Improving emergency obstetric care and reversing the underutilisation of vacuum extraction: a qualitative study of implementation in Tete Province, Mozambique

Tete province is situated in the central region of Mozambique in south-east Sub-Saharan Africa, one of the least developed countries in the world [32]. The province has a surface area of 100,724 km² subdivided in 15 districts, a population density of around 25 inhabitants per km² and a largely underdeveloped transport infrastructure. In recent years Tete has experienced a mineral resources boom, resulting in important socio-economic changes and considerable population growth due to national and international immigration. Despite this, the local economy is still mostly based in subsistence agriculture and poverty levels and inequality are high [33]. Literacy levels are low, especially among women, and access to safe water and sanitation or improved housing limited [34]. The epidemiological profile of the province identifies a multiple burden of diseases, with persisting high levels of malaria, diarrheal diseases, HIV/AIDS and other infections, frequent occurrence of health problems related to reproduction and emerging problems of non-communicable diseases and trauma [35]. A description of the provincial health sector is presented in Table 1.

The Mozambican public health sector is financed by the government. In 2014, it was allocated MT 19.1 billion (USD $635.8 million), 7.9% of the total state budget, equivalent to approximately USD $42 per person, far below the SADC average of USD $266 [36]. However, the sector also receives important financial contributions (estimated at an additional 25%) from a multitude of development partners, and is still considered donor-dependent [36]. Tete provincial health sector receives considerable support from bilateral and multilateral donors, and from national and international non-governmental organisations. An important part of this support is directed at HIV/AIDS (largely funded by the Government of the United States of America), while another part focusses on health system strengthening to improve sexual and reproductive health and rights, within which attention has been directed at the improvement of emergency obstetric care (largely funded by the Government of Denmark). In Mozambique, maternity care is free of charge at point of care, as are most other health services for vulnerable population groups.

Improvement of maternal and perinatal outcomes had been considered a priority in Tete province for years, following the lead of the Mozambican Ministry of Health. Till 2014 most efforts were directed at promotion of institutional childbirth (through expansion and refurbishment of maternity services and health education) and of modern contraception (through health education and the introduction of new methods of long duration, including implants), with considerable success: institutional childbirth coverage increased from around 50% in 2010 to over 66% in 2014, and the number of couple-years protected per 100 women of reproductive age increased from close to 13 in 2011 to over 32 in 2014. Despite this progress, the provincial institutional maternal mortality rate and stillbirth rate remained high at 100–150 per 100,000 births and around 18 per 1000 newborns, respectively. Provincial managers recognised that coverage and quality of maternity care including emergency obstetric care was insufficient and required interventions. In Mozambique, vacuum extraction tended to be considered as a rather risky practice with many potential dangers for mother and child, and was practiced in only a few health facilities. The practice of craniotomy (perforation of the head of a dead foetus to evacuate its content followed by extraction of the foetus) was even less known about, and considered potentially culturally unacceptable. Despite this, Tete’s provincial health management team authorised implementation of four complementary interventions to address the situation from March 2015 onwards: 1) one-week training of a provincial pilot team on the management of prolonged labour, including vacuum extraction and craniotomy, developed and facilitated by a professor considered an expert on obstetric care in low-income countries both internationally and in Mozambique; 2) strengthening of the quarterly emergency obstetric care accreditation process; 3) monthly audit of clinical files of all cases of caesarean section in the provincial hospital (a large majority of all caesarean sections in the province) with specific feedback to clinicians involved, copied to relevant provincial managers; and 4) strict monitoring and evaluation of routine data from all maternities, with specific feedback to district health directors and doctors, copied to all and to relevant provincial managers. These four non-clinical instruments (training, monitoring and evaluation, audit and constructive feedback) have been shown effective in changing behaviour of health professionals [23, 37‐42]. They are presented in more detail in Table 2.

The focus on vacuum extraction was chosen because it is a relatively safe intervention even in inexperienced hands, with ample scientific evidence for its potential to reduce maternal and perinatal morbidity and mortality [5, 10, 13, 28]. Craniotomy and other destructive procedures have also long been considered safer than caesarean section in cases of prolonged and infected labour with intra-uterine foetal death [43, 44]. As no specific budget had been allocated to these interventions during the annual planning process for 2015, all interventions occurred within routine activities and without additional financial incentives, except the training, which was funded through the program for health system strengthening to improve sexual and reproductive health and rights.

