Herbal bitters in routine integrative care: a real-world prescription database study from a Swiss hospital

Abstract

Background

In traditional medicine, natural remedies are often used based on empirical observation and experimental knowledge. Even today, the prescription of phytotherapeutics at major centers for integrative medicine, such as Klinik Arlesheim (KLA) in Switzerland, frequently reflects physician experience rather than a purely guideline-based approach. Therefore, prescribing patterns may offer valuable insights into the clinical logic and cultural framework behind the use of specific remedies. This study examines the use of bitter herbal preparations at KLA as a case example of how routine health data can support hypothesis generation and provide insight into therapeutic culture.

Methods

This real-world data study involved the analysis of KLA’s prescription database between 2018 and 2019. A total of 301,821 prescriptions were screened, focusing on 66,038 prescriptions containing 34 predefined natural bitter remedies. The analysis was narrowed to 10 frequently used mono- and multi-herbal bitter preparations. Indications, forms of administration and prescribing clinical departments were analysed descriptively.

Results

A total of 10,660 prescriptions of bitter preparations were identified. Bitter preparations were most frequently prescribed in psychiatry and oncology particularly for patients experiencing anxiety, stress and stress-related nausea. This pattern diverges from their traditional gastrointestinal use and may reflect evolving clinical applications in integrative care. The study also demonstrates the potential of structured clinical data to surface novel patterns and research hypotheses.

Conclusion

The prescribing patterns of bitter substances observed in this study highlight underexplored areas of integrative care, particularly in psychosomatic and oncological contexts. These findings underscore the value of structured, practice-based data as a window into lived therapeutic traditions. Future research should examine both the clinical effects and efficacy as well as the cultural dimensions of phytotherapeutic use, including the unconventional administration of herbal bitters.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12906-025-05237-7.

Introduction

In traditional medicine, remedies are often selected based on empirical experience. Historically, purely experience-based medicine applied what nature provided, and its effectiveness was evaluated through subjective impressions and observations. Medical myths and rituals were recorded and passed on. Accordingly, the treatments that repeatedly proved effective were mentioned more frequently and gained a permanent place in traditional medicine. Varying existing treatments or trying out new approaches could thus be successful, but sometimes devastating. Depending on the interests of the authorities on traditional beliefs, ineffective or harmful treatments could persist.

In contrast, the goal of contemporary medical science is to clearly quantify the pharmaceutical and non-pharmaceutical effects on physiology and rigorously evaluate their efficacy. However, there has been growing recognition of the importance of synergistic effects, which can amplify, diminish or even enable certain therapeutic outcomes. These interactions are particularly pertinent in integrative medicine, in which the combined effects of different treatments are often central to their efficacy. If traditional remedies show convincing efficacy with tolerable side effects in modern drug trials, they can continue their path to market approval. It is crucial to ensure the safe and effective use of traditional medicine through regulation, research and, where appropriate, the integration of traditional medicine products, practices and applications into health systems.

The Swiss population is open to complementary and alternative medicine (CAM) approaches [1, 2]. Many patients are treated with corresponding therapies, and CAM costs have been covered by corresponding supplementary insurance policies for years. In addition, health policy efforts have led to the inclusion of medical services such as acupuncture, anthroposophic medicine (AM), traditional Chinese medicine (TCM), classical homeopathy and phytotherapy as CAM therapies into the statutory basic insurance benefits package since August 1^st, 2017 [3].

AM has always seen itself as a form of integrative medicine, with the first AM developments taking place in the 1920s in Switzerland at Klinik Arlesheim (KLA), which was founded over 100 years ago. Many AM treatments are influenced by aspects of traditional medicine – not only traditional phytomedicine but also traditional medical methods, such as compresses, external embrocation, medical dietetics and balneological applications, which have been further developed based on a holistic understanding of the anthroposophical view of humans [4, 5]. Today, KLA has a large outpatient centre with various specialties and inpatient areas for internal medicine, oncology, palliative medicine, psychiatry and psychosomatics. In 2016, a digital hospital information system (HIS) was installed at KLA that completely replaced medical records in their previous paper form. This HIS includes digitally recorded prescriptions for medication, which can be viewed and evaluated separately. A wide variety of phytotherapeutic medicines are prescribed at KLA. The reasons for these prescriptions are manifold and are often based on physicians’ practical experience. The prescription activity of the treating physicians alone reflects actual clinical practices regarding the use of certain phytotherapeutics.

