Romania is recognized as one of the most biogeographically diverse countries of the European Union, having Alpine, Continental, Pannonic, Pontic, and Steppic regions. These specific geographic features combined with its modified temperate continental climate provide the conditions for a diverse flora (3,829 vascular and 979 non-vascular spontaneous plant taxa) (Hurdu et al., 2022), also characterized by a great abundance of medicinal and aromatic plants. A recent evaluation concluded that in Romania there are more than 750 species of plants with medicinal properties (Dihoru and Boruz, 2014).

Southern Romania includes two regions, known historically as Walachia and Northern Dobruja (Figure 1). Walachia covers two historical regions, Muntenia (Greater Wallachia) and Oltenia (Lesser Wallachia), and has about 10,784,874 million people in an area of 101,248 km². It is worth mentioning that the South-Eastern Carpathians alone, partially belonging to Southern Romania, represent one of the major biodiversity hotspots in Europe (Bálint et al., 2011; Dihoru and Boruz, 2014; Hurdu et al., 2022).

Near this area, in Northern Dobruja, there is the Danube Delta, a World Heritage (https://whc.unesco.org/en/list/588/). Danube Delta represents a unique biosphere reserve and wetland ecosystem, quoted as the best-preserved delta in Europe, and the third largest biodiversity area in the world, with over 5500 flora and fauna species (Ciocarlan, 1994; Claudino-Sales, 2019).

A survey instrument was developed by a panel of researchers with a medical background and extensive knowledge in phytotherapy and ethnomedicine, in accordance with other studies related to analogous issues (Kennedy, 2005; Raal et al., 2013; Wu et al., 2013). The study population encompassed mothers with a domicile in Southern Romania whose children fell in the age category 0–18 years and was conducted between September 2019 and December 2022 in “Grigore Alexandrescu” Children Emergency Hospital, a tertiary hospital in Bucharest, Romania. The project was approved by the Ethics Commission of ‘Grigore Alexandrescu’ Children Emergency Hospital (no. 3/1736). Ethnobotanical and ethnomedical investigations were carried out following the ethical guidelines outlined by the International Society of Ethnobiology (ISE) Code of Ethics (The ISE code of ethics, International Society of Ethnobiology, 2006).

Data collection was based on structured interviews (Etkin, 1993; Heinrich et al., 2009) applied to mothers of children hospitalized for various diseases. Responses were recorded by a health professional with expertise in phytotherapy and ethnomedicine. The questionnaire was strictly confidential, anonymized, and non-compulsory. The purpose, methodology, and objective of the study were explained to all participants, and written informed consent was obtained subsequently. The interviews occurred in an individual manner and took between 20 and 30 min to complete.

The questionnaire comprises structured open-ended and close-ended questions regarding the medicinal plant species used, their indications and ethnopharmacological activities, sources of knowledge, plant part(s) used, type of preparation, routes of administration, concomitant use of conventional drugs, and treatment outcomes (for further details, please refer to the Table 1, Supplementary Material). The demographic data of each participant (age, gender, area of residence, level of education) were recorded, together with the age when the children first received the herbal remedy. Standardized ethnobotanical procedures that confer transparency (e.g., availability of information on the frequency of citations), important to assess the relative cultural rank of a species, were followed (Heinrich et al., 2009). Since the present study was essentially a survey conducted within a hospital unit, the collection of voucher specimens was not feasible. The specific study design resembled that of other studies, such as surveys evaluating the use of herbal remedies by pharmacy customers (Knotek et al., 2012; Raal et al., 2013; Zeni et al., 2017) or by members of various minorities (Bhamra et al., 2017). The assessment of herbal authenticity was not within the scope of this study. The accuracy of the taxonomy and nomenclature of vascular plants was verified by cross-referencing with World Flora Online (WFO) (www.worldfora online.org).

