STRENGTHS AND LIMITATIONS OF THIS STUDY
-
The retrospective exposure-based study design with adjustment for potential confounders using multivariable analysis had higher rigour compared with commonly used simple pretest/post-test analysis and provides timely results at lower cost for policy-makers, compared with prospective study designs which have stronger rigour.
-
The use of objective questions which assessed changes towards the expected training goals in health workers’ knowledge, attitudes and skills generated non-biased data compared with previous studies that used health workers’ personal perspectives.
-
Some of the study questions such as satisfaction rates and practice patterns were affected by respondent bias and would need further refinement, testing and validation.
-
The absence of data from midwives’ clients limited insights into their perspectives and limited our ability to assess impact of trained midwives’ interventions.
-
Kirkpatrick’s evaluation model was a useful framework to assess training outcomes; however, an ecological framework which recognises the complex interplay between individuals, families, communities and society including the health system would need to be included in future evaluations.
Introduction
Female genital mutilation (FGM) is a harmful practice that affects 14 million women and girls in Sudan.1 This practice involves partial or total removal of tissue or other injuries to the external female genitalia.2 In Sudan specifically, the the most common type affecting 72% of girls and women (15–49 years) living with FGM is classified by the WHO as type 3 or ‘infibulation’. This type involves the cutting of the inner and outer vulvar folds with or without removal of the clitoral glans and closing the outer vulvar folds leaving a small opening for urine and menstrual blood flow.3 Among girls under 15 years, FGM is mainly reported to be performed by midwives (64%) known as ‘FGM medicalisation’,2 followed by traditional practitioners (29%).4 Furthermore, 24% of girls and women (15–49 years) who gave birth in the preceding year report having FGM type 3 repeated or ‘re-infibulation’ performed by midwives.4 Midwives’ involvement not only violates the health professional code of conduct to ‘do no harm’ but also endorses the practice and negatively impacts FGM abandonment efforts.
The high FGM prevalence and the involvement of midwives who make up 23% of the health workforce5 prompted the Ministry of Health (MoH) to introduce FGM-related content into preservice and in-service trainings in efforts improve the quality care to FGM survivors and stop FGM medicalisation. Further, midwives on graduation are required to make an oath not to perform FGM in midwifery schools. This is followed by receiving sensitisation session on FGM medicalisation being a violation of professional code of conduct during the 1-day induction sessions for professional licensure in the National Medical Council for Health Professionals (NMCHP).
The FGM training in preservice and in-service training covers content on FGM epidemiology, typology, health complications, non-linkage of FGM practice with religion, midwives’ role as change agents and FGM medicalisation as a violation of professional code of conduct that carries administrative punitive measures. The FGM-related content emphasis, training methodology and duration varies by the training modality, training institutions and over time. The FGM content is spread across harmful practices/community health and/or clinical modules within the 1 year or 2–4 years preservice curricula in midwifery schools for community or facility-based midwives respectively. In comparison, the FGM-related sessions are provided in 1–2 days during a 10–12 day in-service training by the Federal Ministry of Health (FMoH) and sessions less than 1 hour during the prelicense 1-day induction training. Further, the 1–2 in-service training content emphasis changed over time. Curriculum versions after 2016 were adapted to the first WHO’s guidelines on FGM4 with content mostly on clinical management of FGM-related health complications. In 2018, the in-service training version was revised to focus more on skills for social norm change via communication, community dialogue and advocacy adapted from the United Nations Population Fund and United Nations Child Fund’s manual on social norm change (2017).
Not all midwives receive the same exposure of FGM training content, MoH records indicate that 18%, 31% and 9% of all the midwives (n=16 183) in Sudan received FGM content during prelicence, 2016–2018 version of in-service trainings, and 2018 version of in-service training respectively.6 The effectiveness of the different trainings received by midwives in Sudan was mostly captured in pretests and post-tests which focused on knowledge attainment only. There was no long-term follow-up data on trained midwives’ FGM-related knowledge, attitudes or practices.
To our knowledge, there is no current global literature on the effectiveness of FGM-related trainings implemented at large scale. Much of the existing literature examining the effectiveness for FGM-related training on health workers’ knowledge, attitudes, self-efficacy, skills and patient satisfaction is generated from small scale training interventions in controlled study settings.7 In addition, the study populations in the literature comprised mostly of health workers in settings where either FGM prevalence or FGM medicalisation is low. As such, the training intervention effectiveness might not be replicable in settings where FGM medicalisation and FGM prevalence are high.
