Building consensus on identifying research mentoring gaps and finding ways of addressing the gap in a Kenyan college of health sciences

Background The concept of mentoring in clinical practice has traditionally focused on moving graduates from novice to more respectable positions within the clinical practice hierarchy. With the growing emphasis on evidence-based practice, the role of research in generating evidence for practice cannot be overemphasised. Mentoring in clinical operational research for both students and junior members of academic staff in health professionals’ training colleges is as important as mentoring for clinical skills. Aim This study aimed at building consensus on possible ways of enhancing research mentoring for graduate students and members of academic staff in a college of health sciences. Setting The study was conducted within Moi University College of Health Sciences (MUCHS) in Eldoret, Kenya. Methods The study population was composed of academic staff members and registered graduate students by the end of 2015. All academic staff and graduate students were eligible to participate. The Delphi technique was used to not only collect individual opinions but also build consensus. During the first iteration, questions were sent for which open-ended responses were needed. Responses from the first round were grouped into patterns and themes that guided the writing of questions for the subsequent rounds. Results The response rate was 78%. There was consensus in appreciating that mentoring was fundamental for career growth in clinical practice and research and needed for improving and developing formal structure for effective mentoring. It was crucial to establish training programmes for mentors and for accrediting them. Conclusion Enhancing of current research mentoring in MUCHS was needed and expected by graduate students and academic staff.


Background
Available data on the role of mentoring in medical practice and research reporting it as beneficial at each stage of training and associating it with greater research productivity, career retention and promotion. Most of these data are from high-income countries. 1,2,3,4,5,6,7,8,9 Positive effects of wellstructured mentoring programmes in graduate programmes have been documented. 10,11 There is also evidence that early career mentoring for translational researchers is a process that aims at junior faculty members evolving from novice to expert researchers in a coordinated and monitored programme. This process has been most effective where experts and more experienced researchers apply tasks that are also common in other forms of human relationships. These include recognising compatibility between mentor and mentee, finding time for the needed activities, establishing patterns, agreeing on goals and ensuring that they are achieved. 4,7,12,13,14,15,16 According to the World Development Index (WDI) data on health systems by the World Bank in 2015, in Kenya, the doctor to population ratio was at 0.2 per 1000. 20 In Kenya, nearly all postgraduate medical school training courses have a stringent compulsory research training programme that requires the candidate to write a thesis, which would be determined by graduate studies' examiners, to obtain the postgraduate degree. The period of graduate training in Kenya offers a unique opportunity for research mentoring to nurture researchers in biomedical sciences.
The Moi University School of Medicine has an above average regional rating in research output within Eastern Africa. 21 This is the result of a very fruitful partnership with the 'Academic Model Providing Access to Health Care (AMPATH) Consortium'. The Academic Model Providing Access to Health Care is an educational medical partnership between North American academic health centres led by the Indiana School of Medicine and the Moi University School of Medicine, and has flourished in the time since it started in 1989. 22 This collaborative effort has benefited health service delivery at the regional level in western Kenya where AMPATH and the Moi University collaborate with the county health departments to enhance chronic disease management within level three and four public health facilities. In the Moi University School of Medicine Teaching Hospital, the collaboration in clinical teaching, service and research has resulted in highly rated medical graduates. In research, the collaboration focuses significantly in translational research where research experts from partner universities in the west mentor Kenyans to grow from novices to expert researchers.
A pilot survey that interviewed graduate students, and junior and senior faculty members revealed that there was a significant satisfaction gap in research mentoring that needed to be addressed. It was agreed that the best approach to identifying and addressing the gap would be through consensus building and that the Delphi technique would be most appropriate for identifying and addressing the existing research mentoring gap in the Moi University College of Health Sciences (MUCHS).

Broad objective
The aim of this study was to identify the research mentoring gap and find ways of addressing it for graduate students and faculty members in the MUCHS.

Specific objectives
• To identify the research mentoring gap among graduate students and junior faculty members in MUCHS. • To build consensus on ways to enhance research mentoring among graduate students and faculty members in MUCHS. • To use evidence-based best practices through consensus to propose cost-effective methods to enhance research mentoring in MUCHS.

