THE JOURNAL OF ALTERNATIVE AND COMPLEMENTARY MEDICINEVolume 7, Number 2, 2001, pp. 149–159Mary Ann Liebert, Inc.

Homeopathy and Conventional Medicine: An OutcomesStudy Comparing Effectiveness in a Primary

Care Setting

DAVID RILEY, M.D.,1,2 MICHAEL FISCHER, Ph.D.,3 BETSY SINGH, Ph.D.,4

MAX HAIDVOGL, U.D., Dr.Med.,5 and MARIANNE HEGER, M.D.6

ABSTRACT

Background: Recent meta-analyses of randomized controlled trials in homeopathy have sug-gested that homeopathy is more than a placebo response.

Objective: Comparison of the effectiveness of homeopathy in primary care with conventionalmedicine in primary care for three commonly encountered clinical conditions.

Design: An international multicenter, prospective, observational study in a real world med-ical setting comparing the effectiveness of homeopathy with conventional medicine.

Participants: Thirty (30) investigators with conventional medical licenses at six clinical sitesin four countries enrolled 500 consecutive patients with at least one of the following three com-plaints: (1) upper respiratory tract complaints including allergies; (2) lower respiratory tract com-plaints including allergies; or (3) ear complaints.

Main Outcome Measures: The primary outcomes criterion was the response to treatment, de-fined as cured or major improvement after 14 days of treatment. Secondary outcomes criteriawere: (1) rate of recovery; (2) occurrence of adverse events; (3) patient satisfaction; and (4) lengthof consultation.

Results: Four hundred and fifty-six (456) patient visits were compared: 281 received home-opathy, 175 received conventional medicine. The response to treatment as measured by the pri-mary outcomes criterion for patients receiving homeopathy was 82.6%, for conventional medi-cine it was 68%. Improvement in less than 1 day and in 1 to 3 days was noted in 67.3% of thegroup receiving homeopathy and in 56.6% of those receiving conventional medicine. The ad-verse events for those treated with conventional medicine was 22.3% versus 7.8% for those treatedwith homeopathy. Seventy-nine percent (79.0%) of patients treated with homeopathy were verysatisfied and 65.1% of patients treated with conventional medicine were very satisfied. In bothtreatment groups 60% of cases had consultations lasting between 5 and 15 minutes.

Conclusions: Homeopathy appeared to be at least as effective as conventional medical carein the treatment of patients with the three conditions studied.

149

1University of New Mexico Medical School, Albuquerque, New Mexico.2Integrative Medicine Institute, Santa Fe, New Mexico.3ClinResearch, Cologne, Germany.4Southern California University of the Health Sciences, Whittier, California.5Ludwig-Boltzmann Institute, Graz, Austria.6Research Center HomInt, Karlsruhe, Germany.

INTRODUCTION

Homeopathy, a medical therapy viewedwith skepticism by most medical doctors

and without a scientifically plausible mecha-nism of action theory, has persisted for morethan 200 years. Its use is growing today (Eisen-berg et al., 1998; Jacobs, 1998). The most re-cent of two (Kleijnen et al., 1991; Linde et al.,1997) meta-analyses on treatment with home-opathy was published in 1997 and suggestedthat the clinical effects of homeopathy weremore than a placebo response. The authors ofthis meta-analysis evaluated 89 randomizedcontrolled trials and found that homeopathywas more than twice as likely to be effectivethan placebo. No conclusions were drawnabout the effectiveness of homeopathy for spe-cific clinical conditions. They did howeversuggest using prospective observational stud-ies as a research tool, thereby separating thequestion of whether homeopathy is a usefultool in health care from whether or not it is aplacebo response. This prospective, observa-tional study was designed to evaluate the use-fulness of homeopathy in a real world clinicalsetting.

This was a prospective observational studyof the natural interaction between the partici-pating physicians and their patients. It was nota part of the design for these groups to be iden-tical, matched or comparable at baseline. Thiswas a real-world clinical setting over which wehad no control. Statistical analysis for baselinedifferences would not be appropriate. Wetested for the influences of baseline differenceson treatment outcomes in the adjusted odds ra-tio in the post hoc analysis.

The primary objective of the International In-tegrative Primary Care Outcomes Study (IIP-COS) was to use conventionally licensed healthcare providers and compare the effectivenessof homeopathy in primary care (Jacobs et al.,1998) with that of conventional medical treat-ment in primary care for three commonly seenclinical conditions. Secondary objectives stud-ied were: (1) the rate of recovery; (2) the oc-currence of adverse events; (3) patient satis-faction; and (4) the length of consultation(indirectly related to the cost of service).