The number of health facilities quarterly accredited in basic or comprehensive emergency obstetric care is presented in Fig. 1. After mid 2015 vacuum extraction was no longer the least performed signal intervention, replaced in that position by intra-uterine aspiration, administration of antibiotics or of magnesium sulphate. The performance of all signal interventions improved, as any one was performed each quarter on average by 25 health facilities in early 2015, increasing to 44 at the end of 2016, and to 47 in the first quarter of 2017, after which it reduced somewhat and stabilised around 42 during the remainder of that year. As a result, considerable changes were observed in the coverage of emergency obstetric care at provincial level, although some progress was lost during 2017 (Table 3). The use of vacuum extraction increased considerably from 2014 to 2016, and stabilised during 2017, while the percentage of institutional births assisted by caesarean section reduced (Table 4). Not only the absolute number and the percentage of institutional births assisted with vacuum extraction increased dramatically, but also the number and proportion of health facilities which performed at least one vacuum extraction during each three-month period. The percentage of institutional births assisted by either vacuum extraction or caesarean section therefore increased considerably, especially in district and peripheral health facilities without surgical capacity.

The monthly audit of the clinical files of all caesarean sections in the provincial hospital revealed that, initially, surgery occurred often ‘too little, too late’ but also ‘too much, too soon’ [45]. Similar audit findings have been described for various hospitals in Tanzania [46‐48]. After interventions started in April 2015 changes occurred gradually, accelerating during 2016, as shown through selected indicators in Table 5. The observed reduction of births attended in the hospital appeared to be related to the inauguration of an additional Mother and Child Health (MCH) nurse led maternity close by and to an increased use of peripheral maternities in case of uncomplicated births. Craniotomy and other destructive procedures were introduced as a substitute to caesarean sections in cases of intra-uterine foetal death, without any objections of patients or their families after appropriate explanations with informed consent. The persistently high intra-hospital caesarean section rate was probably related to the changing case-mix of relatively more severely complicated births and referrals, as also suggested by the increasing concentration of maternal deaths in the provincial hospital. It appeared that the shift in indications for caesarean section had reduced at least part of those ‘too little, too late’ as well as of those ‘too much, too soon’, without compromising maternal and foetal outcomes.

At provincial level, maternal and perinatal outcomes after institutional childbirth improved between 2014 and 2017. Institutional maternal mortality per 100,000 births and stillbirths per 1000 newborns reduced considerably (Fig. 2). Case fatality rates of direct obstetric complications reduced from 2.2% in 2014 to 0.4% at the end of 2016 and 0.8% during 2017. The reduction in case fatality rates was similar in each direct complication, prolonged labour, obstetric haemorrhage, (pre-) eclampsia, and sepsis; in 2016, all achieved rates below 1% due to a combination of fewer maternal deaths and more frequent diagnosis and treatment of cases with complications. However, during 2017 the case fatality rates for (pre-) eclampsia and sepsis rose again to well above 1%. The monthly institutional stillbirth rate was inversely associated with the percentage of vacuum extractions among institutional births, but positively associated with the percentage of caesarean sections (Figs. 3 and 4).

Mozambique does not collect reliable routine data on neonatal deaths. Yearly admissions due to asphyxia to the neonatal intensive care unit in the provincial hospital fluctuated around 400–450 between 2014 and 2016, and the case fatality rate ranged from 21% in 2014 to 16% in 2015 and 29% in 2016. Nevertheless, the impression of the provincial paediatrician as well as of provincial programme managers is that perinatal outcomes in general, both stillbirth and neonatal death, had improved in recent years. Taken together with the increasing coverage of institutional births, it is likely that the improvements in institutional maternal and perinatal outcomes extend to population level.

It has been well documented that many interventions to improve health have been scientifically proven effective, but are not generally applied in daily practice, as was the case with Mozambique’s obstetric care guidelines [49]. The lack of implementation may be attributed to barriers at health system level, such as insufficient funding or human resources, or at health facility level, such as a lack of equipment or of capable midwives, a lack of leadership to ensure quality of clinical care or inadequate teamwork [50‐52]. Rather than documenting barriers for implementation, this study chooses to present enabling factors, organised within an implementation science perspective, in factors related to a) the innovation itself, b) the structural and organisational environments in which the innovation is introduced, c) the people involved in the implementation (managers, staff, patients), and d) the process through which the implementation was rolled-out (adapted from Damschroder et al., 2009 and Chaudoir et al., 2013 [29, 30]).