Furthermore, there is a literature gap on trainings’ effectiveness on intentions to change clinical practice or actual changes in clinical practice in either a study setting or for real-world training interventions.8 9 We also found no study assessing training effects beyond 6 months of training. Large-scale trainings are costly (mean of US$296 USD (median: US$157) per midwife)6 and when implemented alone and over long periods of time, this type of intervention raises costeffectiveness questions for governmental and donor spending among decision-makers. It is therefore important to assess long-term effectiveness of scaled FGM training on midwives’ knowledge, attitude and clinical practices.
Our study aims to fill some of the identified gaps in evidence with regards to the long-term effectiveness of a large-scale training programme on knowledge, attitudes and on clinical practice of health workers who are also involved in FGM practice and who work in FGM-prevalent settings. We conducted an exposure based cross-sectional study in Khartoum State, Sudan to identify associations between current FGM-related knowledge, attitudes and practices among midwives who received or did not receive past FGM-related trainings. We hypothesised that midwives who reported having received past FGM related trainings would have a significantly higher knowledge, greater opposition towards FGM medicalisation and improved clinical practice compared with those who have not.
Methods
Setting
Khartoum State houses 11% of all midwives in Sudan (n=16 183).10 Most of girls and women aged 15–49 years (88%) living in Khartoum State have experienced FGM mostly type 3 (72%) and 22% of this age group who gave births in the preceding year underwent reinfibulation. Further, Khartoum State has the highest FGM medicalisation prevalence (89%) in Sudan, performed mainly by midwives.4 During the period 2016–2018, the FMoH and the NMCHP training records for Khartoum State indicate that 75%, 76% and 28% of midwives received 2016 version of FGM-related curriculum in-service training, FGM content during prelicense induction training and 2018 version of FGM-related curriculum in-service training, respectively.
Study design and study population
This exposure based, cross-sectional study compared current FGM-related knowledge, attitudes and practices among midwives who reported to have ever received FGM trainings to those who reported no training.
For the study sample estimation, we used a prevalence of 5% for knowledge on FGM types, 90% for attitudes against FGM practice and 7% for correct knowledge of FGM complications management as a proxy for practice among untrained midwives from previous assessments.11 12 We anticipated a difference of 20% for FGM-related knowledge and practice and 10% for attitude between trained and non-trained arms using 80% for power and 5% alpha error. The sample size generated for these differences ranged between 46 and 71 midwives.
We retrieved registration and training records of midwives in Khartoum State to identify and generate lists of potential trained and non-trained midwives. We used random number generator for sampling. Because of uncertainty about whether records on training status on FGM were complete and up to date, we oversampled by 30% and 100% from the lists of trained and non-trained midwives respectively, so that we could reclassify as needed based on self-reported training status. We then obtained the phone contact details of the selected midwives from the FMoH and Khartoum State Ministry of Health.
Materials and methods
We used Kirkpatrick’s four level training evaluation model as a framework to assess long-term training effectiveness. The first-level ‘Reaction’ focuses on trainees’ perceptions. The second-level ‘Learning’ evaluates whether the trainings’ learning objectives were met. The third-level ‘Behavior’ assesses behavioural change while the fourth ‘Results’ evaluates the training impact on the organisation, quality or user of service. For this study, we used Kirkpatrick’s first three levels since our study included only midwives but not their clients to assess impact.
The primary exposure in the study was any reported previous FGM-related training received by midwives and the training outcome variables for each of Kirkpatrick’s levels are listed in table 1.
Training outcome variables to assess training effectiveness among midwives who received or did not receive FGM content during induction and in-service trainings in Khartoum State
The study tool had 21 questions which included closed and open-ended questions. The first section had six questions on current demographic data (age, sex, midwifery qualification, midwifery work experience, health facility level) which was used for descriptive data and as potential covariates to be controlled for in the multivariable model.
The second section assessed characteristics of training and satisfaction scale (Kirkpatrick level 1), for midwives who received FGM training. This section had eight questions on training type (in-service, induction training, other), FGM training year categorised into versions of the training curriculum received (<2016, 2016–2018, >2018), training duration (days), training institution (FMoH, non-governmental, other) and satisfaction levels (Likert scale) with respect to knowledge and skills gained for FGM prevention and complications management.
The third section had seven questions on current knowledge, attitudes and practices. Open-ended questions focused on describing the FGM types and listing four FGM complications they know of. Closed-ended attitudinal questions assessed agreement with statements about their stances on FGM, FGM medicalisation, efficacy to change practice and experiences providing FGM-related services (see table 1).