METHODS

This was an international multicenter, pro-spective, observational study using 30 investi-gators with conventional medical licenses at sixclinical sites in four countries. We comparedthe effectiveness of homeopathy with conven-tional medicine in a real world primary caresetting. Patients were consecutively admittedinto the study and treated according to the bestmedical practice known to the practitioner andthere were no treatment restrictions placed onthe participating practitioners. The conditionsstudied were: (1) upper respiratory tract com-plaints including allergies: (2) lower respira-tory tract complaints including allergies: or (3)ear complaints.

Inclusion/exclusion criteria

Inclusion criteria were: (1) age older than 1month (irrespective of gender, ethnic origin, orsocioeconomic status); (2) one of the three clin-ical conditions mentioned above; (3) onset ofsymptoms for 0–48 hours or 48 hours–7 days;and (4) informed consent. Patients were ex-cluded from the study if they had histories ofpsychiatric disorders (psychosis, dementia,schizophrenia), spinal cord injury, stroke, renalfailure, liver disease, alcohol or drug abuse,current immunosuppressive treatment, chemo-therapy, or radiation treatment.

Setting

IIPCOS-1 was conducted between July 1996and August 1997 with 30 investigators at sixclinical sites in four countries. Three clinicalsites were in Europe (Berlin, Germany; Bern,Switzerland; Graz, Austria) and three in theUnited States (Albany, CA; Ashland, OR;Santa Fe, NM). All investigators had a con-ventional medical license. There were 24 med-ical doctors (M.D.), 4 physician’s assistants(P.A.), and 2 family nurse practitioners(F.N.P.). The investigators prescribing home-opathy had, in addition to their conventionalmedical qualifications, graduated from ahomeopathic training program and had atleast 5 years experience using homeopathy intheir medical practices.

RILEY ET AL.150

Initial screening

During a 3-month recruitment period (Knip-schild et al., 1991) at each site consecutive pa-tients were admitted to study if they had oneof the three clinical conditions and met the in-clusion/exclusion criteria. At the initial patientcontact each investigator conducted a routinemedical evaluation, obtained informed con-sent, and documented the following: demo-graphic information, concomitant medicalproblems, medications taken during the previ-ous 2 months, chief complaint, onset of symp-toms, clinical diagnosis using International Clas-sification of Diseases, 9th Revision (ICD-9) code,and investigator confidence in diagnosis (0–10scale). Data were collected on the primary treat-ment (homeopathy or conventional medicine),adjunctive therapies, length of consultation,and follow-up recommendations. Informationwas collected on the general health status of allpatients using the Health Status Questionnaire(HSQ-12) (Radosevich and Pruitt, 1995), an in-ternationally validated general health status in-strument. Each patient also completed a HealthComplaint Questionnaire (HCQ-5); a five-itemsubset of the HSQ-12 developed for this studyto measure the severity of their chief complaintat the time of entry into IIPCOS-1. All case re-port forms were translated and back-translatedbetween English and German to check for lan-guage consistency. Investigators were free tochoose any therapy for each patient.

Medications

Homeopathic medications were prepared ina 30C potency according to the German Ho-moeopathic Pharmacopoeia (HAB) and the Home-opathic Pharmacopoeia of the United States (HPUS)by DHU (Deutsche Hom‡opathie-Union) ofKarlsruhe, Germany. Random samples ofhomeopathic medications used in this studywere sent for independent analysis to check forthe presence of contamination; none wasfound.

Patient follow-up

At day 14 and day 28 independent telephoneinterviewers trained in telephone interviewingfor clinical research contacted each patient ask-

ing a series of questions: outcomes includingimprovement or deterioration, when improve-ment was noted, patient satisfaction, and ad-verse events. Data were also collected on com-pliance with prescribed treatment, use, orchanges in adjunctive therapies, and willing-ness to use the prescribed therapy again.Twenty percent (20%) of patients reportingcure on day 14 were contacted on day 28 to ver-ify the accuracy of their initial statements. Noin-person contact by the treating physician wasrequired after the initial patient encounter.

