The gender and race distribution of respondents confirms the under-representation of women and black registrars at UCT. Changes documented in the gender and race profiles of registrars at UCT from 2001 to 2006 are modest but provide evidence of increasing recruitment of black and female registrars. Though not proven, these trends are most likely the result of systematic policies adopted by the Faculty to attract and retain talented black and female trainees [9
]. These policies are similar in many respects to strategies considered and attempted at other formerly racially exclusive training institutions in South Africa to change their race and gender profile [33
]. Given that modest shifts in profiles were achieved from a relatively low base, there is need for continued attention to ensure the implementation of strategies to make the institution environments welcoming for potential registrars, particularly those from disadvantaged groups. Ongoing attention to proactive affirmative action recruitment amongst undergraduates has also been noted in the US as a key strategy to address the health of minority groups [34
As noted, [35
] transformation is more than just about numeric targets. Unless the experience of studying or working at an institution is a positive one, attempts to recruit blacks and females will result in a 'revolving door' situation, where black/female staff members are not retained, or worse, feelings of alienation are exacerbated. About a third of the respondents in this study reported teaching and learning environments as 'poor'.' Although not as extreme as those reported by Winter and others [37
], these findings are sobering, particularly given that registrars' reported ratings of learning and teaching that were better than those for service and research. The reasons for poorer ratings for service and research may lie in declining provincial health budgets in the Western Cape Province over the past 10 years, which have reduced funding available at tertiary level, where much of the training takes place. The budget cuts have led to a reduction in posts, an increased workload and reports that the 'hallways of most hospitals in this country echo with disillusioned, disgruntled, frustrated and fatigued voices overwhelmed by the demands of service delivery' [[38
]: p63]. The importance of an appropriate balance between service demands and educational activities in resident training has been identified in other contexts [39
]. Given increased service pressures, it is therefore usually research activity that is cut to the minimum [38
] whereas teaching and learning, perceived as core functions for clinical teachers, are more likely to be protected [40
Such a distinction may also reflect clinician-teachers' dedication to maintaining high levels of teaching in the face of an increasingly inhospitable service environment [40
]. Indeed, the importance of clinical supervision, the extent and quality of teaching, and the availability of support for registrars' participation in conferences was noted to be an important predictor of South African registrars' willingness to persevere with training under stressful conditions [41
]. The structured availability of learning opportunities was shown to be an important determinant of the quality of residency training in Family Medicine in the US [42
]. Thus, even though the service environment may be pressurised by budgetary shortfalls, there are strategies to ensure that teaching and learning activities continue to meet registrars' academic and educational needs, that do not necessarily require additional resources. Such strategies include the use of shadowing, the allocation of protected teaching and learning time, and systematic mentorship by senior registrars and consultants, who, through a structured relationship, act as educational supervisors and guide registrars to access educational and research resources and networks. In the absence of such strategies, registrars' time risks being overwhelmed by service demands. Indeed, other studies have found that residents report that they have less time for teaching and research than they need or are satisfied with and spend more time than desirable on training in direct patient care [41
In general, black registrars tended to rate teaching, research and service environments worse than white registrars, but the converse applied for the ratings of their learning environment, though these differences were not statistically significant. Although the numbers reporting personal experience of racial discrimination were low (n = 7), it was personal reporting of discrimination that was significantly associated with poorer assessments of learning and research environments. Such findings are in keeping with the international literature that illustrates the impact of gender [24
] and race [21
] discrimination on career development and with studies showing the ongoing salience of race in the undergraduate learning context in post-apartheid South Africa [51
]. The impact on teaching and learning may be direct, in that the experience of discrimination may directly affect learning opportunities, or, alternatively, unrelated experiences at the same institution may flavour the nature of the learning experience for the trainee. The fact that white respondents were amongst those reporting personal experience of race discrimination points to the complexities of implementing Employment Equity policies that remain fair and non-discriminatory, a question raised in broader debates on post-apartheid transformation [52
To what extent did UCT present a welcoming environment to potential trainees? Although there were no statistically significant differences by race in respondents' reporting knowledge of a colleague deciding not to apply to UCT (for any reason), there was a significant difference by race in relation to racial discrimination anticipated by colleagues, which was the most common single reason cited by respondents for others' decisions not to come to UCT (n = 7). Given that no differences were found in actual personal experience of racial discrimination reported by race amongst the same respondents, this most likely reflects an anticipation of institutional discrimination. Indeed, other studies have confirmed there is a real historical precedent for such perception [8
] even if practices have changed in recent times. Further, given the broader patterns of racial stratification in South Africa, still persistent post-apartheid, this result is also a function of registrar applicants' life experiences, in that black applicants are more likely to have black friends who, in turn are more likely to experience discrimination than white colleagues or white friends of white colleagues. As a result, it is not surprising that black respondents, particularly African respondents, were more likely to describe UCT as unwelcoming than white respondents. Measures to address recruitment and retention also therefore need to address the elements of an institutional culture that contribute to making the institution a welcoming environment for all [53
This insider-outside dynamic, however, was not solely based on race. Over a third of respondents reported being alumni of UCT and returning 'home' was an important factor that determined registrars' choice to study at UCT. Hence, being an outsider to both the institution and the city may cause discomfort and anticipation of discrimination. Even among white respondents, registrars who were not originally from Cape Town were about 66% more likely to find UCT unwelcoming than their white counterparts who were from Cape Town. Indeed, it was striking that the registrars who were least likely to describe UCT as unwelcoming were those from outside South Africa, even more so than white registrars from South Africa. Such differences may partly relate to the experience of the culture of Cape Town and not only that of the institution, and confirm the importance of orientation and induction as important strategies for transformation, as much as re-orienting how staff members are made to feel welcome in the course of their attachment.
