Overview and quality assessment of health economic evaluations for homeopathic therapy: an updated systematic review

ABSTRACT

Introduction

Likewise other medical interventions, economic evaluations of homeopathy contribute to the evidence base of therapeutic concepts and are needed for socioeconomic decision-making. A 2013 review was updated and extended to gain a current overview.

Methods

A systematic literature search of the terms ‘cost’ and ‘homeopathy’ from January 2012 to July 2022 was performed in electronic databases. Two independent reviewers checked records, extracted data, and assessed study quality using the Consensus on Health Economic Criteria (CHEC) list.

Results

Six studies were added to 15 from the previous review. Synthesizing both health outcomes and costs showed homeopathic treatment being at least equally effective for less or similar costs than control in 14 of 21 studies. Three found improved outcomes at higher costs, two of which showed cost-effectiveness for homeopathy by incremental analysis. One found similar results and three similar outcomes at higher costs for homeopathy. CHEC values ranged between two and 16, with studies before 2009 having lower values (Mean ± SD: 6.7 ± 3.4) than newer studies (9.4 ± 4.3).

Conclusion

Although results of the CHEC assessment show a positive chronological development, the favorable cost-effectiveness of homeopathic treatments seen in a small number of high-quality studies is undercut by too many examples of methodologically poor research.

Plain Language Summary

To help make decisions about homeopathy in healthcare, it is important, as with other medical treatments, to look at whether this treatment is effective in relation to its costs; in other words, to see if it is cost-effective. The aim of the current work was to update the picture of scientific studies available on this topic until 2012. To this purpose, two different researchers screened electronic literature databases for studies between January 2012 and July 2022 which assessed both the costs and the effects of a homeopathic treatment. They did this according to strict rules to make sure that no important study was missed. They reviewed the search results, gathered information from the studies, and assessed the quality of the studies using a set of criteria. They detected six additional new studies to the 15 already known from the previous work. Overall, they found that in 14 out of 21 studies, homeopathic treatment was at least equally effective for less or similar costs. For the remaining seven studies, costs were equal or higher for homeopathy. Of these seven, two were shown to be advantageous for homeopathy: indeed, specific economic analyses demonstrated that the benefit of the homeopathic treatment compensated for the higher costs. For the remaining five studies, the higher or equal costs of homeopathic treatment were not compensated by a better effect. The quality of the studies varied, with older studies generally being of lower quality compared to newer ones. The authors concluded that although the quality of research on homeopathy’s cost-effectiveness has improved over time, and some high-quality studies show that it can be a cost-effective option, there are still many poorly conducted studies which make it difficult to offer a definitive statement. In other words, while there is some evidence that homeopathy can be effective in relation to its costs, there are still many studies that are not very reliable, which means that interested parties need to be cautious about drawing conclusions.

KEYWORDS:

• Cost-effectiveness
• data extraction
• economic evaluation
• homeopathy
• quality assessment
• systematic literature search

1. Introduction

In many countries the use of homeopathy is rather common in the general population. A systematic review of the 12-month prevalence of homeopathy use based on survey data from 11 countries (U.S.A., UK, Australia, Israel, Canada, Switzerland, Norway, Germany, South Korea, Japan, Singapore) demonstrated that each year a small but significant proportion (median 3.9%, range 0.7–9.8%) of the general population consults homeopaths or purchases over-the-counter (OTC) homeopathic medicines [1]. Among 1,323 parents who visited the pediatric departments of two hospitals in Germany in 2015 and 2016, 40% stated that they were already using therapies of complementary and integrative medicine (CIM) for their children. Homeopathy was the most frequently mentioned CIM therapy with almost 60% of parents confirming its use, followed by osteopathy and phytotherapy. More than 80% of the participants supported the expansion of CIM in hospitals and stated that they would be willing to pay for these treatments if not covered by their health insurance [2]. Similar patient preferences were reported in a forced-choice experiment [3] in 263 subjects in Germany that investigated indirect health benefits in CIM and conventional medicine. Patients in the homeopathy and acupuncture subsets were less cost-sensitive and considered deductible fees to be less important than those in the general medicine group. The authors concluded that acupuncture and homeopathy patients may be more used to out-of-pocket payments.

An economic analysis of the data of the nation-wide EPI3 survey conducted in France in 2007 and 2008 [4] showed that the social-security expenses for treatments by homeopathic general practitioners (GPs) were about 35% lower than for conventional GPs, while supplementary health insurance and out-of-pocket costs were slightly higher for treatments by homeopathic GPs due to higher consultation costs. The authors concluded that the management of patients by GPs certified in homeopathy may be less expensive from a global perspective and may be advantageous from the public health’s point of view. Similar observations were found in Switzerland, where a cross-sectional questionnaire-based investigation of mandatory health insurance claims data was conducted for 562 primary care physicians with and without a certificate for CIM, including homeopathy. From the perspective of the Swiss health system the analysis showed 15.4% lower costs for physicians certified in homeopathy compared to physicians applying conventional medicine [5]. In a small-scale study performed in a real-world setting in UK, more than 90% of 84 evaluable patients treated with homeopathic preparations had lower pharmacy costs than they would have had using conventional medicinal products for the same conditions [6].

An economic analysis of pharmacy-based self-care and self-medication with non-prescription (OTC) drugs was performed from the German societal perspective. It revealed net savings of about 21 billion euros p.a. of expenses for applicable medical services and products and a further 6 billion euros savings due to lowered sick leave-associated losses of productivity and working hours. Overall, each euro spent by consumers on self-medication according to this analysis translates into net savings of 17 euros in resources otherwise needed for the statutory health insurance sector and the national economy [7]. In line with these findings, it is assumed that homeopathic treatment via OTC medications contributes to patient care and relieves the burden on the health care system by conserving scarce resources of health-care professionals as well as money. It may thus be beneficial from a health economic perspective.

Two analyses of data of German health insurance companies came to partly conflicting results. One of them found a reduction in morbidity as well as in the utilization of health insurance resources by patients who took homeopathy in almost all indications studied [8]. Another data analysis compared the health care costs for patients subscribed to the German Integrated Care Contract ‘Homeopathy’ (ICCH) with the costs for those receiving standard care. Health claims data of 44,550 patients from a large statutory health insurance company were analyzed from both the societal perspective and from the statutory health insurance perspective, and both showed higher costs in the homeopathy group as compared to propensity score-matched controls over 18 months [9], which also remained higher after 33 months [10] despite a convergence tendency. While these studies [9,10] compared only costs but not treatment effects, information on cost-effectiveness was not provided.

Homeopathic treatment has been shown to have additional patient-relevant benefits. A recent review [11] suggests that the prophylactic and therapeutic use of CIM, including homeopathic medicines, may be an important contribution to reducing the widespread overuse of antibiotics in a variety of conditions. In the EPI3 survey patients with upper respiratory tract infections who consulted a certified homeopathic GP were significantly less likely to use antibiotic, antipyretic, and non-steroidal anti-inflammatory drugs than those treated by a conventional GP even though their clinical outcomes were similar [12]. In the same survey, patients with anxiety or depression consulting a certified homeopathic GP were significantly less likely to use psychotropic drugs than those treated by a conventional GP, whereas there was a tendency toward better outcomes in the homeopathy group [13]. In a randomized trial that included patients with recurrent tonsillitis, study participants who received the homeopathic preparation SilAtro-5–90 in addition to conventional treatment were at a lower risk of subsequent tonsillitis episodes, exhibited a lower symptom burden, and were less likely to use antibiotics than those who received conventional therapy alone [14].

The objective of health economic research is to provide a scientific rationale for the allocation of limited healthcare and economic resources to consistently assure high-quality care to patients. Health economic considerations based on costs and effectiveness can be applied to homeopathic treatments and to other forms of CIM in the same manner as they are applied to conventional therapies. In 2011, however, it was stated that for CIM ‘more clinical and health service research is needed which includes economic data to provide realistic cost estimates for future healthcare’ [15]. Unfortunately, this conclusion still appears to hold today [16].

Viksveen et al [17] retrieved and assessed 15 economic evaluations of homeopathy in a comprehensive literature search about economic evaluations of homeopathy covering the period until April 2012. Eight found improvements in patients’ health with homeopathy together with cost savings. Five studies found that improvements in homeopathy patients were at least as good as in control group patients at similar costs. Two studies found improvements similar to conventional treatment at higher costs. Although the authors identified evidence of the costs and potential benefits of homeopathy, they concluded that the overall evidence for cost-effectiveness still seemed uncertain and was limited by the substantial heterogeneity of results and methodological issues.