Following the collection and analysis of the questionnaires, we compiled a list of medicinal plants based on informant input. To assess the conformity and/or novelty of our results, we conducted a systematic search of the ethnomedical or bioscientific data on the identified medicinal plants. This comprehensive search was carried out across PubMed and Google Scholar databases. Our strategy involved using specific phrases such as: [(Latin name of the plant) OR (vernacular name of the plant)] AND (children OR toddler OR pediatrics OR adults OR indication OR biological activity). For example, a search query could include (Foeniculum vulgare OR fennel) AND (children OR toddler OR pediatrics OR adults OR indication OR biological activity). This method enabled us to analyze existing literature and evaluate the medicinal properties and potential applications of the identified plants within the age groups of interest across various medical contexts, and in adulthood, specifically in the pathologic spectrum reported by our informants.

In this article, the terms “variable” and “parameter” will be considered equivalent and used interchangeably. The same holds for the terms “association” and “correlation.” The statistical analysis consisted of both descriptive and inferential statistics. Descriptive statistics included the calculation of the median and interquartile range (from the first to the third quartile) to characterize numerical parameters such as mother age, used plants count, treated ailments count, and child age when the first herbal treatment was given.

Based on the number of plants used, the mothers were divided into two groups: one comprising those who utilized one or two plants (n = 191), and the other consisting of those who employed more than two plants (n = 135). Similarly, concerning the number of treated ailments, two groups of informants were defined: one including those who used plants for treating one or two ailments in their children (n = 202), and the other, those who treated more than two diseases (n = 124).

Shapiro-Wilk test was used to assess the normality of data distribution. Inferential statistics encompassed univariate analysis (UVA) and multivariate analysis (MVA). UVA included: 1. regression by Spearman’s method, used to estimate the correlation between two numerical parameters 2. Mann-Whitney (Wilcoxon) test with continuity correction for exploring the correlation of various numerical parameters with binary (i.e., with only two possible values) categorical parameters 3. Kruskal-Wallis test, employed to evaluate the correlation of numerical parameters with multivalent (i.e., with more than two possible values) categorical parameters 4. Fisher’s exact test is used to explore the correlations between various categorical parameters. In particular, Fisher’s exact test was utilized to determine whether there is an association between the social background (rural vs. urban) and the employment of medicinal plants. The strength of the association was estimated by odds ratio. The accompanying 95% confidence intervals for odds ratio are an estimation of the real value in the population: there is a 95% probability that the real value in the population falls within the confidence interval. Of course, it is difficult to decide whether the mothers in this tertiary care pediatric hospital are a representative sample of the entire population of mothers.

When multiple comparisons were performed by Mann-Whitney test, the significance level (commonly set at 0.05) was lowered according to Bonferroni correction: the corrected significance level was determined by dividing 0.05 by the number of comparisons. In cases where Kruskal-Wallis test yielded a significant result, pairwise Wilcoxon tests were subsequently conducted using Benjamini-Hochberg procedure to reduce the false discovery rate.

After having established by UVA the variables significantly associated with used plants count, treated ailments count, and purchase method, MVA was performed to determine the variables independently associated with each of these parameters. The method employed for MVA was backward stepwise regression. For each dependent variable, regression was first performed on all n (supposedly) independent variables pointed out by UVA as being statistically significantly associated with the dependent variable. Among the variables proven to lack a statistically significant (p > 0.05) contribution, the one with the smallest contribution (which was also the variable with the smallest t_value) was eliminated. Subsequently, regression was performed again on the remaining n - 1 variables, and once more, the variable with the smallest contribution was eliminated. This process continued iteratively until all remaining variables were statistically significantly associated with the dependent variable. The steps of the MVA for all three dependent variables are shown in Table 1.

All statistical calculations and resulting graphical representations were performed using R language and environment for statistical computing and graphics (version 4.2.3.). The employed packages were dplyr, tidyverse, ggplot2, ggpubr, mosaic, e1071, and mediation.

We divided the data regarding indications into use categories of diseases (Table 2), according to the International Classification of Primary Care (ICPC)(WHO Second edition ICPC-2, 2012), which was adapted to fit the ethnomedical reality (Staub et al., 2015).

