1) To what extent does regular guidance and feedback influence the retention of healthcare professionals at Kilimanjaro Christian Medical Centre (KCMC), Moshi, Tanzania?

2) To what extent does clinical-skill development through workplace mentorship activities influence the retention of healthcare professionals at KCMC, Moshi, Tanzania?

3) To what extent does improved understanding of organisational culture through workplace mentorship activities influence the retention of healthcare professionals at KCMC, Moshi, Tanzania?

H₀₁: Regular guidance and feedback has no statistically significant effect on the retention of healthcare professionals at KCMC, Moshi, Tanzania.

H₀₂: Clinical-skill development through workplace mentorship activities has no statistically significant effect on the retention of healthcare professionals at KCMC, Moshi, Tanzania.

H₀₃: Improved understanding of organisational culture through workplace mentorship activities has no statistically significant effect on the retention of healthcare professionals at KCMC, Moshi, Tanzania.

Practically, the study provides KCMC management, the Human Resources Department, department heads and clinical supervisors with empirical evidence on how regular guidance and feedback, clinical-skill development and understanding of organisational culture relate to healthcare professionals’ retention. The findings can guide mentor preparation, mentor-mentee matching, protected mentorship time, documentation, equitable access across cadres and integration of mentorship into broader measures addressing recognition, career progression, workload and working conditions.

Theoretically, the study extends Social Exchange Theory to Mentorship training program and healthcare workforce retention in a Tanzanian referral hospital setting. It demonstrates how mentorship operates as a valued organisational resource that healthcare professionals reciprocate through job satisfaction, commitment and willingness to remain. The findings also clarify the theory’s limits by showing that reciprocity through mentorship alone cannot fully explain retention when remuneration, workload, recognition and promotion opportunities also influence employees’ decisions. The study therefore contributes context-specific evidence that may support refinement and application of Social Exchange Theory in healthcare human-resource research.

From a policy perspective, the findings provide evidence for reviewing the KCMC Staff Training Policy, which recognises staff development and retention but does not establish a structured Mentorship training program, reporting framework or performance indicators. The study can inform provisions on mentor and mentee roles, eligibility and matching, frequency of meetings, protected time, documentation, monitoring and evaluation. At the national level, the findings support the Health Sector Strategic Plan V emphasis on strengthening the health workforce, institutional capacity, accountability and evidence-informed health-system improvement ([20]). The evidence may therefore be useful to KCMC and comparable referral hospitals when translating national workforce-development priorities into institution-level mentorship and retention strategies.

Social Exchange Theory (SET) originates in the work of [14], who conceptualised social behaviour as an exchange of activity, tangible or intangible, between at least two parties. [2] extended this into a fuller sociological theory of exchange, distinguishing economic exchange, which involves specified obligations exchanged at an agreed value, from social exchange, which involves unspecified obligations whose returns are not bargained in advance but are left to the discretion of the party who receives the benefit. [7] further developed the theory around the concepts of dependence and power, arguing that exchange relationships persist because each party depends on the other for valued resources it cannot obtain elsewhere. [4] in their widely cited interdisciplinary review, distil the theory into a set of interlocking principles that remain its core today: social interactions generate obligations, resources are exchanged through a process of reciprocity governed by the norm of reciprocity [11], and exchange relationships evolve into trust, loyalty and mutual commitment over time, provided both parties adhere to the rules of the exchange.

Applied to organisational settings, Social Exchange Theory assumes that employees evaluate their relationship with an employer as an ongoing exchange of resources rather than a single transaction. When an organisation provides resources that employees value, such as support, development opportunities, recognition or fair treatment, employees experience a felt obligation to reciprocate, typically through attitudes and behaviours that benefit the organisation, including greater effort, loyalty, satisfaction, commitment and a reduced inclination to leave ([4]). This premise was later formalised specifically for organisational contexts by [6] through the related construct of perceived organisational support, which holds that employees form a global belief about how much the organisation values their contribution and cares about their wellbeing, and that this belief predicts affective commitment and retention.

This study was underpinned by Social Exchange Theory, which provided the framework for explaining how mentorship may influence the retention of healthcare professionals at KCMC. Mentorship training was conceptualised as a valuable organisational resource provided by KCMC through experienced healthcare professionals to their less-experienced colleagues. The three dimensions examined regular guidance and feedback, clinical-skill development and improved understanding of organisational culture represent discretionary benefits characteristic of social exchange. As [2] explains, such benefits are not necessarily contractually owed to employees, and the nature and timing of their repayment are not specified in advance. Based on the norm of reciprocity, healthcare professionals who receive these forms of support are expected to respond with greater job satisfaction, organisational commitment and willingness to remain at KCMC. This theoretical explanation is supported by [23], who found that mentorship strengthened employees’ intention to stay both directly and indirectly through perceived organisational support and affective commitment. It is also consistent with the key-informant findings of the present study, which showed that mentored healthcare professionals developed a stronger sense of belonging and commitment when they felt guided, valued and supported by the organisation.

The principal strength of Social Exchange Theory, relative to alternative explanations, is that it specifies a psychological mechanism rather than a mere input-output association. It explains why mentorship should translate into retention, through felt obligation and reciprocity, rather than simply asserting that it does. This distinguishes Social Exchange Theory from purely economic or productivity-based accounts, which cannot explain why two employees who receive identical mentorship might respond differently depending on how supported or valued they feel.