The underlying assumption of this study is that some therapeutics are perceived as effective in daily clinical work but have not yet been observed or systematically investigated. Thus, the evaluation of the HIS prescription database is therefore an elegant method that can provide insights into the clinical logic and therapeutic traditions that shape integrative care. In this context, practice-based data are not used to prove efficacy, but rather to shed light on patterns of use and generate research questions based on routine care. Descriptive studies of real-world prescribing practices can thus serve as a kind of clinical ethnography—tracking what is done, where, and for whom—and provide valuable starting points for future research.

When looking at the various prescriptions in KLA’s prescription database, it is striking that many were for a large group of bitter substances. In addition, KLA has operated a ward and an outpatient clinic for gastroenterology for several years. In this study, we therefore focused on bitter substance formulations, which have a long tradition in many cultures and are widely used in folk medicine mainly to stimulate digestion and appetite as well as for detoxification. In traditional European medicine, bitter herbs such as Gentiana lutea (yellow gentian), Artemisia absinthium (wormwood) and Cichorium intybus (chicory) are frequently used as tonic preparations for the treatment of digestive problems, liver dysfunction and general fatigue. In TCM, bitter herbs are considered ‘heat reducing’ and ‘moisture expelling’ and are used to treat infections, inflammation and metabolic disorders. In Ayurvedic medicine, bitter plants such as Andrographis paniculata and Swertia chirata are traditionally used to support liver function and treat fever and skin diseases [6–9].

From a pharmacological point of view, bitter substances, including many alkaloids, iridoids, sesquiterpene lactones and phenol derivatives, act on bitter taste receptors (TAS2Rs) belonging to the G protein-coupled receptor family, which have been detected throughout the gastrointestinal tract, in the pancreas, in the immune system and even in the brain [10, 11]. The activation of these receptors has been linked to increased secretion of digestive enzymes, modulation of gut hormones, anti-inflammatory effects and central nervous system responses. Studies have suggested that the activation of extraoral bitter receptors influences gastrointestinal function, appetite regulation and immune signalling and may also influence autonomic balance through the activation of the vagus nerve [12, 13], thereby suggesting broader therapeutic potential than previously thought.

Due to the apparently unconventional and widespread use of natural remedies in AM, we extensively examine prescriptions for bitter substances across all medical areas at KLA and reach surprising conclusions that underline the usefulness of developing a concept using HIS data. The basis of this concept, including its advantages and possible limitations, is described in this article.

Methods

Ethical statement

The ethics committee clarified that this project should not be classified as a research project according to the Human Research Act Art. 2. The decisive factor for this assessment was that the data originated from the clinic’s prescription database, not from patient files. Therefore, the project did not require the approval of the ethics committee (decision of the Ethics Committee Northwestern and Central Switzerland, Basel, 20/10/2021 Req-2021-01176).

Quality and regulatory status of the prescribed substances

The herbal preparations prescribed by physicians at KLA were either officially registered as commercially available medicinal preparations produced by pharmaceutical companies (Weleda, WALA, Ceres and Iscador AG) and approved by Swissmedic or were magistral preparations manufactured by KLA’s pharmacy in a Swissmedic-certified laboratory in accordance with good manufacturing practice guidelines under Swiss drug law.

Data extraction

For this database study, all physicians’ case-related prescriptions for medication within the two-year period from January 1^st, 2018, to December 31^st, 2019 – a period not affected by the Covid-19 pandemic – were extracted from the HIS prescription database into a Microsoft Access 365 database via open database connectivity. This study database contains information on the drugs prescribed, such as the name, the producer, official and internal identification codes, the form of application (e.g. enteral, parenteral or topical), the dosage and the prescription date and period, as well as a number of other application details, such as the reason for need (RFN) and the diagnosis related to the prescription (ICD10 code). Using structured query language, a comma-separated value (CSV) file was created from the MS Access study database, which contained only data from the selected fields. This CSV file, referred to as the prescription table, was imported into R (version 4.3.2) as a working table with all outpatient and inpatient prescription records and contained only data relevant to the analysis. Further data recoding and statistical evaluations were performed using R and Rstudio (version 2023.12.0) statistical open-source software. Each record in the working table consisted of a patient number and a case number, the medication administered, the official diagnosis description followed by its ICD-10 code, the RFN and the form of medication application. These records could be extended to include further details from the original prescription database, if required. Data visualisation was carried out in the form of tables and charts using Excel 2016 and Graphpad/Prism10, respectively.

