Service provision varied across primary level facilities.

The characteristics of the four participating facilities are summarized in Table 1, using data from the first visit, conducted between December 19, 2021, and March 15, 2022. However, a key finding in the ethnographic work – and which emerged in monitoring at other facilities in the mIRA project evaluation – was frequent change in the facilities. During the study, Facility A, was upgraded to a municipal hospital offering caesarean sections and Facility C provided laboratory testing and ultrasound scanning services. In Facility B, the local election resulted in temporary staff cuts and the re-distribution of ANMs to other departments; there was no medical officer toward the end of the study period in Facility D. Changes impacted how ANC care was provided in the facility and also EDSS implementation, with turnover of ANMs trained to use the EDSS, and ANMs sometimes preoccupied with other health priorities outside of maternal and newborn health. Despite purposively selecting clinics with higher client loads, there were relatively few clients accessing ANC at the four facilities. This impacted ANC service provision and implementation of the EDSS. ANC staff sometimes had long periods where they had little to do in ANC, when they either sat around chatting or were deployed to different service areas. Low client load resulted in staff reporting that they rarely saw clients with complications such as hypertension or gestational diabetes and had little confidence to treat them. If ANC clients were seen with any minor abnormalities (e.g., nausea, abdominal pain, slightly raised blood pressure, etc.), ANMs immediately referred to either the doctor at the facility or to a higher-level facility, reflecting risk-averse coping practices.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Facility profile.

https://doi.org/10.1371/journal.pdig.0001340.t001

Organization of antenatal care as routinized and unsystematic practice.

Antenatal care was primarily provided by ANMs who received 18 months training on ANC service in primary health facilities. A staff nurse, who had completed three years training, was also present at one facility. ANMs were either appointed on permanent contracts (for which they had completed an additional national exam through public service commission) or on temporary contracts. ANMs were also responsible for providing delivery and post-natal care. Additionally, they were observed doing family planning, medical abortion, immunization, COVID-19 vaccination, support in the outpatient department, emergency care, medicine distribution, outreach clinics, home visits, nutrition services, recordkeeping, and reporting monthly health management information system data.

We frequently observed two or three ANMs working together to perform the ANC consultation for a single client. One ANM-one client consultation happened sometimes when an ANM was alone in the ANC room, such as when other ANMs were engaged elsewhere. Sometimes one or more ANMs dealt with two clients simultaneously (Box 1). This was mostly the case with Facility A.

In each of these facilities, ANMs exercised discretion in uniquely approaching ANC. ANC consultations were highly routinized, collective, and improvised, rather than structured as individualized, stepwise process as recommended by national guidelines [22]. The service provisions were different in terms of days or daily service provided in those selected facilities. For example, Facility C had designated ANC days where ANMs were under time pressure to examine all ANC clients in a single day. Facility A had more clients attending during a specific time (10 a.m. to 1 p.m.), which kept ANMs busy during this period. One of the reasons for having clients in this specified time was transportation problems in rural settings. Usually, ANMs spent more time (15–30 minutes) with women during their first visit to the facility, including performing the laboratory tests. For subsequent visits, if the women did not report any problems, ANMs spent only a few minutes (4–7 minutes) completing the examination, with slight variations between facilities.

“It gets busy when all the clients come at once … Later, when the client flow is less, they have a lot of time, you might have also noticed they have enough time to chat … When clients come all at once, they are also in a hurry [and] won’t even wait … Sometimes, we don’t have any clients in ANC at all.” (ANM E, Facility A, In-depth interview)

ANMs spoke of history taking and investing time in checkups as crucial components of ANC, although this was not practiced even when they were not under time constraints. However, there were a few instances (such as insufficient ANC staff or additional responsibility beyond ANC duties) where they were clearly overloaded.

“While taking a family history, how is the family environment –supportive family or not? If there is time, we have to do the ANC checkup by utilizing a longer time and communicating, but it’s not like this. Here, you can see that one employee has a lot of work to do. It would be nice if we could at least give 10-15 minutes to examine at least one pregnant woman.” (ANM R, Facility A, In-depth interview)

Contested scope of practice and rule-bending.

