Radiology trainee retention in Auckland, New Zealand: a survey

Background

Each year, New Zealand trainees admitted to Fellowship of the Royal Australian and New Zealand College of Radiologists (RANZCR) are faced with the decision of whether to stay in their home country or pursue employment overseas. Our analysis focusses on recent trainees of the Auckland Regional Training Scheme specifically, the nation’s largest centre.

Although our hypothesis is that a large proportion of recent Auckland trainees are choosing to shift elsewhere, this is yet to be specifically studied or documented.

Particularly at a time where the radiology workforce crisis is taking centre stage and we continue to increase the numbers of registrars in our domestic training programmes, it is even more pertinent for us to look back at our recent local graduates and establish how many of them we are retaining or losing and the reasons why.

Objectives

Our aim was to provide a standardised survey to all RANZCR fellows of the Auckland Regional Training Scheme who commenced their radiology training in the year 2000 or later (a period of approximately 2 decades). We hope to obtain the following primary data:

• Our retention rate: defined as the proportion of fellows physically working within New Zealand at the time of this survey. This includes those who may have completed overseas fellowships and returned home.
• The number and proportion of this cohort who have undergone post-fellowship/subspecialty training, and where.
• The reasons people have selected their current workplace(s), e.g., financial, lifestyle, family, job satisfaction, work-mix, academia.Basic demographic data: age and sex.

A secondary objective was to assess whether these trainees are working solely in a public healthcare system (or equivalent), solely in private practice or in a mixture of both. Other considerations include those working in teleradiology or no longer working at all.

Methods

Study design

An initial contact list was obtained from Auckland City Hospital clerical staff with a list of email addresses of all past and present radiology trainees, dating back to those who commenced training in the year 2000. The list of trainees was used as a base to obtain the best possible email addresses (multiple in some instances) through a multifaceted approach. Hospital directories were also searched to find email addresses of Waitematā/Te Toka Tumai/Counties Manukau radiologists. The Medical Council of New Zealand (MCNZ) registry was used to check whether some individuals still held practicing licences in New Zealand. Social connections and Google searches were used to confirm locations where some trainees were practicing if this was not within the wider Auckland public system; for instance, most private practices list their staff members and their respective training backgrounds.

The survey questions (Appendix 1) were designed in conjunction with the supervising radiologist, Dr Rhian Miranda, with the above study objectives in mind. The questions were sent out via online survey (SurveyMonkey), and the survey was free to complete and submit.

Participants

(See Figure 1 for a visual schematic.)

A list of 155 fellows was obtained for the period of training commencing in 2000, to completion in 2022. Five of these did not complete training and were not contacted; one no longer has a valid practicing licence and was not contacted; and seven fellows were unable to be contacted despite our best efforts using the above methods. Therefore, 149 fellows were eligible to participate in the study. Overall, 142 fellows were contacted, i.e., 95.3% (142/149) of potential email addresses obtained.

Note on ethics/consent

Formal ethics approval was not sought or felt necessary.

A disclaimer was applied at the beginning of the survey once the link was opened: “Although quantitative and qualitative data gathered from this survey may be published and/or presented at a later date, all individual responses to the survey will remain entirely confidential.”

Consent was implied by survey participation.

Bias

We recognise here that we are missing data from several non-responders who work in Australia. Possible reasons for non-response include not having up-to-date email addresses, survey fatigue, participants simply being too busy, or perhaps unwilling to participate, given the lack of personal connection to the author(s) through the local hospital network.

However, based on several sources of data (e.g., MCNZ registry, listed practitioners on private practice websites, word of mouth), we believe that the relative proportion of survey non-responders working in Australia is roughly equal to the proportion that we have in our cohort of responders. Eight of our 39 non-responders (20.5%) are believed to currently work in Australia, which is comparable to the 21/99 (21.2%) of responders that stated they are working in Australia.

Statistical methods

Basic descriptive statistics were yielded from raw data provided via SurveyMonkey.

Results

Participants

(See Figure 1b for a visual schematic.)

Four email addresses had either opted out from or bounced-back SurveyMonkey surveys. Thus, 138 surveys were successfully sent out (135 via direct email invitation and three via web link).

Timeline:

• Survey first sent out: 4 September 2023
• Reminder messages: 10 October 2023
• Final reminder: 20 November 2023
• Survey closed: 1 December 2023.

We received 99 total responses (99/138=71.7% total response rate) (39/138=28.2% non-responders). Ninety-six completed the entire survey (96/138=69.7% complete response rate) (96/99=96.9% survey completion rate). There were 3.1% partially complete responses—note that the responses for the questions they did answer were recorded and included in study data, which explains why the denominator for some response data is not 99. The average completion time was 3 minutes 28 seconds.

Response data

Tables providing complete response data for all survey questions are provided below.

Discussion

The shortage of radiologists in both Auckland and New Zealand at a wider national level are contributed to by a variety of factors, including an ageing population and a disproportionate increase in demand for medical imaging. Health New Zealand – Te Whatu Ora’s profession-specific analysis for radiology and imaging, as part of their 2024 Health Workforce Plan, estimates 90+ full-time equivalent (FTE) shortages for radiologists across the health sector (13.6% of our total need).¹ To this measure, there have been recent increases in medical student numbers, as well as a temporary increase in the number of radiology trainees in New Zealand.

This survey aimed to capture how many of our local Auckland-trained radiology trainees we are retaining, another potentially key factor.