Data grouping and recoding

Complex search list extraction was performed using R’s built-in grepl() function for text pattern analysis and pattern matching and replacement. First, all possible natural bitter remedy–related prescriptions were extracted from the overall prescription table. For this purpose, a list of 34 potential natural bitter remedies was compiled using various sources of information, including well-known manuals, official internet resources and online databases [14–16]. In addition, drug directories helped identify potential bitter preparations that could be included in the prescription table [17–20]. The names of medicinal products with the same active substance could vary due to differences in the concentration of the active ingredient or plant extract, the package size or the manufacturer, among other reasons. In addition, some drug names had been entered manually into the system by medical staff, which may have resulted in inconsistencies and entry errors in the database for the same medication. To overcome these problems, the grepl() function was applied and the various prescription names were structured for evaluation based on the natural bitter remedies mainly used in these preparations. First, the original names of the individual prescriptions were combined into new names, which were then distinguished according to preparations that contained the natural bitter remedies as mono- or multi-herbal preparations. This structuring is shown using yellow gentian (Gentiana lutea) as an example in Supplementary Table ST1. Due to the wide range of possible natural bitter remedies that also contain multiple plant extracts, the six most traditional mono-plant bitter remedies were selected for this study. The number of prescriptions was calculated using the group_by() and summarize() functions from the dplyr package in R. The medical field of application, indication, form of application and dosage were set as the study parameters for the dataset.

The reasons and indications for bitter agent prescriptions were obtained using the RFN data fields. Since the information in these fields could be in the form of free-text entries, the RFNs were categorised using the previously used method for assigning free-text bitter names. In this way, the RFNs were grouped into five main categories (Table 1). The number of prescriptions was calculated as before, using group_by() and summarize(). Table 1 forms the basis of the analysis of symptoms.

Table 1.

Grouping of similar original RFN expressions into five categories. The original expressions were translated from German to English, including errors and abbreviations

CATEGORY	ORIGINAL RFN EXPRESSIONS (ASSOCIATED NUMBER OF PRESCRIPTIONS)
Fear, Stress, Unease	Tension (34)	Restlessness (79)
	Fear (36)	Testlessness, tension (2)
	Tension (8)	restlessness/fear (5)
	for restlessness (2)	restlessness/tension (1)
	inner tension (1)	restlessness/circles of thought (2)
	panic attack, restlessness (1)	State of agitation (1)
	Panic attack (1)	before exam (1)
	Stress situation (1)	center, fear (3)
Pain	Headaches (6)	Headahces (101)
	abdominal pain (1)	Headac (3)
	for headache (6)	Headache (4)
	for headaches (1)	Migraine (4)
	for headache (4)	Pain (34)
	Sore throat (1)	Pains (2)
	Haedaches (3)	pain in the hands (1)
	Headache (48)
Nausea	Nausea (1868) dizziness, nausea, headaches (1) nausea, restlessness (2) nausea, inner restlessness (1) nausea, digestion problems (1)	for nausea (7) headache, übelkeit (2) High bp, nausea, headache, restlessness (1) nausea and bloating (2)
Digestive Complaints	Digestive problems (1)	Diarrhoea (12)
	digestive problems/meteorism (2)	Diarrhea (1)
	Bloating (2)	gastrointestinal problems (1)
		gastrointestinal tract problems (41)
	Meteorism (1)	belly/stomach problems (1)
	Heartburn (1)	Support digestion (2)
	Loss of appetite (1) Inappetence (1)
Others	absence (1)	overwhelming emotions, binge eating, vomiting (1)
	if needed (47)
	before each meal (1)	Shortness of breath (1)
	in case of confusion (2)	for low blood pressure (1)
	Confusion (3)	fever (3)
	in case of dizziness (1)	Heat build up in belly (1)
	Dizziness (1)	Spaticity (1)
	Dizziness/hay fever (1)	Sleep problems (3)

Contrary to the RFNs and drug names, the form of application was selected by the physicians from a drop-down list and was not entered manually. Therefore, no particular cleanup or grouping using the grepl() function, as before, was required to view and count the forms of application. Thus, the data could be easily processed using the group_by() and summarize() functions.

To analyse the medical fields, only inpatient prescriptions were considered, since only in cases of inpatient treatments were prescriptions registered together with the patient diagnoses and corresponding ICD-10 codes. The diagnosis associated with each prescription, indicated by an ICD-10 code, was suitable for assigning the medical areas in which the bitter prescriptions were administered.

