This fieldwork was conducted in the Franco rural community, municipality of Cocal (Fig. 1) (03° 28′ 15″ S–41° 33′ 18″ W), northern Piauí, Brazil, located 268 km from the state capital, Teresina. The municipality has a population of 26.036 inhabitants and a population density of 20.51 inhabitants/km2 [27]. According to the Köppen classification, the climate is Aw’ Tropical, characterized by two well-defined seasons: rainy summers and dry winters, with temperatures ranging between 25 and 35 °C. The average annual precipitation is 900 mm [28]. The Caatinga is the predominant vegetation in the Franco rural community and surrounding regions.

Location of Franco rural community, Cocal, northern Piauí state, Brazil

The community was composed of 125 inhabitants, distributed among 35 family units. The residents mainly rely on family agriculture, cultivating primarily milho (Zea mays L.), feijão (Phaseolus vulgaris L. and varieties), macaxeira or mandioca (Manihot esculenta Crantz). In addition, they raise small, medium, and large animals, such as chickens (Gallus gallus domesticus), ducks (Cairina moschata momelanotus), goats (Capra aegagrus hircus), sheep (Ovis aries), pigs (Sus scrofa domesticus), and cattle (Bos taurus). Apart from activities related to farming, four families also engaged in small local businesses. In two local businesses, people can quickly obtain biomedical drugs for general health issues such as flu, fever, and headache. Some people sporadically engage in hunting and fishing. Many supplement their income by foraging for honey from bee colonies (Apis mellifera) in areas near and far from the community. Home garden management practices characterize the Franco rural community [29], where they cultivate a high diversity of species, especially for food and medicinal purposes, enhancing local food security and sustainability [30]. In most cases, women are responsible for the care and maintenance of these spaces [31], although the men of this community also play an essential role in maintaining knowledge associated with home gardens. The local people also collect medicinal plants from different areas, including anthropized areas, which may include roadside edges, spaces between residences, and areas of primary and secondary forests.

There are no schools in the area, so students need to move to other rural areas or to urban areas, aided by public transportation. The community does not have a Basic Health Unit or Health Unity Center. Residents are attended to by two community health agents; however, they are in irregular and/or deficient services, according to local perceptions. Therefore, people move to other rural communities for medical consultations and vaccination campaigns. In urgent cases, they visit the hospital in the urban area or are sent to the municipality of Parnaíba, 90.5 km from Cocal.

The fieldwork was conducted from December 2019 to April 2021. To ensure greater quality of our collected data, we initially applied a pilot test with 22 local residents (16 men and 6 women) to validate and make possible adjustments to our data collection protocols [32]. The data collection protocols underwent adjustments, and due to this, the data collected in this stage were not included in the data analysis to avoid compromising its quality. The pilot test followed all the guidelines described in the ethical and legal aspects section.

Following the pilot test, to characterize the profile of the research participants, semistructured interviews [33, 34] were conducted with 48 local residents (30 women and 18 men) aged between 18 and 86 years (mean of 41 years), including data on age, gender, level of schooling, individual monthly income, and nature of occupation (number of activities related and unrelated to nature). In general, the research participants had a few years of schooling, with individual monthly incomes ranging from zero to two minimum wages. Most of those earning less than one minimum wage income come from agricultural practices, the commercialization of local products, or activities unrelated to nature. Those earning between one and two minimum wages are retirees or receive some benefit from the federal government. Among the participants, the majority followed the Catholic religion (n = 46). We interviewed all residents over 18 years old who agreed to participate in the research, representing 68.57% of the population (Table 1). This sample represents the population and their behaviors related to health practices involving the hybridization between LMS and biomedicine, according to the reasons described in Table 1.

The free listing technique was applied to collect data on memory LMS options and their interaction with the use of biomedical drugs [34, 35]. Participants were individually invited to list the known and/or used medicinal plants. After the conclusion of the free listing, we applied semistructured interviews [33] to collect complementary data about each item of the LMS recalled in the free listing, based on the following question to guide the application of technique: For which disease(s) or health problem(s) do you know and/or use this plant? and What does the person feel (symptoms) when experiencing this disease or health problem? At this stage, special care was taken to thoroughly document the therapeutic target and differentiate it from general symptoms to avoid underestimating or overestimating our constructed models. For each therapeutic target, participants were asked if, in addition to using plants, there would be another treatment option. From the individual responses, stimulus techniques were applied to identify options such as biomedical drugs, mystical-religious practices, human products, minerals, and animals.

In the second stage, we asked participants to mention health problems treated only with biomedical drugs. For each listed option, we always asked if there would be another option to treat the same disease, applying stimuli to facilitate information recall by local people. In these procedures, we aimed to document as many treatment options as possible for the same disease and to better characterize the interactions between LMS and biomedicine.

