Training Needs Assessment of Health Care Professionals in Reproductive, Maternal and Newborn Health in a Low-Income Setting in Tanzania

Background: Healthcare delivery globally and particularly in low-income setting is challenged by multiple, complex and dynamic problems. The reproductive, maternal and newborn health (RMNH) care is among the most affected areas resulting into high maternal and neonatal mortality and morbidity across the Sub Saharan region and Tanzania in particular. However, under-investment in adequate training and capacity development among health care workers (HCWs) is reported worldwide and contributes to the critical shortages, and lack of adequate knowledge and skills among HCWs. The aim of this study was to assess the training needs among HCWs of RMNH care in selected health facilities of Mwanza, Tanzania. Methods: A cross-sectional descriptive and analytic survey using a self- administered questionnaire was conducted in 36 out of 80 health care facilities included in Improving Access to Reproductive, Maternal and Newborn Health in Mwanza, Tanzania (IMPACT) project within the 8 councils of Mwanza region in Tanzania. The training needs assessment (TNA) tool adapted from the Hennessy-Hicks’ Training Needs Assessment Questionnaire (TNAQ) was used for data collection. The HCWs rated on the importance of their task and their current performance of the task. The differences in scores were calculated to identify the greatest training needs. Results: Out of 153 HCWs who responded to the TNA questionnaire, majority were registered (n=62) and enrolled (n=43) nurses. Ninety percent (n= 137) were from government-owned health facilities, mostly from hospitals 68 (45%). Training needs were high in 16 areas (out of 49) including cervical cancer screening and care; accessing research resources; basic and comprehensive emergency obstetric and newborn care; and sexual and gender-based violence. The overall perceived importance of the training needs was signicantly associated with perceived performance of tasks related to RMNH services (Pearson Correlation (r) = .256; p <001). Conclusions: The study highlights 16 (out of 49) training gaps as perceived by HCWs in RMNH in Tanzania. The utilization of ndings from the TNA has great potential to facilitate designing of effective trainings for local RMNH services delivery hence improve the overall quality of care. a consortium of Aga Development Network Aga Services Aga Khan in and the Aga Khan University East Africa EA), in partnership with the Tanzanian Government of Health, Community Development, Gender, Elderly and Children and President`s Oce Regional Administration and Local Government RALG) at national, regional, district and local communities. The study took place in selected government-owned and designated faith-based in all eight of

health care workers rated on their current perceived importance of a particular task and related performance in that particular task.
Training needs were high in 16 areas (out of 49) including cervical cancer screening and care; accessing research resources; basic and comprehensive emergency obstetric and newborn care skills; and sexual and gender-based violence. The overall perceived importance of the training needs was strongly associated with the total perceived performance of tasks related to reproductive, maternity and newborn care services. The utilization of ndings from this study has great potential to facilitate designing of effective trainings that will strengthen local capacity for services delivery hence improve the overall quality of care along the continuum of reproductive, maternity and newborn care.