The collected data were analyzed using quantitative methods. Use-Value Citation Index, Fidelity Level, and Informant Consensus Factor were calculated (Friedman et al., 1986; Phillips and Gentry, 1993; Hoffman and Gallaher, 2007; Heinrich et al., 2009; Hajdari et al., 2018; Trotter and Logan, 2019).

The Use-value citation index (UV_c) evaluates the relative importance of each species based on its cited uses, and it is calculated for all taxa with the formula: UV c = ∑ U is / N , where U_is is the sum of the total number of all individual use citation reports concerning a given taxa, divided by the total number of informants (N) (Phillips and Gentry, 1993; Hoffman and Gallaher, 2007).

The fidelity level (FL) represents the percentage of informants reporting the usage of a certain medicinal plant for the same major indication (Friedman et al., 1986) and was calculated to determine the most important species used to treat the most frequently reported diseases or ailments. FL  % = N p / N × 100 , where Np is the number of informants reporting the usage of a plant species to treat a particular disease, and N denotes the total number of informants utilizing the plant as a treatment for any given disease (Friedman et al., 1986; Mahomoodally and Sreekeesoon, 2014).

The categories selected for the Informant Consensus Factor (F_ic) analysis are outlined in Table 3. Each taxon’s usage was assigned to the relevant category before analysis, employing the following formula: F ic = N uc ‐ N t / N uc ‐ 1 , where N_uc is the total number of use citations in each category and N_t is the number of taxa used in that category. F_ic estimates user consensus regarding the medicinal plants. The value of this factor ranges from 0 to 1. High F_ic values (close to 1.0) are obtained when relatively few species are reported to be used by a large proportion of informants for a particular nosological category, whereas lower F_ic values indicate that informants disagree upon the taxa to be used in the treatment within a certain category.

Binomial distribution was used to calculate the confidence limits for the various percentages yielded by our study sample.

A total of 550 mothers living in Southern Romania were initially interviewed with a median age of 37 (interquartile range (IQR) of 31–42 and a skewness of −0.2194345). Among them, 224 respondents (40.72%, 95% confidence limits (95% CL): 38.63–42.82) did not utilise herbal remedies based on raw-plant material in treating their children (Figure 2). Reasons for this included: not being aware of this option (187 mothers, among whom 138 were nonetheless interested in using plants to treat their children), having learned from media that plant treatment is not useful or may even be toxic (22 mothers), and considering herbal remedies as inefficient (15 mothers).

Additionally, 74 informants indicated advertising of the pharmaceutic products as one of the factors influencing their decisions. Concerns were raised by 53 informants about plant contamination by pollutants. They declared that due to the lack of time and opportunity to harvest plants of trusted quality, they used pharmaceutical agents based on plant extracts, but mixed with synthetic compounds, vitamins, and minerals that are generally recommended by physicians, pharmacists, friends, and in the media. Due to the mixed nature (herbal and non-herbal/natural and synthetic) of the medicines used, we decided to exclude them from enrollment in the study.

The employment of medicinal plants was declared by 177 and denied by 76 out of 253 mothers living in rural communities. Out of 297 mothers living in urban areas, 149 have confirmed the employment of medicinal plants, while 148 have denied it. Fisher exact test yields a less than 0.00001 p-value, indicating a strong association between social background and the utilization of medicinal herbs. Indeed, mothers of rural background have a higher propensity to use medicinal plants compared to mothers of urban background, as reflected by the odds ratio of 2.31 with a 95% confidence interval of 1.63–3.29.

The remaining 326 mothers (59.27%, CL95%: 57.18%–61.37%), who declared the use of medicinal plants, used plant treatment based on raw vegetal material (self-harvested or bought) for their children’s illnesses and were included in the study (Figure 2).

These herbal therapy users originated from 14 counties of Southern Romania: Argeș (18), Braila (4), Bucuresti (114), Buzau (6), Calarasi (20), Constanta (6), Dambovita (21), Dolj (7), Giurgiu (21), Gorj (6), Ialomita (19), Ilfov (29), Mehedinti (2), Olt (8), Prahova (19), Teleorman (18), Tulcea (3), Valcea (5) (Figure 1).