Although women have outnumbered male medical students at undergraduate level for a number of years at UCT, they still constitute a minority in postgraduate programmes [8
]. Women, in particular; carry a double load of parenting whilst undergoing specialist training [24
]. Saloojee and Rothberg [11
] highlighted the potential for gender discrimination experienced by women in registrar training in paediatrics in South Africa. Findings in this study confirm that having a child during registrarship was significantly associated with poorer rating of the teaching environment. Scores were also worse for learning, research and service, though differences for gender were not statistically significant.
Notably, female gender alone was not predictive of differences in ratings; indeed, women tended to rate their experiences in all four domains better than did men, though not significantly so. This is not to say that gender discrimination is not an issue for improving the quality of registrar training. At least one respondent cited an instance of gender discrimination in her registrar experience, and it is clear in Figure that there is a gender disproportion by department in the faculty. However, it appears that it is the challenge of having a child during training that is the most important determinant of the learning experience, which points to the importance of institutional support in the form of job-sharing or other forms of flexible work arrangements that can accommodate parenting responsibilities [11
]. In other contexts, flexible working arrangements can be as common as comprising 20% of posts in paediatrics and psychiatry [55
]. Such strategies should apply for both women and men, both in the interests of achieving gender equality and because of evidence that the stress of caring for a child whilst undergoing residency training may be more stressful for men than for women residents [56
Of note, as well, is the strong evidence that UCT continues to draw most of its registrar pool from local graduates. This trend has both positive and negative implications. As a way to groom talented students, recruiting one's own alumni offers an opportunity to mentor students from under-graduate years, and points to the importance of senior staff role models who actively encourage students to return for registrar training. The role of the Head of Department (HoDs) is thus critical in this process and HoDs were the most common single source of information about posts other than word of mouth. However, it also means that the institution is perhaps not making optimal use of the widest pool of potential applicants, a particular concern given perceptions of UCT being unwelcoming to applicants. Registrars who did not apply to UCT on the basis of being an alumnus were more than twice as likely to rate the service environment poorly, probably reflecting their lack of familiarity with the difficulties of working in a province where budgetary constraints impact the service platform.
There were many limitations to this study. First, a response rate of 19% means the sample is not representative of the registrar population at UCT, particularly given the over-representation of Indian and under-representation of African registrars amongst the responders compared to the total registrar population. However, included in the respondents were 2/3 of all clinical disciplines and the only large clinical absent amongst responders was the Department of General Surgery. Further, the age range of responding registrars is broadly similar to that anticipated for registrars in the faculty. The low response rate is also not unusual for postal questionnaires and is, in fact, slightly higher than the response to an Institutional Climate survey held at UCT in 2007 [57
]. However, studies amongst equivalent medical trainee populations have achieved much higher response rates, to the order of 50 to 80%,[37
] albeit using different methodologies.
Respondents in this survey may therefore have been systematically different to non-responders, particularly if they had personal experience of discrimination or felt aggrieved by adverse experiences which they wished to voice. Further, responses to the questions may have been less than entirely honest, despite assured anonymity of the questionnaire. The direction of the response bias, if any, may have been either to under- or over-state problems. We can also not discount recall bias affecting the study, since respondents were asked to recall experiences or incidents over periods that may have been up to four years prior to interview. In such cases, it would be likely to underestimate the prevalence of reported racial or gender discrimination. Further, the small sample size may have made it impossible to distinguish large differences in practice and policy between and even within departments, a point raised in feedback discussions with registrars after the study.
Nonetheless, despite the room for selection and measurement error, the findings from this study appear consistent with results from other research in the faculty investigating questions of race, gender and transformation [30
] and from feedback discussions with registrars following the study. The percentage of respondents reporting personal experiences of discrimination was similar to that found in the UCT climate surveys of staff across faculties [57
]. Inasmuch as they point to issues which an institution of higher learning might consider as strategic challenges for transformation, they provide pointers to further action needed, as well as potential baseline data for evaluating the future effectiveness of such strategies. The presentation of the findings to registrars has also spurred interest from the registrars to conduct their own research and to initiate more regular faculty-wide monitoring of the registrar programme.