The theory’s central limitation is that reciprocity depends on subjective perception rather than objective provision: an employee’s response is shaped by how fairly and consistently the exchange is perceived to operate, which Social Exchange Theory does not itself provide instruments to measure, and which can be distorted by factors outside the exchange being studied. Where remuneration, workload, staffing levels, promotion opportunities and recognition are perceived as inadequate, employees may judge their overall relationship with the organisation as unbalanced even where mentorship itself is delivered well, weakening or masking its reciprocal effect on retention. This is consistent with this study’s finding that the three mentorship dimensions jointly explained only 24.2% of the variation in retention, indicating that the broader exchange relationship, rather than mentorship alone, ultimately governs employees’ decisions to stay.

Social Exchange Theory is directly relevant because it maps precisely onto the variables under investigation: the three mentorship inputs function as the resources exchanged, and the retention-related outcomes, job satisfaction, organisational commitment, employee efficiency and willingness to remain, are the reciprocal responses the theory predicts. Because KCMC is a zonal referral, teaching and research hospital dependent on a stable multidisciplinary workforce, sustaining this exchange relationship is of direct practical and theoretical importance. This framing is reflected throughout the study’s conceptual framework and in the interpretation of findings, where mentorship is treated as one component of a broader exchange relationship between KCMC and its healthcare professionals, rather than as a self-sufficient determinant of retention.

The quantitative sample of 231 healthcare professionals was proportionately allocated across the four professional cadres to ensure that each group was represented according to its proportion in the target population. The target population consisted of 254 nurses, 166 doctors, 52 laboratory scientists and 76 pharmacists, giving a total population of 548 healthcare professionals. The proportionate allocation was calculated using the formula (ni = (Ni/N)n), where (ni) represents the sample size for each cadre, (Ni) represents the population of the cadre, (N) represents the total target population, and (n) represents the total sample size. Accordingly, 107 nurses, 70 doctors, 22 laboratory scientists and 32 pharmacists were selected, giving a total of 231 respondents (Table 1).

Table 1. Sampling matrix.

After proportionate allocation, respondents within each professional cadre were selected using a random-selection procedure. A sampling list of eligible healthcare professionals was obtained from the relevant KCMC administrative and departmental records and separated into the four professional strata. Each eligible professional was assigned a unique identification number within his or her cadre. The required number of respondents was then selected from each stratum using simple random sampling, giving every eligible healthcare professional within the respective cadre an equal chance of being included in the study. The selected respondents were then approached and invited to participate in the study. This procedure reduced the possibility of selection bias and ensured that the final quantitative sample reflected the distribution of the four professional groups at KCMC.

The study included healthcare professionals who were doctors, nurses, pharmacists or laboratory scientists working at KCMC during the period of data collection, had sufficient work experience at the institution to provide information about workplace mentorship activities and retention, and were willing to participate by providing informed consent. Professionals who were on extended leave during the data-collection period, temporary personnel without sufficient experience of the KCMC work environment, and those who declined to participate were excluded. The eligibility criteria were applied consistently across the four professional cadres to ensure comparability of respondents and to ensure that participants had adequate exposure to the workplace environment relevant to the study.

For the qualitative component, four key informants were purposively selected from administrative, human-resource, clinical, training, and departmental leadership roles because of their knowledge and involvement in staff development, mentorship and healthcare-professional management. The quantitative and qualitative sampling procedures were therefore complementary: proportionate stratified random sampling ensured adequate representation of the four healthcare-professional cadres, while purposive selection provided contextual information from personnel with relevant institutional responsibilities.

Quantitative data were collected using a structured questionnaire with dichotomous Yes/No items and five-point Likert-scale items. The instrument covered mentorship availability, regular guidance and feedback, clinical-skill development, understanding of organisational culture, job satisfaction, organisational commitment, employee efficiency and willingness to continue working at KCMC. Qualitative data were collected through semi-structured interviews and open-ended questionnaire items focusing on mentorship structures, mentor availability, professional guidance, institutional support, morale and implementation challenges.

The questionnaire and interview guide were developed from the study objective, Social Exchange Theory, the conceptual framework and the variables repeatedly identified in the empirical literature. The [16] Mentorship Training Report were also reviewed to ensure that the items reflected the institutional context. Each questionnaire item was mapped to one study variable: mentorship availability, regular guidance and feedback, clinical-skill development, understanding of organisational culture, job satisfaction, organisational commitment or intention to remain. Yes/No questions were used only where respondents were required to confirm the presence or absence of a specific mentorship practice or experience. Likert items were used where the researcher needed to measure the strength of respondents’ agreement with perceived effects and retention statements. The draft instruments were reviewed by the research supervisors and knowledgeable reviewers for relevance, clarity, wording, coverage and consistency with the objective. Their feedback was used to remove repetition, clarify ambiguous expressions and improve the sequence of questions, thereby strengthening face and content validity.

The revised instruments were pilot-tested at Mawenzi Regional Referral Hospital, a referral-hospital environment with healthcare cadres and working conditions broadly comparable to those at KCMC but outside the final study site. The pilot involved 23 healthcare professionals, equivalent to approximately 10% of the planned quantitative sample of 231, and two key informants. Pilot participants completed the questionnaire under conditions similar to the main study and were asked to identify unclear, repetitive, sensitive or difficult questions. The pilot also assessed the time required, logical flow, suitability of the Yes/No and Likert response options, and whether interview prompts generated information relevant to the objective. Feedback and preliminary response patterns were reviewed before the main survey; unclear wording and question order were revised while the underlying study variables were retained. Pilot participants were excluded from the final KCMC sample.