Background
Healthcare delivery globally and particularly in low-income setting is challenged by multiple, complex and dynamic health problems (1). There is a need for changes in formal education and continuous professional trainings to ensure su cient quantity and quality of healthcare professionals. However, under-investment in quality education and training of healthcare providers (HCWs) is reported worldwide and contributes to the critical shortages, and lack of adequate knowledge and skills among HCWs (2).
The incongruity between education strategies, health system requirements and community demands also affects the quality of maternal and newborn care in healthcare systems (3).
In Sub-Saharan Africa, 36 countries including Tanzania, are considered to have a critical shortage of HCWs to provide minimum coverage of basic services in reproductive, maternal and newborn health (RMNH) and this problem is predicted to exacerbate in the future (4). The Sustainable Development Goals (SDG 3, target 3) emphasizes on the development of HCWs with adequate training, education and skills which are relevant to the needs of the population (5). This investment for better health will also contribute to improved economic growth and achievement of other SDGs (5).
Tanzania has an unacceptably high maternal mortality ratio (556 per 100,000 live births) and neonatal mortality ratio (25 per 1000 live births) (6). The areas bordering Lake Victoria (Mwanza inclusive) and Western zones have the highest rates of mortality (7). In an effort to reduce the problem, the Tanzania government has prioritized the RMNH agenda in its policies and two programmes were launched following this: Sharpened One Plan and Big Results Now (7). These programmes focused on strengthening RMNH services in all regions but with an emphasis on the most affected areas. These government efforts, with support from several stakeholders have reported impressive progress in initiation of basic emergency obstetric and newborn care services (BEmONC) and comprehensive emergency obstetric and newborn care services (CEmONC) in health facilities in Mwanza and other regions (7). However, the evaluation of these efforts particularly in remote areas highlights that suboptimal clinical skills and knowledge of HCWs remain a critical challenge and a potential obstacle for sustaining and consolidating the achievements (7). Furthermore, dissatisfaction with care related to HCWs low competencies, poor communication and lack of respect and dignity to mothers and families are also reported to contribute to negative experiences and uptake of care (8). One of the negative outcomes of undesirable experiences of care could be an explanation for the low rates of childbirth at health facility reported in Tanzania particularly in rural areas (6).
Several interventions to strengthen HCWs' skills predominantly within the scope of midwifery practice have made remarkable contributions in reducing maternal and neonatal mortality and morbidity in countries where the burden was high (9)(10)(11). The need for stakeholders' efforts in strengthening capacity and innovative long-term solutions for human resource is suggested (7).
Historically, the capacity building and professional development has been addressed through short courses and in-service training. This approach is useful mainly where staff shortages are catastrophic. Conversely, evidence show that such trainings, when they are traditional and class-based, are less effective in addressing health problems (12)(13)(14). Onsite short term training (15) and training of HCWs in relevant professional skills is proven successful and better long-term solutions for uplifting and sustaining HCWs' knowledge and skills (9). The Tanzanian government actively encourages upgrading of HCWs in rural areas (16). Thus, the aim of this study was to assess the training needs among HCWs of RMNH care setting in government-owned and designated faith-based health facilities of Mwanza region. The training needs assessment (TNA) was nested in a baseline survey that was conducted to inform the designing and implementation of the Improving Access to Reproductive, Maternal and Newborn Health in Mwanza, Tanzania (IMPACT) project interventions that aim at strengthening human resource capacity and RMNH service delivery in Mwanza. The survey included three components: household coverage survey, health facility survey and training needs assessment. More details of the rst two components of the IMPACT study are provided in another publication (17). This paper reports the results from the training needs assessment component.

Methods
This study employed a cross sectional quantitative survey design. Data for this study were collected as part of the baseline TNA for the IMPACT project conducted in 2017. Health care providers working in RMNH of selected health facilities responded to a self-administered questionnaire to identify individual priority training needs. President`s O ce Regional Administration and Local Government (PO RALG) at national, regional, district levels and local communities. The study took place in selected government-owned and designated faithbased healthcare facilities in all eight councils of Mwanza region: Buchosa, Ilemela, Kwimba, Magu, Misungwi, Nyamagana, Sengerema and Ukerewe (18). This region is located in the northern part of Tanzania with about 3 million inhabitants (18).

Study Population
The participants included all HCWs who were a) working in the labour ward and assisting the deliveries; and 2) working in reproductive and child health clinics (under-ve clinic, family planning, immunization units) at the time of the study. The study included HCWs who could uently communicate and understand either Kiswahili or English language and provided written consent to participate. We excluded non-RMNH care providers and those who were sick during data collection.