Approximately one-third (37%) of the participants resided in rural areas, while the remaining majority (63%) lived in urban areas. In terms of age distribution, 97% of the mothers were above 18 years old, with a significant majority falling into the 35–39 and 40–44 age groups. Regarding educational attainment, 4% of the respondents had no formal education, 18% of them had completed primary education, 57% had attained secondary-level education, and 21% held a college degree (Table 4).

From this point on, the 95% confidence limits were calculated relative to the population of herbal treatment employing women. In 117 cases (35.89%, 95% CL: 33.23%–38.55%) the treatment was recommended by a pharmacist or a physician, in 128 cases (39.26%, 95% CL: 36.56%–41.96%) by family members, in 45 cases (13.80%, 95% CL: 11.89%–15.71%) by friends, and in 36 cases (11.04%, 95% CL: 9.3%–12.78%) the information was found in media or on the Web (Figure 3A). Interestingly, more than one-third of the informants (40.49%, 95% CL: 37.77%–43.21%) stated that herbal remedies were the only treatment used, whereas the others (59.51%, 95% CL: 56.79%–62.23%) associated plants with allopathic drugs. In 196 of the cases, children were given both raw plant-based remedies and pharmaceutical drugs based on plant extracts.

An impressive 88% (95% CL: 86.2%–89.8%) of the informants reported an improvement in the disease course following herbal therapies and consequently declared their intention to use and recommend the use of plant-based remedies in the future.

The collected data revealed that a total of 25 medicinal plant species from 15 families were used to treat pediatric diseases. Plant indications and activities, parts used, way of administration, UVc, FL, and targeted body systems are provided in Table 5.

Apiaceae and Asteraceae were the dominant families with four species each followed by Amaryllidaceae (three species) and Lamiaceae and Rosaceae (two species). The highest proportion of the medicinal plants used is herbaceous (76%), followed by trees (12%), shrubs (8%), and grasses (4%). The remedies were made from fruits, leaves, roots, seeds, stems, buds, flowers, bulbs, or sap. In some cases, a combination of plants was administered.

Only 18% (95% CL: 15.87%–20.13%) of the probands have harvested themselves the plants cultivated in their gardens or growing in the surroundings, whereas the majority of 82% (95% CL: 79.87%–84.13%) have purchased processed plant parts (66%, 95% CL: 63.38%–68.62%) or already prepared remedies (16%, 95% CL: 13.97%–18.03%) from pharmacies, specialized stores, and local markets (Figure 3B).

Various methods of preparation were used for the 25 medicinal plant species: decoction, infusion, ointment, poultice, and syrup, with infusion as the most preferred method (56%), followed by decoctions and poultice (12%). Conversely, ointments, syrups, and the raw administration of plant parts, sap, or seeds were less frequently utilized, accounting for only 8% of the preparations. Some of the medicinal remedies were administered with the meal, either cooked (e.g., garlic bulbs) or raw (e.g., sea buckthorn fruits). Water was the main solvent employed for the herbal treatments.

The remedies were administered orally, as baths with cold or warm infusions, or as external applications (crushed or baked plant materials covering the affected body parts). In one case of intestinal worms (oxyurids), a garlic clove was inserted in the anus.

The diseases were grouped into seven categories, depending on the targeted body systems, among which digestive, skin, and respiratory tract diseases featured predominantly. Informant consensus factor (F_ic) for each nosological category is provided in Table 3.

The most versatile species, having the largest area of indications were chamomile (Matricaria spp) and silver linden (Tilia tomentosa Moench.).

Monotherapy preparations were predominant, although some diseases were treated with polyherbal mixtures. For instance, an infusion of a mixture containing fennel (F. vulgare Mill.), caraway (Carum carvi L.), dill (Anethum graveolens L.) seeds, and chamomile capitulum was used to treat infant colic, beginning at 1.5-month of age. Mint (Mentha spp.) leaves combined with chamomile capitulum were employed to treat diarrhea, while Saint John’s wort (Hypericum perforatum L.) flowers with mint leaves were used to treat gallbladder dyskinesia.

