The purpose of this paper is to examine large, general outcomes studies in homeopathy, to consider the design parameters researchers utilized to gather homeopathic treatment outcomes data and to compare the results. A literature search was carried out using online libraries and databases and a selection of outcomes studies were chosen with geographical diversity in mind. A number of European studies were chosen and one international study that included the USA. Observational studies with a small patient sampling (less than 500) were excluded, as were studies before 1999. Studies were chosen to reflect a range of health care situations (hospitals/medical clinics/private/NHS/non-medical etc.), as well as a range of data collection and analysis methods (methodologies). An Excel spreadsheet was used to view and compare key data (see Appendix 2). The homeopathic outcomes studies (studied in this paper) represent respectable endeavors to clarify the effects of homeopathic care, to reduce complex effects in multi-dimensional subjects to quantifiable, reproducible and/or constructive data, and also to provide hope that there is a way to make research in homeopathy a respectable way to further the profession.
Outcomes study, outcomes research, homeopathy, large cohort, large sample, complementary and alternative medicine, outcomes methodologies
The purposeof this paper is to examine eleven large, general outcomes studies in homeopathy, to consider the design parameters researchers utilized to gather homeopathic treatment outcomes data and to compare the results.
While subjective data is what matters to patients and is what informs their health care choices (Owen et al. 2007), objective data driven research has become the gold standard of researchers and conventional medical clinicians.
Outcomes studies are mostly used to assess treatment efficacy in general based on:
- a reduction or disappearance of symptoms (or not);
- patient satisfaction in general (or lack thereof);
- a patient’s sense of well-being or QoL (Quality of Life);
- miscellaneous other factors such as age and sex demographics and cost effectiveness, changes in medication needs etc.
In recent years homeopathic outcomes studies have come under repeated attack by scientists espousing evidence-based medicine and clinical trials as the ‘gold standard’ of medical (conventional and CAM) research. “Public policy decision-makers and third party payers are likely to view an outcome such as a greater sense of well-being as unhelpful or insufficient in deciding whether or not to pay for homeopathic treatment for all patients with a given conventional diagnosis” (Owen et al. 2007)
Homeopathy, along with many other CAM therapies, is resistant to controlled trials for numerous reasons including:
- the individualized nature of the therapy itself i.e. more unusual for patients with similar symptoms to receive the same treatments;
- the healing influences (including empathy) that occur in the relationship;
- the fact that multiple healing interventions are often applied. (Lewith et al. 2002).
I have a special interest in outcomes studies having conducted an annual audit of my own (private) practice for twenty-two years in order to quantify my results, to identify weaknesses or areas needing improvement and to track the results of those changes. Ten years ago I designed a software program to carry out outcomes studies automatically. In all these years of examining my own practice I have only made a cursory study of the more formal world of observational studies, so I have used this paper to become more acquainted with the territory.
A literature search was carried out using online libraries and databases including GoogleScholar, AMED, Embase, Ovid, Medline (PubMed), BioMed Central, Wiley InterScience, and Science Direct. The following search strings and words were used: “outcomes study” or “observational study” and “CAM” (or complementary and alternative medicine) and “homeopathy” (or homoeopathy) and “large sample” (or cohort or sampling).
A selection of outcomes studies were chosen with geographical diversity in mind. Observational studies with a small patient sampling (less than 500) were excluded, as were studies before 1999. I looked for studies that would reflect a range of health care situations (hospitals/medical clinics/private /NHS/non-medical etc.), and a range of data collection and analysis methodologies.
I excluded two large studies (Sevar 2005) because only one practitioner was involved, there was not enough information about the design of the study, it was not clear how the data was collected, nor did the paper provide an explanation of why a second (similar) study was carried out five years after the first one.
An Excel spreadsheet was used to view and compare key data (see Appendix 2).
These were large scale studies that were inclusive of all patients (regardless of complaint) i.e. there were few to no exclusions or restrictions. They mostly examined patients with chronic diseases in private or hospital out- patient practices. They were carried out within the conventional and homeopathic medical community in: hospitals, clinics and private practices. Eight out of eleven studies involved homeopathic medical doctors, one involved homeopathic dentists, one homeopathic veterinarians and one study involved professional (non doctor) homeopaths (Steinsbekk & Lüdtke 2005) (Steinbekk et al, 2005). Objective measures (reductions in conventional meds/laboratory tests and other measures) were not evaluated.