Reliability of the questionnaire. Responses from the 23 pilot questionnaires were coded and analysed in SPSS. Yes, was coded as 1 and No was coded as 0 for the dichotomous items, while the Likert items were coded from 1 (Strongly Disagree) to 5 (Strongly Agree), with negatively worded items reverse-coded before reliability analysis. Internal consistency was assessed using the Cronbach’s alpha reliability procedure. For items scored 0/1, Cronbach’s alpha is mathematically equivalent to the Kuder-Richardson Formula 20 (KR-20); therefore, the same analysis appropriately assessed whether the Yes/No items consistently represented the mentorship construct. The pilot questionnaire produced an overall reliability coefficient of .784, exceeding the commonly accepted minimum of .70 and indicating acceptable internal consistency. After data cleaning, the main-study questionnaire produced an overall coefficient of .826, indicating good internal consistency. Reliability was therefore established from the consistency of responses across related items, not from the proportion of respondents who answered Yes or No. A high percentage of Yes responses alone would not demonstrate reliability.

Two response formats were used in the mentorship section of the questionnaire, and it is important to explain how each was established. For items assessing the existence and provision of mentorship (for example, whether the facility has a formal mentorship arrangement, whether regular guidance and feedback are provided, and whether mentorship improved understanding of organisational culture), respondents answered on a dichotomous Yes/No basis, because these items measure the presence or absence of a mentorship attribute rather than a degree of agreement. For items assessing perceived effects and retention outcomes (for example, clinical-skill development, job satisfaction, commitment and long-term intention to remain), a five-point Likert scale was used, ranging from strongly disagree to strongly agree. During data preparation, the Likert responses for the mentorship-effect items were dichotomised into agreement (Agree/Strongly Agree = Yes) versus non-agreement (Neutral/Disagree/Strongly Disagree = No) to produce comparable Yes/No proportions, while the full five-point distributions were retained for the retention-outcome items reported. This dual approach allowed the study to report the prevalence of each mentorship attribute clearly while preserving the richer Likert distribution for the outcome variables.

Completed questionnaires were checked, coded, cleaned and analysed using SPSS version 27. Frequencies and percentages described respondents’ characteristics and mentorship experiences. The three independent variables, regular guidance and feedback, clinical-skill development and improved understanding of organisational culture, were entered into a multiple linear regression model to test their combined and individual contributions to retention. Healthcare-professional retention was measured using items on job satisfaction, job commitment, employee efficiency, long-term commitment and perceived effects of turnover; negatively worded items were reverse-coded, and favourable responses were used to compute a composite retention mean score ranging from 0 to 1, where values closer to 1 indicated stronger retention. Multiple linear regression using the Enter method tested the independent contribution of each mentorship factor at p < .05. Y = β₀ + β₁X₁ + β₂X₂ + β₃X₃, where Y represents healthcare-professional retention, X₁ = regular guidance and feedback, X₂ = clinical-skill development, and X₃ = improved understanding of organisational culture.

The predictor variables in this study were intended to measure respondents’ reported exposure to and experience of workplace mentorship activities, rather than respondents’ general perceptions that mentorship had improved their outcomes. Specifically, regular guidance and feedback measured the extent to which healthcare professionals actually received professional guidance, advice, and feedback from experienced colleagues in their routine work. Clinical-skill development measured respondents’ reported experience of developing or strengthening clinical skills through interaction with experienced colleagues, while improved understanding of organisational culture measured respondents’ experience of gaining knowledge about institutional values, procedures, expectations and workplace practices through mentorship. Thus, the predictor items focused on the mentorship-related experiences reported by respondents rather than asking whether mentorship was generally beneficial.

The distinction was maintained between exposure to mentorship activities and retention outcomes. Items measuring regular guidance and feedback, clinical-skill development and understanding of organisational culture were treated as indicators of workplace mentorship activities, whereas retention was measured separately using indicators such as job satisfaction, organisational commitment, willingness to continue working at KCMC and related retention statements. Therefore, the regression model examined whether respondents who reported greater exposure to these mentorship activities also reported stronger retention. The study did not interpret the predictor variables as objectively verified improvements in clinical performance or organisational outcomes because such outcomes were not independently assessed through clinical-performance records, supervisor ratings or administrative retention records.

Nevertheless, the study recognises the possibility of common-method bias because both the mentorship predictors and retention outcomes were obtained from the same respondents using the same questionnaire. In particular, respondents who had more positive experiences at KCMC may have been more likely to report both favourable mentorship experiences and stronger intentions to remain. Consequently, the observed associations should be interpreted as relationships between self-reported workplace mentorship experiences and self-reported retention, rather than as definitive evidence that mentorship objectively caused improvements in clinical skills or actual employee retention. The inclusion of qualitative interviews with key informants provided additional contextual evidence on how mentorship was implemented and how it was perceived to support healthcare-professional retention.

To minimise ambiguity in the interpretation of the findings, the terminology throughout the manuscript should therefore distinguish between “reported experience of clinical-skill development through workplace mentorship activities” and “actual clinical-skill improvement”, and between “retention intention” and “actual employee retention”. This clarification ensures that the conclusions remain consistent with the cross-sectional, self-reported nature of the data.