Design
We used the Delphi technique, which is a qualitative study design that facilitates a group communicator process that aims to achieve convergence of opinions on a specific real-world issue. This technique can use mixed qualitative and quantitative methods. During a feasibility survey, it was observed that some potential participants significantly controlled opinion during group discussions, when conducting in-depth interviews on faculty colleagues and it was realised that there was a possibility of compliance bias in the responses. The Delphi technique suited this study because there was no direct interaction among the respondents or even with the interviewers.

Site
The study setting was MUCHS, Eldoret, Kenya.

Study population
The study population comprised senior faculty members, junior faculty members and graduate students from MUCHS.

Inclusion criteria
• Academic members of staff from different departments and schools on the December 2015 payroll in the MUCHS. • Graduate students on the graduate students' roll for the different departments and schools on the MUCHS in December 2015.

Exclusion criteria
• Visiting academic staff from other universities at the time of the study. • Academic members of staff and graduate students who declined to participate.

Sampling procedures
The following three categories of participant included: • Senior faculty members: senior lecturers, associate professors and professors • Junior faculty members: lecturers and tutorial fellows • Graduate students on the graduate students' roll for the different departments and schools on the MUCHS in December 2015.
The three categories of graduate research work (senior faculty members, junior faculty members and graduate students) were conveniently selected to participate. This was guided by a survey carried out before the study when it was observed that all stakeholders (experts and learners) of research had differing opinions on ways to enhance research mentoring.
All senior members (senior lecturers, associate professors and professors) and junior academic staff from the different departments and schools were eligible to participate. A total of 460 participants (55 senior academic staff, 160 junior academic staff and 245 graduate students) were eligible to participate in the study.

Data collection methods
Potential participants expressed support for the use of the email, with a Word document attachment, for receiving and sending responses. The responses of the first iteration would be grouped into themes within the three separate categories of participants and only those responses that had more than 50% concordance, among participants, would be used for the second iteration.
For the second and third iterations, consensus was set at an agreement of 80% and above. This was to minimise ambiguity.
For the third iteration, the same questions were sent out again to the same participants with highlights on themes that lacked consensus during the second iteration.
Regular reminders (by email, short mail messages by phone) were sent out every fortnight to participants who delayed responses during the first iteration, and those who did not respond after three reminders were declared as not consenting and lacking interest to participate in the first round.
The responses to the second and third iterations were prompt with minimal need for sending reminders.

Data management
The responses to the first iteration question were analysed using pattern matching to develop themes that were used to prepare questions for the second and third iterations.
The responses to the questions of the second and third iteration questions were analysed as quantitative data and results presented in percentages.
Signed informed consent for all participants was a precondition to participation for the first iteration.

Study implications
Building of consensus among research experts and teachers, junior faculty members and graduate students on cost-effective ways for research mentoring in MUCHS will influence policy and practice in MUCHS and other similar institutions.

Study limitations
Our study aimed at interviewing all participants and so we did not have a scientifically representative sample. Without scientific sampling, it is possible that those who declined to participate may have a differing opinion, but with an above 75% response rate, the majority opinion can be considered as generally representative.

Ethical considerations
Ethical approval was obtained from the Moi University Institutional Research Ethics Committee (IREC) and permission to conduct the study was obtained from the Principal of the Moi University College of Health Sciences (Formal approval number: FAN: IREC 1488).

Results
The first iteration question was sent out to 55 senior faculty members, 160 junior faculty members and 245 postgraduate students.
Those who consented to participate and send responses back were 356 (45, 123 and 188) for senior, junior faculty members and graduate students, respectively. This was a response rate of 82%, 77% and 77% for the three categories of participants.
The responses from the three categories of participants to the open-ended questions of the first iteration are summarised in Tables 1-3.
Only responses that had above 50% consensus were included in developing the themes that guided the second and third iterations. This was done to minimise inclusion of equivocal responses.
For Tables 1-3, the following apply: • The different rows present collated themes of responses collected • Different columns represent collated themes of responses about what may be needed to enhance research mentoring for the specific group • Tables 1-3 represent collated themes of responses on what each group proposed should happen to enhance mentoring for each of the stakeholders.
These responses were then merged along common themes and used to develop closed questions for the second and third iterations. The summary of these responses is summarised in Tables 4 and 5.
The three items were presented as a summary of the responses in the second round for the third round (iteration) By the end of the third iteration, we had realised the consensus from the participants.

