Study design, recruitment and collection of data

The EPSHILS project group used a qualitative design employing semi-structured group interviews. The interview guide was developed by the EPSHILS project group, including contributors from both Tanzania and Norway.

Thirty-two nurse-midwives and five doctors were recruited through purposive sampling by the head nurse-midwives of the obstetric and labor wards at each hospital. The inclusion criteria were: 1) registered nurse-midwife or doctor with experience from the labor ward, 2) currently working within the department of obstetrics and gynecology, and 3) willing to participate in the study. The nurse-midwives’ working experience ranged from recently educated to thirty years, two nurse-midwives had a Master’s degree in midwifery. The doctors had 3–8 years of working experience and represented resident doctors, registrars, and obstetricians and one of them was specialized in Obstetrics and Gynecology. The interview groups consisted of either nurse-midwives (seven groups) or doctors (two groups). In general, the informants had worked mainly in the institution where they were currently employed at the time of the interviews. We did not collect information regarding the informants’ age. In the nurse-midwives’ groups there were 4–5 respondents, the doctors were in groups of 2– 3.

Nine group interviews were conducted between August and December 2018, moderated by the local research assistant of the EPSHILS project, DAM, who is a male clinical officer known to the staff members from a previous research project. Five interviews were conducted at the Regional Referral Hospital (RRH) and four at the Zonal Consultant University Hospital (ZCUH). Each interview was recorded and lasted between 60 and 105 minutes.

Preparing the material

The material was transcribed and translated from Swahili to English. The transcribed material was anonymized before it was passed on to the authors, AIH and JMEH, who kept a dialogue with the moderator of the interviews, clarifying uncertainties in the material.

Data analysis

We performed an inductive, qualitative content analysis²⁴. AIH and JMEH read through the material several times getting a sense of the whole, and meaning units relevant for our research question were identified²⁵. HyperResearch (ResearchWare, version 4.0.3) was used to organize the data. Meaning units were condensed and coded and an iterative process of categorizing was performed, resulting in mutually exclusive, exhaustive, and saturated categories (Supplementary file)^26,27. All co-authors were given access to the anonymized material and the analysis in order to ensure a mutual understanding of the content.

Study setting

Tanzania’s healthcare system is decentralized and divided into national, zonal, regional and district levels²⁸. The context of this study was one zonal consultant university hospital and one regional referral hospital in Northern Tanzania. Both hospitals receive women referred from district hospitals.

The partograph is started at cervical dilation of 4 cm in the ZCUH and 3 cm in the RRH while the woman is in the antenatal ward. Women are admitted to the labor ward when they have reached cervical dilation of 7 cm. CSs account for 15–20% of the births in the RRH and 35–40% in the ZCUH. Approximately 5% of the laboring women in the RRH are referred to the ZCUH. Both hospitals use fetoscopes (steel/plastic) called Pinard horn, and electronic dopplers for fetal monitoring. There is no access to continuous fetal heart rate monitoring. Oxytocin augmentation is at both facilities administered by infusion, drip rate being manually adjusted due to lack of infusion pumps. The birthing women have no access to adequate pain relief besides opiates. Few women have a companion in the labor ward, although their presence is permitted at both hospitals.

Various ways of understanding prolonged labor

The respondents expressed variations in expected duration of labor. Prolonged labor was understood as going beyond ‘expected hours’. Respondents described the partograph as essential in diagnosing prolonged labor. In the RRH, many respondents operated with a timeframe of twelve hours. One doctor said:

‘It is labor that goes beyond its usual time which is maximum allowed without having delivered. Let us say twelve hours, beyond that then it is prolonged labor.’ (Doctor, RRH)

At the ZCUH, the expected duration of the active stage for nulliparas ranged from 8 to 24 hours. Respondents in both hospitals differentiated between primiparous and multiparous women for both first and second stage of labor. Nurse-midwives expected the second stage of labor to last no more than twenty minutes to one hour.

Three terms were described explicitly and implicitly – prolonged labor, poor progress in labor and obstructed labor. For medical terminologies, health workers in Tanzania often mix English words with Swahili words. Prolonged labor in Swahili (uchungu uliochukua muda mrefu) may be translated to ‘labor which takes long time’, poor progress in labor (uchungu usioendelea) to ‘labor which does not progress’ and obstructed labor (uchungu kinzani/pingamizi) to ‘labor that is resistant/blocked’. Poor progress was described as an early stage of prolonged labor, but also as an indication for emergency CS, as expressed by this nurse-midwife:

‘According to my opinion the management of poor progress and prolonged labor are similar, for example in case of prolonged labor I have to establish the causative factor and the management will go from there. But in poor progress, in my management I will directly prepare her for CS ...’ (Nurse-midwife, ZCUH)