Data collection tool
The TNA tool was adapted from a validated WHO/Hennessy-Hicks tool (21) with 49 items focusing on capacity and training needs of RMNH health personnel. Prior to actual data collection, the researchers visited the eld to familiarize with the context and gain insights on the planning of the study and the intervention. Field visit notes and the existing evidence from the study context inspired the modi cation of the Hennessy-Hicks training needs analysis questionnaire. This questionnaire has been internationally validated and considered successful in identifying and prioritizing training needs at the individual, group or organizational level (21). Inspired by this tool, researchers developed a questionnaire with 49 items to assess individual training needs of HCWs in reproductive healthcare in general for maternal, adolescents and newborns (36 items), leadership and management skills (9 items), and research skills (5 items). The reliability and validity of the developed TNA questionnaire was determined and the reliability of the adapted was found to be 0.954. The indexes for construct validity indicated that Comparative Fit Index was equal to 1, minimum discrepancy per degree of freedom (CMIN/DF) was equal to 0.000 and the root mean square error of approximation (RMSEA) was equal to 0.185. This indicates that the TNA questionnaire has acceptable psychometric properties.

Data collection procedure
Data was collected from HCWs working in 36 sampled government-owned and designated faith-based healthcare facilities (7 hospitals, 12 health centres and 17 dispensaries) using the modi ed and an internationally validated questionnaire. The self-administered questionnaires were distributed to 153 participants by research assistants supervised by one of the researchers from the research team. The eligible RMNH providers were identi ed by the in-charge of the health facility.
Four research assistants with experience in data collection in RMNH were involved. They were trained for nine days around the best practices on data collection, obtaining informed consent, adherence to ethics as applied to human subjects' research and data security measures.

Pilot study
A pilot was conducted in one government hospital that was selected based on how similar its characteristics were with the facilities that were to be included in the TNA baseline survey and this facility was excluded in the actual survey. The results from the test were discussed with research assistants, clari cations made, and minor modi cations done including adjustment of the ow of the questions.

Sampling and sample size
All health care workers responsible for RMNH were all eligible to participate in this study. Participants were selected from 36 out of 80 strati ed random sampled health facilities supported by the IMPACT project. The healthcare facilities involved 7 out of 8 hospitals in eight districts in the region, 12 out of 19 health centres, and 17 out of 53 dispensaries in Mwanza. All healthcare workers in RMNH who were available at the time of the survey were included in the study. There were no any refusals and some who did not meet the inclusion criteria were excluded.

Data analysis
The Statistical Product and Service Solutions (SPSS, version 25.0) was used for data entry and statistical analysis. Both descriptive and inferential statistics were analyzed. The HCWs provided ratings on: a) how important is the task to their caring role (Rating A) and b) how well the task is currently performed (Rating B) (21), and the differences in scores were calculated to identify the greatest training needs among the studied tasks (21). The greater the difference between rating A and B, the greater the training need and was categorized as a top priority i.e. important task, but not well performed. Where a task scored low on both A and B, the task was considered a low priority i.e. unimportant task, but not well performed. Where both ratings were rated high, the task required no training i.e. an important task, well performed. The difference in aggregate scores were analyzed indicating the percentage of participants who scored at least ve for A minus the percentage of participants who scored at least 5 for B. The Chisquare test and correlation coe cient was used to determine the relationship between sociodemographic and task performance and perceived importance between HCWs.

Results
The TNA questionnaire was completed by 153 eligible HCWs. The age of the participants ranged between 21-62 years with mean age of 38.70 ± 10.89 years. The majority of HCWs in the RMNH units who participated in this study were Registered Nurses (n = 62) and Enrolled Nurses (n = 43) ( Table 1).  Most of the participants were from the government-owned health facilities 137 (90%) mostly from the hospital 68 (45%). Few participants were from faith-based designated district hospitals 16 (10%). All the participants worked in RMNH units with more than ve years of working experience. The total number of health facilities and participants in this study is summarized below ( Table 2). The difference between the care items with the highest score in rating A and the lowest score in rating B (important task, not well performed) were obtained. The care items with a score of greater than 40 were marked as a high priority area for training. The training needs were the highest in 16 (out of 49) care items including cervical cancer screening and care; research; basic and comprehensive emergency obstetric and newborn care (BEmONC & CEmONC); and sexual and gender-based violence (Table 3).   The correlation between demographic characteristics (gender, duration of employment, age) with perceived task importance was generally found to be not statistically signi cant. However, perceived importance of task was positively correlated with overall performance of task related to RMNH services (Pearson Correlation (r) = .256; P < 001) (Fig. 1)

