Vacuum-assisted vaginal delivery is not used widely in sub-Saharan Africa (4, 8, 9, 20). Despite measured success to re-introduce the practice in Kigoma (14), we knew that uptake of vacuum extraction was highly variable and we knew little about how providers perceived the procedure. This paper explored factors – individual, facility-based and contextual – that potentially influenced the practice of vacuum extraction in the last 3 months. Several factors stood out as more influential than others such as the reporting of having been exposed to different types of training, knowledge of indications and contra-indications for the use of vacuum extraction, type of facility in which the provider worked and health worker’s sex.
Much has been learned about the art and science of training health workers, for example, we have observed that centralized training is expensive and disruptive to the delivery of services because of absent health workers (21, 22). Also, didactic trainings are less effective than practice-based learning (21). The training modality that stood out above others in our study was the experience of hands-on practice of vacuum-assisted birth during in-service training. Two-thirds of those who reported in-service solo experience and half of those with pre-service solo practice had recently performed a vacuum delivery. Our findings confirm what researchers found in 3 regions in Senegal: 64% of the providers interviewed were trained to perform vacuum but only 30% did so as routine practice. The authors attributed this low level to the lack of or limited hands-on practice during training and limited knowledge of the indications, contra-indications and complications of vacuum extraction (20). Another recent study, this one conducted at the Muhimbili National Hospital in Tanzania, the country’s largest teaching hospital, described an environment of minimal vacuum extraction practice, extremely high cesarean delivery rates, poor vacuum extraction knowledge, reports of dysfunctional equipment and the need for updated clinical guidelines. The authors concluded that the primary solution would be to support a well-structured training program with hands-on practice for residents and midwives that would help counter a climate of providers’ fear of being blamed for unfavorable outcomes (23).
Practice-based learning in the field of emergency obstetric care is challenged by the rarity of some obstetric emergencies. This was addressed in Kigoma by ensuring that the majority of trainees had access to the use of simulation or mannequins to supplement their learning, and these are known to be effective training aids. A systematic review of operative vaginal delivery demonstrated that vacuum-assisted simulation training decreased many perineal lacerations and newborn injuries, and increased residents’ knowledge regarding vacuum-assisted delivery as well as their comfort performing the procedure immediately following the simulation training, at 4 and at 12-months post-training (24).
One training modality that has been reported to significantly influence the uptake and performance of other maternal and neonatal life-saving procedures is the use of low-dose high-frequency skills-building sessions. Studies in Ghana, Tanzania and Uganda reported improved knowledge and skills of providers and better maternal and perinatal outcomes when providers were trained through simulation, case-based learning and small content packages spread over short time intervals (25–28).
Multifaceted and frequent interventions tend to be more successful than one-off interventions (22). We also found this to be true, as the exposure to different types of training increased, so did the likelihood that the health worker had performed a recent vacuum extraction. Furthermore, the ability of an individual provider to seek additional training on their own (like with E-learning) was also associated with greater performance (22). This aligns with our finding of a strong correlation between recent vacuum extraction and experience with the Thamini Uhai E-learning platform.
The program’s primary focus was to support rural health centers to provide comprehensive EmONC, which meant that most supported facilities were in districts other than Kigoma Municipality. The fact that those facilities had more (consistent) interventions overtime may be one reason that recent performance of vacuum delivery was higher (personal communication with authors SD and MK). Another possible reason for this finding could relate to the high turnover and internal rotation of maternity ward staff to other departments at Maweni Regional Referral Hospital, located in Kigoma Municipality. Staffs from Maweni who were trained in vacuum extraction may not have remained long enough in maternity units for the practice to be sustained. Health centers have fewer staff with less frequent turnover and that may have contributed to higher performance at that level facility (personal communication SD and MK).
We found that recent performance of a vacuum-assisted birth was twice as likely among male providers as female providers but the reasons for this differential are not entirely clear. The decentralization of the EmONC program depended heavily on task-sharing, which meant extending training of vacuum extraction to nurse officers, assistant nurse officers, registered and enrolled nurses. However, the inclusion of these cadres was new and occurred mostly in the later years of programmatic support. They made up 44% of the health workers interviewed for this study and almost two-thirds were women. Assistant medical officers, largely male, were the most likely professional category to have recently conducted a vacuum delivery. However, the bivariate relationship between professional category and recent practice showed no association. When cadre was added to the multivariable model, it also showed no independent effect on practice. Female disadvantage may be a function of gender dynamics in society and the health system resulting in less confidence or assertiveness. In a study on health provider confidence in obstetric clinical skills conducted in Uganda and Zambia, researchers found that being female was associated with lower confidence compared to males after adjusting for cadre, age, and other covariates (29). Another study conducted with health providers in Nigeria also found that females were less confident than males in their clinical skills and providing obstetric care (30). However, neither study looked at relationships between confidence, training modalities and actual performance.
Based on this study’s findings, future pre-service and in-service training in vacuum extraction must combine various approaches that allow for frequent hands-on practice and self-directed learning ensuring health providers are knowledgeable, skilled and confident. This could include the development of a package that includes low-dose high-frequency training sessions with simulation and mannequins that are reinforced by accessible self-directed courses (e.g., e-learning modules), and backed up with regular supportive supervision, mentorship and routine clinical audits. To ensure more women are empowered and more confident to take advantage of hands-on solo practice during training, trainers should develop facilitation strategies that address gender dynamics in their training sessions. Ultimately, gender-integrated training should positively impact the quality of decentralized life-saving services.
There were several imitations to this study. The study’s sample was not representative but purposefully selected to learn more about the knowledge, practice and attitudes toward vacuum extraction in an environment where vacuum extraction had been widely introduced, with extensive access to resources (equipment, training, and on-site supportive supervision). Had resources for this study not been an issue, a more complex study design would have allowed us to compare providers who worked in supported facilities with providers working in facilities that received no external support (and where vacuum extraction is rarely used). This might have led us to a different set of findings. Another limitation of our study arose during analysis because we were unable to pull out facility managers and district health officials from the overall sample to see if they differed in knowledge, practice or attitudes.