1.1. Aims

The aim of this systematic review was to update the previous review [17] and provide a current overview of the cost-effectiveness of homeopathy. Additionally, the quality of all economic evaluations published until July 2022 was assessed.

2. Methods

2.1. Literature search

The literature search for the period starting in January 2012 was conducted in July 2022 and followed the methodology described previously [17]. The literature databases MEDLINE, EMBASE, the Allied and Complementary Medicine Database (AMED), the Centre for Reviews and Dissemination (CRD) databases DARE, NHS EED, HTA, and Cochrane Library were searched for variations of the words ‘homeopathy’ and ‘cost,’ using the Boolean operator ‘AND’ and wildcard symbols as appropriate for each database. No filter for language was set. In addition, the major homeopathy journal Homeopathy was searched for the word ‘cost,’ and a manual search was performed of relevant literature known to the authors.

2.2. Selection of studies

All hits were screened by two authors independently. Each individual title was considered for relevance, and if in doubt, the abstract or full article was reviewed. Studies were included when they met three criteria: they were identified as a study in humans; they were published in a peer-reviewed journal; and they described full economic evaluations (i.e. analysis of homeopathic treatment in comparison to another health intervention in terms of both costs and outcomes with comparators including do nothing, placebo, conventional drug or other health intervention or before to after homeopathic treatment). Studies on animal, plant or in vitro investigations were excluded, as were studies assessing CIM treatment but not specifying results for homeopathy. Furthermore, studies were excluded as well if no data either on costs or clinical outcomes were reported; if there was only a single intervention without a comparator; or if only a congress abstract was available.

2.3. Data extraction

A data extraction form was developed to ensure a standardized, uniform process for the assessment of the literature. It accounted for the following different elements and steps required in economic evaluations [18,19]: study population characteristics, intervention and comparator, analysis technique and perspective, health outcome measures, scope of costs, currency and price year, time horizon and discount rates, main results as well as sensitivity analysis and results.

The number of participants analyzed was used to determine study population. Analysis techniques included cost-effectiveness analysis, which was defined as an economic evaluation that measures costs in a monetary unit and quantifies a single outcome in a physical or natural unit (e.g. number of successfully treated patients, number of life years gained, number of symptom days averted) as well as cost-utility analysis. The latter was defined as an economic evaluation that measures costs and outcomes by means of specific health-related quality of life measures such as quality-adjusted life years (QALYs). For price year, the year of manuscript submission or publication was assumed when no price year was specified, and sensitivity analyses were taken into consideration where information on cost and outcome or incremental cost-effectiveness/cost-utility ratio were reported.

Data extraction of one study [20] was carried out as a pilot, and outcomes were discussed between all authors to reach consensus about completing the form. Afterwards all other studies were assessed independently by two authors. Any disagreements between the reviewers were resolved by a third reviewer and, if there was still no consensus, a final assessment was obtained through discussion among the entire team of authors.

2.4. Quality assessment of included studies

The Consensus on Health Economic Criteria (CHEC) list was applied for quality assessment [18] of each study by two authors independently. The CHEC list focuses on the methodological quality of economic evaluation aspects. It includes 19 criteria which are to be answered with ‘yes’ (i.e. sufficient attention is given to this aspect), ‘no’ (i.e. insufficient attention is given to this aspect), or ‘not applicable’.

2.5. Data analysis

A descriptive synthesis of the extracted data was performed by summarizing the characteristics and the results of the included studies in overview tables. To compare costs between studies, costs were updated to euros (€) values in 2022 using the CCEMG-EPPI-Centre Cost Converter online tool (https://eppi.ioe.ac.uk/costconversion/); purchasing power parities were based on data from the International Monetary Fund.

Based on a format suggested by Nixon et al [21], a graphic depiction summarizing the correlation between observed costs and health outcome effects was created.

To investigate changes in the quality of health economic studies in homeopathy over time, linear regression analysis was performed with the CHEC score (number of positively answered items) as the dependent variable and publication year as a predictor.

3. Results

The literature search from January 2012 to July 2022 resulted in 408 records after removal of 72 duplicates (Figure 1). After screening, six studies were newly identified which met the selection criteria [20, 22–26]. Together with 15 studies already included in the previous systematic review [27–41], 21 studies were available for analysis.

Figure 1. PRISMA 2020 flow diagram of study selection.

3.1. Analysis of included studies

3.1.1. Characteristics of studies

The characteristics of newly included studies are presented in Table 1. Updated and extended information on the studies covered by the previous review [17] and reflecting our own work is shown in Table 2.

Table 1. Data extraction from newly identified studies.