Prescriptions containing multiple bitter substances were included in the dataset but were not analysed as separate therapeutic combinations. The focus of this study was on the frequency of prescriptions and the clinical contexts of individual substances rather than on the evaluation of a combination of prescriptions. Since many of the combinations were highly variable and tailored to individual patients, no standardised selection criteria could be derived. A separate analysis was necessary to systematically investigate multiple substance patterns and their potential therapeutic rationales.

Statistics

Data from the KLA database were used to perform a descriptive analysis. Based on the selection of bitter preparations compiled in the bitter preparation table, the absolute numbers of prescriptions and cases of bitter preparations, as well as their RFN, application form and medical field of use, were determined. These values were compared to the overall prescription table to obtain insights into the frequency of use of these bitter preparations.

Results

An analysis of the KLA prescription database was carried out for the 2018–2019 period, during which 301,821 prescriptions were recorded (Fig. 1). This study identified 66,038 prescriptions for 34 natural bitter remedies known to be used in bitter preparations (Fig. 1, steps ①–③). Of these prescriptions, 24,162 were recorded for the 11 most commonly used mono-herbal bitter preparations in the KLA prescription database (Fig. 1, step ④). However, they contained numerous different and unusual bittering compositions. For this reason, the selection, based on frequency of occurrence, was limited to six mono-herbal preparations and four multi-herbal preparations, resulting in 10,660 identified prescriptions (Fig. 1, step ⑤; Tables 2 and 3).

Fig. 1.

Data extraction steps, from the prescription table to the final bitter preparation table. Step 1: List the natural bitter remedies that could potentially be found as bitter preparations in the prescription database. Step 2: Identify and list all bitter preparations containing these natural bitter remedies using drug directories. Step 3: Search for these bitter preparations in the prescription database. Matches are linked to their natural bitter remedies and are combined in Bitter Preparation Table 1. Step 4: First sort of bitter remedies, which leads to Bitter Preparation Table 2. Step 5: Final sort, which leads to Bitter Preparation Table 3. This is used to investigate the reason for need and the form of application of these bitter remedies as well as the medical field in which they were employed

Table 2.

Number of prescriptions for mono-herbal and multi-herbal bitter preparations

PLANT	BITTER PREPATATION	NUMBER OF PRESCRIPTIONS	TOTAL
Multi-herbal	Wormwood comp. (KLA)	1478	3786
	Gentiana comp. (KLA, WALA)	1027
	Amara drops (Weleda)	672
	Absinthium/Resina laricis (KLA, Weleda)	609
Strychnos nux vomica	Nux vomica semen (KLA)	1625	2507
	Nux vomica D3-6 (Weleda)	797
	Nux vomica D4-30 (Omida)	70
	Nux vomica e semine (WALA)	15
Gentiana lutea	Gentiana lutea radix (KLA)	2140	2244
	Gentiana urtincture (CERES)	83
	Gentiana lutea eth. decoc. (Weleda)	17
	Gentiana lutea radix (WALA)	4
Cichorium intybus	Cichorium intybus pl. tota (KLA, WALA)	324	651
	Cichorium plumbo cultum (Weleda)	119
	Cichorium (Weleda)	71
	Cichorium stanno cultum (Weleda)	71
	Cichorium intybus urtincture (CERES)	66
Taraxacum officinale	Taraxacum stanno cultum (Weleda)	302	573
	Taraxacum officinale Planta tota (KLA/WALA)	219
	Taraxacum urtincture (Ceres)	52
Cetraria islandica	Cetraria praep. (KLA)	276	551
	Cetraria praeparata (Iscador AG)	275
Angelica archangelica	Archangelica ointment (KLA)	237	348
	Archangelica 10% ointment (Weleda)	87
	Angelica urtincture (CERES)	24

(KLA Klinik Arlesheim, WALA Weleda CERES Omida Natural Products Pharmaceutical Company)

Table 3.

Proportion of inpatients and outpatients by bitter Preparation prescriptions and patient cases

	PRESCRIPTION TABLE			BITTER PREPARATION TABLE
PARAMETER	OUTPATIENT	INPATIENT	TOTAL	OUTPATIENT	INPATIENT	TOTAL
Prescriptions	125,115	176,706	301,821	3685	6975	10,660
	(41.5%)	(58.5%)	(100%)	(2.9%)	(3.9%)	(3.5%)
Cases	15,956	3145	19,101	2708	2305	5013
	(83.5%)	(16.5%)	(100%)	(17.0%)	(73.3%)	(26.2%)

In the bitter preparation table columns, the proportions are always relative to the values in the prescription table columns, while in the prescription table columns, the proportions are relative to the total number in the same part of the table

