Study design

Our study consisted of a qualitative emergent-theory design based on ethnographic fieldwork [51].

Study site and population

This study was carried out in eight villages of Bac Ai and Ninh Son districts, situated in the hilly and forested part of the Ninh Thuan province (South-Central Vietnam). In 2022, Bac Ai counted a population of approximately 31,000 inhabitants, predominantly of Raglai ethnicity, in addition to Cham and Kinh ethnicities, the latter being the Vietnamese majority population [52]. According to 2022 local statistics, the Ninh Son district had a population of over 61,000 inhabitants, predominantly of Kinh ethnicity in addition to various other ethnic groups including Cham, Raglai, Co Ho, and Nung [52]. Ninh Son is approximately 28 km from Bac Ai district.

Our interviews and observations were in majority with people who identified as being from Raglai ethnicity and refered to themselves as (the) “Raglai”. The Raglai are a diverse group of people, and when we will mention “(the) Raglai” in this work, we only refer to the study population we interviewed. Our Raglai informants took a certain pride in saying they were Raglai, and always asked us from where we came. Therefore, in referring to these informants and their communities in that way, we are trying to respect their identity.

People of Raglai ethnicity are categorized by the government as an ethnic minority group. Despite the government’s socio-economic interventions in the past 20 years, people of Raglai ethnicity are among the poorest groups in Vietnam [53]. Raglai people engage in small-scale subsistence slash-and-burn agriculture in the forests on the mountains, occasionally combined with livestock, hunting, and logging [54, 55]. In the last two decades, Raglai were relocated from ‘old villages’ in their traditional forested territory to ‘new villages’ constructed by the government as part of its policy for resettlement and socio-economic development of ethnic minority groups [56]. Public healthcare and schools are made available in ‘new villages’. Despite the government’s interventions and the gradual integration with the market economy, more than half of the district’s population–estimated at 58,6%–lives under the national poverty line [57].

Health studies found that people of Raglai ethnicity had a higher risk of malaria infection and malnutrition rates among children under 5 years of age in comparison to those who were of Kinh ethnicity [54, 58–60].

District hospital data from 2016–2018 reported an increasing trend of dispensed antibiotics at public health facilities [61]. Although there are no granular AMR data from Bac Ai district, a small hospital-based study at the provincial reference hospital for our study setting reported a high prevalence of ABR amongst bloodstream infections in 2017 (e.g. 100% resistance to ampicillin, 92.7% to trimethoprim/sulfamethoxazole, 72.7% to ciprofloxacin, and 72.0% to ceftriaxone in 144 E. coli bloodstream infections, preserving only carbapenems to treat these infections) [62]. Malaria parasite resistance was found in some provinces in south-central Vietnam including Ninh Thuan [63] and delayed parasite clearance has been spreading [64–66].

Positionality

The research team was composed of two field researchers (TTN–female, MR–female). The Vietnamese researcher (TTN) was from Kinh ethnicity and had already been conducting fieldwork among predominantly Raglai communities for several months, acquiring familiarity with informants of Raglai ethnicity, gatekeepers in the field, the Raglai culture, and the Raglai accent in Vietnamese. Prior to the study, TTN conducted four periods of fieldwork in the study sites, between July 2016 and April 2018, as part of her PhD research on malaria transmission among Raglai population. Each fieldwork was between four and six weeks. Being from Kinh ethnicity, TTNs ethnographic approach was a requirement to gain trust and allow proper data collection among this ethnic minority. With the exception of a few people who only spoke Raglai, TTN could communicate directly with most informants in Vietnamese. The second Belgian researcher (MR) faced language barriers, needing translation, and initial increased socio-cultural distance. The prolonged stays in the communities, in combination with MR’s precedence from a “neutral” country–not involved in the Vietnamese war–and not part of inter-cultural socio-political tensions at national level, helped to gradually build trust and lead to informants opening-up about certain topics in a different way. Members of the larger study team (XXN and KPG) have been doing research embedded in this study area and population since 2004.

Data collection

Two researchers conducted ethnographic fieldwork between April-August 2018 (MR and TTN) and May-July 2019 (TTN). Collected data consisted of semi-structured in-depth interviews (IDI), (focus) group discussions, and participant observation including informal conversations. The field researchers participated in everyday activities including observations at communal health centres and private practices as well as residing with community members. This allowed to build rapport and trust, and gain insights into living conditions, socio-cultural practices, people’s health-seeking behaviours, as well as their perceptions, use, and/or prescription of medicines. When IDIs and group discussions were recorded, this was done using a dictaphone. Sometimes only notes could be taken. For informal conversations, notes were written down as soon as possible after the conversation. All data were transcribed in either English or Vietnamese by TTN and MR with the assistance of a local field assistant.