Author/year/country	Study population	Intervention and comparator	Technique and perspective	Health outcome measure	Scope of costs and currency/price year	Time horizon and discount rates	Main results^a	Sensitivity analysis
Ostermann et al [20]/2021/Germany	Hypothetical (simulated) population of children (age 6 to <12 years) and adolescents/adults with moderate, recurrent tonsillitis (≥3 acute throat infections in the prior 12 months/≥2 in each of the previous 2 years)	SilAtro-5–90 (complex homeopathic medicinal product) as adjuvant homeopathic therapy (3 times 8-week cycles of 1 (children)/2 (adults) tablets 3 times a day) to STC vs. STC, comprising self-care with antiseptics and/or anesthetics, plus discretionary use of antibiotics and/or analgesics/antipyretics for the treatment of acute throat infections	Cost-effectiveness analysis Model-based economic evaluation (4- state Markov model) Societal perspective	Number of acute throat infection episodes avoided	Drugs, GP consultations, surgery, travel time, productivity loss €/2019	2 years 3% discount for costs and outcomes	Adults/adolescents over 12 years: SilAtro-5–90 + STC: 317.94 €, 0.46 acute throat infections. STC alone: 241.16 €, 0.92 acute throat infections. ICER: 167.42 € per acute throat infection avoided. Children 6 to <12 years: SilAtro-5–90 + STC: 234.87 €, 0.46 acute throat infections. STC alone: 241.16 €, 0.92 acute throat infections. SilAtro-5–90 + STC dominated STC alone. Assessment summary: health outcomes +, costs +	Univariate sensitivity analysis with values changed by 50%, threshold sensitivity analysis and probabilistic sensitivity analysis Cost-effectiveness results were most sensitive to the probability of having at least 1 acute throat infection with STC, the number of previous acute throat infections, and the expected number of acute throat infections with STC. The probability that SilAtro-5–90 + STC is cost-effective was 65% for adults and adolescents and 71% for children at a threshold value of 1,000 € per acute throat infection avoided.
Kass et al [22]/2020/Germany	2,524 adult or adolescent patients with allergic rhinitis, asthma, atopic dermatitis, migraine/headache, or depression	Treatment given by certified homeopathic physicians acc. to ICCH (HOM, n = 1,543) vs. STC (n = 981, matched on diagnosis, sex, and age)	Cost-utility analysis Comparative, prospective, observational study Primary endpoint: healthcare payer perspective; secondary analysis: societal perspective	Quality-adjusted life years (QALYs) based on Short Form Health Survey (SF-12) data	Primary analysis: costs acc. to ICCH (HOM group), medication costs, outpatient care costs, inpatient costs, other costs (for home healthcare, patient transportation, remedies and aids) Secondary analysis: productivity loss €/year not specified (submitted 2020)	1 year Discounts not applicable	Adjusted (for age, sex, education, residential state, complementary medicine affiliation, per protocol initial anamnesis and baseline utility value) total costs and adjusted mean QALY were significantly higher in the HOM group for migraine/headache (∆ costs: +442.16 €, p < 0.001; ∆ QALY: +0.034, p < 0.001) and for depression (∆ costs: +439.59 €, p < 0.001; ∆ QALY: +0.035, p < 0.001). For atopic dermatitis, adjusted total costs were non-significantly higher for HOM (+145.76 €, p = 0.250) whereas adjusted mean QALY was significantly higher (+0.038, p < 0.001).HOM was cost effective for migraine/headache (ICER: 12,863 € per QALY [95% CI: 12,490–13,235 €], depression (ICER: 12,413 € [95% CI: 12,101–12,726 €], and atopic dermatitis (ICER: 3,930 € per QALY [95% CI: 3,589 –4,271 €]) (assumed cost-effective threshold: ICER < 50,000 € per QALY).Assessment summary: health outcomes +, costs +	Sensitivity analyses on the variances of diagnosis-specific cost-effectiveness differences were operationalized using nonparametric bootstrapping. Based on these bootstrap results, cost-effectiveness acceptance curves were calculated to show probabilities of cost-effectiveness according to different values of the willingness to pay for a QALY gained. For migraine/headache, depression and atopic dermatitis, probability of cost-effectiveness was 100% at a threshold value of 50,000 € per QALY.
Colombo et al/2018 [23]/Italy	455 adult (n = 286) and pediatric (n = 169) patients from a single center, with ≥3 episodes of respiratory tract infections during the year preceding study inclusion	Oscillococcinum (Anas barbariae hepatis et cordis extractum 200k, 1 dose a week for 8 months per year) in addition to STC, comprising conventional treatment with symptomatic drugs (n = 246) vs. STC alone (n = 209)	Cost-effectiveness analysis Retrospective, observational study Health care payer perspective	Average number of episodes of respiratory tract infections per patient per year	Costs for medication and consultations €/year not specified (published 2018)	1 to 10 years (mean ± SD: 5.3 ± 2.6 years) No discounts reported	All patients: 0.66 ± 0.67 infections/year and costs of 23.77 € ± 28.78 € for Oscillococcinum + STC; 1.16 ± 1.18 infections/year and costs of 52.27 € ± 71.84 € for STC. Pediatric patients: 0.93 ± 0.80 infections/year and costs of 21.70 € ± 17.74 € for Oscillococcinum + STC; 1.69 ± 1.06 infections/year and costs of 43.97 € ± 34.69 € for STC. Assessment summary: health outcomes +, costs -	None reported
Frei [24]/2015/Switzerland	43 patients ≥20 years of age, ≥3 concurrent diagnoses (‘multimorbid’); single practice	Homeopathic single remedies prescribed according to polarity analysis Comparators for health outcome: pre-post treatment comparison costs: hypothetical patients with identical diagnoses treated with STC	Cost-effectiveness analysis Prospective case series study Perspective not reported	Patients’ rating of symptoms intensity (0=absent to 10=maximal intensity) and healing (0 to 10=complete healing)	Costs for medication and consultations €/year not specified (submitted 2014)	1 year Discounts not applicable	43 of 50 patients (86%) achieved an improvement of symptoms ≥80% and were counted as responders while 6 patients stopped treatment for lack of efficacy or other reasons. 1 patient showed improvement <80% and was thus also counted as treatment failures. The total costs were 657 € for HOM vs. an estimated 1,619 € for STC for the same diagnoses, i. e. HOM reduced the treatment costs by 59%. Assessment summary: health outcomes +, costs -	None reported
Roll et al [25]/2013/Germany	135 children aged 1–14 years, treated by homeopathic GP and conventional GPs) with mild to moderate atopic eczema (3 out 4 Williams’ criteria, Three-Item severity score between 2 and 7) for ≥6 months	Classical HOM based on homeopathic history case taking and prescription of single remedies according to the simile law (n = 48, 10 homeopathic GPs) recommendation of emollient creams and corticosteroids as rescue medication only vs. STC acc. to German disease management guidelines (n = 87, 16 conventional GPs)	Cost-effectiveness analysis Non-randomized, prospective, multicenter, open, comparative observational study Societal perspective	Main outcome: SCORAD (SCORing Atopic Dermatitis) score (range: 0–103 points; lower score indicates better outcome)	Costs of drugs, hospitali-zation, GP consultations, medical aids and adjuvant therapies, productivity loss (of parents) €/year not specified (submitted 2012)	36 months for health outcomes, 6 months (months 31-36) for costs Discount: 3% for costs	HOM: SCORAD total score 13.7 [95% CI: 7.9–19.5] after 36 months, total costs 261.14 € [95% CI: 185.46–336.85 €]. STC: SCORAD total score 14.9 [95% CI: 10.4–19.4] after 36 months, total costs 120.26 € [95% CI: 64.63–175.83 €]. Assessment summary: health outcomes 0, costs +	Non-parametric bootstrapping adjusting for confounders. Results support those of the main analysis.
Basili et al [26]/2011/Italy	Stratified sample of 82 patients treated for the common cold, influenza or para-influenza, who asked for reimbursement of 2 or more prescriptions during the follow-up. Participants were taken from CASAGIT, an Italian health insurance company for journalists and their families.	HOM including purely homeopathic, anthroposophical, and homeotoxicological remedies vs. STC. Participants were free to choose HOM, STC, or a combination of both interchangingly during follow-up.	Cost-effectiveness analysis Prospective cohort study Perspective not reported	Time between consecutive prescriptions	Costs of drugs Currency not reported but assumed to be €/year not specified (submitted 2008)	40 months No discounts reported	Intervals between consecutive HOM prescriptions were on average shorter than for STC. On the other hand, patients who favor homeopathy were at a reduced risk of re-prescriptions. Older patients required more frequent treatment while sex had no influence on treatment re-occurrence. Overall costs for HOM and STC were assessed to be similar for the indications studied (assessment based on intra-subject differences between HOM and STC presented graphically on the subject level). Assessment summary: health outcomes 0, costs 0	None reported

^aAll amounts from the original publication converted to € and a target price year of 2022. HOM: homeopathic treatment; STC: standard care (i.e. conventional medical care); ICER: Incremental cost-effectiveness ratio; QALY: Quality-adjusted life years; ICCH: German Integrated Care Contract ‘Homeopathy’ for the reimbursement of homeopathic treatments by statutory health insurance companies; GP: general practitioner;Assessment summary, HOM compared to applicable comparator: +: better health outcomes / higher costs; 0: similar health outcomes / similar costs; ‑: poorer health outcomes / lower costs.

Table 2. Data extraction of previously identified studies, updated and extended.