An important key to success is the choice of the correct intervention for the problem which needs solving. In this case, the main problem to be addressed was the inadequate management of prolonged labour as principal cause of institutional maternal death and stillbirth. Vacuum extraction was re-introduced within a generally improved management of labour, including completion of the partogram and its use for timely decision-making, conservative interventions such as artificial rupture of membranes and augmentation with oxytocin intravenously, with caesarean section as last recourse. This course of action is in accordance with national and international guidelines, and was recommended by a senior expert with longstanding experience in Mozambique and good connections with several high-level personnel of the Ministry of Health, which made the advice unreservedly acceptable to Tete’s health personnel and managers.

Other positive aspects of vacuum extraction are directly related to the instrument. It is safe even in less-experienced hands, it offers the opportunity to regularly practice in non-urgent situations, for example in tired nulliparous women with a slight delay in the second stage of labour. It is a procedure of little complexity, which can be guided by few simple criteria and rules. The components of the instrument may be put together in a variety of combinations, which makes functional availability at peripheral level more likely. The positive effect of its successful use is immediate and rewarding for the practitioner (and the patient). This offers a more favourable option in the periphery compared with referral and summoning an ambulance or facing the possibility of an asphyxiated newborn or intra-partum stillbirth. The new handheld model vacuum extractor, received from national level, is small and easy to use, which encouraged its utilisation. Even after the model tended to break down with repeated use, MCH nurses were already convinced that vacuum extraction was an intervention they could perform and was beneficial to them and their patients. This meant they would look for, assemble, and use alternative equipment, which was already available to some extent in the province. Many additional resources were thus not required: checking inventories and subsequent redistribution was largely sufficient. Some experience with vacuum extraction was already present among the MCH nurses, which facilitated in-service training within existing teams. Similarly, some experience with augmentation with oxytocin or misoprostol, vacuum extraction, craniotomy and other destructive procedures, existed among obstetricians in the provincial hospital.

The Mozambican government and health sector are structures with a strong hierarchy, which consider the reduction of maternal mortality an important priority. Every month, data on maternal deaths must be presented directly from the provincial health directorate to the provincial governor and to the minister of health, which creates considerable pressure to find ways to success. This pressure is then translated down the chain of command, via province and districts to health facilities and individual staff, creating an environment where the reduction of maternal deaths must be on everybody’s agenda. It also sets the stage for a certain competition, for peer pressure between staff, health facilities, districts, and even provinces, to find answers and produce positive results, in the expectation to be recognised and rewarded for them. In a similar fashion, this hierarchical pressure also stimulated Tete’s provincial hospital to try harder to prove its status as tertiary referral hospital with a ‘model maternity’, especially after peripheral health facilities took the lead in improving emergency obstetric care and vacuum extraction.

Based on local research and evaluations, a common understanding had been created within the provincial health directorate on the main contributing factors towards the lack of progress in maternal and perinatal mortality, through which ‘inconvenient truths’ such as poor quality of care within health facilities could be assumed and subsequently addressed. The concurrent programme aimed at health system strengthening focussed on results-based management with evidence-based decision-making, using a ‘complex adaptive systems’ perspective. This led to improvements in health system management, governance, financial management, logistics and transport, human resources management and monitoring and evaluation throughout the provincial health sector. The process involved attention for stronger collaboration, promotion of transformational leadership and iterative learning. Taken together, this permitted the emergence of ‘champions of vacuum extraction’ at peripheral health facilities, whose positive experiences and good initial results aided in a gradually expanding practice of vacuum extraction at provincial, district and health facility level.

Staff: Many MCH nurses were afraid to perform a vacuum extraction, worried about possible damage to mother or child and of repercussions in case of failure to save the baby. Once the process of vacuum extraction and its components were demystified through simple rules and multiple possible assemblies, and once regular performance of vacuum extraction was presented as a routine professional practice of any competent MCH nurse, more staff were gradually willing to try out the procedure in practice. Any attempt to perform vacuum extraction was always praised, even when unsuccessful, because trying and failing was considered better than not trying at all. Many MCH nurses discovered quickly that vacuum extraction is not a difficult procedure, that it is largely safe for mother and child and that it forms a powerful tool for any MCH nurse to resolve difficulties in the second stage of labour without need to obtain outside assistance, avoiding unnecessary referrals or caesarean sections. In peripheral health facilities, MCH nurses realised that it was often easier for them to perform vacuum extraction than to convince the patient and her family that referral was necessary and to subsequently arrange an ambulance. The following quote from a group of MCH nurses in a provincial meeting on maternal and perinatal mortality summarises this process of empowerment:

“We knew that a vacuum extraction is an instrumental childbirth which might reduce perinatal mortality and also reduce the anxiety of the mother in a situation of prolonged labour. But in the beginning, we were afraid and apprehensive. We needed individual change, to gain courage and start performing vacuum extractions when we as MCH nurses considered it necessary to save a foetus in danger of perinatal mortality. Now, when we do a vacuum extraction, it means a victory and an achievement to us.”

In the provincial hospital, staff gradually gained the confidence to promote vaginal births rather than quickly deciding on a caesarean section. The understanding emerged that caesarean section is not always the best solution, that a vaginal birth may be safer for mother and baby, and that caesarean section might be best seen as a last resort if all else fails. Better collaboration within the multidisciplinary team, with clearer divisions of roles, more dialogue and second opinions aided in this process.

Patients: Although direct contributions from patients were not available, staff reports suggested that assistance with vacuum extraction during prolonged labour was gratefully received by patients, especially when it saved the baby and avoided referral for caesarean section in a far-away hospital. For socio-cultural reasons, vaginal births are traditionally better accepted than caesarean sections among Tete’s population. Referral to a health facility with surgical capacity is not popular, due to difficult travel, return transport costs for the patient and her family, fear of surgical intervention, and the unfortunate fact that often the baby cannot be saved even through surgery due to unavoidable delays related to long distances and poor transport infrastructure. Consequently, many women in Tete’s rural communities know sad histories of acquaintances or family members who experienced difficulties when giving birth, were referred, had a caesarean section but lost their baby and never got pregnant again. Vacuum extraction appears thus as a potentially acceptable tool to promote vaginal birth resulting in a live, healthy mother and baby at peripheral health facilities.

While a common understanding of the problem of prolonged labour and its possible solutions existed in the province, planning of the four interventions was very limited. Only the training was formally planned, the other interventions were suggested during the training or were included spontaneously shortly afterwards. The roll-out of the innovation did not follow the usual process in the Mozambican health system, consisting in widespread formal training of all staff involved with elaborate written guidelines and protocols. Rather, the implementation process in Tete was flexible and responsive to changes and resistance, relying mostly on personal initiatives of MCH nurses and other maternity staff, guided by few simple rules and criteria within their day-to-day routines, supervised through existing hierarchies by district medical officers and health directors. By largely avoiding classroom teaching, the impression that vacuum extraction is a difficult procedure which must be formally learned was avoided. This also prevented passive resistance from MCH nurses who had not benefitted from formal training and its accompanying per diem, which tends to be widespread in Mozambique [53].

Most encouragement was directed towards those maternities attending the largest numbers of institutional childbirths, ensuring enough cases to permit regular practice to gain experience and develop skills. Improved practices in these maternities also ensured an impact on a considerable number of childbirths, enough to provoke an impact at provincial level. Early positive results stimulated MCH nurses and district doctors to continue and expand the practice, creating a snowball effect, increasing practice and increasing benefits for maternal and perinatal outcomes, resulting in the creation of a virtuous circle.

Initial results were presented to the provincial government and the Ministry of Health, which created a pressure to move forward, as permitting a resurgence of maternal and perinatal mortality after its reduction was shown possible would mean a severe loss of face for Tete’s provincial managers. The practice of vacuum extraction was therefore included in the list of priority activities on which monthly feedback was required via phone message from districts directly to the provincial chief medical officer. Pressure for accountability on results via email, WhatsApp, phone messages, reports, meetings, etc. was strong and mostly public among provincial and district staff, as were prizes and praise for those performing well. As such, peer pressure aided in the establishment of a new cultural norm in which vacuum extraction is now considered a normal part of day-to-day practices of any competent MCH nurse or district doctor. However, this peer pressure was interrupted in 2017 due to staff changes at provincial level affecting the chief medical officer, chief public health officer, MCH manager and locally-based international technical advisor, which explains the partial relapse noted during 2017, especially in the implementation of all signal interventions as required for accreditation. Determined to maintain previous advances, steps to resume the regular accreditation and feedback system have been taken by relevant managers of the Provincial Health Directorate at the start of 2018.