Sampling

We selected informants following theoretical and snowball sampling strategies. Both communities and participants were gradually selected within the study area on the basis of emerging information. Informants included community members (both from “old” and “new” villages); human health service providers (public and private sector, as well as informal providers); animal health service providers (public, private sector, and informal); representatives of local government (local authorities, forest protection guards etc.); and traditional power structures (shaman, respected “wise” persons). We aimed for maximum variation in occupational profiles (farmers, doctors, informal providers, grocery shop owners); ethnicities (Raglai, Cham, and Kinh); gender (female, male); and age groups (older people, adults, and children). Further details on the categories of informants are provided in S1 Table. Data collection resulted in 43 informal conversations in combination with participant observation, 71 IDIs, and 10 group discussions.

Analysis

Analysis followed the principles of retroductive analysis [67]. We analysed data intermittently during, between, and after data collection phases. We carried out thematic content analysis and analysed transcribed data with the help of NVivo (11) software (QSR International Pty Ltd. Cardigan UK), based on themes pre-identified from the initial research questions in addition to new themes that emerged during intermittent analysis [68]. Data was coded and analysed by two authors (TTN and MR). A theoretical framework on vulnerability underpinned this work (cf. discussion).

Ethical approvals

Ethical approval was granted by the Institutional Review Board of the Institute of Tropical Medicine (ITM), Antwerp (Belgium)–Ref 1234/18 –in addition to the Ministry of Health and the National Institute of Malariology, Entomology, and Parasitology (NIMPE), Hanoi (Vietnam)–IRB decisions 1648/QD-VSR and 3731, QD-BYT. We obtained oral consent from all informants. Participants had the opportunity to ask questions, voice concerns, or withdraw from participation at any time during the research. Specific attention was given by the researchers to detecting culturally specific “polite/indirect” signs of refusal to answer questions and/or to participation, in which cases researchers stopped the data collection. As an example, some individuals responded "I do not speak Vietnamese" as a means of declining to participate in the study, even though they had basic knowledge of the language. In other situations, people expressed interest in hearing our introduction, but politely declined when invited to participate in the study.

Inclusivity in global research

Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the S1 Checklist.

Factors influencing patients’ access to healthcare

Factors driving patients’ choice of providers

Even though more than 90% of the population in the studied villages was estimated to have access to a basic health insurance provided by the government to populations living in poverty and remote areas [69], public health facilities were not always the first choice of care. Therapeutic itineraries for one health problem commonly combined several types of care, either in parallel or in alternation, which was influenced by a complex interplay of factors.

In this study setting, the range of healthcare options was relatively wide, offering many therapeutic options to communities, ranging from primary and secondary biomedical healthcare–including public and private commune health centres (CHC); hospitals; home practices by nurses, doctors, pharmacy technicians; sole dispensing of medical products by private pharmacies, grocery shops, and mobile sellers–to shamanic healing. See Table 1 for an overview of the available therapeutic options and the perceived comparative advantages in our study context.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Therapeutic options and comparative (dis)advantages as perceived by informants in the study setting.

https://doi.org/10.1371/journal.pgph.0002982.t001

Time and work

In subsistence farming, which determined most Raglai’s life rhythm, time meant money and resources. As such, seeking care became a priority only when the symptoms compromised the ability to work. In this case, striving for time efficiency appeared to weigh most heavily in the decision where and with which provider to seek healthcare. Consequently, waiting lines in public health centres were perceived as an important impediment. In addition, while in theory there should always be staff on duty, in some health centres staff occupied post-office hours for emergencies only, and not for minor healthcare problems. For some community members, however, the perception of what was urgent or not was different from staff’s perceptions, leading to occasional disagreements that impacted on the relationship of trust between provider and patient. These experiences would then influence subsequent health-seeking itineraries.

Distance and geographical accessibility

The time required to travel to a health facility was particularly important in remote and old villages where it takes a considerable amount of time to get out of the forest, down the mountain, on bad quality roads, to a pharmacy or a public health centre. Home practices, grocery shops, and mobile sellers provided a comparative advantage in this regard. Poorest families had to mainly travel by foot, while better off families had motorbikes. The motorbikes were often owned and used by men, which sometimes made it difficult for women to access care quickly in case of emergencies.