Author/year/country	Study population	Intervention and comparator	Technique and perspective	Health outcome measure	Scope of costs and currency/price year	Time horizon and discount rates	Main results^a	Sensitivity analysis
Kooreman and Baars [27]/2012/The Netherlands	Data of patients of a Dutch health insurance and contributed by 1,992 GPs, including 28 homeopathic GPs	HOM by specialized GP (n = 1,181) vs. STC by conventional GP (n = 151,952)	Cost-effectiveness analysis Retrospective insurer dataset-based economic evaluation Health care payer perspective	Mortality rate	Costs of consultations, hospital care, medication, and paramedic care €/year not specified (submitted 2010)	4 years (2006–2009) No discounts reported	Crude mortality rates were 3.8% and 2.6% for HOM and STC, respectively (p < 0.05). In a regression model adjusting for confounders, HOM showed a non-significant reduction of mortality rates as compared to STC. Quarterly costs were 554.80 € for HOM compared to 589.12 € for STC. Assessment summary: health outcomes 0, costs -	None reported
Thompson et al [28]/2011/UK	35 children, 7–14 years old, with asthma (≥step 2 of the BTG (British Thoracic Society Asthma Guidelines)) recruited from 2 secondary care respiratory hospital clinics	Individualized HOM + STC (n = 17) vs. STC alone (n = 18)	Cost-effectiveness analysis Pragmatic randomized, controlled study Societal perspective	Pediatric Asthma Quality of Life Questionnaire (PAQLQ), Asthma Control Questionnaire (ACQ), measures of morning and afternoon peak flow	Healthcare costs (primary care consultations, secondary care and ambulance services, medication); personal costs (travel, loss of earnings) £/year not specified (submitted 2010)	16 weeks Discounts not applicable	No evidence for differences between groups for asthma control (ACQ) and asthma quality of life (PAQLQ); no clinical meaningful differences for peak flow or any other outcome measure. Average healthcare costs: HOM 1,299.66 €, STC 448.02 €; personal costs: HOM 149.80 €, STC 31.90 € Assessment summary: health outcomes 0, costs +	None reported
Kneis and Gandjour [29]/2009/Germany	113 adult patients aged 18–59 years with acute maxillary sinusitis (AMS, confirmed by radiography with mucosal thickening >6 mm, air-fluid level or complete opacification on the most AMS-affected side), 6 ear-nose-throat clinics and outpatient departments	Complex HOM (Sinfrontal®, n = 57) vs. placebo (n = 56) for 22 days. Concomitant paracetamol and saline inhalations were allowed.	Cost-utility analysis Double-blind, randomized controlled trial Societal perspective	Quality-adjusted life years (QALYs) gained, based on EQ-5D questionnaire	Costs of drug and productivity loss €/2005	3 weeks (main analysis), observational 8 week follow-up Discounts not applicable	Main analysis: compared to placebo, Sinfrontal® led to cost savings of 358 €/patient (95% CI: 134; 563 €) Utility gained: Sinfrontal® 0.0410 QALYs, placebo 0.0323 QALYs, incremental utility 0.0087 QALYs (95% CI: 0.0052–0.0123). Follow-up results support the main analysis. Assessment summary: health outcomes +, costs -	One-way sensitivity analyses based on different healthcare outcomes and cost estimates. Sinfrontal® led to cost savings and utility gains in 4 out of 5 sensitivity analyses and dominated placebo in 99.8 to 100% of cases. Non consideration of indirect costs led to a 100% cost -effectiveness at a willingness to pay of 3,600 € per QALY gained.
Pomposelli et al [30]/2009/Italy	61 patients with type-2 diabetes mellitus and diabetic polyneuropathy, single clinic with 4 participating physicians	Individualized, classical HOM (>10 different homeopathic remedies) + STC (diet, insulin or oral hypoglycemic agent, physiotherapy; n = 32) vs. STC alone (n = 29)	Cost-effectiveness analysis Non-randomized, prospective, observational study Perspective not reported	Diabetic Neuropathy Symptom (DNS) score, SF-36 quality-of-life questionnaire	Costs of medication €/2003 and 2004 (2004 used for cost conversion)	12 months Discounts not applicable	The DNS improved from 1.40 ± 1.21 to 1.07 ± 1.25 points at month 6 (p = 0.016) and to 1.22 ± 1.27 points at month 12 (p = 0.146) for HOM + STC. For STC alone, the DNS improved from 1.26 ± 1.06 to 1.06 ± 1.15 points at month 6 (p = 0.350) and to 0.94 ± 1.21 points at month 12 (p = 0.182). Significant improvements in SF- 36 scales were observed for HOM + STC. First 6 months’ costs for STC were 159 ± 157 € at baseline and were reduced to 132 ± 89 € in the HOM + STC group (p = 0.078) but increased to 109 ± 82 € (starting from 96 ± 77 € at baseline) for STC alone (p = 0.139). During months 7–12, STC costs in the HOM + STC group tended to increase but remained below the baseline level whereas costs in the STC-only group were comparable to the first 6 months. No between-group test statistics were reported. Assessment summary: health outcomes +, costs 0	None reported
Witt et al [31]/2009/Germany	126 children aged 1-14 years with mild to moderate atopic eczema (duration of at least 6 months, fulfilling 3 of 4 Williams’ criteria, Three-Item Severity score between 2 and 7), 26 physicians	Individualized, classical, single-remedy HOM (41 different homeopathic remedies prescribed in total), emollient creams recommended, steroids as rescue therapy (n = 46, 10 physicians) vs. STC acc. to German disease management guidelines (n = 80, 16 physicians)	Cost-effectiveness analysis Prospective, multicenter, comparative observational study Societal perspective	Main outcome: SCORAD (SCORing Atopic Dermatitis) score (range: 0–103 points; lower score indicates better outcome) Quality-of-life (QoL) questionnaires assessing children and their parents	Costs of consultations, hospital stays, medication, number of sick leave days, medical aids and adjuvant therapies €/year not specified (submitted 2009)	12 months Discounts not applicable	Baseline-adjusted SCORAD scores were 22.5 ± 3.0 for HOM vs. 18.2 ± 2.3 for STC at month 6 (p = 0.290) and 17.4 ± 3.0 for HOM and 17.3 ± 2.3 for STC at month 12 (p = 0.974). QoL results in both groups were similar. Average 6 months costs were 1,162 € (95% CI: 813–1,512 €) vs 639 € (95% CI: 344–935 €) for HOM and STC, respectively (p = 0.026) and 12-month costs were 1,895 € (95% CI: 1,489–2,300 €) vs. 896 € (95% CI: 528–1,265 €) for HOM and STC, respectively (p = 0.001). Main cost drivers were consultations. Assessment summary: health outcomes 0, costs +	Per-protocol analysis, different missing data imputation strategies, non-parametric bootstrapping (1,000 times) adjusting for confounders. Bootstrapping results for 6 month-period: 8.1% probability of superior cost-effectiveness for HOM. For 12-month period probability is 35%. ≥94% probability of additional costs in the HOM group.
Sevar [32]/2005/UK	455 consecutive patients aged 1 month to 87 years from a single homeopathic clinic, various chronic conditions	Individualized, classical, single-remedy HOM (187 different single remedies prescribed in total), comparison of pre- vs. post-treatment	Cost-effectiveness analysis Prospective observational study Perspective not reported	Glasgow Homoeopathic Hospital Outcome Scale (GHHOS), successfully treated patients (GHHOS ≥1)	Costs of drug (savings of conventional drugs) £/2003	5 years total observation period Mean treatment /follow-up per patient: 11 months No discounts reported	304 patients were successfully treated (66.8%). In adults, success rates were higher in females (71.3% of n = 296) than in males (58.5% of n = 159). Age was not systematic associated with response. 148 patients (32.5%) stopped or substantially reduced their conventional medication, leading to average conventional drug cost savings of 65.14 €/patient/year. Assessment summary: health outcomes +, costs ­-	None reported
Witt et al [33]/2005/Germany	120* adults with diagnosis of headache, lower back pain, depression, insomnia or sinusitis and 68* children aged 1–16 years with diagnosis of bronchial asthma, atopic dermatitis or allergic rhinitis, treated by 101 homeopathic GPs and 59 conventional GPs *number estimated from available data in publication	HOM administered by homeopathic GPs vs. STC administered by conventional GPs	Cost-effectiveness analysis Prospective, multicenter, parallel group, comparative cohort study Health care payer perspective	Severity of symptoms assessed by patients and physicians (VAS scale from 0 to 10); adults: quality of life (MOS-SF-36)	Costs of medication, consultations, physiotherapy, hospitalization, productivity loss €/year not specified (published 2005)	12 months Discounts not applicable	Self-rated symptom improvement under HOM was significantly more pronounced than for STC in both adults (p = 0.002) and children (p = 0.029). In the physicians’ symptom rating HOM was superior to STC in children (p < 0.001) but not in adults (p = 0.251). Improvements in the SF-36 physical component were more pronounced for HOM during the first 6 months (vs. STC: p = 0.016) but not during the second 6 months (p = 0.649). No significant differences for the mental component (p = 0.273 favoring HOM). No significant differences for over-all adjusted costs: adults HOM 2,802 €, STC 2,617 €, p = 0.856; children HOM 1,912 €, STC 1,022 €, p = 0.137 Assessment summary: health outcomes +, costs 0	None reported
Trichard et al [34]/2005/France	431 children aged 18 months to 4 years with recurrent, acute, viral rhinopharyngitis (≥5 episodes during last 6 months)	Patients having received at least 1 homeopathic drug and no antibiotic (n = 241, 73 homeopathic GPs) vs. patients with at least 1 antibiotic and no homeopathic drug (n = 190, 62 non-homeopathic GPs). Other drugs such as mucus liquifiers, trace elements, corticosteroids, or antipyretics were allowed in both groups	Cost-effectiveness analysis Pragmatic, prospective observational study Societal perspective, health care payer perspective, patient perspective	Number of episodes of acute rhinopharyngitis, complications linked to episodes of acute rhinopharyngitis, adverse events; quality-of-life questionnaire (Par-Ent-Qol®) for parents of children with ear, nose, and throat infections	Costs of medication, consultations, complementary tests, productivity loss (number and duration, not economically evaluated) €/2000	6 months Discounts not applicable	Compared to patients on antibiotics, those on HOM had fewer episodes of rhinopharyngitis (2.71 vs. 3.97 on average, p < 0.001), fewer complications (1.25 vs. 1.95 on average, p < 0.001), and their parents made more favorable global quality-of-life ratings (21.4 vs. 30.4 points, p < 0.001). Adverse event rates were similar (4.6% vs. 4.2%, p = 0.86). Average direct costs were comparable (HOM 235 €, antibiotics 233 €, p > 0.05), but HOM caused less social security expenses (119 € vs. 134 €, p < 0.05) and more out-of-the-pocket costs (116 € vs. 99 €, p < 0.05). Fewer parents in the HOM group had to take leave to care for their sick child (9.5% vs. 31.6%, p < 0.001). Assessment summary: health outcomes +, costs 0	None reported