Silver fir (Abies alba Mill.), onion (Allium cepa L.), black radish (Raphanus raphanistrum subsp. sativus (L.) Domin), dog rose (Rosa canina L.), wild thyme (Thymus serpyllum L.) and silver linden were indicated for the treatment of respiratory diseases. Anise (Pimpinella anisum L.), chamomile, fennel, mint, caraway, Saint John’s wort, garlic (Allium sativum L.), pumpkin (Cucurbita pepo L.) and dill were administered to treat digestive ailments, whereas for skin conditions, chamomile, common marigold (Calendula officinalis L.), greater celandine (Chelidonium majus L.), and wheat (Triticum aestivum L.) were prescribed as plant remedies.

The highest number of mentions was registered for mint. It was used by 165 out of 326 herbal treatment employing women (50.61%, 95% CL: 47.84%–53.38%), with a mean administration time of 4.8 days. Interestingly, FL is 93.3 for diarrhea and only 6.7 for bloating. The youngest patient treated in this manner was a 3-month-old. Chamomile the next most frequently referenced plant, was reported 128 times (39.26%, 95% CL: 36.56%–41.97%), with the highest FL (68) recorded for infantile colic.

Common marigold was identified in 127 cases (38.96%, 95% CL: 36.26%–41.66%), being indicated in the treatment of various skin ailments, such as diaper dermatitis (FL 72.4), eczema (FL 8.7), etc. Prepared as an ointment, it was applied to the affected area since the first month of life. Another plant administered in ointments was the mountain arnica (Arnica montana L.) with 87 mentions (26.69%, 95% CL: 24.24%–29.14%). The indications were traumatic lesions (FL 100), such as concussions, sprains, and hematoma. The youngest patient that received this preparation was 2 years old.

Sea buckthorn (Hippophae rhamnoides L.) fruits, with 74 mentions (22.7%, 95% CL: 20.38%–25.02%), were indicated as a tonic in treatment of weakness and as an immunostimulant agent in acute respiratory diseases (FL 100). A three-year-old child was the youngest recipient. Also, ramson (Alliumm ursinum L.) was indicated as an immunostimulant, being eaten raw in salads (FL- 100). Silver linden flowers were referenced 54 times (16.56%, 95% CL: 14.51%–18.62%) and were utilized in the treatment of respiratory disorders, notably cough (with a FL- 92.5), as well as agitation and anxiety. The youngest patient treated was 3 years of age.

Greater celandine latex was mentioned in 37 cases (11.35%, 95% CL: 9.59%–13.11%), for the treatment of skin verruca (Fl- 100), with an anti-infective role. The patient at the lowest end of the age spectrum was 2 years old. Common nettle (Urtica dioica L.) young leaves were employed in 34 cases (10.43%, 95% CL: 8.74%–12.12%) to treat anemia (FL- 61.8) and weakness (FL-38.2), and a 1.5 years old was the youngest among the patients. Boiled wheat bran was applied as a poultice on dry dermatitis lesions (FL-100) with very good results.

Garlic bulbs and pumpkin seeds were indicated in intestinal parasites (FL-50 garlic) and worm treatment (FL-50 garlic, FL-100 pumpkin), whereas black radish sap with honey and dog rose was administered in acute respiratory diseases and cough (FL-100), in children aged over 3 years old.

Our search in PubMed and Google Scholar for available data on the listed medicinal plants, comprising 25 in total, revealed the following results (for more details, please refer to the Table 2, Supplementary Material):

1) 22 plants were reported to have ethnopediatric practices in other countries. The plants lacking documented use in this regard are black radish, wheat, and wild thyme.

Based on this data, it appears that although the herbal species mentioned here for ethnopediatric practices are quite common, only part of them have been extensively researched regarding the illnesses reported by the respondents interviewed herein. Furthermore, the majority of the medical uses documented by the mothers enrolled in our survey align with the scientific literature.

