RESUMO
Objectives. To compare numbers of applications to the University of the Witwatersrand Human Research Ethics Committee (Medical) for clearance by the full committee during 2011 - 2013; to see the proportion of clinical Master's applications in 2013 and to look for the influence of eight variables in applications reviewed from January to June 2013.Methods. A retrospective extraction of data from committee minutes (2011-2013) and application forms (January - June 2013) was done. Statistical analysis was completed using SAS for Windows (version 9.4). Variables examined were committee decision; choice of research method; supervision or not; supervisors' research degree; supervisors' publication group; university administrative entity; registered degree and month of approval after first review.Results. Total numbers were 685 (2011); 845 (2012; a 23.4% increase from 2011) and 769 (2013; a 9.0% decrease on the previous year). In 2013; 22% of applications were for clinical Master's degrees required by the Health Professions Council of South Africa (HPCSA) for specialist registration. A number of cross-tabulations of variables are presented. Logistic regression analysis (Proc Catmod) showed that three variables significantly influenced the committee's initial review decision; namely school (p=0.03); applicants' registered degree (p=0.01) and the research method chosen (p=0.03). The month of committee approval was also significantly affected by school (p=0.002). Preferred research methodologies for supervised and independent research applications differed within and between schools.Conclusion. A predicted continuous increase in number of applications from 2011 to 2013 did not happen for unknown reasons. Research method; school; and registered degree significantly influenced the committee's decision at the initial review of applications. For clinical Masters; and other Masters; a supervisor's degree had no effect on the committee's decision at the initial review of applications; however undergraduate and honours applicants having supervisors without a research degree had more than double the approval rate at first review than when supervisors had either a Doctorate or a Master's (p=0.008). Supervisors' possession of a research degree did not increase approval rate of applications nor did a supervisor's publication grouping