The following six natural bitter remedies predominantly used in mono-herbal preparations were selected for this study: Strychnos nux vomica, Gentiana lutea, Cichorium intybus, Taraxacum officinale, Cetraria islandica and Angelica archangelica. Accounting for almost half of all prescriptions, Strychnos nux vomica (2507 prescriptions) and Gentiana lutea (2244 prescriptions) were the most frequently prescribed mono-herbal bitter preparations (Table 2). The other four natural bitter remedies – Cichorium intybus, Taraxacum officinale, Cetraria islandica and Angelica archangelica – were slightly less represented, with 651, 573, 551 and 348 prescriptions, respectively (Table 2). In addition, four commonly used multi-herbal bitter preparations (Supplementary Table ST2) were included: wormwood comp., gentiana comp., amara drops and Absinthium/Resina Laricis. These were prescribed most frequently, with a total of 3786 prescriptions (Table 2). For this purpose, a list of 34 potential natural bitter remedies was compiled by using various sources of information, including well-known manuals, official Internet resources, and online databases.

This dataset of bitter preparations was analysed by indication, type of administration, ratio in which they were administered and medical area in which they were used (Fig. 1, step ⑥).

Inpatient and outpatient bitter agent prescriptions

During the study period (2018–2019), prescriptions for the six selected natural bitter remedies accounted for only 3.5% of the overall 301,821 prescriptions, with 10,660 prescriptions (Table 3). However, considering the medical cases, 5013 (26.2%) of the 19,101 identified patient cases were treated with bitter preparations at least once (Table 3). Looking only at inpatients, 2305 of the 3145 inpatient cases (73.3%) were treated at least once with bitter preparations (Table 3). For outpatients, the number of prescriptions for bitter preparations in relation to all outpatient prescriptions was significantly lower, at 17.0% (Table 3). The significantly lower proportion of outpatient prescriptions for bitter preparations can be explained by the fact that outpatients and inpatients were not treated for the same medical conditions and did not receive the same proportion of prescriptions. Indeed, outpatients received eight prescriptions per case, whereas inpatients received 56 prescriptions per case (Table 3).

Reasons and indications for bitter agent prescriptions

Most of the selected bitter preparations are indicated for nausea and digestive complaints, appetite loss and infections of the digestive tract (e.g. gastroenteritis). However, C. islandica is used in cancer treatment and A. archangelica ointments are used for lymph node inflammation or congestion [17]. As expected, 1885 of 10,660 (17.7%) bitter preparation prescriptions were issued for nausea, of which 1768 (16.6%) were issued for S. nux vomica. However, only 67 (0.6%) prescriptions were issued for digestive complaints (Fig. 2A). Looking exclusively at the multi-herbal preparations, there were 99 prescriptions (0.9%) for nausea and 57 (0.5%) for digestive problems (Fig. 2B). Other indications were grouped as pain, fear/stress/unease and others, with 219, 178 and 68 prescriptions recorded for all preparations and 135, 81 and 43 recorded for only multi-herbal preparations, respectively (Fig. 2A and B). Wormwood comp. was most frequently represented in these three categories (Fig. 2B). Unfortunately, only one in four bitter preparation prescriptions (22.7%) had an RFN. For C. islandica and A. archangelica prescriptions, the RFM indications were the most incomplete, with 0 and 4 RFNs, respectively. Consequently, important prescription information was sometimes missing from the database.

Fig. 2.

Number of prescriptions of bitter preparations by RFN category, considering all bitter preparations (A) or only multi-herbal bitter preparations (B). The numbers above the bars indicate the total number of prescriptions for the respective preparations. Mono-herbal preparation plant names and multi-herbal preparations names are shortened as follows: Strychnos = Strychnos nux vomica, Gentiana = Gentiana lutea, Taraxacum = Taraxacum officinalis, Angelica = Angelica archangelica, Cichorium= Cichorium intybus, Cetraria = Cetraria islandica, Wormwood cp = Wormwood comp., Gentiana cp = Gentiana comp., Amara = Amara drops and Abs./Res. = Absinthium/Resina laricis

Forms of application

With more than 8740 prescriptions, most of the bitter preparations were administered enterally (Fig. 3). Almost all mixtures (3732 prescriptions), as well as mono-herbal medications of G. lutea, S. nux vomica, C. intybus and T. officinale (2040, 1833, 623 and 477 prescriptions, respectively), were included in this category (Fig. 3). S. nux vomica was also used subcutaneously and parenteral subcutaneously (357 and 290 prescriptions, respectively). The RFN for the use of S. nux vomica in all three application forms was nausea (not shown). A total of 307 prescriptions, almost exclusively A. archangelica ointments, were administered topically (Fig. 3). Psychiatric physicians stated that many applications were topical, using compresses. Unfortunately, this could not be verified from the database. One preparation, G. lutea as a mono-herbal preparation, was prescribed by inhalation (162 prescriptions, Fig. 3). The anti-cancer drug C. islandica was prescribed only for subcutaneous use (Fig 3).