Several nurse-midwives associated prolonged labor with severe findings of dry vaginas and restless, febrile mothers. One doctor described a spontaneous vaginal birth as a reason to abandon the diagnosis of prolonged labor:

‘… I will give her more time, but sometimes we succeed and after some time the mother delivers, you can see it is not prolonged labor.’ (Doctor, RRH)

Descriptions of poor progress, prolonged labor and obstructed labor varied. The three terms were often overlapping and at other times, contradictory. Absence of progress was labelled poor progress by some, and prolonged labor by others. A respondent from the ZCUH said that they often encounter poor progress but stated that prolonged labor usually was referred from other hospitals involving poor outcomes for mother and child. Prolonged labor was also said to cause obstructed labor. Some respondents explicitly addressed confusion among the terms poor progress and prolonged labor. The term ‘early prolonged labor’ (in the latent phase) was also used.

Assessing progress in labor

The respondents expressed confidence in the partograph – as a guide, indicating time for intervention, avoiding prolonged labor, and helping to prevent poor outcomes. A nurse-midwife said:

‘Partograph is the first, this is the important tool which I use to know prolonged labor, because it shows from when the mother enters active phase, it shows how contraction goes, it shows fetal heart rate, it shows everything. So, the partograph is a tool we depend on very much and have confidence in.’ (Nurse-midwife, ZCUH)

All groups emphasized the importance of filling the partograph properly. A nurse-midwife in the RRH said that even though they know how to fill in the partograph, often it is not done properly. Some doctors implied that nurse-midwives sometimes filled the partograph in a way that did not indicate prolonged labor. Both doctors and nurse-midwives reflected on how an erroneous first plot in the partograph may wrongly indicate prolonged labor. Nurse-midwives expressed uncertainty in establishing the start of active labor. Referral cases provided additional challenges due to undocumented anamnesis. Despite their confidence in the partograph, doctors at the ZCUH reported that the partograph occasionally forced them to intervene. According to their experience, they sometimes avoided CS by not following the indications of the partograph. Nurse-midwives also experienced the partograph as having a narrow normal range.

Nurse-midwives described vaginal examination and abdominal palpation as a challenge and advocated for training. The respondents used the terms level, descent and station and there was a lack of common understanding of their differences and similarities as this nurse-midwife said:

‘In measuring spine level, it is a problem … even in interpreting, when you mix interpreting engagement and then comes station, you mix it with level. You have to think twice, which one do you write down? It is zero in spine level but where is it? In plus one, plus two or more? Where is plus two? Where is minus two? It is a problem altogether.’ (Nurse-midwife, RRH)

Nurse-midwives also explained how different finger sizes would lead to differences in cervical dilation measurements. In the ZCUH, nurse-midwives found it challenging how inexperienced doctors sometimes lacked skills in cervical dilation measurement.

Malposition was acknowledged as a cause of prolonged labor. Nurse-midwives in both hospitals regarded it as important, however challenging, to determine the child’s position and presentation. In detecting malposition, respondents also described pain as an indicator. Furthermore, severe pain, lack of pain and unexpected pain made the respondents alert to complications like uterine rupture. Detecting the cause of pain was found to be challenging but essential, as expressed by a doctor:

‘Why so much pain? … I will go and find out why, why, why.’ (Doctor, RRH)

The respondents described how different expressions of pain influenced how they intervened, sometimes resulting in unnecessary interventions.

Appropriate intervention at the appropriate time

‘We must take action’ was a frequently used phrase among the nurse-midwives, often synonymous with CS. Respondents said that action must be taken when vaginal birth was not regarded as possible, there was a lack of progress for some time, the child was too big or if the ‘partograph said so’. A nurse-midwife said:

‘If you use the partograph well, you can know if the things are going well or not, so you can take action. If it is a big baby that definitely cannot come out in a normal way, you must take action as early as possible.’ (Nurse-midwife, RRH)

Furthermore, nurse-midwives described how women asked for CS after being in labor for a long period. However, doctors in both hospitals said that without danger signs, they would not perform a CS solely on request from laboring women. Some doctors said that the fear of performing vacuum extraction was prominent among doctors, leading to an underuse of the procedure.

The respondents mentioned some referral cases where there was said to be prolonged labor and the partograph indicated CS, and still the membranes were not ruptured. When artificial rupture of the membranes (ARM) was performed, the women gave birth. One doctor expressed frustration with guidelines that may indicate early interventions and CS when unnecessary:

‘… we discuss the difficulty at hand but at the end of the day if I follow my decisions, I find I have helped the mother to have her baby normally; I have not followed guideline.’ (Doctor, ZCUH)

However, for referral cases presenting with dry vagina of high temperature and with the child’s head high, the respondents found that vaginal birth was unlikely to take place.