There was some contestation around the scope of practice of ANMs. This was notable in the extent of training and provision of ultrasound (USG). We observed debates over whether ANMs needed additional training to provide the initial ultrasound used primarily for calculating week of gestation. Doctors always carried out later abnormality scans that were often done at other facilities.

“I have not received USG training. Even the doctor keeps telling me again and again not to do it because I am not fully capable. I have not received complete training, but I have been doing it. People come from far away to get free USG, and I feel pity for them, so I do it.” (ANM B, Facility A, In-depth interview)

Maternal health services in government facilities are officially free of charge for service users in Nepal. Women are also given incentives to attend at least four ANC visits and to deliver in health facilities. No under-the-counter payments were observed; however, we observed that clients were charged for some aspects of care, including small or no charges (NPR 0–20) for registration (preparing ‘OPD’ card), NPR 350–675 for ANC blood test package (blood group, hemoglobin, random blood sugar, HIV, HBsAg, VDRL), and NPR 300 for an ultrasound scan when conducted by doctor. There also appeared to be some staff incentives to carrying out some procedures, such as: doctors for conducting USG, ANMs for dressing, conducting deliveries, post-natal home visits, and other staff (conducting lab test and x-ray) as clients paid for those services. Amounts were normally displayed in the facility, although they varied between facilities.

Interviews about and observations of EDSS use.

A clear disjunction existed between how ANMs spoke about the EDSS in interviews and how they used it in practice. Views on and use of the EDSS intervention differed between the formal interviews and observations, as summarized in Table 2. Generally, in interviews, ANMs expressed positive attitudes toward the EDSS, indicating that they understood what the EDSS was for, and they valued it. A staff nurse, who was rarely observed using the EDSS, explained how she valued its decision support prompts:

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. ANMs’ response to EDSS with methodological differences.

https://doi.org/10.1371/journal.pdig.0001340.t002

“Many points are being missed when we do [ANC checkup] without using tablets. When we use the tablet, it reminds us, and if there is any problem, the tablet suggests advice. If the blood sugar is high, it suggests what to do further, and if there is a minor disorder like nausea, it suggests the statement about food [diet].” (Staff nurse, Facility B, In-depth interview)

However, we observed very few times where ANMs used the EDSS for decision support; most used it as a recordkeeping application, if it was used at all. If the EDSS suggested that an ANM do something that was not what they usually did, they ignored it. So, for example, the EDSS would make prompts and recommendations for counselling or additional tests based on data entered, and ANMs were observed either ignoring EDSS prompts or entering that recommended counselling was done even when it was not performed. A range of explanations for nonuse or misuse of the EDSS were both observed and discussed. We observed ANMs not using the EDSS, with justifications like the client’s expected delivery date was close, the tablet was not available in the ANC room or not charged, or sometimes not using the EDSS when women visited the facility for not a full ANC visit (e.g., women attended either for conducting blood tests or having an ultrasound).

ANMs also commented that if they fully followed the instructions of the EDSS, which included all recommended aspects of ANC care, for example, counselling at every visit – it would make the ANC visits too long. ANMs explained that pregnant women were impatient to finish the visit and get home when transport was available. We observed some women walking for up to three hours from rural areas to the facilities, and we witnessed the problem of limited transportation or vehicles traveling on a specific schedule in the morning and evening. Women’s time constraints and transport limitations further reinforced pressure to shorten ANC consultations.

“I have so much works left to do at home. The vehicle will be leaving soon [from the bus station]. Sister, please do my check-up quickly.” (Pregnant woman, Facility A, Fieldnotes)

Some ANMs were frustrated by initial technical problems with the EDSS that resulted in the software being slow when the ANM wanted to enter information. Mostly, we observed the WHO EDSS tablets freezing when ANMs entered the value of test results, and this made them feel irritated about using the EDSS.

ANMs spoke about the increased recordkeeping workload imposed by the EDSS, which was implemented in this study alongside, instead of replacing, paper-based records. ANMs were observed having to enter, for example, vital signs or test results in the woman’s handheld record, the facility’s register and in the EDSS. The EDSS was not seen to help with recordkeeping, including for the computerized reporting of monthly services statistics to the district.