Fig. 3.

Number of prescriptions of bitter preparations by form of application. The numbers above the bars indicate the total number of prescriptions for the respective preparations. Mono-herbal preparation plant names and multi-herbal preparations names are shortened as follows: Strychnos = Strychnos nux vomica, Gentiana = Gentiana lutea, Taraxacum = Taraxacum officinalis, Angelica = Angelica archangelica, Cichorium = Cichorium intybus, Cetraria = Cetraria islandica, Wormwood cp = Wormwood comp., Gentiana cp = Gentiana comp., Amara = Amara drops and Abs./Res. = Absinthium/Resina laricis

A total of 148 bitter preparations were administered in other forms (data not shown), such as intravenously (30 prescriptions) and buccally/sublingually (36 prescriptions). However, for 63 prescriptions, the form of application was unclear, since they were categorised as ‘general’ and ‘other’.

Medical field

To evaluate the bitter preparation prescriptions by medical field, the relative frequency per patient case was of particular interest. For this purpose, the total number of prescriptions and the number of bitter preparation prescriptions for the associated medical cases were related to the different medical fields (Table 4). In addition, the number of patient cases per medical field was analysed for bitter preparations (Fig. 4). The proportion of patient cases treated with bitter preparations was the highest in psychiatry and oncology/immunology, with 663 and 765 cases, respectively. In both fields, more than 80% of the patient cases received bitter preparations at least once. Most bitter preparations were prescribed in psychiatry (2635 prescriptions), with an average of four prescriptions for bitter preparations per patient case. In this field, multi-herbal preparations were mainly used (Fig. 4). By far the most frequently used multi-herbal preparation in this field was Wormwood comp. (not shown). S. nux vomica was the most commonly used mono-herbal bitter preparation (Fig. 4), which was frequently employed in oncology or for the treatment of therapy-associated nausea (Fig. 2A). C. islandica was also prescribed exclusively for cancer therapy (Fig. 4).

Table 4.

Number of prescriptions, patient cases and prescriptions per case by medical field

	PRESCRIPTION TABLE			BITTER PREPARATION TABLE			BITTER PREPARATION TABLE/ PRESCRIPTION TABLE
Medical Field [ICD-10 letter]	Prescriptions	Cases	Prescriptions/Case	Prescriptions	Cases	Prescriptions/Case	Prescriptions (%)	Cases (%)
Psychiatry [F]	47,309	804	58.8	2635	663	4.0	5.6	82.5
Oncology/immunology [C–D]	58,062	938	61.9	2239	765	2.9	3.9	81.6
Respiratory system [J]	15,369	269	57.1	486	193	2.5	3.2	71.7
Circulatory system [I]	16,285	297	54.8	350	160	2.2	2.1	53.9
Digestive system [K]	7591	164	46.3	270	103	2.6	3.6	62.8
Infection [A–B]	5060	101	50.1	192	79	2.4	3.8	78.2
Bones and muscles [M]	8022	127	63.2	177	84	2.1	2.2	66.1
Uncategorised symptoms [R]	3970	122	32.5	158	67	2.4	4.0	54.9
Urinary tract and genitals [N]	4755	88	54.0	153	58	2.6	3.2	65.9
Nervous system [G]	4893	95	51.5	135	59	2.3	2.8	62.1
Injuries [S–T]	2126	59	36.0	61	30	2.0	2.9	50.8
Endocrinology/ metabolism [E]	1794	36	49.8	59	22	2.7	3.3	61.1
Eyes and ears [H]	859	23	37.3	28	12	2.3	3.3	52.2
Pregnancy [O]	179	4	44.8	20	3	6.7	11.2	75.0
Skin [L]	356	8	44.5	10	5	2.0	2.8	62.5
Health-harmful factors [Z]	59	9	6.6	2	2	1.0	3.4	22.2
Genetic [Q]	17	1	17.0	0	0	0.0	0.0	0.0
	176,706	3145		6975	2305

The ratios of prescriptions and cases in each medical area between the bitter preparation table and the prescription table are shown. The table ranges from the highest prescription values to the lowest prescription values in the bitter preparation table

Fig. 4.