Regarding oxytocin augmentation, respondents agreed that it should not be administered too early, and that there should be normal fetal heart rate (FHR) in advance. A nurse-midwife at the ZCUH said that the FHR should be monitored ‘most of the time’ when a laboring woman is augmented with oxytocin. Unnecessary induction and inaccurate administration of oxytocin were reported as causes of prolonged labor among the respondents at both hospitals. They expressed a challenge related to the practical administration of oxytocin – whether the valve was sufficiently opened. A doctor explained:

‘… you will find oxytocin dripping but there is no change in the situation of the mother. It is you who have made the decision of applying oxytocin while she had mild contractions and you have not opened the valve for oxytocin enough for the mother to have strong contractions.’ (Doctor, ZCUH)

The nurse-midwives expressed a need for guidelines regarding the use of oxytocin augmentation as well as when ARM should be performed, although experience had taught them that performing ARM too early would result in prolonged labor, whereas delayed ARM may prevent it. When discussing ARM, nurse-midwives expressed fear of endangering the child by cord prolapse or increased risk of infection, especially transmission of human immunodeficiency virus (HIV).

Discussing the timing for interventions, nurse-midwives addressed the right time for admission to the labor ward. They said that too early admittance may result in misdiagnosed prolonged labor. Too late admittance, as for many referral cases due to repeated delays, made them anxious for both mother and child. A nurse-midwife said:

‘If prolonged labor happens far away in the rural areas, it takes time to diagnose and to organize. To transport the patient takes time as well. By the time she reaches here, she is tired and also the baby is tired.’ (Nurse-midwife, ZCUH)

Monitoring fetal heart rate

Respondents from both hospitals conveyed challenges related to finding and interpreting fetal heart rate (FHR). Differentiating FHR from the maternal pulse was found to be difficult, as explained by a nurse-midwife:

‘It is possible you have listened, but you listened to the maternal heartbeat. You can say there is fetal heartbeat, but if you have not incorporated your colleague or you have not asked the mother if the baby is kicking inside and confirmed it with ultrasound - at the end of the day you come up with such results [stillborn/macerated].’ (Nurse-midwife, RRH)

Several nurse-midwives mentioned incidences where there was said to be FHR, but the child was stillborn. They understood FHR as of either low, normal, or high frequency, and with strong or weak sound. A nurse-midwife described continuous monitoring as listening to the FHR every half hour. Obese or non-cooperative mothers and cases of malpositioned children, represented additional challenges in finding FHR. Some respondents described that they faced no challenges related to the interpretation of FHR, only difficulties in monitoring. Difficulties were closely related to distrust in all available fetal monitoring equipment, as explained by a nurse-midwife:

‘… there is no answer which I can precisely give if I can rely on either pinard, fetal scope, doppler or even ultrasound - all has its challenges. We have an experience of being told there is fetal heartbeat but on delivery you get a fresh stillbirth.’ (Nurse-midwife, ZCUH)

Nurse-midwives described how dopplers might give faulty readings and mislead them in their work, resulting in children with surprisingly low Apgar score at birth.

Working as a team

Among the respondents, the perception of urgency varied. Teamwork was perceived as challenging when the doctor gave the laboring women time but forgot to follow up or ‘gave time’ repeatedly. Nurse-midwives in the RRH described how doctors gave more time although they were informed that FHR was negatively affected, as this nurse-midwife explained:

‘I am the one who knows the patient, I stay with the patient, maybe I have already taken one or two actions. Maybe FHR is 100 or 90 and I can see it is in distress and contractions has slowed down and child cannot come out. I have called the doctor who says we should give her another one hour. Personally, I will tell the doctor … it is not possible to give mother another one hour.’ (Nurse-midwife, RRH)

The doctors said that they only gave time if no danger signs were present for mother or child. They described being available after allowing additional time as a necessity.

Teamwork seemed to be challenged by a mutual distrust between experienced nurse-midwives and inexperienced doctors. In decision-making, doctors were sometimes bypassed if the nurse-midwife did not agree with the decisions made. Nurse-midwives described this bypassing as a way of advocating for the women and children when they felt their conditions were not taken seriously. They argued that increased independence for nurse-midwives, especially when managing uncomplicated labors, would benefit the childbearing women. At the same time, they described fear of being accused of wrongdoing. When fearful of a poor outcome, they put effort into documenting the dialogue with the doctor to minimize the risk of blame.

Strengths and limitations

The main strength of this study is that it conveys both nurse-midwives’ and doctors’ perspectives on their management of prolonged labor and what they find clinically challenging. Cross-cultural research carries a risk of missing or distorting information. Attempts were made to limit misinterpretations through recurrent dialogue and close cooperation with the project group of EPSHILS in Tanzania.