Cost

To evaluate the affordability of a provider the following needed to be taken into account: (i) direct non-medical costs such as the financial costs of transport, (ii) direct medical costs of healthcare and medicines, as well as (iii) indirect costs such as the loss of income and food due to not being able to work; as well as the missed chores and (temporary) abandonment of relatives (children, older people, disabled) in need of care at home. All Raglai communities were entitled to free basic health insurance provided by the government due to their designation as ‘poor’ and remote residence, covering selective medical services in the public sector [70]. To obtain a health insurance card, people needed to officially register for a new card annually, except for children under six years of age who were exempt from the renewal requirement). Given the Raglai’s poverty, families that for various reasons did not obtain a health insurance card or did not go to the public health sector needed to use a high proportion of their total income to purchase medicines and health services. This sometimes led to the need to sell domestic animals, to borrow money from relatives or to take an individual loan from a grocery shop (often owned by someone from Kinh ethnicity) as a financial coping mechanism, which would subsequently become a major burden on the household. In such a context, the option of buying partial drug regimens, according to the means of the patient, or having the option of paying at a later time, were considered important advantages for health providers to attract clientele.

Perceived diagnostic capacities of the provider

The possibility for receiving a diagnosis, an explanation for the cause of the disease and a description of the problem itself, was a valued advantage of biomedical healthcare providers, both in health centres and in private practices. Nevertheless, the speed of recovery was often deemed to be more important and therefore previous experiences relating to speed of service and recovery prognosis often outweighed the benefits of receiving a concrete diagnosis. People therefore often sought out those drug dispensers that were perceived to be more accessible and were known to have more effective and fast-acting drugs, regardless of their diagnostic capacities.

Provider-patient relationship

Being welcoming, taking time to listen to the patient and giving clear information in a respectful way and comprehensible language (preferably Raglai), showing cultural proximity or sensitivity (more likely to be achieved by Raglai staff) were elements that were seen to foster trust from the patients and contribute to a good reputation of the healthcare provider. These perceptions were often circulated and fed back within the community, leading to a reputation of a provider which would either attract more patients in the future, or to the contrary, lead to avoidance by patients in case of negative experiences. Anecdotes were shared of patients being addressed by staff in disrespectful ways, such as being asked to sit further away during a consultation because of their unpleasant smell. References were also made to unequal inter-ethnic relations as healthcare workers were often from a different ethnicity than the Raglai patients in our study.

“…In general the instructions [given by the district health centre] are easy for patients to understand. I am Raglai and they can understand what I say… Patients find your people [Kinh] annoying. My people rarely raise our tone when we speak to the patient, we always handle the conversations with them gently, we explain every little detail to them. Your people [Kinh] often give short explanations to the patients [which makes them] think they were being scolded.” (IDI, healthcare worker at a public health centre, 16/05/2018)

Perceived aetiology of the illness symptoms

In Raglai cosmology, humans are intrinsically connected to the environment they are born in, to the land that hosts living beings as well as their ancestors. All these connected elements are believed to have spirits. These spirits, including the own spirit of a person, can be offended in case of antisocial behaviour or disrespect towards essential traditional values, unnecessary destruction of natural resources, and trespassing of sacred spaces. Such behaviours can trigger ill health. “Bad” shamans were believed to be able to afflict illness as well. Both biomedical and spiritual aetiologies allowed for biomedical (“western”) treatment and/or spiritual shamanic healing (involving rituals, prayers, offerings/sacrifices) as appropriate courses of action as both could be intertwined. Herbal medicine is also part of traditional care, though it was mentioned that a lot of this knowledge was being lost among the younger generations. Biological diseases could be associated with “worms” in the body (imagery stemming from biomedical healthcare personnel ranging from viewable to invisible organisms in the body causing illness). Additionally, sudden ecological changes (e.g. in the weather and climate) in combination with intensive working conditions were also perceived to place a heavy burden on the body, leading to sickness or other disruptions in a natural balance such as changes in diets etc. Sudden changes in weather could refer to sudden rain, wind, and cold or on the contrary, to very high temperatures and sun exposure. However, this was also associated to perceptions of climate change, showing increasingly unpredictable changes in weather and seasons which was said to have a significant impact on agriculture practices and working conditions in the fields. Spiritual illness could be associated with conflicts between spirits and people (e.g. offenses to the ancestors) or between people (e.g. jealousy). The longer biomedical treatments remained ineffective in relieving symptoms, the higher the probability that people perceived either the medicines not to be appropriate in relation to the severity or the “strength” of the disease; or interpreted the illness to be caused by something which “Western medicine” could not solve. This was the case for complicated, relapsing, and chronic diseases, where aetiological associations with spirits and/or sorcery often became the final diagnosis, dedicating all efforts towards spiritual healing.