The relationships between categorical variables like age groups, urban/rural residence, education levels, and numerical variables (e.g., mother’s age, count of used plants, count of treated ailments) were explored by conducting correlation and multivariate analyses, as outlined in the Data analysis section.

Shapiro-Wilk test demonstrated that none of the numerical variables followed a normal distribution. Therefore, non-parametric tests were used to assess the associations between numerical and categorical variables.

The associations between categorical parameters as evaluated by Fisher’s exact test are shown in Table 6.

The associations between numerical and binary categorical parameters as appraised by Mann-Whitney test are shown in Table 7, while those between numerical and multivalent categorical parameters as estimated by Kruskal-Wallis test, are displayed in Table 8.

The majority of medicinal plants utilized by the mothers participating in our study to treat their children were employed for similar ailments as reported in ethnobotanical surveys, clinical studies, and case reports conducted worldwide in both children and adults, as revealed by our literature search in PubMed and Google Scholar (summarized in Table 2, Supplementary Material). This indicates their accurate utilization by our informants and the preservation of a solid knowledge base regarding the curative properties of the botanical remedies. Another issue highlighted by our search of the ethnomedical or bioscientific data on the identified medicinal plants was the scarcity of clinical studies on pediatric treatments with botanical remedies.

Mint and chamomile, the plants with the highest number of mentions in our survey were largely employed for ethnopediatric uses in other countries. The efficacy and safety of common marigold as a topical remedy for treating diaper dermatitis, an indication mentioned in the present study, was reported in a randomized comparative trial that included infants less than 3 years old. In topical applications, marigold yielded better outcomes than aloe (Panahi et al., 2012). Additionally, in another study conducted on 171 children aged 5–18 years with otalgia due to acute otitis media (an indication not mentioned in our study), Sarrell et al., 2003 found clinical benefits for naturopathic herbal extract ear drops containing marigold flowers, among other ingredients.

Mountain arnica has been utilized for treating various types of traumatic lesions. Scientific evidence supports its use in combination with echinacea to facilitate the detachment of the umbilical cord (Perrone et al., 2012) and for soft-tissue bruising in children aged 0–8.5 years (Thompson et al., 2010).

Ramson is most commonly consumed as food or spice in many countries (Abbet et al., 2014; Pawera et al., 2017; Mustafa et al., 2020). However, caution should be exercised, particularly in vulnerable groups such as young children, due to reported gastrointestinal toxicity in the literature (Fuchs et al., 2011; Lüde et al., 2016). Notably, a search with the phrase “Allium ursinum OR ramson OR wild garlic AND children” in PubMed and ScienceDirect databases provided no clinical studies. Silver linden’s flowers were employed to treat respiratory diseases, agitation, and anxiety. Some scientific evidence for anxiolytic activity is available (Gutiérrez et al., 2014), but it is not focused on pediatric use. Furthermore, another search using the phrase “T. tomentosa OR silver linden AND children” in PubMed and ScienceDirect databases revealed no clinical study.

Greater celandine fresh latex was reported here to be applied on skin verruca. Interestingly, in our previous historical review on Romanian medicinal plants with ethnopediatric use (1860s–1970s), another plant part and indication were identified: the root of greater celandine was administered as a bath for general strengthening (Petran et al., 2020).

Boiled wheat bran was found to be used mainly externally in dry dermatitis, whereas wheat in spray formulations was reported to be efficient for controlling gingival inflammation in schoolchildren (Bello et al., 2020). In vitro studies showed that wheat extract has modulated the expression of inflammation-associated molecules (Funel et al., 2020), and speeded up keratinocyte healing and tissue regeneration (Tito et al., 2020; Morretta et al., 2022). Clinical studies indicated wound healing effects as well, including the curing of venous leg ulcers (Romanelli et al., 2015) and the moisturizing effects of this plant extract (Boisnic et al., 2019).

Black radish, used for the treatment of respiratory disorders and cough in our study, was not yet mentioned in the scientific literature for pediatric use.