Healthcare-professional retention in this study was defined as the intention of healthcare professionals to remain employed at KCMC and continue working at the institution in the foreseeable future. This definition was adopted because the study used a cross-sectional questionnaire and did not have access to sufficiently complete administrative records on actual employee turnover during the study period. Therefore, intention to remain was used as the measurable indicator of retention rather than combining job satisfaction, organisational commitment, employee efficiency and perceptions of turnover into a single retention score.

The retention outcome was measured using questionnaire items that directly assessed respondents’ intention to remain at KCMC. The items focused on whether respondents intended to continue working at KCMC, whether they expected to remain employed at the institution, and whether they were willing to continue their professional career at KCMC. Responses were measured using a five-point Likert scale ranging from 1 = Strongly Disagree to 5 = Strongly Agree. Where negatively worded items were included, they were reverse-coded so that higher scores consistently represented stronger intention to remain. A composite intention-to-remain mean score was then calculated from the relevant items, with higher scores indicating stronger intention to remain at KCMC.

Job satisfaction and organisational commitment were therefore not included in the retention score. Instead, they were treated as related but conceptually distinct employee outcomes that may help explain why healthcare professionals intend to remain with an organisation. Similarly, employee efficiency was excluded because it represents a work-performance outcome rather than employee retention. Perceptions concerning the effects of staff turnover were also excluded because they describe respondents’ views about the consequences of turnover rather than whether an individual respondent intends to remain employed at KCMC. This separation provides a clearer distinction between the predictors, related workplace outcomes and the actual retention construct.

The revised measure aligns with the study definition of retention because an employee who expresses a strong intention to continue working at KCMC demonstrates a behavioural intention consistent with future retention. However, intention to remain is not equivalent to verified administrative retention. Since the study was cross-sectional and did not directly track subsequent employee departures, the findings are interpreted as evidence regarding healthcare professionals’ intention to remain, rather than confirmed actual retention. Accordingly, the dependent variable in the multiple linear regression analysis was the intention-to-remain mean score, while regular guidance and feedback, clinical-skill development and understanding of organisational culture were entered as the three workplace mentorship predictors.

Table 2.Response rate by professional cadre.

Source:Field Data (2026).

A total of 231 questionnaires were distributed to healthcare professionals across the four cadres.

Table 2 shows that nurses constituted the largest proportion of respondents, with 102 returned questionnaires, followed by doctors with 67, pharmacists with 30, and laboratory scientists with 21. The response rate exceeded 93% across all professional cadres, indicating strong participation and reducing the risk that the findings were influenced by non-response bias within any particular group.

This section presents respondents’ demographic and employment characteristics, including age, gender, education level, professional cadre, working experience, and supervisory or managerial responsibility.

Table 3. Demographic characteristics of respondents (n = 220).

Source:Field Data (2026).

The results in Table 3 show that most respondents were aged 30 - 39 years (43.6%), followed by those aged 20 - 29 years (25.9%). Most respondents held bachelor’s degrees (41.8%) or master’s degrees (30.5%), indicating that the study involved a professionally qualified workforce capable of assessing mentorship practices and retention. Furthermore, 33.6% of respondents held supervisory or managerial positions, which helped capture views from staff involved in guidance and institutional decision-making.

This section presents the effects of the mentorship training programme on healthcare-professional retention at KCMC. The findings are summarised in Table 4.

Table 4.Effects of the MTP in retaining healthcare professionals.

Source:Field Data (2026).

The results in Table 4 show that respondents responded positively to the mentorship training programme at KCMC, with positive responses ranging from 59.2% to 75.1%. This implies that mentorship contributes to professional development, job satisfaction, organisational commitment and willingness to remain, while the lower proportions for commitment to stay and continued employment imply that mentorship is only one of several factors influencing retention. This pattern is consistent with [12], who found that a mentorship programme for newly graduated nurses in the United States positively influenced most mentees’ decision to remain in the profession, with the effect strongest among mentees mentored for one to two years before diminishing among those mentored beyond two years, indicating that the contribution of mentorship to retention intentions is not indefinitely cumulative and depends on factors beyond the mentoring relationship itself.

On formal mentorship, 155 (71.8%) reported that KCMC had a formal mentorship arrangement, while 61 (28.2%) reported otherwise. This perception should be interpreted against the institution’s Staff Training Policy ([10]), which provides for formal and sponsored training but does not establish a dedicated mentorship programme with defined mentor and mentee roles, matching and documentation. The gap between the majority who perceived mentorship as formal and the absence of a formal mentorship policy implies that respondents were describing well-established but informal supervisor-to-junior mentorship, consistent with the policy provision that supervisory staff have a duty to assist in developing junior staff.

The results also show that 143 (69.1%) received regular guidance and feedback, while 64 (30.9%) did not. This implies that most employees benefited from mentor support, although the 30.9% without regular contact implies that guidance was not uniformly accessible. This corresponds with [24], who found that mentor approachability (mean = 3.84) and the provision of adequate supportive guidance (mean = 3.98) were positively associated with mentees’ outcomes, and that a substantial share of mentees had infrequent contact with their mentors, a pattern likely to weaken guidance where it occurs.