Outcomes research collects data from a clinical setting and the challenge for clinicians and researchers is to collect data that is valid and verifiable. Because most homeopathic outcomes studies have not included a “concurrent (randomized) control group these kinds of studies can merely point to a possible association between treatment and outcome.” (Mathie 2003)
I chose eleven studies (see Appendix 2 for details), mostly from European countries with a preponderance of UK studies plus one international study involving the USA. I was surprised not to find more studies from North America and could find none from South America, Canada, Australia or New Zealand.
Table 1: Studies by country
|1: Witt, 2008|
|2: Spence, 2005|
|3: Steinbekk, 2005|
|4: Wassenhoven, 2004|
|5: Attena, 2000|
|6: Guthlin, 2004|
|7: Mathie, 2006|
|8: Riley, 2001|
|9: Rossi, 2009|
|10: Mathie, 2007 (dental)|
|11: Mathie, 2007 (animals)|
All studies collected various demographical data including age and sex. Most studies included the percentage of female patients and various ages (although each study chose different age groupings making it impossible to summarize this information.) Some studies included the average age of patients. One study (Attena et al. 2000) included the educative levels of patients and one (Witt et al. 2008) mentioned that the majority of patients were “highly educated female adults … below the age of 60.”
|1: Witt, 2008||73%||Adults: 48 yrsChildren: 14 yrs|
|2: Spence, 2005||64%||n/a|
|3: Steinbekk, 2005||63%||30 yrs|
|4: Wassenhoven, 2004||67%||39 yrs 8 mo|
|5: Attena, 2000||64%||n/a|
|6: Guthlin, 2004||55%||30 yrs|
|7: Mathie, 2006||71%||41 yrs 6 mo|
|8: Riley, 2001||68%||n/a|
|9: Rossi, 2009||66%||n/a|
|10: Mathie, 2007 (dental)||63.4%||42.6 yrs|
|11: Mathie, 2007 (animals)||n/a||n/a|
A variety of outcomes or ratings scales are being used in observational studies. Many studies create or develop a new scale or adapt a scale to fit their own outcomes objectives. The different measures or scales used in the observational studies selected for this paper made comparisons difficult (see Table 2 and Appendix 1).
Table 2: Outcomes/Ratings Scales
|1: Witt, 2008|
|2: Spence, 2005|
|3: Steinbekk, 2005|
|4: Wassenhoven, 2004|
|5: Attena, 2000|
|6: Guthlin, 2004|
|7: Mathie, 2006|
|8: Riley, 2001|
|9: Rossi, 2009|
|10: Mathie, 2007 (dental)|
|11: Mathie, 2007 (animals)|
* CS Complaint severity
The studies mostly used scales that involved patient’s subjective evaluations – whether patients felt generally better or not (Quality of Life measures), whether their symptoms (physical or psychological) were better or not (Complaint Severity/Improvement), and whether they were generally satisfied with homeopathic treatment or not. Some studies compared practitioners’ evaluations with patients. Studies that independently compared practitioners’ ratings with patients’ found them to be similar.
Most studies included information about whether data was patient- (with or without the practitioner present) and/or practitioner-generated, collected by an independent researcher and how (in the office or taken home and returned by mail etc.)
|Patient||Practitioner||Patient with Practitioner||Independent Assessor|
|1: Witt, 2008||X|
|2: Spence, 2005||X|
|3: Steinbekk, 2005||X|
|4: Wassenhoven, 2004||X||X|
|5: Attena, 2000||X|
|6: Guthlin, 2004||X||X||X (insurance company)|
|7: Mathie, 2006||X|
|8: Riley, 2001||X|
|9: Rossi, 2009 *|
|10: Mathie, 2007 (dental)||X|
|11: Mathie, 2007 (animals)||X|
* Study did not describe exactly how the data was gathered
Ethics committee approval
Four out of eleven studies stated an ethics committee had been approached for approval: Witt et al. (2008), Mathie et al. (2007, 2 studies), and Riley et al. (2001)
Two studies listed the remedies given and included those most commonly prescribed (Riley et al. 2001 and Wassenhoven and Ives, 2004). One study (Attena et al. 2000) stated pluralist homeopathics had been prescribed. It is not clear whether homeopaths in the other studies were classical prescribers (administering one remedy at a time) or pluralist prescribers.