Outcomes criteria

The primary outcomes criterion was the re-sponse to treatment, which was defined ascured or major improvement after 14 days ac-cording to the Glasgow Homoeopathic Hospi-tal Outcome Score (GHHOS), a nine-point out-comes scale from 14 to 24 using sequentialquestions. Secondary outcomes criteria wererate of recovery, occurrence of adverse events,length of consultation, and patient satisfactionwith treatment using the Santa Fe Patient Sat-isfaction (SFPS) rating scale. This was a 12 to22 defined as follows: 12 5 very satisfied,11 5 somewhat satisfied, 0 5 neutral, 21 5somewhat dissatisfied, 22 5 very dissatisfied.The GHHOS and the SFPS scale have not beenvalidated.

Monitoring

Monitoring, including source data verifica-tion was performed by an independent clinicalmonitor according to Good Clinical Practice(GCP) guidelines at each site. The study wasapproved by the Freiburg International EthicsCommittee in Europe and by the Bastyr Uni-versity Institutional Review Board in theUnited States. It was conducted in accordancewith the Helsinki Declaration, GCP guidelines,and legal requirements in the participatingcountries.

Statistical analysis

Statistical analysis was conducted using uni-variate, bivariate, and multivariate statisticalmethods by the Institute for Numerical Statis-

HOMEOPATHY AND CONVENTIONAL MEDICINE 151

tics (IFNS) in Cologne, Germany. A total of 500patients were enrolled in the study. Forty-fourpatients (8.8%; 30 receiving homeopathic treat-ment and 14 receiving conventional treatment)were excluded from the statistical analysis.Forty-one (41) had no follow-up data and 3 hadnot met the inclusion/exclusion criteria. Fourhundred and fifty-six (456) patient outcomeswere suitable for comparison. Homeopathywas prescribed for 281 patients and 175 pa-tients received conventional medicine. Com-parisons between the effectiveness of home-opathy and conventional medicine regardingprimary and secondary outcomes criteria wereperformed using the two-sided Mann-WhitneyU test for rank ordered data and the two-sidedFisher’s exact test for dichotomous variables(no a adjustment was made due to the ex-ploratory character of the study). Response totreatment (the primary outcomes criterion was

analyzed according the following prespecifiedsubgroups: gender, age, whether or not the pa-tient was known to the practitioner, practicesetting, duration of consultation, clinical con-ditions, onset of symptoms, concomitant med-ical problems, adjunctive therapies for chiefcomplaint, and initial HSQ sum score. Unad-justed odds ratios, with 95% confidence inter-vals, were determined for the total sample andthe subgroups. In addition, multiple regressionanalysis was conducted to adjust for potentiallyconfounding factors affecting response tohomeopathic or conventional treatment.

RESULTS

Demographic data

The demographic data are presented in Table1. In the group of patients receiving homeopa-

RILEY ET AL.152

TABLE 1. DEMOGRAPHIC DATA

Homeopathic treatment Conventional treatmentn 5 281 n 5 175

GenderMale 91 (32.4%) 67 (38.3%)Female 190 (67.6%) 108 (61.7%)

Age,2 years 29 (10.3%) 5 (2.9%)02–11 years 109 (38.8%) 11 (6.3%)012–17 years 16 (5.7%)0 13 (7.4%)018–64 years 125 (44.5%) 136 (77.7%)$65 years 2 (0.7%)0 10 (5.7%)0

Duration of consultation,5 minutes 23 (8.2%)0 49 (28.0%)5–15 minutes 169 (60.1%) 115 (65.7%)16–30 minutes 71 (25.3%) 10 (5.7%)0.30 minutes 6 (2.1%)0 —

Geographic regionUnited States 143 (50.9%) 37 (21.1%)Europe 138 (49.1%) 138 (78.9%)

Chief complaintUpper respiratory complaints 148 (52.7%) 102 (58.3%)Lower respiratory complaints 103 (36.7%) 79 (45.1%)Ear complaints 50 (17.8%) 8 (4.6%)0

Onset of symptoms0–48 hours 114 (40.6%) 81 (46.3%)48 hours–7 days 163 (58.0%) 92 (52.6%)

Concomitant medical problemsYes 80 (28.5%) 59 (33.7%)No 193 (68.7%) 116 (66.3%)

Health Status Questionnaire(HSQ-12) sum score (X 6 SD) 24 6 7 24 6 7

HSQ, Health Status Questionnaire; SD, standard deviation.

thy 49% (138) of the patients were childrenyounger than 12 years versus 9% (16) in thosereceiving conventional medicine. Patients be-tween the ages of 18 and 64 years comprised45% (125) of the adults in the homeopathictreatment group and 78% (136) of the adults inthe conventional treatment group. Upper res-piratory tract complaints were the most com-mon (.50%) in both groups followed by lowerrespiratory tract complaints and ear com-plaints. The two groups appeared similar priorto treatment as measured by the following pa-rameters: gender, onset of symptoms (0–48hours or 48–72 hours), HCQ-5 sum score, HSQ-12 sum score, and the number of concomitantmedical problems.