On clinical-skill development, 138 (66.0%) indicated that mentorship improved their clinical skills, while 71 (34.0%) reported otherwise. This implies that mentorship supported practical learning through observation, supervised practice and feedback, and that clinical-skill transfer is a central channel through which mentorship contributes to competence and retention. This is consistent with [8], whose rapid review of six studies found that mentorship-based training produced a larger improvement in healthcare workers’ neonatal-resuscitation skills (pooled mean difference = 6.03) than simulation-based training (3.45), indicating that direct mentorship may be particularly effective for developing practical clinical skills.

Regarding understanding of organisational culture, 142 (67.9%) respondents reported improvement, whereas 67 (32.1%) reported no improvement. This suggests that mentorship served as a mechanism for professional and organisational socialisation by helping employees understand institutional values, communication practices, professional expectations and teamwork, beyond the acquisition of technical competencies. Similarly, [21] found that mentees in a structured Tanzanian research-mentorship programme reported strong satisfaction with their mentoring relationships and with the programme’s contribution to their careers, suggesting that well-structured mentorship can help healthcare professionals integrate into their professional and institutional environment. The finding therefore indicates that mentorship may help healthcare workers integrate into the organisational culture and function effectively within institutional teams.

On job satisfaction, 141 (67.5%) reported that mentorship increased their satisfaction, while 68 (32.5%) reported otherwise. This implies that mentorship strengthened workplace experience and confidence, although the sizeable minority implies that satisfaction is also shaped by factors beyond mentorship, such as remuneration, workload and promotion. This is consistent with [22], who found that supervisory support and participation in decision-making were significantly associated with job satisfaction among KCMC health workers, reinforcing mentorship’s role as a form of supervisory support capable of strengthening satisfaction.

The results also show that 127 (61.4%) reported that mentorship strengthened their commitment to remain at KCMC, while 122 (59.2%) reported that it influenced their decision to continue working at the institution. These proportions, lower than those for guidance, clinical skills and job satisfaction, imply that mentorship contributes to retention but that decisions to remain are also driven by salaries, career opportunities, recognition and workload. This corresponds with [22], whose KCMC-based study found that pay, career-advancement opportunities and nature of work were significant determinants of job satisfaction alongside supervisory support, indicating that development-oriented strategies such as mentorship are most effective when combined with adequate remuneration and welfare provisions. These findings suggest that mentorship contributes to retention by improving professional support and job satisfaction, but it cannot independently address structural factors such as remuneration, workload, staffing levels and promotion opportunities. This interpretation is consistent with [21], who found that structured mentorship contributed to measurable career progression, including publications, PhD enrolment and research grants, among mentees in Tanzanian health-sciences institutions.

The highest positive response concerned willingness to recommend KCMC, with 160 (75.1%) respondents indicating that they would recommend the institution to other healthcare professionals because of its mentorship. This suggests that mentorship enhanced KCMC’s reputation as a teaching and professional-development institution and encouraged positive advocacy among employees. However, the 24.9% who would not recommend KCMC indicates that mentorship benefits were not experienced equally across staff. This finding is consistent with [12], who found that mentorship supported newly graduated nurses’ professional communication, self-confidence, transition into practice and decisions to remain in nursing. These benefits may strengthen employees’ connection to their workplace and consequently their willingness to recommend it to other professionals.

This section presents respondents’ open-ended responses on the organisational support available for mentorship at KCMC. The responses were coded into themes.

The results in Table 5 show that training, clinical guidance and professional development was the most frequently reported theme, mentioned by 29 (17.0%). Respondents referred to lectures, workshops, clinical supervision, continuous education and mentorship opportunities. This finding is consistent with [8], who found that mentorship improved healthcare workers’ practical skills more than simulation-based training, and with [24], who found that approachable, supportive mentors were associated with improved learning and performance outcomes among mentees. Therefore, KCMC should continue strengthening mentor preparation, clinical supervision and workplace learning opportunities.

Table 5.Organizational support for mentorship training program (n = 171).

Source: Field Data (2026).

Formal structure, policy, monitoring and programme integration was reported by 27 (15.8%), while regular interaction and supportive mentor-mentee relationships were identified by 24 (14.0%). Respondents mentioned mentorship policies, schedules, documentation, progress monitoring, continuous communication, feedback and emotional support. This agrees with [21], whose structured mentorship programme for young researchers in Tanzania featured clearly defined mentor-selection criteria, regular mentor-mentee meetings and quarterly institutional reporting, arrangements that were associated with strong mentee outcomes including research publications and grant success. Therefore, KCMC should establish a formal mentorship policy, define mentor and mentee responsibilities, promote regular meetings and ensure appropriate matching of mentors and mentees.

Provision of financial, physical and digital resources was reported by 23 (13.5%). Participants mentioned funding, teleconferencing facilities, internet access, meeting rooms and learning materials. This shows that effective mentorship requires institutional investment and cannot depend only on individual mentors. Therefore, KCMC should allocate adequate financial, technological and physical resources for mentorship activities.

Recognition, leadership and supportive organisational culture were mentioned by 22 (12.9%), while 16 (9.4%) reported limited, absent or uncertain organisational support. This implies that mentorship becomes more meaningful when management values it, recognises mentors and supports implementation across departments. However, unequal support shows that mentorship was not experienced consistently by all staff. This finding supports [24], who similarly found that supportive, approachable mentorship was associated with stronger mentee outcomes, and is consistent with Social Exchange Theory, which explains that organisational support, recognition and consistent mentorship may strengthen commitment, loyalty and willingness to remain through reciprocity.