The co-Principal Investigator (Co-PI), MA, translated the study tool into Arabic that was cross-checked by PI (WA). The co-PI used the Arabic version to develop an online survey administration software (google form survey) with constraints and skip patterns to minimise errors and trained four data collectors with experience in health survey data collection on research ethics and the electronic questionnaire. We used phone interviews because of the high network coverage in Khartoum state. Second, midwives are community-based (83%)5 and own cellphones as a means to be accessed by their clientele who are dispersed in large geographical areas. Third, because of the nature of their clinical practice, we anticipated difficulties for in-person interviews. Data collectors contacted midwives over the phone, explaining the study’s objectives, participation as a voluntary exercise and their rights to refuse or stop interviews at any point before obtaining verbal consent. On obtaining verbal consent, data collectors then interviewed and entered electronic data. The data collection period started on October 2022 and ended in January 2023.
Data analysis
The PI (WA) and co-PI (MA) reviewed, cleaned and coded the autogenerated data set from google survey into Microsoft Excel (2018). The open-ended responses on FGM types and complications were coded according to WHO definitions and categories.3 The data set was then imported into StataCorp. 2021. Stata: Release V.17. The scale of missing data was less than 5% and randomly missing that did not require any imputation. We conducted descriptive analysis of study population and the trainings. Student’s t and χ2 tests were conducted for continuous and categorical variables, respectively.
We conducted univariable and multivariable logistic models to examine the association between ever being trained (independent variable) and outcomes on knowledge, attitude and practices (dependent variables) as primary analysis. For the multivariable models, we controlled for age, a characteristic that was significantly different between the two groups and was also related to training exposure, midwifery expertise and training outcomes. Because we conducted multiple testing, we used Holm-Sidak corrections for p values. We conducted exploratory analysis to examine associations between training outcomes and training year as a proxy for training curriculum content, as well as number of trainings received.
The study was considered as programmatic research not subject to human research protections; nevertheless, we conducted the research following research ethics guidelines including obtaining verbal consent before data collection.
Public patient involvement statement
This research did not involve patients in the development of research questions and outcomes. It was not appropriate or possible to involve patients or the public in the design, or conduct, or reporting, or dissemination plans of our research as we did not conduct the study at health facility level to retrieve patient records or have access to them. We note this limitation in method and discussion section and recommend their involvement in future studies.
Results
Among the sample of 246 midwives, we interviewed 182 (74%); 74 (30%) were not traceable because of non-valid phone numbers, 5 (2%) were not eligible because they were not midwives, 2 (1%) were dead and 2 (1%) had left Sudan. All the 182 midwives reached agreed to participate and completed phone interviews lasting 10–15 min.
Most midwives (70%) reported having received an FGM training by the time of data collection (table 2). All the midwives interviewed were female, most trained as village midwives (92%) and most worked in health centres (89%) (table 2). The mean age was 51 years (SD=10), and mean years of midwifery experience was 23 years (SD=12). Both study populations had similar demographic and professional characteristics, but the groups differed in mean age. Midwives who did not receive any FGM training were older (M=56 years, SD=11) than those who were trained (M=49 years, SD=9) and this difference was statistically significant (p≤0.001).
Characteristics of midwives who received and did not receive trainings on female genital mutilation in Khartoum State, Sudan
The largest group of trained midwives (46%) reported receiving two trainings (table 3). There were similar proportions of midwives who received each of the three FGM training curriculum versions that is, prior to 2016, 2016–2018 and after 2018 versions. The mean duration of in-service training was 3 days (SD=1) during 2016–2018, 3 days (SD=1) after 2018 and 2 days (SD=1) for trainings prior to 2016. Most of in-service training was conducted by the Ministry of Health (94%) followed by non-governmental organisations (6%).
The reported type, year and number of trainings with female genital mutilation content received by 127 midwives in Khartoum State, Sudan
FGM training content satisfaction levels (Kirkpatrick level 1)
Generally, the satisfaction levels on the knowledge and skills gained on FGM prevention and care management were high for in-service or midwifery induction trainings. Overall, most of the trained midwives (89%–100%) reported being either ‘very satisfied’ or ‘satisfied’ on FGM prevention and health complications management knowledge and skills. A higher proportion (12%) of midwives who received midwifery induction trainings (n=26) reported less satisfaction on skills on FGM complications management compared with midwives (5%) who received in-service training (n=64). This finding was affirmed by all midwives (n=24) who received both types of trainings.