Prescribed medications

The most commonly prescribed medicationsfor both treatment groups are presented inTable 2. Eleven homeopathic medications ac-counted for 71.1% of all prescriptions in thegroup of patients receiving homeopathy and70.9% of the patients treated with conventionalmedicine received antibiotics. Adjunctive ther-apies for the chief complaint were used in bothgroups. Of the group treated with homeopa-

thy, 49.5% received adjunctive therapies. Ofthese 23% used herbal treatments, primarilyechinacea. In this group treated with home-opathy, 16% received conventional treatment:cough and cold preparations, 5.7%; bron-chodilators, 2.5%; nasal sprays, 2.1%; antibi-otics, 1.8%; and analgesics, 1.4%. Acupuncturewas used as an adjunctive therapy in 2.8% ofpatients treated with homeopathy.

Of the group treated with conventional med-icine, 49.7% received adjunctive therapies. Ofthese, 40.6% used conventional treatments(analgesics, 22.3%; cough and cold prepara-tions, 8%; nasal sprays, 7.4%, etc.), 4.6% usedherbal treatment, and 3.4% used homeopathictreatment. Acupuncture was used as an ad-junctive therapy in 1.1% of patients in thegroup treated with conventional medicine.

Patient outcomes

The response to treatment defined as curedor major improvement after 14 days of treat-ment according to the GHHOS revealed thatpatients treated with conventional medicinenoted a 68% response to treatment whereasthe group treated with homeopathy noted an82.6% response to treatment. The difference

HOMEOPATHY AND CONVENTIONAL MEDICINE 153

TABLE 2. TREATMENTS PRESCRIBED

Homeopathic treatment, n 5 281, most frequently prescribed medicationsa

n %

Pulsatilla 35 12.5Hepar sulphuris 21 7.5Lycopodium 21 7.5Sulphur 20 7.1Belladonna 19 6.8Ferum phosphoricum 19 6.8Kali bichromicum 16 5.7Mercurius jodatus ruber 16 5.7Phosphorus 12 4.3Rhus toxicodendron 10 3.6Spongia 10 3.6

Conventional treatment, n 5 175, most frequently administered medications*

n %

Antibacterials for systemic use 124 70.9Cough and cold preparations 58 33.1Antiasthmatics 11 6.3Nasal preparations 11 6.3

aMultiple responses possible.

between these groups was significant (p 50.0058, two-sided Mann-Whitney U test; Table3). The unadjusted odds ratio was 2.23 (95%CI; 1.43–3.47) in favor of homeopathy. Ad-justment for age as a potential confoundingvariable in the logistic regression model re-duced the odds ratio to 2.07 (95% CI;1.24–3.47). Further adjustment for all prespec-ified covariables reduced the odds ratio to 1.96(95% CI; 1.04–3.70). Odds ratios in the pre-specified subgroups of response to treatmentdid not favor homeopathy in the followinggroups: patients younger than 2 years, pa-tients older than 65 years, and patients notpreviously known to the practitioner (Fig. 1).The calculation of odds ratios in the sub-

groups of patients less than 2 years of age andin patients older than 65 years were of limitedusefulness due to the small sample size in oneof the treatment groups. Odds ratios of re-sponse to treatment were essentially the samefor practitioners offering both homeopathicand conventional treatment as compared withpractitioners who utilized either only home-opathy or only conventional medicine.

