of Staphysagria in postcoital cystitis. Br

Hom J 1974; 63: 276–277.

6. Linde K, Clausius N, Ramirez G, et al. Are the clinical effects of homeopathy

placebo effects? A meta-analysis of placebo-controlled trials. Lancet 1997; 350:

834–843.

7. NHS Centre for Reviews and Dissemination. Homeopathy. Effective Health Care

Bulletin 2002; 7(3): 1–12.

8. Harrison H, Fixsen A, Vickers A. A randomized comparison of homoeopathic and

standard care for the treatment of glue ear in children. Complement Ther Med

1999; 7: 132–135.

9. van Haselen RA, Fisher PAG. A randomized controlled trial comparing topical

piroxicam gel with a homeopathic gel in osteoarthritis of the knee. Rheumatology

2000; 39: 714–719.

10. Gmünder R, Kissling R. The efficacy of homeopathy in the treatment of

chronic low back pain compared to standardized physiotherapy. Z Orthop Ihre

Grenzgeb 2002; 140: 503–508.

11. Fisher P, Scott DL. A randomized controlled trial of homeopathy in

rheumatoid arthritis. Rheumatology 2001; 40: 1052–1055.

12. Kaziro GS. Metronidazole (Flagyl) and Arnica montana in the prevention of

post-surgical complications, a comparative placebo controlled clinical trial. Br

J Oral Maxillofac Surg 1984; 22: 42–49.

13. Reilly D. Randomised controlled trials for homoeopathy: when is useful

improvement a waste of time? Double positive paradox of negative trials. Br Med

J 2002; 325: 42.

14. Gassinger CA, Wunstel G, Netter P. A controlled clinical trial for testing

the efficacy of the homeopathic drug Eupatorium perfoliatum D2 in the treatment

of common cold. Arzneimittelforschung 1981; 31: 732–736.

15. Maiwald VL, Weinfurtner T, Mau J, Connert WD. Therapy of common cold with a

homeopathic combination preparation in comparison with acetylsalicylic acid. A

controlled, randomized double-blind study. Arzneimittelforschung 1988; 38:

578–582.

16. Lüdtke R, Wiesenauer M. A meta-analysis of homeopathic treatment of

pollinosis with Galphimia glauca. Wien Med Wochenschr 1997; 147: 323–327.

17. Barnes J, Resch K-L, Ernst E. Homeopathy for postoperative ileus? A

meta-analysis. J Clin Gastroenterol 1997; 25: 628–633.

18. Jonas WB, Linde K, Ramirez G. Homeopathy and rheumatic disease. Rheum Dis

Clin North Am 2000; 26: 117–123.

19. Linde K, Jobst KA. Homeopathy for chronic asthma. Cochrane Database Syst Rev

2000; 2: CD000353.

20. Vickers AJ, Smith C. Homoeopathic Oscillococcinum for preventing and

treating influenza and influenza-like syndrome. Cochrane Database Syst Rev 2000;

2: CD001957.

21. Ernst E, Barnes J. Are homoeopathic remedies effective

 

 

 

 

 

Article continues at site. Use clickable link at top.

 

 

Gettingwell- / Vitamins, Herbs, Aminos, etc.

 

To , e-mail to: Gettingwell-

Or, go to our group site: Gettingwell

 

 

 

SBC DSL - Now only $29.95 per month!