A 2022 MCNZ workforce survey² demonstrated a snapshot of New Zealand medical graduates from the year 2000 that were retained in New Zealand with every subsequent year. It shows that the medical graduate retention rate slowly declined with every year post registration. For instance, at 8 years post-registration (i.e., graduation from medical school), an average of 81% of medical graduates had been retained in New Zealand. For our interest’s sake, this point in time may arbitrarily be when someone may have completed radiology training. The number of retained medical graduates slowly declined and plateaued over subsequent years; for instance, for the cohort graduating between 2000–2006, their retention rates ranged between 59–74% at 15 years post-registration. The survey found that most New Zealand-trained doctors practicing outside of New Zealand are in Australia (2,187 in 2019).

Of the 99 Auckland-trained radiologists who completed our survey (all having commenced training from 2000, to completion by 2020), 75.8% (75/99; Table 3) answered that they currently work in New Zealand. Hence 75.8% is our retention rate, arguably better than what was found in the 2022 MCNZ survey.

Referring again to Health New Zealand – Te Whatu Ora’s recent analysis,¹ there is a clear acknowledgement of the need for relative growth across the radiology workforce going forwards, as this is where the greatest impact of the 24.2% of “lost” radiologists highlighted in our study would be felt. Perhaps bolstering workforce numbers through targeting improvements in retention rates would be an effective strategy, alongside the increase in trainee numbers.

Beyond the retention rate, the following data were also obtained through the survey:

Seventy-eight of our responders (78.8%; Table 4) are fellowship-trained, with 39.7% (n=31) having completed this training in New Zealand, and the remainder overseas (61.3% [n=61]).

Of those that completed an overseas fellowship, 51.3% (n=40; Table 5) did so in Australia (note that some people completed more than one fellowship).

A total of 41.0% (n=32; Table 6) completed a fellowship in musculoskeletal radiology, far outweighing any other subspecialty.

A total of 83.7% of our ex-trainees felt their registrar training in Auckland prepared them well for their first senior role (Table 16). Although training can always be improved, this is a reassuring statistic and was not felt to be the main cause for loss of retention.

The top three quoted reasons for current choice of workplace(s) (calculated by weighted average; Table 10) were “geography/lifestyle”, “family/friends” and “work-mix”. “Family/friends” had the highest proportion of people rating it as an “extremely important” factor.

The top five reasons included “flexibility” and “financial” reasons. With nation-wide medical specialist strikes taking place in New Zealand in late 2023,³ the ongoing cost of living crisis and established salary gap to Australia,⁴ financial factors are likely to continue to be a key factor.

More than half (58.2%; 57/98) of our survey responders work in a mixture of both public and private settings. Only 17.3% (17/98) work solely in public. Although none of our responders currently do pure teleradiology, 11.1% (11/98) are incorporating some form of teleradiology work, a number that may rise in future.

When considering how the landscape of radiology may change given the rising demands and pressures of the specialty, it is worth noting that more than half (56.2% [54/96]; Table 13) of responders are “definitely” or “possibly” considering significantly reducing their work hours over the next 1 to 5 years. “Parenting/family-related” factors were the most frequently selected reason (Table 14).

Half of these individuals considering cutting down their hours are “feeling overworked” and/or “have poor work–life balance”. Twenty-seven percent have “declining job satisfaction”. Only one person selected “approaching retirement” as their reason for wanting to reduce work hours; given that 89.9% of our responders were under 50 (Table 2), it is assumed that there would be more individuals wanting to reduce hours within the older cohort of radiologists (not surveyed here).

A total of 68.1% (66/97) of responders state they are overall “satisfied” or “very satisfied” with their current work (Table 9), and 97.9% (95/97) still felt satisfied with radiology as their choice of career (Table 15).

Limitations

As with any survey, an improved response rate could have strengthened the validity of our results. We were, however, pleased with our response rate of 71.7% (99/138). For reference, the 2020 RANZCR Clinical Radiology Workforce Census Reports of New Zealand⁵ and Australia⁶ had response rates of 55.7% and 47% respectively.

The primary statistic that we wanted to minimise bias in was the overall retention rate. As aforementioned, the proportion of non-responders (20.5%, n=8) suspected or known to be working in Australia was fortunately equivalent to our proportion of responders working in Australia (21.2%, n=1).

It is feasible that our non-responders would have provided different responses to the more subjective study questions (e.g., reasons for selecting workplace, considerations made to reducing hours).

Generalisability

External validity of study results cannot be determined, given we have selected a very specific demographic and timeframe, i.e., ex-Auckland trainees from the year 2000 to 2022. Retention of these trainees to New Zealand was our chosen metric, as opposed to retention within the Auckland region specifically. This survey also only assesses current incidence at the time the survey was conducted. It would be interesting to follow up with another study of ex-Auckland trainees in the future to assess evolving trends, and to compare retention rates with other regions of New Zealand.

Conclusion

How we define an acceptable number of retained versus “lost” trainees is open to debate, and is of growing importance given the radiology workforce crisis within New Zealand. The survey results hope to aid in providing some perspective on this, as well as some of the underlying factors that contribute to workplace selection. The success of a recent pilot initiative to increase the number of radiology training positions will rely on adequate retention of Auckland’s large number of fellowship-trained radiologists.

View Figure 1, Table 1–16.

View Appendix.