One study (Wassenhoven and Ives, 2004) mentioned whether one or more remedies had been prescribed. A total of 333 remedies were prescribed with 21 accounting for more than half the prescriptions. Prescribing strategies and percentages were shown with most prescriptions based on totality of symptoms.
In one study (Attena et al. 2000) the researchers ask patients whether the ‘therapy had been carried out either properly, not properly or not at all.’ Interestingly 14.3% had not carried out the therapy properly (taken their homeopathic medicines) and 2.9% hadn’t started the therapy at all. Curiously, homeopaths assume their patients take their medicines as prescribed when more than 50% of allopathic patients apparently do not do so. (Greenberg 1984).
General Improvement Responses
70.7% – 90.1% of patients, including the veterinarians study and the (mostly acute) dental study, experienced improvements: either some improvement or significant improvements (feeling better or much better). Between 5% and 23.1% experienced no changes and between 0.0% and 6.1% were worse.
Younger patients experienced more improvements in two studies (Rossi et al. 2009) & (Witt et al. 2005) & (Spence et al. 2005), and longer treatment times created improved outcomes in one study (Rossi et al. 2009)
A lack of follow up data was reported in some studies. This figure tended to be more common in the longer term studies where patients had moved or died etc. This figure ranged from 0.0% to 4.8%.
Adverse events were minimal to none in most studies, but 7% of homeopathy patients in Guthlin et al’s study (2004) reported adverse events. These were not investigated to find out whether they were an initial aggravation, a return of an old symptom, an unintentional proving or something else. 5% of the acupuncture patients in the same study experienced adverse events but many of these were minor reactions like needle sensations. The study did not make mention of the disparity in numbers: 5% of 5000 acupuncture patients = 250 whereas 7% of 900 homeopathic patients = 63.
Other Healing Factors
Most studies appeared to make the tacit assumption that any improvements were due to the homeopathic treatment, however long a period. Only one asked at follow-up whether patients had experienced any other healing factors, and none asked the patients whether those other factors could have accounted for the improvements they had experienced (Riley?)
One Study (Witt et al. 2005) mentioned that 41% consulted another CAM therapist during the study period (not a homeopath) and 40% were treated with conventional medicine. We don’t know how many of those patients overlapped. Interestingly patients who used other treatments had a “smaller chance to improve relevantly.”
(Güthlin et al. 2004) discussed the difficulties of identifying the source of therapeutic effects as it is almost impossible to screen out the following other healing influences including:
- Patient’s expectations in general
- Patient’s expectations because of the alleged effects of a particular treatment
- The setting where consultations are performed
- Non-specific healing elements in the homeopath/patient relationship
- Non-specific healing effects of any (overt or inadvertent) therapeutic rituals
Cost Effectiveness and Reduction of Drugs
Wassenhoven & Ives, 2004 assessed patients’ discontinuation of conventional drug treatments (52% discontinued). Consultation and prescription costs were measured as well as patient satisfaction and improvements.
Key points from each study’s conclusion. The three studies involving Mathie all came to the same conclusion.
|1: Witt, 2008||Patients treated homeopathically improved considerably and improvements persisted for up to 8 yrs.|
|2: Spence, 2005||An important strand of evidence in favor of the effectiveness of homeopathy in the management of a wide range of chronic diseases.|
|3: Steinbekk, 2005||Patients using conventional medications reduced them from 39% to 16%.|
|4: Wassenhoven, 2004||Costs were significantly lower: 52% of patients discontinued their prescribed drugs.|
|5: Attena, 2000||Improvements were higher in those who were satisfied with the care and in those who completed therapy.|
|6: Guthlin, 2004||Acupuncture & homeopathy seem to be effective practices that satisfy patients and help the majority to a sufficient degree.|
|7: Mathie, 2006||Systematic recording of clinical data in homeopathy is feasible and capable of informing future research.|
|8: Riley, 2001||Homeopathy appeared to be at least as effective as conventional medical care in the treatment of patients with the three conditions studied.|
|9: Rossi, 2009||Homeopathic therapy is associated with improvement in a range ofchronic and recurring pathologies. Certain characteristics of patient and pathology influence the outcome.|
|10: Mathie, 2007 (dental)||A systematic recording of clinical data in homeopathy is both feasible and capable of informing future research|
|11: Mathie, 2007 (animals)||A systematic recording of clinical data in homeopathy is both feasible and capable of informing future research|
Clinical outcomes studies are a valuable contribution in the field, adding to the evidence base without the demand of an experimental trial and although some large scale studies have taken place, they have not been carried out with further research studies in mind (Haselen and Fisher, 2004). They have tended to ask basic questions about whether patients were getting better and if so what percentage are improving and by how much. They have tended not to raise questions for future research studies to take observational research to the next step.