The HCQ-5 sum score, used as an indicatorfor the severity of the chief complaint, de-creased similarly for both groups as expected(p 5 0.3534, two-sided Mann-Whitney U test;Table 3). The rate of recovery was differentamong treatment groups. Of the group receiv-ing homeopathy, 16.4 % improved in less than 1

RILEY ET AL.154

TABLE 3. PATIENT OUTCOMES AND SATISFACTION

Homeopathic treatment Conventional treatmentn 5 281 n 5 175

Health Complaint Questionnaire(HCQ-5) sum score (X 6 SD)

Initial contact 15 6 4 15 6 4After 14 days 8 6 3 9 6 4

p 5 0.3534 (Mann-Whitney-U-Test, two-sided)

Glasgow Homeopathic HospitalOutcomes Scale (GHOOS)

Cured, back to normal 175 (62.3%) 91 (52.0%)Major improvement 57 (20.3%) 28 (16.0%)Slight/moderate improvement 32 (11.4%) 42 (24.0%)No change 2 (0.2%) 5 (2.9%)0Deterioration 8 (2.8%) 4 (2.3%)0No remark 7 (2.5%) 5 (2.9%)0

p 5 0.0058 (Mann-Whitney-U-Test, two-sided)

How soon after the initialcontact was improvementnoted? 46 (16.4%) 10 (5.7%)0

Less than 1 day 143 (50.9%) 90 (51.4%)1–3 days 68 (24.2%) 47 (26.9%)4–7 days 11 (3.9%) 20 (11.4%)8–14 days 4 (1.4%) 4 (2.3%)0.14 days 9 (3.2%) 4 (2.3%)0No improvement/no remark

p 5 0.0011 (Mann-Whitney-U-Test, two-sided)

Santa Fe Patient Satisfaction(SFPS) Rating Scale

Very satisfied 222 (79.0%) 114 (65.1%)Somewhat satisfied 34 (12.1%) 39 (22.3%)Neutral 12 (4.3%) 13 (7.4%)0Somewhat dissatisfied 8 (2.8%) 3 (1.7%)0Very dissatisfied 3 (1.1%) 6 (3.4%)0No remark 2 (.7%)0 0 (0%)00.

p 5 0.0010 (Mann-Whitney-U-Test, two-sided)

day and 50.9% noted improvement in 1 to 3days. For the group treated with conventionalmedicine, 5.7% noted improvement in less than1 day and 51.4% first noted improvement be-

tween days 1 and 3. This difference betweentreatment groups for patients with improve-ment in less than 1 and in 1 to 3 days was sig-nificant and is consistent with the response to

HOMEOPATHY AND CONVENTIONAL MEDICINE 155

FIG. 1. Subgroup analysis. Odds ratio and 95% confidence intervals of response to treatment (cured or major im-provement after 14 days of treatment) in prospectively defined patient subgroups. Odds ratio below 1 indicate thatthe specific patient subgroup responded better to conventional treatment. Odds ratios above 1 indicate that the sub-group responded better to homeopathic treatment.

treatment at 14 days (p 5 0.0011, two-sidedMann-Whitney U test; Table 3).

Adverse events and patient satisfaction

Adverse events were significantly lower inthe group treated with homeopathy; 7.8% ver-sus 22.3% (p , 0.0001, Fisher’s two-sided exacttest; Table 4). The adverse events in the grouptreated with conventional medicine appear tobe side effects–related to antibiotic therapy.The adverse events in the group treated withhomeopathy were primarily headaches.

Patient satisfaction was high in both groups.Seventy-nine percent (79.0%) of patientstreated with homeopathy were very satisfiedand 12.1% were somewhat satisfied. Of pa-tients treated with conventional medicine,65.1% were very satisfied and 22.3% weresomewhat satisfied. This difference betweentreatment groups for those very satisfied wassignificant (p 5 0.0010, two-sided Mann-Whit-ney U test; Table 3). Further analysis of the cor-relation between adverse events and the re-sponse to treatment as well as patientsatisfaction revealed that as anticipated pa-tients without adverse events had a better re-sponse (p 5 0.0172, two-sided Mann-WhitneyU test) and reported a higher rate of satisfac-tion (p , 0.0001, two-sided Mann-Whitney Utest) than patients who experienced adverseevents.

Length of consultation

Both treatment groups had consultationslasting between 5 and 15 minutes in approxi-mately 60% of cases. Longer consultation timeswere more common in the homeopathic treat-ment group (25% versus 5%) and shorter con-sultation times were more common in the con-ventional treatment group (28% versus 8%).