Trichard et al [35]/2004/France	499 children aged 18 months to 4 years with recurrent, acute, viral rhinopharyngitis (≥5 episodes during last 6 months)	Treatment by specialized homeopathic physician (HOM; n = 268, 73 GPs) vs. treatment by non-homeopathic physician (STC; n = 231, 62 GPs). In both groups, physicians were free to prescribe any treatment according to their medical judgment	Cost-effectiveness analysis Pragmatic, prospective, observational study Societal perspective, health care payer perspective, patient perspective	Number of episodes of acute rhinopharyngitis, complications linked to episodes of acute rhinopharyngitis, adverse events; quality-of-life questionnaire (Par-Ent-Qol®) for parents of children with ear, nose, and throat infections	Costs of medication, consultations, complementary tests, productivity loss (number and duration, not economically evaluated) €/1999 and 2000 (2000 used for cost conversion)	6 months Discounts not applicable	Compared to patients on STC, those on HOM had fewer episodes of rhinopharyngitis (2.77 vs. 3.83 on average, p < 0.001), fewer complications (1.40 vs. 1.81 on average, p = 0.009), and their parents made more favorable global quality-of-life ratings (21.979 vs. 28.125 points, p = 0.010). Adverse event rates were similar (4.9% vs. 3.0%, p = 0.42). Compared to STC, HOM was associated with higher average costs to society (262 € vs. 223 €, p = 0.012) and to the patient (129 € vs. 93 €, p < 0.001) but with similar costs to social security (133 € vs. 130 €, p = 0.73). Fewer parents in the HOM group had to take leave to care for their sick child (13.1% vs. 27.4%, p = 0.001). Assessment summary: health outcomes +, costs +	None reported
Slade et al [36]/2004/UK	97 consecutive patients aged 0–90 years with various complaints, referred to 1 homeopathy service run by 1 GP	HOM by qualified practitioner (no details reported), pre- vs. post-treatment comparison	Cost-effectiveness analysis Prospective observational study Perspective not reported	Self-rated symptom severity, activity limitation, and wellbeing were scored using the MYMOP (Measure Yourself Medical Outcome Profile, 7-point scale) tool	Costs of savings on conventional drug £/year not specified (submitted 2003)	Health outcomes: Primary symptoms individual follow-up: mean duration of 134 (25–222) days. Secondary symptoms, activity limitation and wellbeing: mean duration of 126 (25–222), 141 (27–222) and 132 (25–222) days respectively Consultation/medication fees: 6/12 months before vs. 6/12 months after initial consultation Conventional medication costs savings: mean duration of 146 (19–222) days Discounts not applicable	Average improvements between 1.41 and 2.49 points for primary and secondary symptoms, ability to undertake activities, and well-being between initial visit and final assessment (p < 0.0001 for each item). The average number of consultations decreased from 6.33 during 6 months before inclusion to 5.26 during 6 months after inclusion (p = 0.004). 28 of 49 evaluable patients (57%) reduced or stopped the use of conventional medication after the start of HOM. For the 22 patients, for whom clear information was available, the annual cost saving was 4,582.81 €. Assessment summary: health outcomes +, costs -	None reported
Van Wassenhoven and Ives [37]/2004/Belgium	782 patients with mainly chronic diseases of all major organ systems; 80 specialized homeopathic practices	Classical HOM, including a total of 333 different single remedies, 21 of them accounted for about half of the prescriptions, pre- vs. post-treatment comparison	Cost-effectiveness analysis Prospective observational study Perspective not reported	Patients and physicians’ ratings of severity of physical and psychological symptoms; patient’s perceived effectiveness of current and previous treatment; satisfaction with current and previous treatment	Medication and consultation fees €/year not specified (submitted 2003)	Health outcomes estimated by patients and physicians on 1 specific day comparing before and after homeopathic treatment. Cost calculations based on a 1-year period Discounts not applicable	The proportion of patients with no appreciable symptom burden was 18.2% before and 64.2% after HOM for physical symptoms and 24.8% and 66.6% for psychological symptoms, respectively. 89% of patients stated that HOM had improved their physical condition. 95% of patients reported to be fairly or much satisfied with homeopathic treatment, compared to 18.9% for previous conventional treatment. Patients rated the effectiveness of their homeopathic treatment as very effective. Patients’ estimated average annual expenditure on drugs was 377 € during HOM, compared to 486 € before HOM. Homeopathic GPs’ (data from 47 GPs) expenditure on conventional drugs was on average about ^1/3 of the national GP average amongst them ^1/5 for antibiotics Assessment summary: health outcomes +, costs -	None reported
Paterson et al [38]/2003/UK	60 adult patients aged 29–84 years with dyspepsia (upper abdominal or retro sternal pain, discomfort, heartburn, nausea, vomiting ≥2 weeks), single clinic	Classical, individualized HOM by qualified GP (n = 21) vs. treatment by conventional GP (STC; n = 19) or acupuncture (n = 20)	Cost-effectiveness analysis Open, randomized clinical trial Health care payer perspective	Measure Yourself Medical Outcome Profile (MYMOP) and General Well-being Index (GWBI), SF-36 health survey	Costs for medication, consultations, referrals (National Health Service, NHS costs) £/1998 and 1999 (1999 used for cost conversion)	Health outcomes at 6 weeks (primary), 4 and 6 months (trial period) and 12-month follow-up (MYMOP only); costs during 6-month trial period and 12- month follow-up Discounts not applicable	MYMOP score mean changes at week 6 (primary outcome measure) were 0.44 ± 1.41 points for HOM and 0.53 ± 1.76 points for STC, mean difference −0.09 (95% CI: −1.19; 1.01) points. No significant differences at week 6 or at any other time point. Little change over time and no appreciable group differences for GWBI and SF36. Average total dyspepsia-related NHS primary care costs per patient for 6-month period: HOM 104.22 ± 132 €, STC 129 ± 148 €. Assessment summary: health outcomes 0, costs -	None reported
Frei and Thurneysen [39]/2001/Switzerland	230 patients between 0 and 16 years with acute otitis media (acute ear pain, hearing loss, signs of inflammation of the eardrum), single clinic	Classical, individualized HOM (n = 230, 27 different remedies given).Comparator for health outcome: placebo group; comparator for economic analysis: costs for STC	Cost-effectiveness analysis Prospective, observational study Perspective not reported	Proportions of patients free of pain at 6 h after start of HOM needing a 2^nd HOM and free of pain at 12 h after start of treatment	Costs for medication and consultations CHF/price year not specified (submitted 2000)	12 h Discounts not applicable	90 patients (39%) pain-free at 6 h after 1^st HOM, 76 patients (33%) pain-free after 12 h and 2^nd HOM; 64 patients (28%) not recovered within 12 h and received antibiotics. Based on literature comparison, improvement with HOM 2.4 times faster than with placebo. Total costs/patient 14% cheaper for HOM treatment than for STC (69 € vs. 80 €) Assessment summary: health outcomes +, costs -	None reported
Bachinger et al [40]/1996/Germany	114 patients with chronic arthritis of the knee with arthrosis therapeutics	HOM (n = 57): Zeel® comp. (homeopathic complex remedy) intra-articular injections (3 cycles of 10 injections for 5 weeks each) vs. STC (n = 57): hyaluronic acid intra-articular injection (3 cycles of 5 injection for 5 weeks each). In both groups: concomitant therapy: heat applications and movement therapy	Cost-effectiveness analysis Model-based economic evaluation (decision tree) Societal perspective	Treatment response, defined as pain reduction of 30% compared to baseline after 5 weeks (visual analogue scale)	Costs within the healthcare system, costs of intervention, travel, sick leave, adverse effects DM/year not specified (published 1996)	1 year Discounts not applicable	Average total direct + indirect costs, HOM: 788 € per patient, 1,728 € per successfully treated patient (26 out of 57); STC: 1,454 € per patient, 2,688 € per successfully treated patient (34 out of 57). Assessment summary: health outcomes 0, costs -	Probabilistic sensitivity analysis varying the probability of discontinuation of therapy Based on different proportions of premature treatment discontinuation the adjusted costs per successfully treated patient were always lower for HOM as compared to STC.
Feldhaus [41]/1993/Germany	Dental surgery patients with statutory insurance, single dental practice, sample size not reported	HOM: Arnica 12X before and after every planned surgical treatment and at least after every unplanned surgical treatment (1 tablet per day for 3 days before and up to 1 week after the procedure, other homeopathic single remedies as indicated), comparison of pre- vs. post-treatment STC for costs: averages from other dental practices acc. to reimbursement plan of statutory insurance	Cost-effectiveness analysis Retrospective analysis of health records Health care payer perspective	Frequency of post-operative treatments and sick certificates before and after the introduction of Arnica 12X	Costs related to postoperative follow-up treatment (medication, consultations, surgery) and due to productivity loss (number of sick certificates) DM/year not specified (originally published in German 1992)	Time horizon not reported (analysis of health records from 1983 to 1990) No discounts reported	After the introduction of HOM using Arnica 12X, the number of post-operative follow-up treatments decreased by an average of 54.7% compared to before the introduction, resulting in average annual savings of 934 € for the insurance company. The average number of sick certificates issued by the practice per quarter was reduced from 7.0 to 2.6 compared to pre-HOM. Assessment summary: health outcomes +, costs -	None reported