Number of cases treated with bitter preparations by selected medical fields (with the ICD-10 letter code for the corresponding primary disease categories in square brackets). The numbers above the bars indicate the total number of prescriptions for the respective preparations. A Mono-herbal preparation plant names and (B) multi-herbal preparations names are shortened as follows: Strychnos = Strychnos nux vomica, Gentiana = Gentiana lutea, Taraxacum = Taraxacum officinalis, Angelica = Angelica archangelica, Cichorium = Cichorium intybus, Cetraria = Cetraria islandica, Wormwood cp = Wormwood comp., Gentiana cp = Gentiana comp., Amara = Amara drops and Abs./Res. = Absinthium/Resina laricis

Discussion

In this study, we analysed a prescription database to demonstrate a way of using an enormous amount of data generated by digitisation. We hope that this contribution inspires new ideas for creatively and compliantly handling the vast amounts of data generated in today’s healthcare system, while also highlighting additional useful ways in which data collection can support research and development.

The value of optimised data collection and the use of health-related data is briefly explained in the following paragraphs, using the example of bitter substances.

Overall, research on the therapeutic potential of natural bitter remedies is still in its infancy. However, clinical experience of the possible positive effects of natural bitter remedies, such as at KLA, can make a valuable contribution. In particular, the HIS implemented at KLA in 2016 gave us the opportunity to closely examine routine clinical prescription practices. We were particularly surprised that most bitter substances were prescribed in psychiatry and oncology. The frequency with which bitter preparations were prescribed in these two medical fields shows a clear extension of the traditional use of these substances. In oncology, patients are often confronted not only with physical complaints but also with profound psychological stress. Understanding the potential role of bitter compounds in alleviating such complex symptoms could provide new avenues for supportive cancer care.

Similarly, bitter remedies may be prescribed in psychosomatic and psychiatric departments based on the therapeutic assumption that improving gastrointestinal function can help stabilise emotional and autonomic regulation. In integrative medicine, and especially in anthroposophic approaches, digestion and metabolism are seen as central to restoring inner balance and physical resilience. Bitter substances are often used to reduce nervous tension, promote restful sleep and strengthen the patient’s ability to cope with psychological stress. This may explain why prescriptions cluster in fields in which patients frequently suffer from both metabolic and psychological dysregulation.

The potential use of bitter compounds in the treatment of mental health problems such as anxiety, depression and post-traumatic stress disorders highlights a broader therapeutic spectrum of these substances than originally assumed. Further research in this area could provide valuable insights into the psychophysiological interactions of bitter compounds and support this practice with evidence if it is found to be effective.

To gain deeper insights into the practical application and implications of our findings, we engaged in a detailed discussion with physicians in KLA’s psychiatric department. Their firsthand experiences and observations provided valuable perspectives on the use of bitter preparations in psychiatric treatments. The physicians noted that the prescription and dispensing of medications to patients could not necessarily be compared with the patients’ actual intake of the medication. This is because many prescriptions for bitters are issued on an as-needed basis. The medication is provided by nursing staff when patients are admitted to prevent nausea, indigestion, restlessness, fear, panic and sleep problems, among others. Consequently, not all prescribed medications are actually administered to patients, since they may be discontinued or not taken at all. This study focuses on prescribing practices, while medication adherence is a separate but important aspect, particularly in outcome studies. Information on the cancellation of prescriptions and refused receipts is included in the electronic patient records in the HIS and could be deserving of future analysis.

In summary, the observed prescribing patterns – in particular, the use of bitter herbal remedies such as Strychnos nux vomica, Gentiana lutea and Cichorium intybus for symptoms such as anxiety, restlessness, nausea and digestive complaints – are partly supported by new pharmacological and preclinical findings. TAS2Rs, originally described in organs with a connection to digestion, are now known to be present throughout the body, including the gastrointestinal tract, skin, central nervous system and even the skin [10, 11]. For example, bitter substances from Gentiana lutea (e.g. amarogentin) bind to TAS2R1 and TAS2R38 outside the oral cavity, where they stimulate intracellular calcium signalling, epithelial differentiation and lipid synthesis in human skin and intestines [21]. In vivo experiments have shown that the activation of TAS2R in the smooth muscles of the gastrointestinal tract can delay gastric emptying and modulate motility – mechanisms relevant for nausea and digestive tone [22]. In addition, certain bitter compounds have been linked to immunomodulation and mild mood enhancement through the release of enteroendocrine hormones, thereby suggesting communication pathways between the gut and brain [23, 24]. Although robust clinical studies are limited, these mechanistic findings and longstanding traditional uses lend plausibility to the clinical expansion of bitter remedies in integrative settings and underscore the need for targeted, hypothesis-driven clinical research.