DISCUSSION

This study illuminates aspects relating to theclinical practice of homeopathy and its use inprimary care. It is the first observational studyusing conventionally licensed health care prac-titioners to compare the effectiveness of home-opathy in primary care with that of conven-tional medicine. In this study, 84% of patientsin the group treated with homeopathy receivedno conventional medications, suggesting thathomeopathy is used as a stand-alone treatmentmodality in primary care, even by convention-ally licensed practitioners. We were surprisedthat 11 homeopathic medicines covered ap-proximately 70% of the prescriptions in thegroup treated with homeopathy. Despite theindividualized nature of homeopathic treat-ment, it appears that clinical pathways to a spe-cific prescription exist. This suggests thathomeopathy can be evaluated in clinical trials.

RILEY ET AL.156

TABLE 4. ADVERSE EVENTS

Homeopathic treatment Conventional treatmentn 5 281 n 5 175

# of Patients with AEs (%) 7.8 22.3thereof witha

Headache 2.1 4.0Diarrhea 0.0 5.1Fatigue 0.4 2.3Abdominal pain 0.4 2.3Nausea 0.0 2.9Dyspepsia 0.4 1.1Pruritus 0.0 1.1Rash 0.0 1.1Dizziness 0.0 1.1Somnolence 0.0 1.1Allergic reaction 0.0 1.1Gastrointestinal disorder 0.0 1.1

aMultiple responses possible.AEs, adverse events.

We also noted that homeopathic prescribing for acute illnesses was possible within the time constraints associated with a conventionalmedical consultation. These two points suggestthat homeopathy could be integrated into theprimary care setting.

The wide difference in the rate of adverseevents in this study between treatment groupswas significant (p , 0.0001) and of importancegiven the concern about drug safety and sideeffects. Nyquist and coworkers (1998) recentlydocumented the use of antibiotics in childrenfor colds (44%), upper respiratory tract infec-tions (46%), and bronchitis (75%), all condi-tions that typically do not respond to antibi-otics, all conditions seen in this study. A recentmeta-analysis of prospective studies evaluat-ing adverse drug reactions reported a high in-cidence of adverse events among hospitalizedpatients receiving conventional medicine(Lazarou et al., 1998). The potential side effectsof antibiotics could have been responsible forthe high rate of adverse events in the conven-tional treatment group and may have bluntedthe effectiveness of conventional medicine. Analternative explanation is that homeopathy iseffective for the three conditions seen in thisstudy.

Several factors could have contributed to thehigher response rate in the group treated withhomeopathy. Patient satisfaction or dissatis-faction with treatment appears to be related tothe growing interest in alternative medicine(Campion, 1993; Sutherland and Verhoef, 1994;Hentschel et al., 1996). A survey in the UnitedKingdom found that patients consulting aphysician using homeopathy cited the fact thatthe physician incorporated homeopathy intotheir medical practice as the main reason forseeing that physician (Vincent and Furnham,1996). Another factor in this survey was the be-lief that complementary therapies would be ef-fective in treating their complaints and that pa-tients appreciated being treated as a wholeperson and playing an active role in their ownhealth. This survey also documented that theperceived side effects of conventional medicineas well as communication between patient andphysician were important to patients in theirchoice of practitioners. Astin (1998) recently

noted that “the majority of alternative medicineusers appear to be doing so because they findcomplementary and alternative medical thera-pies to be more congruent with their own val-ues, belief, and philosophical orientations to-ward health and life.”

Other reasons may favor success with con-ventional medicine. Homeopathic philosophytends to emphasize self-healing and patientsusing homeopathy appear to regard their prac-titioners and the treatments they prescribe ashaving less influence on their health than pa-tients treated with conventional medicine(Kaiser, 1997). In addition homeopathy claimsmore success in the treatment of chronic as op-posed to acute illness.

A study with this level of complexity has lim-itations. One of the potentially confoundingvariables was that investigators could enrollpatients in a particular treatment group. Otherpotentially confounding factors such as age ofthe patients differed considerably between thetreatment groups; these differences may be ex-pected in a nonrandomized trial. However,baseline difference did not affect the outcomesas demonstrated by the fact that the adjustedodds ratios showed that these potentially con-founding variables had no significant effect onthe overall results of the study.

It is challenging to evaluate the treatment ofacute illnesses where it may be difficult to dis-tinguish among spontaneous recovery and theresults of a therapeutic intervention. Expecta-tion bias is another potentially confoundingvariable. Patients often try please their physi-cians (Bischoff and Zeitler, 1997) and practi-tioners may overestimate the effect of treat-ment (König and Nemeth, 1994). We usedindependent telephone interviewers in orderto minimize expectation bias on the part of thepatient and the practitioner (Jhni, 1994). Fi-nally practice pattern variations are a con-founding variable in all of primary care andcomplicate clinical research, particularly in areal world medical setting (Horton, 1999).Even with these limitations we believe that thepractice of medicine in the United States andEurope has many things in common, not theleast of which is the growing use of alterna-tive medical therapies.