The homeopathic outcomes studies (studied in this paper) represent respectable endeavors to clarify the effects of homeopathic care, to reduce complex effects in multi-dimensional subjects to quantifiable, reproducible and/or constructive data, and also to provide hope that there is way to make research in homeopathy a respectable way to further the profession.
There’s a huge gap between homeopathic practitioners’ and their patients perception of the value of homeopathy and some scientists’ refusal to accept homeopathy as anything other than quackery or worse. There’s a bigger gap between the millions of patients who use homeopathy – as home-prescribers and as patients in clinical settings (hospitals and clinics) – and the refusal of a hard core group of doctors and scientists who have written homeopathy off as an elaborate scam or worse. Positive research studies such as those listed by the European Network of Homeopathic Researchers (2005) are viewed in a poor light by those seeking to discredit homeopathy, but well-designed observational studies could pave the way for better acceptance, especially those with case controls or cohort design – and also if combined with economic research on costs and feasibility.(Lewith et al. 2002) suggests that patient input regarding the design of a research study in the planning stages might “radically affect” the design and interpretation of a study.
Although observational methods are deemed inferior to RCT designs, recent findings show that studying outcomes using cohort or case-control designs (i.e. without placebo or randomization) does not in fact lead to (the presumed) overestimation of treatment effect. Such designs are also less prone to creating paradoxical conclusions than RCTs. (T. D. Thompson 2004)
I learned a lot in the process of getting more acquainted with this territory, although it is fraught with a tremendous amount of information and data, frustratingly more than I had time to get to grips with. Homeopathic practitioners – myself mostly included – appear to experience encouraging outcomes in their clinics and it is encouraging to see similar trends (in terms of percentages) confirmed in these large scale observational studies. For homeopathic clinicians “the songbird of homeopathy sings often enough to keep us engaged in a demanding discipline.” (Trevor Thompson, 2004)
Attena, F. et al., 2000. Homoeopathy in primary care: self-reported change in health status. Complementary Therapies in Medicine, 8(1), 21–25.
Bell, I.R. et al., 2004a. Development and validation of a new global well-being outcomes rating scale for integrative medicine research. BMC Complementary and Alternative Medicine, 4(1), 1.
Bell, I.R. et al., 2004b. Development and validation of a new global well-being outcomes rating scale for integrative medicine research. BMC Complementary and Alternative Medicine, 4(1), 1.
Claudia, W. et al., How healthy are chronically ill patients after eight years of homeopathic treatment?–Results from a long term observational study. BMC Public Health, 8.
Concato, J., Shah, N. & Horwitz, R.I., 2000. Randomized, controlled trials, observational studies, and the hierarchy of research designs,
Greenberg, R., 1984. Overview of patient compliance with medication dosing: a literature review. Clinical therapeutics [0149-2918], 6(5), 592 -9.
Güthlin, C., Lange, O. & Walach, H., 2004. Measuring the effects of acupuncture and homoeopathy in general practice: an uncontrolled prospective documentation approach. BMC Public Health, 4(1), 6.
Heger, M. & Haidvogl, D.S.M., 2000. International integrative primary care outcomes study (IIPCOS-2): an international research project of homeopathy in primary care. British Homoeopathic Journal, 89, 10–13.
Jonas, W.B. et al., 2001. A systematic review of the quality of homeopathic clinical trials. BMC Complementary and Alternative Medicine, 1(1), 12.
Lewith, G., Jonas, W.B. & Walach, H., 2002. Clinical Research in Complementary Therapies, London: Harcourt Publishers. Available at: [Accessed August 8, 2009].
