Career choice is a process with many interacting factors. A literature review published in 1995 by Bland and colleagues, and recently updated,¹⁶ features among the most comprehensive summaries of this process to date for medical career choice.¹⁷ It shows how the process of career choice is essentially a balance between expected future career needs and the perception of the characteristics of a specialty. Career needs are determined by the preferences and values of a student, affected by student characteristics and educational program characteristics. The stronger the perceived similarity between career needs and the perceived characteristics of a specialty, the stronger the desire for this specialty as a career choice.¹⁷ These career needs change during medical school under the influence of personal development, curriculum, life experience, and specialty experience.^16,17 Although this model provides valuable information about the relationship between factors affecting decision and preference of specialty, it was designed to address shortages and geographic maldistribution, which were predicted in the wake of a decrease in the number of medical school graduates entering primary care in the United States in the early 1990s. Secondly, the model was proposed more than two decades ago, and its current applicability is unclear. Nevertheless, the model still provides a valuable lens for investigating factors that influence career choice.

Most career preference and choice studies in medicine are specialty specific.^18-21 Moreover, many take a quantitative approach²² and lack deep insight into students' perspectives on career choice.

This study aims to provide insight into the factors that influence career choice specifically during the final phase of medical school. To reach a deeper understanding of the participants’ views, a qualitative approach was chosen.^23,24 The research question we attempted to answer is: Which factors affect career choice of medical students in their final study year?

An exploratory qualitative study was conducted using thematic analysis. Qualitative methodology was used since this study focuses on the experiences and perceptions of medical students; highlighting how this is affected by multiple factors and therefore how their career preferences might be influenced. We conducted 24 individual in-depth interviews to explore senior medical students’ opinions about their career preferences and to relate this to the Bland model.^25,26

The Dutch educational model allows for variation in study timelines, and graduations in medical schools occur several times per year. During the final-year rotations, students assume increased levels of responsibility in patient care, as they are expected to participate as a junior doctor under strict supervision. In this last year of undergraduate medical training, the medical student in most Dutch medical schools is called a ‘semi-physician’, somewhat similar to a foundation doctor in the UK system or a sub-intern or intern in the USA.²⁷ However, successful completion of this year does not guarantee a residency position in a preferred specialty. Dutch graduates need to apply for a residency in an open job market.

The study was conducted at the University Medical Center (UMC) Utrecht in the Netherlands with two groups of final-year medical students. The duration of UMC Utrecht’s undergraduate medical curriculum is six years, and the final year is created as a transitional year towards residency and contains predominantly elective clinical and science rotations.^28,29 Each sub-cohort starts with a six-week non-clinical block, with the starting dates for these blocks spread out over the final year. The study was explained during lectures in the first week of two such sub-cohorts, once in May and once in October. The lectures were attended by a total of 69 students and all these students were sent a follow-up information email and were offered the opportunity to sign up for the study. After one week, a reminder was sent to all non-responders. This convenience sample included all students starting in these two modules. (24) The study was approved by the Netherlands Association for Medical Education Ethical Review Board (ERB number 308). The participants were informed that participation was voluntary, that confidentiality was secured and that non-participation would not be held against them. They could withdraw from the study at any time without giving a reason. Written informed consent was obtained from all the participants. To preserve anonymity, every participant was asked to choose a pseudonym that was used for data storage and analysis.

We conducted semi-structured, in-depth interviews stimulating respondents to talk freely about sensitive matters. The interviews were held in a quiet studio at UMC Utrecht and lasted one hour. The interviews were planned following participants' availability options. The first interviews were conducted by two researchers (SQ and SB), to encourage similar interview styles in subsequent ones. Furthermore, the interviews were equally divided among the two researchers and were performed face to face. All interviews were audio recorded and during the interview field notes were made for reference during the interview. To further increase the study’s credibility, member validation was performed by writing summaries of each interview and sharing them with the participants.²⁴ This did not lead to essential changes; three students made minor changes at the sentence and linguistic level. The topic-list with questions that guided the interviews was based on what was known from the literature on career choice. The list was piloted with six medical students and was refined prior to the interviews with the sample group subjects. The items focused on the participants’ career preference and included questions such as, what are your career preferences? Can you explain these? How familiar are you with these specialties? Which are your electives during the transitional year and why did you choose these? Follow-up questions were used to probe explanations of the answers more deeply.

The researchers (SQ and SB) had several meetings to discuss the developing analysis. This promoted alignment of the researchers’ individual interpretations and enhanced reflection.^26,30,31 The interviews were transcribed verbatim using a transcription company with participants’ personal data being de-identified. All transcripts were checked against audio recordings by the first (SQ) or second (SB) author.