HOMEOPATHY AND CONVENTIONAL MEDICINE 157

CONCLUSION

Regardless of evidence about the safety ofhomeopathy or its potential usefulness in pri-mary care, the lack of a testable mechanism ofaction theory contributes to the scientific skep-ticism about this medical therapy (Sampson,1995) and the controversial nature of these re-sults (Horton, 1998; Belon et al., 1999). Never-theless, an estimated 75% of the world’s popu-lation use alternative medicine and the use ofhomeopathy is growing in the United States(Eisenberg et al., 1998) and Europe. There maybe aspects of the way homeopathy is used inprimary care that offers insights into what itmeans to be an effective practitioner in this set-ting. Its usefulness should continue to be eval-uated with further practice-based research; ex-ploring issues such as practitioner preference,and prescribing patterns (including dosing fre-quency). We believe that the information fromthis study will improve clinical trial design inthe real world medical setting and offer a bet-ter understanding of how homeopathy andother alternative medical therapies are and canbe integrated into the practice of medicine.

CONTRIBUTORS

D. Riley conceived and coplanned the study,was a principal investigator, supervised thedata collection, participated in the interpreta-tion of the data analysis, wrote the first draft ofthe manuscript, participated in all of the man-uscript revisions, and wrote the final version ofthe manuscript. M. Heger coplanned the study,was a principal investigator, supervised thedata collection, participated in the interpreta-tion of the data analysis, participated in themanuscript revisions, and approved the finalversion of the manuscript. M. Fischer inCologne, Germany, was responsible for thedata extraction and statistical analysis, wrotethe statistical analysis section, and approvedthe final draft of the manuscript. B. Singh par-ticipated in the data analysis and interpreta-tion, participated in the manuscript revisions,and approved the final version of the manu-script. M. Haidvogl supervised the data collec-tion in Austria, participated in the interpreta-

tion of the data analysis, participated in themanuscript revisions, and approved the finalversion of the manuscript.

ACKNOWLEDGMENTS

This study was funded by a grant fromHomInt in Karlsruhe, Germany. We would liketo thank Carlo Calabrese, N.D. for his adviceon outcome studies and Roger A. Edwards forhis advice during the initial planning phase.We would like to thank Helga Sager, Ann Seipt,Gerda Kempf, and Wolfgang Mayer for theirtechnical support.

We would like to acknowledge the investi-gators for their participation in this study: IraBerkowitz, Mary Beth Burton, Peggy Chipkin,Christine Ciavarella, P. Klaus Connert, MaryAlice Cooper, Marsha Cummings, LucindaDykes, Walter Fiala, Deborah Gordon, NancyHerrick, Richard Jenkins, Kathryn Keith, Jo-hannes Klement, Alexander Kozel, GloriaKozel, Elisabeth Krainer, Roger Morrison,Monika Mutici, Willi Neuhold, Susanne Prgger, Alan Rogers, Anton Rohrer, JonathanShore, AndrJ Thurneysen, Oskar Vogel, UlrichWolschner, and Aimee Zagon.

REFERENCES

Astin JA. Why patients use alternative medicine: Resultsof a national study. JAMA 1998;279:1548–1553.

Belon P, Cumps J, Ennis M, Mannaioni PF, Sainte-LaudyJ, Roberfroid M, Wiegant FAC. Inhibition of human ba-sophil degranulation by successive histamine dilutions:Results of a European multi-centre trial. InflammationRes 1999;48 (suppl. I):17–18.

Bischoff F, Zeitler H. Laienkonzepte, Krankenbilder undTherapieerwartungen von Patienten. Z Allg Med1997;73:1019–1023.

Campion E. Why unconventional medicine? N Engl J Med1993;328:282–283.

Eisenberg D, Davis R, Ettner S, Appel S, Wilkey S,Rompay M, Kessler R. Trends in alternative medicineuse in the United States, 1990–1997. JAMA 1998;280:1569–1575.

Hentschel C, Kohnen R, Hauser G, Lindner M, Ernst E,Hahn E. Entscheidung zur Komplementärmedizin: Sa-chorientiert oder irrational? Dtsch Med Wschr1996;121:1553–1560.