In general, the physicians were able to confirm that the indications for bitter preparations were very wide-ranging and could easily be reconciled with the quantitative prescription results identified.

Using insights from clinic data to design new research projects is in line with the growing interest in integrative medicine, which seeks to combine traditional and modern therapeutic approaches for holistic patient care.

However, the data processing carried out in this study showed that the quality, accuracy and efficiency of data collection for meaningful, secure and efficient documentation of healthcare data needs to be improved. For example, it is crucial that the diagnoses of both outpatients and inpatients are correctly recorded as ICD-10 codes [25, 26].

In this study, we showed how prescription data can be used for research and that evaluating prescription data can provide information about hidden treatment practices and generate ideas and hypotheses. Since no health-related patient data from electronic medical records are used, the analysis of prescription data can be used to plan clinical studies without ethical concerns.

In addition to prescriptions, indications should be recorded. This would greatly enhance the value of the prescription database for researchers, administrators and practitioners. Indications and other recurring entries should be selectable via pull-down menus whenever possible to avoid input errors.

Limitations

This study has several limitations. First, it was conducted at a single centre with a strong integrative and anthroposophical background orientation, which may limit the transferability of the results to other conventional clinical contexts. The prescribing patterns observed were influenced by the therapeutic paradigms and the training of the physicians at KLA and do not necessarily reflect the practice in conventional settings. Second, while the study provides descriptive data on prescription frequency and associated diagnoses, it does not allow for causal inferences or direct evaluations of clinical outcomes. Third, due to its retrospective design, the data quality depended on the accuracy and completeness of documentation in the hospital information system.

Conclusions

Scientific progress in the areas of health promotion, prevention, diagnosis, and therapy depends not only on basic and clinical research, but also on feedback from real-world clinical practice. This study contributes to the growing recognition that observational data from routine care can generate meaningful research hypotheses. Practice-based documentation, as demonstrated here, provides large volumes of health-related data that - when accurately recorded - can serve as a foundation for future targeted clinical trials.

Our analysis of prescribing patterns at the hospital Klinik Arlesheim, a leading center for integrative medicine in Switzerland, revealed that bitter herbal remedies were frequently used not only for gastrointestinal complaints, but also in psychiatry and oncology. These findings point to areas of current medical practice that have been little explored in the research literature. Rather than proving efficacy, our study shows how structured observation of prescribing behavior can reveal common therapeutic traditions and stimulate future investigations into their clinical and cultural dimensions.

Abbreviations

AM

 Anthroposophic medicine

CAM

Complementary and alternative medicine

CSV

comma-separated values

KLA

Klinik Arlesheim

HIS

Hospital information system

RFN

Reason for need

TAS2Rs

Bitter taste receptors

TCM

Traditional chinese medicine

Authors’ contributions

DK and CG: conceptualisation; SN: data curation; SN, DK and AZK: formal analysis; SN: investigation; SN, DK and CG: methodology; AZK, DK and CG: project administration; DK and CG: resources; SN and AZK: software; CG: supervision; DK and CG: validation; SN and AZK: visualisation; CG: funding acquisition; SN, DK and AZK: writing – original draft; SN, DK, AZK and CG: writing – review and editing.

Funding

SN, AZK and CG are supported by PRIAM-BS (Verein Stiftungsprofessur für Integrative und Anthroposophische Medizin an der Universität Basel). DK has received research funding from Weleda AG over the past 5 years.

Data availability

The data analysed in this study are subject to the following licenses/restrictions: The dataset contains information that could compromise the privacy of hospital patients. Any request to access these datasets should be directed to Prof. Carsten Gründemann at carsten.gruendemann@unibas.ch.

Declarations

Ethical approval and consent to participate

The ethics committee clarified that this project should not be classified as a research project according to the Human Research Act Art. 2. The decisive factor for this assessment was that the data originated from the clinic’s prescription database, not from patient files. Therefore, this project did not require the approval of the ethics committee (Decision of the Ethics Committee Northwestern and Central Switzerland, Basel, 20/10/2021 Req-2021-01176).

Consent for publication

None.

Competing interests

The authors declare no competing interests.