Protected time, scheduling and permission to participate were reported by 13 (7.6%), while accessibility, inclusion and awareness were mentioned by 12 (7.0%). These findings suggest that workload conflicts, limited awareness and unequal access may restrict mentorship participation, consistent with [24], who similarly found that a substantial proportion of mentees experienced infrequent mentor contact, most commonly monthly or less, which the authors suggested could limit the effectiveness of a mentorship programme. Therefore, KCMC should provide dedicated mentorship time, flexible schedules, clear communication and equal access to mentorship opportunities. These findings from the quantitative and open-ended responses were supported by statements from key informants as follows:

“Mentorship is important for the retention of healthcare staff because it helps junior staff to become more competent and confident in their professional practice.Through mentorship,experienced staff provide guidance,feedback,and support to junior employees,which enables them to understand their responsibilities,improve their clinical skills,and adapt to the organizational culture.Mentorship also has positive impacts on staff commitment and morale because employees feel valued,supported,and recognized by the institution.When staff feel that there is someone to guide and support their professional growth,they are more likely to develop a sense of belonging and remain committed to the organization.” (Respondent 1, 12:47 p.m.)

This statement indicates that mentorship supports retention by improving competence, confidence, morale, belonging and commitment. From the perspective of Social Exchange Theory, mentorship is a valued resource extended by the organisation, which mentees reciprocate through improved morale, a sense of belonging and willingness to remain. This is consistent with [12], who found that mentorship improved mentees’ self-confidence and positively influenced their decision to remain in the profession.

“Although we do not have concrete written evidence,we can observe the impact of mentorship through staff performance.At KCMC,staff performance is monitored through the electronic health record system,where the number of patients attended by each individual staff member can be tracked.Because the hospital has specialists and super-specialists,junior doctors have gained confidence in managing conditions they previously could not handle,and patients are improving.However,mentorship can be strengthened by documenting and formalizing it,for example by advertising mentorship opportunities through the website,notice boards,or social media.” (Respondent 2, 12:05 p.m.)

This statement shows that mentorship was perceived to improve clinical confidence, service capacity and staff performance, and it confirms that mentorship at KCMC is currently delivered informally and without concrete documentation.

“Mentorship helps,although it is not the whole thing.It builds capacity,enhances the ability to provide services,and improves staff skills and competence.Some healthcare professionals have directly benefited from mentorship and are now working for the institution while also mentoring others.However,retention is not only a question of mentorship; if people are not paid well,they may not stay longer or remain committed.” (Respondent 3, 11:33 a.m.)

This section presents respondents’ participation in Mentorship training program at KCMC, showing whether they were currently participating, had participated before, or had never participated. The analysis helps establish staff exposure to mentorship opportunities within the institution, as presented in Figure 1.

Source:Field Data (2026).

Figure 1. Participation in mentorship training program at KCMC (n = 217).

The results in Figure 1 show respondents’ participation in mentorship training program at KCMC. Among 217 respondents, 81 (37.0%) had previously participated, while 62 (29.0%) were currently participating. In total, 143 (66.0%) had either participated before or were currently participating, while 74 (34.0%) had never participated. This indicates that mentorship is practised at KCMC and has reached the majority of surveyed healthcare professionals. However, the fact that about one-third had never participated suggests that mentorship opportunities are not yet available to all staff.

This finding is consistent with the methodological heterogeneity noted by [8], who found that mentorship approaches, training content and study populations varied considerably across the studies they reviewed, suggesting that mentorship reach and consistency often differ across institutions and settings. Therefore, KCMC should establish clear eligibility procedures and ensure that mentorship is available across all professional cadres and departments. The finding also relates to [24], who found that mentor approachability was strongly associated with positive mentee outcomes; expanding mentorship participation at KCMC should therefore involve both enrolling more mentees and ensuring mentors are adequately prepared and accessible.

This section presents the duration of respondents’ participation in Mentorship training program at KCMC and assesses how long respondents had been involved in mentorship activities, as presented in Figure 2.

Source:Field Data (2026).

Figure 2.Duration of participation in mentorship training program.

The results in Figure 2 show that 61 (40.8%) had participated in mentorship training program for one to two years, while 45 (37.4%) had participated for three to four years. In total, 106 (78.2%) had participated for between one and four years. In comparison, 21 (12.6%) had participated for five to six years, and only 16 (9.2%) had participated for seven years or more. This indicates that most participants had short- to medium-term mentorship experience, while only a few had long-term mentorship involvements.

These findings suggest that mentorship opportunities at KCMC may have expanded in recent years or are mainly provided during early career development and professional transition. However, the small proportion of long-term participation suggests limited continuity of mentorship relationships. This is consistent with [21], whose four-year mentorship programme in Tanzania showed that sustained mentee-mentor engagement was associated with strong outcomes, including publications, PhD enrolment and research grants. It is also consistent with [12], who found that the perceived benefits of mentorship on nurses’ decision to remain in the profession peaked among mentees mentored for one to two years and diminished thereafter, suggesting that the relationship between mentorship duration and its benefits may not be strictly linear. From the perspective of Social Exchange Theory, sustained mentorship signals continued organisational investment in the exchange relationship, which is necessary to maintain reciprocal commitment; therefore, KCMC should strengthen mentorship continuity by establishing clear programme durations, follow-up schedules and progression pathways, and by preparing experienced mentees to become mentors.