Knowledge of FGM types and health complications (Kirkpatrick level 2)
Overall, less than a third of all midwives knew four FGM types (WHO definitions for FGM types: type 1: ‘Partial or total removal of the clitoral glans and/or prepuce’, type 2: ‘Partial or total removal of the clitoral glans and the labia minora, with or without excision of the labia majora’, type 3: ‘Narrowing of the vaginal opening with the creation of a covering seal by cutting and appositioning the labia minora or labia majora with or without excision of the clitoral prepuce and glans (infibulation)’, type 4 ‘All other harmful procedures for example pricking, piercing, incising, scraping and cauterization’) or four health complications. However, 44% and 80% were able to name more than two FGM types and health complications respectively. Of the FGM types named, type 3 and type 1 were the most common. While obstetric and chronic complications were the most reported FGM health complications. FGM-related acute, psychological and sexual complications were the least reported.
The difference in knowledge on FGM types and FGM health complications between midwives who reported receiving FGM training to those who did not, was not statistically significant.
FGM-related attitudes and practices (Kirkpatrick level 3)
Most of the midwives who reported ever receiving FGM training (99%) and most of those who did not (96%) were supportive of the abandonment of FGM and its medicalisation. A high proportion of midwives thought that FGM is a harmful practice that needs to stop (99%) and not a religious requirement (95%). They also believed that FGM medicalisation does not make the practice safer (97%) and is a violation of professional code of conduct (71%).
With regards to practice-related knowledge, 95% of all midwives reported that they always provided FGM counselling. In contrast, only 11% cited the correct deinfibulation surgical procedure during labour.
We found two statistically significant differences in FGM-related attitudes and practice-related knowledge between the two groups. Midwives who received FGM training were more aware that FGM medicalisation violates their professional code of conduct (p=0.001) and reported “always” on counselling patients to abandon FGM (p<0.001) compared with midwives who did not receive FGM training.
Table 4 provides detailed findings on training outcomes among midwives who received and did not receive FGM content during induction and in-service trainings in Khartoum State.
Training outcomes among midwives who received and did not receive female genital mutilation content during induction and in-service trainings in Khartoum State, Sudan
Univariable, multivariable and exploratory logistic model analysis
The univariable logistic model analysis (table 5) showed that trained midwives were more likely to be aware that FGM medicalisation was a violation of professional code of conduct (OR: 3.3, 95% CI 1.6 to 6.7, p=0.001) compared with non-trained midwives. However, after adjustment for age, this association was no longer statistically significant.
Female genital mutilation-related knowledge, attitudes and practices among trained compared with non-trained midwives in Khartoum State, Sudan
Exploratory analysis for training outcomes by reported training type (induction, in-service), training year (<2016, 2016–2018 and >2018) and training dose (1–3 trainings) showed higher OR in overall knowledge, correct attitude and practice among midwives who reported receiving FGM during in-service training before 2016 (online supplemental table 1). In particular, the knowledge on four correct FGM health complications (aOR 5.1, 95% CI 1.9 to 13.5, p=0.001) and awareness that FGM medicalisation as a violation of professional code of conduct (aOR 15.6, 95% CI 3.7 to 66.1, p<0.0001) was significant after adjustment for age. Similarly, midwives who received one training (aOR 27.0, 95% CI 3.4 to 211.8, p=0.002) or two trainings (aOR 6.2, 95% CI 2.1 to 18.1, p<0.0001) were more likely to be aware that FGM medicalisation as a violation of professional code of conduct. Interestingly, midwives who received FGM content during induction training in 2016–2018 (aOR 13.8, 95% CI 2.1 to 88.6, p=0.01) and those who reported FGM training before 2016 (aOR 11.9, 95% CI 2.0 to 70.5; p=0.01) had comparable results on knowledge of four FGM types.
Supplemental material
Discussion
This exposure-based cross-sectional study found that there were no significant long-term differences between trained and non-trained midwives with regards to the levels of knowledge, opposition towards FGM practice and its medicalisation and clinical practice in FGM prevention and care during labour. We did not find existing literature on long-term effectiveness for small-scale or large-scale FGM trainings to make direct comparisons. The closest comparable study we found assessed effectiveness after 4 years elapsed post training, showed no differences in domestic violence knowledge levels between trained educators and non-trained educators.13
The long-term effectiveness of large-scale trainings could be affected by the training modality used, by the quality and amount of FGM content received, or by training fidelity. Most midwives reported receiving in-service trainings, a method which has been shown in a systematic review to improve health workers practice in low-income settings.14 Our exploratory analysis compared the effectiveness of different curricula versions which had different content emphasis and suggests gaps in content especially after 2016 but the sample size was too small to impact regression model results and would need further investigation. Moreover, the global low knowledge on practical skills for the deinfibulation procedure among all the midwives in the study indicates training gaps in midwifery schools that would need to be investigated further.