^aAll amounts from the original publication converted to € and a target price year of 2022. HOM: homeopathic treatment; STC: standard care (i.e. conventional medical care); QALY: Quality-adjusted life years; GP: general practitioner;Assessment summary, HOM compared to applicable comparator: +: better health outcomes / higher costs; 0: similar health outcomes / similar costs; ‑: poorer health outcomes / lower costs.

The studies were published between 1993 and 2021 with data from Germany (six national studies [22, 25, 31, 33, 40, 41], two international studies [20, 29]), the United Kingdom (four studies [28,32,36,38]), Italy (three studies [23, 26, 30]), France (two studies [34, 35]), Switzerland (two studies [24, 39]), Belgium (one study [37]), and the Netherlands (one study [27]).

Primary health outcome measures were highly heterogeneous and included various symptom scores (e.g. SCORing Atopic Dermatitis [25, 31] or Diabetic Neuropathy Symptom score [30]), as well as self-rated or observer-rated clinical status (e.g. symptom severity [24, 33, 37], Measure Yourself Medical Outcome Profile (MYMOP) [36, 38]) or quality of life ratings (e.g. Pediatric Asthma Quality of Life Questionnaire (PAQLQ) [28], SF-12 [22] or EQ-5D [29]). Furthermore, the number or proportion of some clinical events of interest observed or avoided (recurrence of acute rhinopharyngitis [34, 35], of acute throat infections [20] or of respiratory tract infections [23]), the cure rate [32, 39, 40], or the time between consecutive events of interest [26] were chosen as health outcomes. Except for stratification by age group (e.g. children, adolescents, and adults, where applicable), only a minority of studies reported results that were adjusted for participants’ characteristics.

In terms of economic outcomes, all studies reported medication costs, with four studies presenting costs of medication only [26, 30, 32, 36]. Two studies reported that patients spent less money on conventional medication during homeopathic treatment but did not specify homeopathic treatment costs [32, 36]. Consultation fees were included in 16 studies [20, 22–25, 27, 28, 31,33–35, 37–41] and hospital costs (if applicable for the study indication) as well as other direct health-care costs in seven studies [20, 22, 25, 27, 28, 31, 33] while four studies considered also direct non-healthcare costs, e.g. transportation to consultations [20, 22, 28, 40]. Indirect costs such as those attributable to sick leave were included in 11 studies [20, 22, 25, 28, 29, 31, 33–35, 40, 41]; however, some of these studies reported only on changes in work absenteeism but presented no data on the financial implications [34, 35, 41].

Two studies were model-based economic evaluations (i.e. 4-state Markov model [20] or decision tree [40]). In one study a retrospective insurer dataset-based economic evaluation was performed [27], and one study retrospectively analyzed health records [41]. The remaining publications were clinical study-based economic evaluations of which 13 were based on data from observational studies [22, 23, 25, 26, 30–37, 39], two on data from open-label randomized controlled trials (RCTs) [28,38], one on data from a double-blind RCT [29] and one was a case series [24].

Two publications [22, 29] were cost-utility analyses; all others reported on cost-effectiveness analyses. Six studies used a societal perspective [20, 25, 28, 29, 31, 40], and five used a health-care payer perspective [23, 27, 33, 38, 41], while the remaining publications used either multiple perspectives [22,34,35] or did not report the perspectives used [24, 26, 30, 32, 36, 37, 39].

Nine studies were monocentric [23, 24, 28, 30, 32, 36, 38, 39, 41], 11 were multicentric [20, 22, 25–27, 29, 31, 33–35, 37], and one publication did not mention the number of sites [40]. Sample sizes varied considerably, ranging from fewer than 20 to several thousand patients per treatment group.

Four studies [20, 23, 29, 40] assessed homeopathic combination products, and one study assessed one particular single remedy [41], while all others investigated individualized homeopathic treatment. In four studies, homeopathic medicines were given in addition to standard care [20, 23, 28, 30]. Most publications compared homeopathic treatment to standard care or, in case of add-on homeopathy, to standard care alone (14 of 21 papers [20, 22, 23, 25–28, 30, 31, 33–35, 38, 40]). Two studies were based on pre- vs. post-treatment comparisons in a single group of patients [32, 36], four used different comparator groups for health outcomes (pre- vs. post-treatment or published data) and costs (published data) [24, 37, 39, 41], and one [29] was placebo-controlled.

Indications for homeopathic treatment included respiratory tract and ear disorders (eight studies [20, 23, 26, 28, 29, 34, 35, 39]), atopic eczema (two studies [25, 31]), chronic arthritis [40], dyspepsia [38], diabetic polyneuropathy [30], and dental surgery [41]. The remaining seven studies investigated patients with various complaints or multimorbid patients [22, 24, 27, 32, 33, 36, 37]. Four studies assessed only adults [24, 29, 30, 38], six only children and/or adolescents [25, 28, 31, 34, 35, 39], while the remaining studies assessed children, adolescents and adults or did not indicate the age range of their patient population [20, 22, 23, 26, 27, 32, 33, 36, 37, 40, 41].

Due to their heterogeneity, a statistical analysis of clinical outcomes and health economic parameters across all studies was not possible.

3.1.2. Summary of main results

Tables 1, 2 provides a summary of the main clinical and economic results of the studies, including an assessment as to whether the effect of homeopathic treatment in comparison to the control intervention was associated with a better (+), similar (0), or poorer (-) health outcome, as well as with decreased (-), similar (0), or increased (+) costs.

Figure 2 shows that among the 21 identified studies, 14 studies (66.7%) reported more favorable clinical outcomes in the homeopathy group than in the control group; seven (33.3%) observed similar treatment effects. Regarding economic outcomes, 11 studies (52.4%) found that the costs for sole or adjuvant homeopathic treatment were lower than for the comparator. Four studies (19.0%) reported similar costs for both treatments; six (28.6%) described increased costs with homeopathy.

Figure 2. Summary of economic evaluation of homeopathic treatment vs. comparator.

*studies with health outcomes + and costs -: Colombo 2018, Frei 2015, Kneis 2009, Sevar 2005, Slade 2004, van Wassenhoven 2004, Frei 2001, Feldhaus 1996; studies with health outcomes + and costs 0: Pomposelli 2009, Witt 2005, Trichard 2005; studies with health outcomes 0 and costs -: Kooreman 2012, Paterson 2003, Bachinger 1996; studies with health outcomes + and costs +: Ostermann 2021, Kass 2020, Trichard 2004; studies with health outcomes 0 and costs 0: Basili 2011; studies with health outcomes 0 and costs +: Roll 2013, Thompson 2011, Witt 2009.

**studies with health outcomes + and costs -: Kneis 2009; studies with health outcomes + and costs +: Ostermann 2021, Kass 2020, Trichard 2004; studies with health outcomes 0 and costs +: Roll 2013, Witt 2009.

Read the detailed description of this figure

HOM: homeopathic treatment; +: better health outcomes / higher costs for HOM; 0: similar health outcomes / similar costs for HOM; ‑: poorer health outcomes / lower costs for HOM; green colored boxes: favors HOM; yellow colored boxes: needs further consideration, i.e. incremental analysis; red colored boxes: favors comparator.