This section presents respondents’ rating of retention-related outcomes among healthcare professionals at KCMC. The responses are summarised in Table 6.

The results in Table 6 show that respondents generally had positive perceptions of retention outcomes at KCMC. Job satisfaction was reported positively by 141 (66.2%), while 36 (16.9%) were neutral and 36 (17.0%) disagreed or strongly disagreed. This finding is similar to [22], who found that only 49% of KCMC health workers were satisfied with their jobs, with dissatisfaction concentrated among nurses and doctors; the higher satisfaction level found in the present study may reflect the contribution of mentorship and other improvements since that earlier assessment.

Table 6. Retention-related outcomes among healthcare professionals at KCMC.

Key: SD = Strongly Disagree, D = Disagree, N = Neutral, A = Agree, SA = Strong Agree. Source: Field Data (2026).

Regarding job commitment, 140 (66.0%) gave positive responses, while 44 (20.8%) were neutral and 28 (13.2%) disagreed or strongly disagreed. This agrees with [24], who found a strong positive relationship between supportive, engaged mentorship and improved performance outcomes among mentees (r = .746, p < .001), suggesting that similarly supportive relationships may strengthen healthcare professionals’ engagement with their institution.

The study further found that 149 (69.9%) reported being more efficient as employees, which was the highest positive response among the retention outcomes. This suggests that most respondents believed they could perform their duties effectively, possibly due to improved knowledge, competence, confidence, teamwork and familiarity with institutional procedures. This is consistent with [8], who found that mentorship improved healthcare workers’ practical skills, and with [24], who found that supportive, approachable mentorship was strongly associated with improved performance among mentees.

In relation to long-term commitment, 139 (65.3%) indicated willingness to remain committed to KCMC, while 45 (21.1%) were neutral and 29 (13.7%) disagreed or strongly disagreed. This is consistent with [21], who found that structured mentorship was associated with strong career progression and satisfaction among mentees in Tanzanian health-sciences institutions. However, it contrasts somewhat with [22], who found only moderate overall job satisfaction (49%) among KCMC health workers, suggesting that retention intentions may differ depending on institutional context, workload, employment conditions and professional opportunities.

The finding on employee turnover was more divided. A total of 92 (43.4%) disagreed that turnover had affected retention, while 70 (33.0%) agreed and 50 (23.6%) were neutral. This suggests that turnover effects may not be experienced equally across departments or cadres. Therefore, KCMC should not rely on mentorship alone, but should combine it with adequate remuneration, manageable workloads, recognition, career progression and supportive working conditions to achieve sustainable retention. The findings are also supported by key informants:

“Mentorship has an impact on staff commitment and morale,although at KCMC the main reason people leave is monetary.They leave for financial reasons because salaries here are lower compared to other hospitals of our level.” (Interviewee #3, 12 May 2026, 12:05 p.m.)

This statement indicates that mentorship improves morale and commitment but cannot independently prevent turnover. This explains why quantitative findings may show positive perceptions of mentorship while commitment to stay and intention to continue working remain relatively lower, and it is consistent with [22], who similarly found that pay was a major driver of dissatisfaction among KCMC health workers despite reasonably positive perceptions of supervisory support. Therefore, mentorship should be considered as one component of a broader retention strategy, together with improved remuneration, recognition, career progression and supportive working conditions.

Multiple linear regression analysis was conducted to determine the combined and individual contributions of regular guidance and feedback, clinical-skill development and improved understanding of organisational culture to the retention of healthcare professionals at KCMC. A null hypothesis was rejected when its p-value was less than .05 and was not rejected when its p-value was equal to or greater than .05. As explained in the methodology, the analysis was based on the 220 respondents with complete data on all model variables.

5.8.1. Regression Assumption Tests

Regression assumption tests were performed to determine whether the data met the basic requirements for reliable regression interpretation. Normality of the residuals was assessed using the histogram and the Normal P-P plot of the regression standardised residuals. As shown in Figure 3, the histogram displayed an approximately bell-shaped distribution with a mean of approximately 0 and a standard deviation of approximately .993, and in the Normal P-P plot (Figure 4) the observed cumulative probabilities followed the diagonal line closely; the assumption of normally distributed residuals was therefore satisfied. Linearity and homoscedasticity were assessed using the scatterplot of standardised residuals against standardised predicted values (Figure 5), which showed a random scatter of points with no clear pattern or funnel shape, confirming that both assumptions were met. The residuals statistics are summarised in Table 7.

Table 7. Collinearity diagnostics for the regression model.

Source:Field Data (2026).

The results in Table 7 confirm that the condition index values ranged from 1.000 to 4.794, with the highest value (4.794) being well below the recommended cut-off of 30. This indicates that there were no harmful linear relationships among the three independent variables; the multicollinearity assumption was therefore satisfied, and all three predictors could be retained in the regression model.

Table 8.Residual statistics and outlier assessment.

Source:Field Data (2026).

Source:Field Data (2026).

Figure 3.Histogram of regression standardised residuals.

The results in Table 8 show that the residuals had a mean of zero, indicating that the model did not systematically overestimate or underestimate retention. The standardised residuals ranged from −2.719 to 2.583, which were within the acceptable limits of −3 to +3. Therefore, no extreme outliers were identified, and the regression assumption concerning residual outliers was satisfied.