Discussion
The training need assessment was designed to collect information on training gaps in various clinical RMNH care and identify training needs among health personnel working in sampled health care facilities under the IMPACT project in Mwanza region. The ndings of the TNA were critical in guiding the design of the interventions for capacity building among HCWs in the region. The ndings highlight the training gaps as perceived by HCWs in RMNH that indicate training priorities for the study setting in Tanzania.
The ndings of this study show that the training needs were the highest in 16 (out of 49) care items.
Amongst the identi ed gaps were: cervical cancer screening and care; basic and comprehensive emergency obstetric and newborn care (BEmONC & CEmONC); sexual and gender-based violence; providing information, education, counselling on family planning to adolescents; and implementing the maternal, infant and young child nutrition programmes. These ndings are supported by research in India, Pakistan and Nigeria where gaps were identi ed in maternal and newborn care, skills in emergence and basic obstetric care among health providers (22)(23)(24). However, for the Pakistan study, the gaps were categorized by cadres where medical doctors poorly performed in MNCH in comparison with other cadres. Our study did not perform such a comparison, because this was a perception study, we did not observe HCWs' performance. The highlighted training gaps correspond with areas in RMNH with poor indicators in Mwanza and the country in general (25), it is therefore not surprising that HCWs perceived these areas as gaps. For example, there is high maternal mortality rate, neonatal mortality rate and perinatal mortality rate (556 per 100, 000 live births, 25 per 1000 live births and 39 per 1000 live births respectively), only 32% use of modern contraceptives methods among women of reproductive age, and 42% of ever married women have experienced spousal violence, whether physical or sexual in Tanzania (20). To enhance providers' competencies, refresher continuing medical educational programmes relevant to speci c health care cadres are required.
Previous studies have highlighted barriers preventing HCWs from attending continued education that include negative attitudes (26), time constraints and limited nance to support training (26-28), child care and home responsibilities (27), lack of opportunity and previous negative experience(28). It is critical to have mechanisms in place that ensure health care professionals are continuously supported to attend relevant on-job trainings that will translate in improved RMNH outcomes. Furthermore, the ndings of our study call for a collaborative work between professional training institutions and the government to design innovative continuing educational programmes that not only would respond to the needs of HCWs but also include blended short courses that allow exibility and enhance their clinical practice for improved quality of care. Additionally, health care managers need to understand factors hindering and motivating HCWs from attending continuing educational programmes and set up strategies to promote and warrant participation. Addressing the training needs of HCWs may contribute towards achieving reduced maternal and neonatal deaths and morbidities especially in low resource settings.
Surprisingly, accessing research resources (i.e. time, money, information, equipment) and identifying research needs and designing locally relevant research were among the topics that received the highest ratings. This ndings contradicts a common stereotype on lack of interest for research among clinicians reported in United Kingdom and Bahrain (29,30). However, the ndings from other previous studies indicate a low self -assessed research capacity among nurses (31,32) and a need for training in research methods among nurses and other HCWs (31,33,34). In alignment with our study ndings, lack of research resources like time and funding were found to be the most commonly perceived barriers to undertaking research among practitioners in the UK, USA and Bahrain (29,30,35), increased paper work and disruption to work ows (35) and lack of statistical support (30). Among the motivating factors for participating in research included providing bene ts to patients, hoping to create knowledge relevant for patient care, nding solutions to di cult health problems and as an opportunity for professional development among clinical staff (35) and research supportive environment such as a research mentor Overall, the perceived importance on the speci c tasks was positively correlated with the overall perceived performance of tasks related to RMNH services. Similarly, ndings of a study in South Africa indicated that, the more comprehensively professional nurses were trained, the more competent they felt, and they expressed more negativity towards their work if they perceived they were inadequately trained (37). The fact that the HCWs' perceived importance of the tasks correlated with their perceived performance underscores the need for ensuring strategies are in place to address the identi ed training gaps. This implies, HCWs may lack con dence in their performance in the areas that were perceived important but with less ability to perform the tasks that might hinder their overall performance. Conversely, in a study among Tanzanian enrolled nurse midwives, more than 90% of the participants demonstrated both high performance and perceived competence in provision of primary health care services related to family planning, maternal and child care (38). One explanation could be that our study involved HCWs at various levels of care including referral facilities that require advanced skills for provision of RMNH including emergence care as opposed to the above study that only explored the elements of primary health care. Nonetheless, there is a need to tailor the trainings to the needs of the providers, this study was undertaken to determine and contextualize the speci c training needs of HCWs in Mwanza.
As expected, majority of HCWs in the studied setting were registered and enrolled nurses constituting 67 percent of all study participants. These ndings re ect the overall number of nurses and midwife in Tanzania who constitute more than 60 percent of the total health care workforce in Tanzania(39).
Nevertheless, the results may be confounded by higher numbers of registered and enrolled nurse midwives that may not be representative of the training needs among other HCWs.