In a synthesis of the results for both health outcomes and costs, 11 studies (52.4%) found more favorable clinical effects for homeopathic treatment at costs which were lower [23,24,29,32,36,37,39,41] or similar [30,33,34], and three studies (14.3%) reported similar health outcome results at lower treatment costs [27,38,40]. Consequently homeopathic treatment was at least equally effective for less or similar costs than control treatment in a total of 14 out of 21 studies (66.7%). One study found similar health outcomes and similar costs for both treatment groups [26]. Three other studies [20,22,35] found improved clinical outcomes for homeopathic treatment at higher costs. The three remaining studies [25,28,31] (14.3%) observed similar outcomes but at higher costs for homeopathic treatment and thus favored the comparator.

3.2. Result of quality assessment of included studies

Details of the quality assessment for each study using the CHEC list for health economic studies can be found in Table 3. Out of a total of 19 criteria, between two and 16 were met by the studies included in the review. When placed into proportion to the total number of criteria which applied (i.e. all criteria excluding those which were answered with n/a), this corresponds to 11% and 89%. Six studies (28.6% of 21) met more than 60% of the quality criteria [20, 22, 25, 29, 31, 35], six (28.6%) met between 50 and 60% [23, 28, 33, 34, 38, 40], and the remaining nine studies (42.9%) met less than 30% of the criteria [24, 26, 27, 30, 32, 36, 37, 39, 41]. Studies with a CHEC quality assessment of at least 60% are also presented on Figure 2.

Table 3. Quality assessment of included studies by means of the CHEC list.

		Ostermann 2021 [20]	Kass 2020 [22]	Colombo 2018 [23]	Frei 2015 [24]	Roll 2013 [25]	Kooreman 2012 [27]	Basili 2011 [26]	Thompson 2011 [28]	Kneis 2009 [29]	Pomposelli 2009 [30]	Witt 2009 [31]	Sevar 2005 [32]	Witt 2005 [33]	Trichard 2005 [34]	Trichard 2004 [35]	Slade 2004 [36]	v. Wassenhoven 2004 [37]	Paterson 2003 [38]	Frei 2001 [39]	Bachinger 1996 [40]	Feldhaus 1993 [41]
1	Study population	✓	✓	✓	⃠	✓	⃠	✓	✓	✓	✓	✓	⃠	✓	✓	✓	✓	⃠	✓	⃠	⃠	⃠
2	Competing alternatives	✓	⃠	⃠	⃠	⃠	⃠	⃠	⃠	✓	✓	✓	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	✓	⃠
3	Research question	✓	✓	✓	⃠	✓	⃠	⃠	✓	✓	⃠	✓	✓	✓	✓	✓	✓	⃠	✓	✓	✓	✓
4	Study design	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
5	Time horizon	✓	✓	✓	✓	⃠	✓	✓	⃠	✓	⃠	✓	⃠	⃠	✓	✓	⃠	⃠	⃠	⃠	✓	⃠
6	Perspective	✓	⃠	✓	⃠	✓	⃠	⃠	✓	✓	⃠	✓	⃠	⃠	✓	✓	⃠	⃠	✓	⃠	✓	✓
7	Cost identification	✓	✓	⃠	⃠	✓	✓	⃠	✓	⃠	⃠	⃠	⃠	✓	✓	✓	⃠	⃠	⃠	⃠	✓	⃠
8	Cost measurement	✓	⃠	✓	⃠	⃠	⃠	⃠	⃠	✓	⃠	⃠	✓	⃠	⃠	⃠	⃠	⃠	✓	⃠	⃠	⃠
9	Cost valuation	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠
10	Outcome identification	✓	✓	✓	✓	✓	⃠	⃠	✓	✓	✓	✓	⃠	✓	✓	✓	⃠	⃠	✓	⃠	⃠	⃠
11	Outcome measurement	✓	✓	⃠	⃠	✓	⃠	⃠	✓	✓	✓	✓	⃠	✓	✓	✓	✓	⃠	✓	⃠	✓	⃠
12	Outcome valuation	n/a	✓	n/a	n/a	n/a	n/a	n/a	n/a	✓	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a
13	Incremental analysis	✓	✓	⃠	⃠	⃠	⃠	⃠	⃠	✓	⃠	✓	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠
14	Discounting	✓	n/a	⃠	n/a	⃠	⃠	⃠	n/a	n/a	n/a	n/a	⃠	n/a	n/a	n/a	n/a	⃠	n/a	⃠	n/a	⃠
15	Sensitivity analysis	✓	✓	⃠	⃠	✓	⃠	⃠	⃠	✓	⃠	✓	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	✓	⃠
16	Conclusions	✓	✓	✓	⃠	✓	✓	⃠	✓	⃠	⃠	✓	✓	✓	⃠	⃠	✓	✓	✓	⃠	✓	⃠
17	Generalizability	✓	✓	⃠	⃠	⃠	✓	✓	✓	✓	⃠	✓	✓	✓	✓	✓	⃠	✓	✓	⃠	✓	⃠
18	No conflict of interest	⃠	✓	✓	✓	✓	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	⃠	✓	⃠	⃠	⃠	⃠	⃠	⃠
19	Ethical/distributional issues	✓	⃠	✓	⃠	✓	⃠	⃠	⃠	⃠	⃠	✓	⃠	✓	✓	✓	⃠	⃠	⃠	⃠	⃠	⃠
	Number of ✓	16	13	10	4	11	5	4	9	13	5	13	5	9	10	11	5	3	9	2	10	3
	Proportion of ✓*	89%	72%	56%	24%	61%	28%	22%	53%	72%	29%	77%	28%	53%	59%	65%	29%	17%	53%	11%	59%	17%

✓: Sufficient attention is given to this aspect; ⃠: insufficient attention is given to this aspect; n/a = not applicable; *proportions based on number of quality criteria that apply.

The average number and proportion of quality criteria met was 6.7 ± 3.4 (or 39.0% ± 20.6% of the applicable criteria; mean ± SD) for the ten studies published before 2009, 9.4 ± 4.3 (52.9% ± 23.8%) for the 11 studies published in or after 2009, and 8.1 ± 4.0 (46.3% ± 22.9%) for all studies. The linear regression analysis found a significant association between publication year (predictor) and CHEC quality score (dependent variable; β = 0.46, p = 0.026, R² = 0.214) and thus confirmed an increase in the quality of the studies (Figure 3).

Figure 3. Association between publication year (predictor) and CHEC quality score – linear regression analysis results (regression line and 95% confidence interval; data points are studies included in the analysis).

During their review of the study publications, the authors found various methodological shortcomings. Several studies had no ‘real’ comparator group and used simple pre-/post-treatment comparisons to assess health outcomes [24, 32, 36, 37, 41]. Four studies had different comparators for costs and health outcomes [24, 37, 39, 41], whereas one study described the type of intervention based only on the specialty of the patients’ GP [27]. Some studies analyzed costs and outcomes over different time horizons [25, 36–38], while in one study the time horizon was even not reported [41]. Sensitivity analyses were either not done at all [23, 24, 26,27, 28, 30, 34–39, 41], or not properly done for both costs and outcome (e.g. [32,33]). Furthermore, data presented in tables and/or figures were not consistent with the accompanying text [32, 38] or results were contradictorily described [26]. Finally, various studies considered only medication costs for cost assessment [26, 30, 32, 36] which is far from exhaustive.

4. Discussion

The first systematic review on economic evaluations of homeopathic treatments by Viksveen et al was published online in 2013 [17]. To the best of our knowledge, no other comprehensive reviews have since been published in this area. Now, approximately ten years later, the aim was to provide an updated systematic review by including six new studies. In addition, we created a synthesis and summary of the complete clinical and economic data in a hierarchical decision matrix (Figure 2) [21]. Lastly, we added a quality assessment of all studies using a standardized health economic tool based on expert consensus criteria [18].

Beyond the six newly included studies, we also found other publications which offered information about costs and/or clinical outcomes but which did not meet our selection criteria. This included for example a publication describing six case reports which had no comparator group for the assessment of health outcomes and had therefore to be excluded [42]. Another study had to be excluded because it examined only the costs for one comparator group [43] whereas another study was not included because it only assessed costs for both the intervention and comparator groups [44].