“There is a thing that needs to be added in the app [EDSS]. What needs to be added is that we keep recording, but it would be good if monthly reports can be extracted from it. It would be better if everything done in the paper would be the same in this app and we don’t have to turn in papers and helps in online reporting.” (ANM B, Facility A, In-depth interview)

During our first visits, when EDSS had not yet been implemented in these facilities, four ANC protocol visits were performed. However, by the third visits, during the implementation of EDSS, these facilities started following the WHO recommendation of eight protocol visits, endorsed by the Nepal government countrywide. All ANMs tried to comply with the government’s protocol, in contrast with EDSS use. Generally, they viewed EDSS implementation as a temporary change and more of a research project. Although there were a few times where clients had difficulty complying with the eight ANC visits due to the rainy season and transportation issues, ANMs were flexible in calling clients for follow-up visits, considering their difficulties.

“This app is separate; now it’s not in the government protocol, and we are operating it from a non-governmental organization. We must send a report to the municipality. We are afraid that the details might get missed in the register, so we have to maintain our register first.” (ANM J, Facility C, In-depth interview)

In informal conversations, and in interviews, ANMs spoke about whose responsibility it was to use the EDSS. The mIRA project model was that one person from each facility was externally trained to use the EDSS and given the tablet for the facility. Onsite support by fieldworkers in the first month of implementation was intended to support both the trained ANM and untrained ANMs in using the tablets and software. The interviews revealed that some untrained ANMs felt that it was the responsibility of the person who had been trained to use the EDSS, but it was not their responsibility, or they had not had sufficient training. We observed similar tensions around other trainings that took place in the facility: the person who was trained was expected to implement the practice and this was not necessarily seen as the responsibility of all ANMs.

“We have a mindset. Even I follow this mindset that, who went for training should be responsible and should work on it.” (ANM G, Facility B, Informal conversation)

“Tablet is being used [by] who have received the training, it is the focus for her [EDSS trained ANM].” (ANM I, Facility B, Informal conversation)

Belief that they were providing good quality of care.

The service delivery practices were underpinned by ANMs’ beliefs about what constituted “good quality” ANC, which we analyzed using street-level bureaucracy theory. In both formal interviews and informal conversations, ANMs consistently perceived the ANC they provided as good quality, often equating quality with service availability such as USG and laboratory test, rather than complete clinical assessment or counselling as recommended by guidelines. ANMs also explained that they were seeing relatively few women and those that they saw were generally in normal condition. If there was any sign of any issues, then the women were immediately referred either to a doctor working in the facility or the nearest hospital.

When we interviewed ANMs about their perceptions of the quality of ANC, they emphasized the importance of maintaining women’s privacy. However, during our observations at Facility A, we observed that up to four women were examined simultaneously in the same room, including one seeking a termination of pregnancy. Although ANMs defined quality in terms of privacy and respectful care, observations revealed a gap between these ideals and actual practice.

“We should clear their confusion regarding the quality of services being provided or not. Another thing is, we should maintain their confidentiality. Some of the women might be in fear due to family pressure. We should create an environment for them to share their problems without any fear. I think this is also quality service.” (ANM B, Facility A, In-depth interview)

Most women seemed satisfied with the ANC services provided when we enquired. However, we observed that ANMs often missed components of ANC, such as complete history taking, head-to-toe and abdominal examinations, proper counselling and basic tests that should be performed during relevant visits (such as urine test, hemoglobin test, blood sugar test, etc.).

“It is much better now. Previously ANM were old and they didn’t used to communicate properly and didn’t provide any advice, they just used to provide medicines but now these ANM are different. They provide us information about what I asked, provide appropriate advice and listen to us.” (Pregnant woman, Facility A, informal conversation)

ANM behaviors to move towards and away from ANC clients.