Horton R. The last battle: Medicine’s struggle with com-plementary therapy. Integrative Med 1998;1:35–38.

RILEY ET AL.158

Horton R. Evidence and primary care. Lancet 1999;353:609–610.

Jacobs J. Homeopathy should be integrated into main-stream medicine. Altern Ther Health Med 1995;1(4):48–53.

Jacobs J, Chapman E, Crothers D. Patient characteristicsand practice patterns of physicians using homeopathy.Arch Fam Med 1998;7:537–540.

Jüni P. Randomisierte kontrollierte Studien in der Kom-plement,rmedizin. Forsch Komplement,rmedizin 1994;1:166–176.

Kaiser S. Vergleich von Patienten der Schulmedizin undder Hom‡opathie unter der Verwendung psychologis-cher Konstrukte. Braunschweig, Germany: Diplomar-beit, Institut für Psychologie der Technischen Univer-sit,t, 1997.

König P, Nemeth G. Documentation in homoeopathicpractice. Homoeopathic Links 1998;2:102–106.

Kleijnen J, Knipschild P, ter Riet, G. Clinical trials of ho-moeopathy. BMJ 1991;302: 316–323.

Knipschild P, Leffers P, Feinstein AR. The qualificationperiod. J Clin Epidemiol 1991;44:461–464.

Lazarou J, Pomeranz B, Corey P. Incidence of adversedrug reactions in hospitalized patients: A meta-analy-sis of prospective studies. JAMA 1998;279:1200–1205.

Linde K, Clausius N, Ramirez G, Melchart D, Eitel F,Hedges L, Jonas W. Are the clinical effects of homoeop-

athy placebo effects? A meta-analysis of placebo-con-trolled trials. Lancet 1997;350:834–843.

Nyquist A, Gonsales R, Steiner J, Sande M. Antibiotic pre-scribing for children with upper respiratory tract in-fections, and bronchitis. JAMA 1998;279:875–877.

Radosevich D, Pruitt M. Twelve-Item Health Status Ques-tionnaire: HSQ-12 Verion 2.0. Minneapolis: Health Out-comes Institute, 1995.

Sampson W. Homeopathy does not work. Altern TherHealth Med 1995;1(4):48–53.

Sutherland L, Verhoef M. Why so many patients seek asecond opinion or alternative medicine? J Clin Gas-troenterol 1994;19:194–197.

Vincent C, Furnham A. Why do patients turn to comple-mentary medicine? An empirical study. Br J Clin Psy-chology 1996;35:37–48.

Address reprint reqests to:David Riley M.D.

Integrative Medicine InstituteP.O. Box 4310,

Santa Fe, NM 87502

E-mail: dsriley@integrativemed.org

HOMEOPATHY AND CONVENTIONAL MEDICINE 159

This article has been cited by:

1. Iris R. Bell. 2005. All Evidence Is Equal, but Some Evidence Is More Equal than Others: Can Logic Prevailover Emotion in the Homeopathy Debate?. Journal of Alternative and Complementary Medicine 11:5, 763-769.[Citation] [PDF] [PDF Plus]

2. Harald Walach, Wayne B. Jonas, John Ives, Roel Van Wijk, Otto Weingärtner, Dr.Phil.Nat.. 2005. Research onHomeopathy: State of the Art. Journal of Alternative and Complementary Medicine 11:5, 813-829. [Abstract][PDF] [PDF Plus]

3. Iris R. Bell, Daniel A. Lewis II, Gary E. Schwartz, Sabrina E. Lewis, Opher Caspi, Anne Scott, AudreyJ. Brooks, Carol M. Baldwin. 2004. Electroencephalographic Cordance Patterns Distinguish ExceptionalClinical Responders with Fibromyalgia to Individualized Homeopathic Medicines. Journal of Alternative andComplementary Medicine 10:2, 285-299. [Abstract] [PDF] [PDF Plus]

4. Iris R. Bell, Mary Koithan, Margaret M. Gorman, Carol M. Baldwin. 2003. Homeopathic Practitioner Viewsof Changes in Patients Undergoing Constitutional Treatment for Chronic Disease. Journal of Alternative andComplementary Medicine 9:1, 39-50. [Abstract] [PDF] [PDF Plus]

5. E. Ernst. 2002. A systematic review of systematic reviews of homeopathy. British Journal of ClinicalPharmacology 54:6, 577. [CrossRef]