In the cases in which therapeutic targets where participants mentioned options from both LMS and biomedicine origin, we investigated the preference between using options from both systems. We considered preference to be the human choice of one therapeutic option over others that are also offered and could be used equally [36]. Participants were asked which options were preferred if they were experiencing the mentioned health problem. We recorded the individual responses in writing, which included (1) preference for LMS options over biomedical drugs, (2) preference for biomedical drugs over LMS options, and (3) sequential use of LMS and biomedicine options or vice versa during the same disease cycle experienced.

The collection and processing of botanical material followed the guidelines proposed by Santos et al. [37]. For their identification, consultations were made with specialized literature, visits to online databases, and the collection of the HDELTA-UFDPar Herbarium, using dichotomous keys and consulting botanical family specialists [38]. Botanical synonyms were updated using the online database of Flora e Funga do Brasil [39]. For some species not found in this database, the World Flora Online database was used (http://www.worldfloraonline.org/). The botanical families are organized alphabetically, following the proposal and taxonomic treatment of the Angiosperm Phylogeny Group IV [40], except for the Turneraceae Kunt ex DC. family, which was not considered a subfamily of Passifloraceae Juss. ex Roussel. The botanical material voucher was incorporated into the collection of the R Herbarium, Museu Nacional/UFRJ (R Herbarium), with duplicates sent to the HDELTA Herbarium, Universidade Federal do Delta do Parnaíba (UFDPar).

The animals indicated as medicinal resources were identified using photographs, illustrated guide analysis, and consultations with specialists in different classes. We updated the nomenclature using data available in the following databases: reptiles in The Reptile Database (https://reptile-database.reptarium.cz/), birds in Birds of the World (https://birdsoftheworld.org/bow/home), insects, arachnids, gastropods, and mammals in iNaturalist (https://www.inaturalist.org/home).

The collected information was compiled, categorized, and standardized into a digital database. We considered gender (m for male and f for female), age, level of schooling, individual monthly income (transformed logarithmically), and total occupations related to nature (e.g., farmer, fisherman, hunter, home garden maintainer, and house care) and unrelated to nature (e.g., mason, carpenter, and mason’s assistant) as socioeconomic variables. To estimate each participant’s preference for different medical strategies, we separated therapeutic targets where the participant indicated both local and biomedical options. The level of individual preference (P) for the LMS was calculated as the ratio of the “total therapeutic targets treated preferentially by LMS strategies” divided by the “total therapeutic targets in which strategies from both systems are indicated.” This ratio was then multiplied by 100. A higher P value indicates a greater preference for LMS options over biomedicine.

To quantify each participant’s status in a hybridization context, a score (IS) was calculated, defined as the ratio of the “total therapeutic targets in which elements from both systems are indicated” divided by the “total therapeutic targets mentioned.” This ratio was then multiplied by 100. For example, for an individual who mentioned 20 therapeutic targets treated only with medicinal plants, 3 treated with mystical-religious practices, and 21 treated with animals, plants, or biomedical drugs, the IS was 21/(21 + 3 + 20) × 100 = 52.5. The maximum value that IS can obtain is 100. The closer it is to this value, the more the individual indicates local and biomedical strategies for most therapeutic targets they know. The IS, then, assesses each individual’s contribution to generating more significant interactions between elements of different systems in a hybridization context.

To answer questions 1 (what is the role of socioeconomic (gender, age, level of schooling, nature of occupation, and individual monthly income) and the preference for LMS health treatments in structuring hybridization of local knowledge in human health-seeking behaviors?) and 2 (What is the role of the same socioeconomic variables on the preference for LMS health treatments over the use of biomedical drugs?), we checked whether the explanatory variables presented a normal distribution using the Shapiro‒Wilk test, which is recommended for small samples [41]. For both hypotheses, multiple regressions were employed; however, before developing the model, the correlation between the explanatory variables was evaluated using the Spearman test. One of these variables was excluded from the same model if a correlation was identified. For both response variables (IS and preference), several models with different combinations of predictor variables were constructed. These models were compared with a null model, from which we selected those most essential for explaining IS and preference values (data variance), including the null model, based on the Delta AIC—Akaike Information Criterion values, Δi. Lower AIC values indicate that less information was lost and that the model fit better [42]. When comparing models to the best-fitting model, Δi values < 2 suggest significant evidence for the model, Δi values between 3 and 7 indicate weak support for the model, while Δi values > 10 demonstrate that the model is improbable [43]. For all experiments, the significance level considered was p ≤ 0.05. All analyses were performed using R Software version 4.0.0 [44].

This study was approved by the Research Ethics Committee (REC) of the Universidade Federal do Rio de Janeiro (Ethical Approval number 3.912.909). Meetings were held with local residents to clarify the methodological procedures to be carried out, objectives, importance, benefits, and associated risks [45]. All participants read and signed the Informed Consent Form [46, 47]. The results of this research were registered in the Sistema Nacional de Gestão do Patrimônio Genético e do Conhecimento Tradicional Associado (SISGEN) under Registration Number AD5B2EC.