Mathie, R.T., 2003. Clinical outcomes research: contributions to the evidence base for homeopathy. Homeopathy, 92(1), 56–57.
Mathie, R.T. & Farrer, S., 2007. Outcomes from homeopathic prescribing in dental practice: a prospective, research-targeted, pilot study. Homeopathy, 96(2), 74–81.
Mathie, R.T. et al., 2007. Veterinary: Outcomes from homeopathic prescribing in veterinary practice: a prospective, research-targeted, pilot study. Homeopathy, 96(1), 27–34.
Mathie, R.T. & Robinson, T.W., 2006. Outcomes from homeopathic prescribing in medical practice: a prospective, research-targeted, pilot study. Homeopathy, 95(4), 199–205.
Owen, D., Leckridge, B. & Fisher, P., 2007. Principles and Practice of Homeopathy, Elsevier Health Sciences. Available at: [Accessed August 9, 2009].
Riley, D. et al., 2001. Homeopathy and conventional medicine: an outcomes study comparing effectiveness in a primary care setting. The Journal of Alternative & Complementary Medicine, 7(2), 149–159.
Sevar, R., 2005. Audit of outcome in 455 consecutive patients treated with homeopathic medicines. Homeopathy, 94(4), 215–221.
Sevar, R., 2000. Audit of outcome in 829 consecutive patients treated with homeopathic medicines. British Homeopathic Journal, 89(4), 178–187.
Spence, D.S., Thompson, E.A. & Barron, S.J., 2005. Homeopathic treatment for chronic disease: a 6-year, university-hospital outpatient observational study. Journal of Alternative & Complementary Medicine, 11(5), 793–798.
Steinsbekk, A. & Lüdtke, R., 2005. Patients’assessments of the effectiveness of homeopathic care in Norway: a prospective observational multicentre outcome study. Homeopathy, 94(1), 10–16.
Thompson, T.D., 2004. Can the caged bird sing? Reflections on the application of qualitative research methods to case study design in homeopathic medicine. BMC Medical Research Methodology, 4(1), 4.
Van Wassenhoven, M. & Ives, G., 2004. An observational study of patients receiving homeopathic treatment. Homeopathy, 93(1), 3–11.
Witt, C.M. et al., 2005. Homeopathic medical practice: long-term results of a cohort study with 3981 patients. BMC Public Health, 5(1), 115.
Outcomes/Ratings Scales (used in the studies listed in this paper)
Source: OUTCOMES DATABASE: http://www.outcomesdatabase.org/
AIOS: Arizona Integrative Outcomes Scale (Bell et al. 2004)
Patients use a visual analog scale (VAS) – basically a 4” or 10 cm line – to self-rate their sense of well-being: spiritual, social, mental, emotional and physical (and spiritual?).
CD-RISC: Connor-Davidson Resilience Scale (Connor et al 2003)
The Connor-Davidson Resilience Scale (CD-RISC) consists of 24-items that measure the ability to cope with stress and adversity. The items reflect several aspects of resilience that include a sense of personal competence, tolerance of negative affect, positive acceptance of change, trust in one’s instincts, sense of social support, spiritual faith, and an action-oriented approach to problem solving. A shortened 2-item version is also available.
GHHS: Glasgow Homeopathic Hospital Scale (Reilly 1997)
Not listed at the Outcomes Database. Rates (positive or negative) changes in overall condition.
GHRS: Global health rating scale (Idler & Benyamini 1997)
Not listed at the Outcomes Database. Physical health rating scale.
MYMOP: Measure Your Medical Outcome Profile (Paterson 1996)
This is an individualized measure that aims to measure three or more symptoms or complaints the patient considers the most important, recorded in their own words. Symptoms are rated on a scale of 1-6 (at baseline) and 1-7 (at follow up)
HRQL: Health Related Quality of Life Scale (Patrick et al 1988)
Not listed at the Outcomes Database. Rates health-related quality of, or satisfaction in life i.e. not limited to (physical) health.
SF-36: Short-Form Health Survey (Ware et al, 1994)
This is a 36-item multi-dimensional questionnaire that assesses 8 dimensions of health status: physical functioning; role limitations due to physical health problems; bodily pain; social functioning; general mental health; role limitations due to emotional problems; vitality, energy or fatigue; and general health perceptions. (Ware 1990). The SF-12 is a shortened version.