Our research has identified a considerable decline in ethnobotanical knowledge compared to the earlier period examined by our team from 1860 to 1970, with only 25 medicinal plants identified, as opposed to the 153 previously documented (Petran et al., 2020). Several potential explanations for this trend may include the following:

• The information was collected from subjects residing in a relatively narrow geographical area i.e., Southern Romania.

Statistical analysis revealed intriguing and valuable findings regarding the relationships between categorical and numerical variables documented in this study. Fisher’s exact test (Table 6) demonstrated that only the education level (among the categorical parameters) was significantly associated with both the number of employed plants and the number of treated ailments (the higher the educational level, the greater the number of employed plants and the number of treated ailments) (Figure 4).

The manner herbs were procured is also correlated with education level—somehow counterintuitively, mothers with higher educational levels had a greater propensity to harvest the plants themselves. A rural background is also associated with a higher probability of using plants harvested (rather than purchased) by the mothers (Figure 5). Remarkably, education level was not associated with residency status.

Fisher’s exact test (Table 6) demonstrated that the preparation method was also influenced by education level: while infusion was by far the most prevalent preparation method in both subgroups (lower and higher educated), higher-educated mothers utilized a more diverse range of preparation methods, in contradistinction to lower educated mothers, which used almost exclusively infusion as a preparation method (Figure 6).

Education level also influenced the recommendation source employed by the mothers (Table 7): for lower educated mothers, medical professionals, and family members (in this order) were the most important sources, while for higher educated mothers, family members (not medical professionals) were by far the most important source, while web and media were also a significant source (absent in lower educated mothers) (Figure 7).

A similar result was found for the association between residency and the source of information: for mothers originating from a rural environment, family members and medical professionals (in this order) being by far the most important sources, while for the mothers dwelling in urban areas the information sources were more uniformly represented, with the web and media having greater importance (Figure 4).

Kruskal-Wallis test too revealed an interesting and rather unexpected result (Table 8): mothers using the web and media as an information source were generally older, while those seeking the advice of medical professionals tended to be younger (Figure 4)

Mann-Whitney test (Table 7) confirmed that a higher educational level was associated with a greater number of both employed herbs and treated diseases, as well as with a later start in treating the children with herbs. In other words, lower-educated mothers had a tendency to initiate herbal treatment for their children at an earlier stage. Harvesting (as opposed to purchasing) medicinal plants was also associated with a higher count of employed herbs and treated disorders, and a later start in the employment of herbal treatment in children. This association can be attributed, at least partially, to the correlation between a higher education level and a more pronounced inclination to harvest (rather than purchase) the plants.

In the MVA, age, residency, education level, and procurement method were all entered as putative independent parameters in the determinism of both the count of employed plants and the count of treated disorders (dependent parameters). MVA revealed that education level was the only independent determinant for both of these parameters. Interestingly, both residency and education level emerged as independent determinants of procurement method, while age did not show significant independent influence.

The number of informants included in the study was relatively reduced, mainly due to the specific restrictive conditions during the SARS-COV19 pandemic. Furthermore, the majority of informants resided in urban areas, whereas traditional knowledge regarding the therapeutic use of herbal remedies is typically better preserved in rural zones. Consequently, the conclusions drawn from our survey may not apply to the entire population of Romania, as our informants were exclusively from Southern Romania, and other regions may exhibit variations in medicinal flora and/or folk medical traditions.

Additionally, no herbal authentication was conducted, with the results solely relying on self-reports from our informants. Although botanical authentication would have been ideal, it was deemed impractical for commercial plant products subjected to processing, such as fragmentation or powdering, which eliminates morphological characteristics necessary for accurate macroscopical analyses (Pawar et al., 2017; Ichim et al., 2020).

Furthermore, while all vernacular names of medicinal plants provided by informants were common, there were two instances where these denominations failed to differentiate between several similar taxa in the genus Mentha and Matricaria. As a result, a definitive botanical identification of these two particular taxa cannot be asserted.