Source:Field Data (2026).

Figure 4. Normality assumptions.

Figure 3 and Figure 4 present the histogram and Normal P-P plot of the standardised residuals. The residuals had a mean of .000, a standard deviation of .993 and ranged from −2.719 to 2.583 for 220 observations. The histogram approximately followed the normal curve, while the points in the P-P plot generally followed the diagonal reference line, with only minor deviations. Therefore, the regression residuals were approximately normally distributed, and the normality assumption was reasonably satisfied.

Source:Field Data (2026).

Figure 5. Linearity and homoscedasticity assessment.

As presented in Figure 5, the standardised predicted values ranged from −1.981 to .882, while the standardised residuals ranged from −2.719 to 2.583. The residuals were distributed above and below zero without a clear curved or funnel-shaped pattern. Therefore, the assumptions of linearity and homoscedasticity were reasonably satisfied.

5.8.2. Overall Model Summary

Table 9. Multiple regression model summary.

Source:Field Data (2026).

The results in Table 9 show a moderate positive combined relationship between the three mentorship dimensions and healthcare-professional retention (R = .492). The three mentorship dimensions jointly explained 24.2% of the variation in retention (R² = .242; adjusted R² = .232). This indicates that mentorship makes a meaningful but partial contribution to retention, and that the remaining 75.8% of the variation is explained by factors outside the model, such as remuneration, workload, recognition and career progression. Mentorship should therefore be treated as one component of a broader staff-retention strategy rather than as the sole determinant of employees’ decisions to remain.

5.8.3. Overall Significance of the Regression Model

Table 10.ANOVA results for the multiple regression model.

Source:Field Data (2026).

The results in Table 10 show that the overall regression model was statistically significant, F (3, 216) = 23.017, p < .001. This confirms that the three mentorship dimensions jointly predict healthcare-professional retention, and it was therefore appropriate to interpret the individual regression coefficients and test the study hypotheses.

5.8.4. Individual Effects of Mentorship Dimensions on Retention

As presented in Table 11, the estimated regression equation showed that regular guidance and feedback, clinical-skill development and improved understanding of organisational culture all made positive contributions to healthcare-professional retention: Y = .301 + .120X₁ + .173X₂ + .143X₃. The constant of .301 represents the predicted retention mean score when all three mentorship variables are coded as zero. Regular guidance and feedback made a positive and statistically significant contribution to retention, increasing the retention mean score by .120 points among respondents who received such support (B = .120, β = .176, t = 2.652, p = .009); the first null hypothesis was therefore rejected. This agrees with [24], who found that consistent, approachable mentor guidance was strongly associated with positive mentee outcomes.

Table 11.Regression coefficients and hypothesis decisions.

Source:Field Data (2026).

Clinical-skill development was the strongest positive and statistically significant predictor, increasing the retention mean score by .173 points among respondents who reported improved clinical skills through mentorship (B = .173, β = .262, t = 4.073, p < .001); the second null hypothesis was therefore rejected. This agrees with [8], whose rapid review found that mentorship produced a larger improvement in practical clinical skills (pooled mean difference = 6.03) than simulation-based training (3.45) in the immediate post-training period.

Improved understanding of organisational culture also contributed positively and significantly to retention (B = .143, β = .213, t = 3.216, p = .001), indicating that mentors support not only clinical learning but also professional socialisation, team acceptance and adjustment to institutional expectations, as supported by [21], who found that structured mentorship helped mentees integrate into their professional and institutional environment; the third null hypothesis was therefore rejected.

5.8.5. Robustness Check: Sensitivity of Results to the Definition of Retention

Job satisfaction and organisational commitment can be understood as intermediate mechanisms (mediators) between mentorship and retention, rather than as components of retention itself, a robustness check was conducted on the full sample of 220 respondents to test whether the results depended on how the retention outcome was defined. In the main model reported above, the composite retention score combined five outcome items (job satisfaction, job commitment, employee efficiency, long-term commitment and perceived turnover effect). For the robustness check, the model was re-estimated using a narrower retention measure that excluded the job-satisfaction and job-commitment items and retained only the items most directly reflecting intention to remain, namely long-term commitment, employee efficiency and the reverse-coded turnover item. The three mentorship predictors were kept unchanged.

Under this narrower specification, the overall model remained statistically significant, F (3, 216) = 16.752, p < .001, with the three predictors jointly explaining 18.9% of the variation in retention (R² = .189). Clinical-skill development remained the strongest and a highly significant predictor (B = .159, β = .231, p < .001), and improved understanding of organisational culture remained significant (B = .137, β = .196, p = .005). Regular guidance and feedback remained positive but became weaker (B = .105, β = .148, p = .032); in a stricter specification using only long-term commitment and turnover, it was no longer statistically significant (p = .330).

Two conclusions follow. First, the contributions of clinical-skill development and understanding of organisational culture to retention are robust: they hold regardless of how retention is defined, which strengthens confidence in these two findings. Second, the contribution of regular guidance and feedback is more sensitive to the definition of retention; part of its apparent effect in the main model operated through job satisfaction and commitment, consistent with a conceptual ordering in which satisfaction and commitment mediate between mentorship and retention. This refines rather than overturns the main findings: it indicates that the effect of guidance and feedback on retention is likely to be partly indirect. Future research using continuous composite scores and mediation or structural-equation modelling could formally test these indirect pathways.