Junior faculty expectations of graduate students Junior faculty expectations of senior faculty Both junior and senior faculty
• The learners should be involved in choosing their research mentors unlike the current state where they supervisors are assigned to them • Assigned projected time for mentoring interaction with students • Students should actively seek and express interest in getting involved in research projects that individual faculty members are involved in • All should understand that mentoring is fundamental for career growth, promotion and retention in clinical practice and research • All should understand that mentoring is fundamental for career growth, promotion and retention in clinical practice and research • The need for formal and structured training on mentoring • Assigning mentors to junior faculty members as soon as recruited into schools and departments • Set aside small grants for faculty members to help with mentoring skills • Enhance collaborations to purposefully target mentoring of faculty members • All should understand that mentoring is fundamental for career growth, promotion and retention in clinical practice and research • Explicitly adopt a research culture in departments and schools • Training on mentor and mentee relationship • Continuous and structured two-way feedback on mentoring progress • Building relationships that enable a conducive mentoring environment • Research should be a guided process where the supervisor and the graduate student have a friendly learning environment that is non-threatening with the supervisor offering professional guidance while allowing the registrar to bring out his or her thoughts. The mentor should offer support during challenging times, for example, offer feasible suggestions on addressing difficult reviewer comments. The mentor should make available any possible research grants to the mentee or advise when the research topic can attract research funds. The supervisor should be accessible during the research period and get constant briefs on the research process • At the completion of the research process with data and results analysed, the supervisor and mentor should maintain professional contact and be in constant communication on research matters and any other relevant professional assistance more so on the processes involved in publishing results of the study • Departments should at least have a research agenda where graduate students can engage and participate in research with their mentors. Departmental research is one way the graduate students can observe their mentors actively engaged in research and learn from their mentors how they manage challenges encountered during the research process

Discussion
Consensus by participants drawn from the graduate students, junior and senior faculty members of the MUCHS in Western Kenya was remarkable. By the end of the third iteration, the potential benefactors (senior faculty members) and beneficiaries (graduate students and junior faculty members) had unanimously agreed on ways needed to enhance research mentoring in this college that is located in a sub-Saharan African country.
By the third iteration the participants involved concurred that mentoring in research was requisite for career growth, promotion and retention in clinical practice and research. This crucial observation has been documented in studies that looked into mentoring in research and clinical practice. Most of the data are from developed countires. 1,3,4,5,6,7,8 The important role played by mentoring in faculty development was supported by all junior faculty members, senior faculty members and graduate students and proposed as one that needed enhancing and structuring. There are available data that also support mentoring as crucial in faculty development. 6,23,24,25,26 Decastro proposes a focus on developing a network of mentors geared towards individual needs of mentees instead of a theoretically assumed hierarchy of needs. This was also ranked high in our study. 23 The participants in our study acknowledged that the current general assumption is that mentoring was going on in the MUCHS. There, however, was consensus on the need for structured mentoring that suited both the mentors and the mentees. Sambuco presented it more explicitly as '[an] Academic medical faculty often lacks the skills and knowledge necessary for successful negotiation, especially early in their careers, for effective mentoring to take place'. 27 There are also other available data that also emphasised the need for structured mentoring programmes. 10,11,28,29 Mentoring like any other academic engagement requires formal training and preparation of the mentors. The assumption that the successful careers of men and women make them effective mentors for those who look up to them for guidance has not worked to the satisfaction of the mentor and the mentee. One of the causes attributed to the poor results is the lack of well-trained and well-supported mentors. Abedin et al. proposed the need for validation and development of competencies for mentors. 16,30,31,32 Our study participants proposed mentoring programmes that involved mentoring retreats, a mentoring consultation online (that included Kenyan and foreign experts). The same was presented by Fieldman et al. among other available documents on the subject. 16

Conclusion
Our study findings on methods that would enhance research mentoring concurred with findings documented in developed countries. Although MUCHS encourages mentoring as part of training and faculty development, there was a need for formal structuring of mentoring programmes and finding ways to appreciate successful mentor and mentee programmes.