Both authors read each transcript twice to familiarize themselves with the contents. The authors analysed all data using the lens of the Bland model. The known "student factors" (as opposed to factors that are determined by the educational institution, i.e., status, salary, work-life balance, interesting work, altruism, experiences, role models)¹⁶ formed the basis for a coding template for describing factors influencing medical career choice.^32,33 The quotations were first identified as expressions of a particular factor and further categorised into the sub-themes (e.g., incoming values, career needs to satisfy). The raw data were coded line by line in an open coding process to identify any factor that seemed to influence a student's medical career choice.^32,33 Coding and analysis of all interviews were performed by the first researcher (SQ). For analytical rigor, the second researcher (SB) also performed an analysis of four interviews selected to reflect variation in gender, transitional-year start date (May or October), and career preference (intramural or extramural medical specialties). Thus, both researchers were familiar with all different codes and themes.

The coding scheme, as well as the themes, were discussed with the researchers’ team (SQ, SB, MdR, LW, OtC) during several meetings throughout the analysis. This allowed for alignment of the researchers’ individual interpretations and enhanced reflection, to increase credibility in the interpretations of the data.^30,31 The Bland model was used as a starting point for analysis to understand factors of influence for medical career choice, both to provide guidance and to be supplemented with new findings during the process. Data analysis continued by importing all data in a qualitative software application (Dedoose).³⁴ With the number of interviews, saturation was reached to understand the range of considerations of medical students with respect to their career preferences.

The two researchers (SQ, SB) who conducted the interviews were both trained in interview methods. SQ has a Master of Science degree in Health Care Management and was a policy advisor at the Royal Dutch Medical Association. She was not known to the participants and could work primarily as a researcher with an outsider perspective, with knowledge of career options, curricula, and the theoretical framework. SB graduated from medical school at Utrecht University and works as a medical educator and researcher at University Medical Center Utrecht. She was not acquainted with any of the participants but had inside knowledge of the Utrecht medical education program from her own experience. Two co-researchers (MdR and LW) had an outsider perspective, not being involved with University Medical Center Utrecht’s medical school, and one co-author (OtC) had an insider view, being the director at the Center for Research and Development of Education at University Medical Center Utrecht at the time. All three supervising researchers (MdR, LW, OtC) have theoretical knowledge of career choice, to ensure there was consensus for the use of the codebook. This insider/outsider status of the researchers allowed us to understand and interpret the scope of career options as well as the way they emerge in clinical education.

The students talked about various personal aspects such as where they live, their age, their personality, their relationship status, and so on. With the Bland model in mind, we categorized these under the theme of student characteristics. We identified three student characteristics of note in the interviews: geographic origin, parental education and personality.

All medical students have needs that they wish to fulfil in their future careers and their professional lives and these differ from person to person. These needs may be categorized in three: (1) personal: features that match personal desires around lifestyle, location/type of practice, working hours and work-life balance. Also included here are career features that are important to students because they line up with preferences of other people important to students, sub-themes are career counselling, career options, opinion of others and parental preferences; (2) features that match broader societal needs, the students mentioned altruism and (3) features that align with the content and habits of the specialty, such characteristics of team and colleagues, intellectual satisfaction, interesting work and patient characteristics. Some examples are described below.

When asked about previous knowledge and experiences interviewees recalled what they had experienced or heard and how this influenced their image of the speciality. The perception of specialty characteristics may be categorized in three: (1) Experience with the specialty, such experience before study, experience in courses & rotations, extracurricular experience, experience with role model or mentor and personal experience; (2) Information about the specialty such representation in the media and student-initiated information collection and (3) Market dynamics of the specialty as chances to obtain a residency position. Some examples are described below.

This study has some limitations. First, this study was performed at a single institution, one medical school in the Netherlands. We have to bear in mind that the institutional environment can influence results. The Dutch situation may differ from other countries, by its transitional year and admission process for residency.^28,43 Furthermore, the results of the study are predominantly a reflection of female students’ experiences. The proportion of male and female students in Dutch medical schools (about 65% female)^44,45 is different from the proportion in our study population (83.3% female) which could result in bias in the data. In addition, participation was voluntary which may have led to a non-representative sample of the student population as a whole. However, we had a representative mix of participants as we look to their age and specialty preferences.

The authors like to thank H. Carrie Chen, MD, PhD, Georgetown University, Washington DC, Marieke F. van der Schaaf, PhD, director of the Centre for Research and Development of Education at the Education Centre at University Medical Centre Utrecht and Anneke van Enk, PhD, researcher, Centre for Health Education Scholarship, University of British Colombia, Vancouver, Canada for their suggestions with an earlier version of the manuscript and textual suggestions.

The authors declare that they have no conflict of interest.