Factors influencing patients’ choice and use of medical products

Administration route.

Parenteral medical products, such as transfusions and injections seemed widely preferred as routes of administration by patients, as they were thought to lead to rapid symptom improvement and required more dedicated care from healthcare workers, which in turn fed into the perceived well-being of a patient. These options were also prescribed by both private and public health practitioners as they believed these administration routes could enhance the patient’s ability to recover. Nevertheless, they were more expensive and therefore the most common type of medical products consumed were oral medicine, in tablets or more seldom in instant powder. Clear distinctions were made between medication for children and for adults. Tablets were often delivered in mixed-medicine plastic bags and could contain (a combination of) labelled tablets (from cut blister packages), unlabelled tablets (in bulk), or all ground to powder, depending on the dispensers. Some dispensers (e.g. at CHCs) prepared different sachets for each type of medicine, with separate instructions per type of drug (and sachet). All these medicines were then grouped into a bigger plastic bag that was given to each patient. Other dispensers packed the mixed drugs in sachets per daily intake, while others again combined all drugs purchased with different instructions in one bag (Fig 1). Patients often had little or no idea which medicines were included in the package as it was often referred to as “medicine bag to treat fever”, “belly pain”, or “headache”, etc.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Picture of medicines bags purchased by a community member in a private pharmacy for stomach-ache and fatigue symptoms.

Image credit: TTN.

https://doi.org/10.1371/journal.pgph.0002982.g001

Perceived efficacy to treat illness symptoms

The efficacy of a treatment was judged in relation to the speed with which symptoms were alleviated. Raglai informants often explained this in terms of “fast” or “strong” medicines versus “slow” or “mild” medicines. When medication was started, patients tended to take medication for one to three days in line with their capacity to pay and the other factors described hereabove. During this period, the patient evaluated the drug efficacy against the rapidity of symptom relief. If symptoms abated, treatment was considered effective and pursued (with the potential purchase of additional dosages when means allowed) until symptoms were experienced to have sufficiently decreased. The course was then considered complete and further doses were stored for future use. Treatment failure was often interpreted initially as a “bad prescription”, i.e. the wrong answer to the problem. Such a “bad prescription” could be associated with (i) poor quality of the drugs; (ii) a wrong diagnosis, i.e. the prescribed/purchased regimen was inappropriate for the disease the patient actually had; (iii) the disease being too “strong” for the prescribed medicine. The latter explanation was often linked to spiritual aetiologies, such as the spirit of the sick person being too weak or travelling too far from the physical body for the medicine to take effect.

"… At the communal health centre, they gave me medicines, but sometimes I would still be sick after taking them for a week. I am not sure why, but I think the medicines they prescribed were too mild. Unfortunately, I could not afford to buy stronger medications, so I had to rely on the ones provided at the communal health centre. Occasionally, they would refer me to the district health centre, where they have stronger medicines and highly trained health professionals." (IDI, patient, 11/08/2018)

Perceived quality of medical products

Informants sometimes referred to medicines sold in grocery shops as thuốc bậy bạ or “nonsense medicine” implying that this kind of medicine was considered a temporary solution while waiting to see if symptoms would decline or preparing to seek care from a private or public provider further away. Informants and shop owners named these medicines “fever medicine” or “belly pain medicines”.

Among both Raglai and Kinh people, the quality of drugs from public health centres was often perceived to be lower than in private practices or pharmacies. While some people believed that the health centre medicines healed at a slower pace, others also believed these medicines could only partially heal the disease. The perception about “slow”-working domestic medicines in health centres versus “strong” and “fast”-working imported (foreign) medicines in private pharmacies or healthcare practices was shared by some staff working in health centres, pharmacies, and Raglai key informants with higher socio-economic status. These perceived benefits of foreign medicines were promoted by private providers and used as an argument to justify the higher costs.

Factors driving diagnosis and prescribing practices of providers

Limitations

While we looked at several social factors underlying use of medicines and potentially impacting AMR, more data on social factors related to antibiotic use in animals, prescription practices and more specific data on antibiotics is needed to gain a more holistic understanding, especially if combined with interdisciplinary data on AMR in this context (e.g., microbiological and epidemiological data). However, given the current scarcity of AMR data from this context and the time and budget limitations, this was not achievable within the scope of the current study.