Our observations revealed that ANMs exhibited behaviors that diverged from ANC guidelines, including those recommended by the mIRA project EDSS intervention or from the government’s new eight ANC visits policy. Using street-level bureaucracy theory, we explore how ANMs adapted the intervention or policy to suit ANC clients (moving towards) but also to suit themselves and their values and context (moving away) and rationalized these differences (Table 3). ANMs rationalized routine ANC practice, categorized or prioritized the clients, or distanced themselves from care provision. For instance, ANMs in these facilities generally prepared the woman’s handheld ANC card after three months of pregnancy. Prior to the end of the first trimester, ANMs used OPD cards to record information. ANMs reasoned that some pregnancies were not wanted and would be aborted, so the woman’s handheld card was not needed. However, in one instance, an ANM wanted to skip preparing the woman’s handheld card even after three months and was observed attempting to convince the client to visit a nearby health post, which was closer to the woman’s home, though the woman wanted to seek ANC in that ANM’s facility (Table 3, example 5). In some cases, ANMs considered the financial implications for women from poor economic backgrounds when performing investigations, if the women appeared normal (Table 3, example 2). Moving away behaviors reflected other strategies such as over-referring minor abnormalities to doctors or higher centers (Table 3, example 7) or skipping routine checks like blood pressure measurement (Table 3, example 8). These practices show how ANMs rationalized their actions to suit their own context. The ultimate implications were that they compromised on providing good care, tried to short-cut routine practice, and did not provide equal care to every pregnant woman.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. ANMs behavior: Moving towards and away from ANC clients.

https://doi.org/10.1371/journal.pdig.0001340.t003

Overall, these practices demonstrate how ANMs exercised discretion in balancing formal guidelines, EDSS recommendations, and perceived client constraints. While such adaptations were often justified as protecting women from financial burden or unnecessary procedures, they also resulted in selective implementation of ANC components and inconsistent adherence to standardized care pathways. These discretionary practices reflect classic street-level bureaucratic coping strategies, with implications for equity and quality of ANC delivery.

The content and extent of counselling in ANC visits were adapted based on perceived need. This further illustrates how discretionary judgment, rather than algorithmic guidance in the EDSS, shaped ANC service delivery. Generally, ANMs, instead of giving the same information to all women, made judgements about what and when to give information to ANC clients. ANMs explained that in early ANC visits, when preparing the woman’s handheld card, they interacted with women more because they had to ask more questions. In later visits, ANMs focused on calculating gestational age and checking weight, blood pressure and fetal heartbeat, and only offered counselling if women complained of any problems.

“If women do not have any problems, then what should we explain to them? They are all okay. In the initial visit, we asked about having headaches and blurred vision and then counseled them to visit facilities as soon as possible; in the other visits, if they did not complain about any problem, we felt that counselling was not required.” (Validation meeting, Facility C)

Care provision was also shaped by ANMs’ perceptions of risk and liability. ANMs often referred women when ANMs identified minor complaints such as nausea, diarrhea, or back pain that could be manageable at their level. ANMs did not want to take a risk if a doctor was present in the facility, or they tended to refer pregnant women to higher-level facilities. When discussing caring for complications, ANMs stated that these risks depended on circumstances. Some ANMs did not want to take the risk if higher-level facilities were nearby, but in more remote areas, if doctors or a higher-level facility was not accessible, ANMs would have to take the risk and manage these more minor complications. Also, in some scenarios, pregnant women preferred to be referred to higher-level facilities for minor complications.

“Sometimes clients do not want to be referred; sometimes they choose to be referred; it is their right to choose; sometimes clients are so complaining that we are threatened, if something happens, then they will create a hassle in the facility and damage things.” (Validation meeting, Facility D)

The ethnographic work revealed how social structures and entrenched practices shaped provision of ANC and how little the EDSS intervention did to alter these systems. The research participants recognized that the system that healthcare providers were used to was shaping their behavior and resulted in the lack of effect of the EDSS on changing practice. One facility-in-charge explained:

“The healthcare providers are operating with a 19th-century mindset in a 21st-century system. They continue to practice as they always have and do not attempt to change because they believe this is simply how things are done. Changing both the prevailing mentality and the system is difficult.” (Validation meeting, Facility D)