The low training effectiveness could also be explained by the absence of supportive mentoring and supervision. Studies have shown that monitoring and evaluation (M&E) is an effective strategy in improving health workers’ practice in low-income settings.15 16 A study which assessed Sudan’s health programmatic interventions found no FGM-related data at facility level,6 suggesting limited tracking for M&E and low accountability for health workers on the quality of care and FGM medicalisation.
The finding on midwives reporting reinfibulation as a correct step in clinical management is concerning because reinfibulation increases genital tissue scarification in subsequent deliveries, thus increasing the risks to a wide range of gynaecological, obstetric and neonatal health complications. Furthermore, the reported reinfibulation as correct surgical procedure is incongruent with midwives’ high knowledge of FGM health complications, with their awareness that FGM medicalisation as a violation of professional code of conduct, and with their reported high frequency of counselling on FGM abandonment. This inconsistency in findings could be explained not only in training gaps but possibly due to deeply ingrained values17 that are harder to change. The anticipated findings of a current randomised cluster trial testing the effectiveness of a training targeting midwives in FGM prevalent settings using value clarification exercises18 may shed light on effective training content that can potentially change midwives’ value systems and clinical practices.8 Midwives in our study also reported having low influence in changing FGM practice, a finding similar to a study among trained Somali midwives who were not able to provide interventions because they were apprehensive in going against prevailing culture and religion.19 Finally, the financial incentive to perform reinfibulation may have outweighed individuals moral compass and professional code of conduct, in the prevailing context of high unemployment rates and low pay if employed.6 20 Financial incentives have been repeatedly found to be a driver for health workers in performing reinfibulation or involvement in FGM in general.21–23
Our study brings in new evidence on the long-term FGM training outcomes among health workers involved in the practice or working in a high FGM prevalent setting. Though our study design was cross-sectional, the use of comparative arms and multivariable analyses strengthened rigour which was noted to be absent from previous pre–post study designs.7 The retrospective, exposure-based study design was a practical approach to inform policy-makers and programme managers who require timely data for decision-making but may not have the financial budgets to conduct prospective studies.24 Finally, we used objective questions to assess actual knowledge instead of midwives’ perceptions as was done previously9 to generate non-biased data for decision-makers involved in training programmes and quality of care.
One of our study limitations was the absence of qualitative data to complement our survey data as well data from midwives’ clients to determine the impact of midwives’ knowledge, attitudes, and practices on the quality of care received by their clients or their quality of life (Kirkpatrick level 4). Additional qualitative data from midwives could have strengthened data triangulation and provided contextual depth and additional understanding of the factors affecting the training outcomes, such as translating their opposition towards FGM medicalisation and their current reinfibulation, that is, FGM medicalisation practice. Future evaluations may need to complement Kirkpatrick evaluation model with an ecological framework which recognises the complex interplay between individual, family, community and societal25 and the health system factors on midwives’ knowledge, attitudes and practices on FGM. Finally, some of the study questions would need further improvements, testing and validation to build on the existent gap on validated comprehensive knowledge, attitude and practice assessment tools for health workers.26 For instance, midwives’ high response rate in providing FGM abandonment counselling did not align with their belief that health workers have no time to provide FGM prevention services or influence to change FGM practice. Their high response may have been due to social desirability bias to meet the Ministry of Health expectation.
In conclusion, our study findings highlight the need to review all existing FGM training curricula content to assess adequacy in addressing midwives’ value systems on FGM and reinfibulation, self-efficacy in changing their clients’ stance on reinfibulation, and practical skills in performing deinfibulation and managing FGM health complications. Appropriate evaluation tools will need to be developed and used pretrainings and post-trainings and during supportive M&E. Finally, the influence of health system and societal factors on midwives’ practices should not be ignored in future evaluations to identify modifiable intervention areas that will enable midwives to translate their newly acquired knowledge and skills into practice.
Data availability statement
Data are available upon reasonable request. The datasets generated and/or analysed are available from the corresponding author on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by University of Washington Institutional Review Board (STUDY00012584) and Sudan’s FMoH National Health Research Ethics Review Committee (P2-3-21). Participants gave informed consent to participate in the study before taking part.
Acknowledgments
The authors would like to appreciate Mohamed Babiker from the Federal Ministry of Health and Mohamed Modber from Médecins Sans Frontières in Sudan and Amira Adam from the WHO Sudan country office during training data retrieval phase and responding to queries. Gabrielle O’Malley for guidance in study tool development. Data collection team and study participants who dedicated time and effort to generate important data for use.
This post was originally published on https://bmjopen.bmj.com