Limitations
This study was conducted in only one region and therefore might limit the generalization of the ndings in broader contexts. However, the sample was obtained randomly thus enabling generalization to the study area. The study used a WHO tool that has not been previously validated within the country. However, in this study, the reliability of the adapted TNA questionnaire was tested and found to be 0.954.
Relatedly, indexes for construct validity was found to be at acceptable level with CFI equal to 1, (CMIN/DF) equal to 0.000 and RMSEA of 0.185, suggesting the tool to having acceptable reliability and validity. As this is a cross sectional study, it does not provide a causal-effect relationship that would provide more insights on health outcomes. However, as explained earlier, this study was set to guide the designing of the IMPACT project interventions. Future publications would focus on the intervention and its effect on HCWs perceptions of their performance on various RMNH care items.

Conclusion
The results of the current study highlight the training gaps as perceived by HCWs in RMNH in Tanzania, where 16 (out 49) care items were identi ed as priority for training. Furthermore, this study demonstrates that the perceived importance of the training needs to be signi cantly associated with performance of tasks related to RMNH services.

Implication For Practice
The utilization of ndings from TNA assessment has great potential to improve local service delivery hence improve the overall quality of care in a particular setting. The ndings of this study might be helpful in identi cation of the training gaps that has potentials to guide the design of continuing professional educational programs among HCWs in Tanzania. The ndings could also be used by higher learning educational institutions to strengthen pre and post quali cation postgraduate curricula for various professional programmes. Furthermore, since the tool has been validated for use in some low resource countries (Hennessy, D., Hicks IX/2517 on 9 th June, 2017. Research permits for all researchers and study clearance were obtained from The Tanzania Commission for Science and Technology (COSTECH). The permission to conduct this survey was sought from the regional, district and health facilities authorities. Detailed information regarding the study was provided to all the participants before embarking on the study. Data was collected only after the informed consent was obtained and consent forms signed. Questionnaires were distributed and left with participants for them to ll at their convenient time and place.

Consent for publication
Not applicable

Availability of data and materials
Data to support the ndings of this study are available and may only be accessed by those interested who will be required to obtain special permission from the Aga Khan University, Monitoring and Evaluation Research Unit (AKU-MERL).

Competing interests
Authors declare no competing interest. Correlation between perceived importance and performance of task