In the 21 reviewed studies, we found a great diversity of therapeutic indications, treatments, methodological approaches, and reported outcomes, which meant that an overall summary and evaluation was not obvious. It appears, however, that most of the studies indicated either an improvement of health outcomes under homeopathic treatment, a reduction of costs, or both and thus favored homeopathic treatment over control treatment. Lower proportions of investigations clearly favoring homeopathic treatment were found among studies meeting at least 60% of the applicable CHEC quality criteria, even though four out of the six studies with a CHEC score above 60% ascertained a positive effect of homeopathic treatment (adjuvant or alone) on clinical outcomes [20,22,29,35]. Of these one also showed lower costs for homeopathy [29] and two showed homeopathic treatment to be cost-effective by incremental analysis [20,22].

Health economic studies are important for all therapeutic fields including CIM [15,16], and high quality is a prerequisite for using these data meaningfully [17]. For the scope of this review, all 21 studies identified were assessed as economic evaluations, since they measured both costs and outcomes for two or more interventions (study selection criterion). With the CHEC list, a face-valid health economic tool [18] was applied to assess their methodological quality. The quality of some identified studies was very low, which was reflected by their low score on the CHEC list (<30% for two-fifths of them [24,26,27,30,32,36,37,39,41] compared to an average of 46.3% for all studies), and various methodological deficits were found. Encouragingly, however, a clear trend toward higher quality standards in more recently published research was seen. Indeed, a look at the two most recent publications showed that they belonged to those with the highest quality ratings according to CHEC (89% and 72%). One re-analyzed the data of a randomized, controlled trial [20] while the other used the data of a large cohort study [22]. In addition to being multicenter studies, they both performed a comprehensive analysis of costs, taking into account direct health-care costs (costs of medication, consultation fees, hospital and other care), direct non-medical costs (e.g. for transportation), as well as indirect costs such as sick-leave payment and reduced productivity. Also, in both studies, the research question was clearly stated, the economic study design well-chosen, important outcomes identified and measured appropriately, the time horizon suitably chosen, and incremental analyses as well as sensitivity analyses performed. One study [22] even considered quality-of-life measures as outcomes for calculating quality-adjusted life years associated with homeopathic treatment.

The positive trend in health outcomes and costs found in our results are consistent with those of a review on economic evaluations reporting cost-effectiveness of CIM therapies, particularly in higher-quality studies. Even if CIM included homeopathic treatment, the authors of this review did not report outcome data separately for homeopathy [45]. The methodological issues described in our review are almost identical to quality issues recently identified in studies of other therapy concepts such as acupuncture [46] or Traditional Chinese Medicine (TCM) [47]. Also, in these fields there is an important need for high-quality economic research to meet the needs of health decision makers [46,48].

The limitations of the present investigation lie within the nature of systematic reviews: the search procedure was broad and delivered a high number of hits. Nevertheless, it cannot be ruled out that the search terms ‘cost’ and ‘homeopathy’ may have missed studies if the key words were not indexed accordingly. For example, the study by Colombo et al [23] could only be found by manual search, since ‘homeopathy’ does not appear as an indexed term. Moreover, relevant cost-effectiveness data may not be publicly accessible (e.g. register data: information in trial registers of planned or ongoing studies with cost-effectiveness as an endpoint) or only available as congress abstract (e.g [49,50]). Also, it cannot be ruled out that economic evaluations which report that homeopathy is not cost-effective are less likely to be published. This problem, also known as publication bias, affects all areas of medical research and is one of the elements which can distort evidence by overestimating the benefits of a treatment [51]. Finally, the external validity of our conclusions is limited by methodological deficits of the publications included.

5. Conclusion

In addition to the systematic review of Viksveen et al [17], our systematic review identified six further studies that contribute to the overall evidence on cost-effectiveness of homeopathy. In 14 out of 21 studies, homeopathic treatment favored control treatment in a synthesis of costs and effectiveness. For two studies showing better health outcomes but at higher costs, additional incremental analysis indicated cost-effectiveness of homeopathic treatment. Studies were, however, highly heterogeneous in term of the conditions investigated, the treatments administered, the design of the investigations, and the clinical and economic outcomes assessed.

Although most recent studies yield promising data about the benefit of homeopathy for the public health system, the conclusion drawn by Viksveen et al ten years ago is still applicable today: based on the existing evidence, no firm conclusion about the cost-effectiveness of homeopathy can be drawn. Further high-quality cost-effectiveness studies are needed so that more robust statements may be made.

6. Expert opinion

Ten years after the work of Viksveen et al [17], the favorable cost-effectiveness assessment of homeopathic treatments from a small number of existing high-quality studies continues to be undercut by too many examples of methodologically limited research. Thus, the lessons to be learned from the data and the recommendations for further research into the cost-effectiveness of homeopathic treatments remain essentially the same: the use of properly planned randomized controlled trials would greatly improve the internal validity of the data. Moreover, studies should focus on a well-defined condition (or should be sufficiently large to allow separate assessments for multiple conditions), and study results should yield claims with direct relevance to the respective cohort. This in turn should be ensured by appropriate inclusion and exclusion criteria. For generalizability, single-center studies should be avoided. As far as costs are concerned, studies should investigate all relevant expenses (direct and indirect) and should not be limited to medication costs, in part because particularly in homeopathy, consultation costs may be many times higher than the expenses for medicines. Finally, outcomes should be selected to (a) enable researchers to reliably distinguish treatment success from failure, and to (b) provide a standardized assessment of cost-effectiveness, e.g. based on QALYs.

In 2011, Ostermann et al [52] concluded that more high-quality studies are needed for CIM for a valid overall assessment of cost-effectiveness. Here it is important that health economic aspects are considered in an early planning stage of studies, which would facilitate economic assessments alongside clinical trials as is the case in trial-based economic evaluations. Gold standard trial-based economic evaluations should have an adequate power (among other things, so they can detect differences in costs), a judicious time horizon and sensibly chosen outcomes. Health economic modeling is another powerful tool which can aid decision makers to identify the cost-effectiveness of CIM interventions using different scenarios [53]. Only when research is performed under these conditions can meaningful conclusions about the cost-effectiveness of homeopathic treatments be made.

Over the next five years, we expect to see a further quality improvement in economic evaluations of homeopathic treatments, in line with what we have already witnessed in the most recent studies. This is a necessary prerequisite to persuade policy and decision makers to consider data and results about the cost-effectiveness of homeopathic treatments when they make pricing and/or reimbursement decisions. Furthermore, such research plays an important role in convincing physicians and patients about the added value of homeopathic treatments.

Article highlights

• Evaluations of the cost-effectiveness of homeopathic treatments are needed as a basis for rational healthcare decision-making.

• A total of 21 studies published between 1993 and 2021 were reviewed which assessed the costs and health outcomes of homeopathic therapy for indications such as respiratory tract and ear disorders, atopic dermatitis, chronic arthritis, dyspepsia, and diabetic polyneuropathy.

• Compared to control treatment, most of the studies indicate either an improvement of health outcomes under homeopathic treatment, a reduction of costs, or both, and they thus favor homeopathic treatment.

• The favorable cost-effectiveness assessment of homeopathic treatments results from a few high-quality studies. Due to this low number, no firm conclusion about the cost-effectiveness of homeopathy can be drawn.

• A quality assessment based on the CHEC list shows a trend toward higher-quality research over time. Nevertheless, the methodological quality of more than two-thirds of the studies reviewed was moderate or poor, and more high-quality health economic studies are needed. Shortcomings included poor description of study populations and competing alternatives, failure to capture and evaluate all relevant costs, and failure to perform incremental and sensitivity analysis.

Declaration of interest

J Burkart and S De Jaegere are employees of Deutsche Homöopathie-Union DHU-Arzneimittel GmbH & Co. KG Karlsruhe, Germany. The other authors received honoraria of Deutsche Homöopathie-Union DHU-Arzneimittel GmbH & Co. KG Karlsruhe, Germany.

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Author contributions

All authors equally contributed to the conceptualization of the review. Literature review was performed by J Burkart and C Raak. Data extraction and quality assessment was performed by J Burkart and C Raak. Here C Raak was supported by T Ostermann and J Burkart was supported by S De Jaegere. Disagreements between reviewers were resolved by S Simoens. All authors read, revised, and edited the drafted manuscript and approved the final version.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

The authors thank Andreas Völp of Psy Consult Scientific Services for providing statistical support and medical writing services and the Information Services & Library team of Deutsche Homöopathie-Union DHU-Arzneimittel GmbH & Co. KG Karlsruhe, Germany for conducting the literature search.

Additional information

Funding

This manuscript was funded by Deutsche Homöopathie-Union DHU-Arzneimittel GmbH & Co. KG Karlsruhe, Germany.