Modernising physician resource planning: a national interactive web platform for Canadian medical trainees

Healthcare accounts for a significant portion of public expenditure in many countries globally, and human resources occupy a majority of this spending [1]. Imbalances in healthcare workforce supply and demand negatively impact the provision of health services and ultimately strain overall public expenditure [2‐4]. Even under universal healthcare systems, such as the Canadian framework, these imbalances result in adverse sequelae. For example, between 2005 and 2019, the physician to 100,000 population ratio increased from about 190 to 240, representing one of the greatest rises of physician supply in Canadian history, while the proportion of Canadians without a family physician increased from 13.6 to 14.5% during this period [5‐7]. Ultimately, patient morbidity and mortality are known to rise when the supply of healthcare providers does not match population needs [8‐10]. This mismatch may be attributed to decades of overtraining subspecialist physicians, which has resulted in undersupply of primary care specialists [11, 12]. Furthermore, there is a geographic dichotomy: only 10% of the total physician workforce and only 2% of specialized physicians practice in rural Canada where 18% of the population is situated [13]. Therefore, strategies to improve patients’ access to care in Canada are necessary.

To address this, Health Canada invested $1.8 million in 2010 to review postgraduate medical education in Canada [14]. One of their recommendations was to iteratively re-evaluate medical training programs to ensure the right “physician mix, distribution, and number” of generalist and specialist positions to serve the Canadian population. At a national level, Canadian medical residency training has historically seen a deficit of students applying to primary care programs resulting in unfilled training positions; for example there were 169 vacant family medicine positions in 2020 [15]. This may be due to the perception that it is not a prestigious specialty and lacks innovative technology or academic opportunities [16, 17] In contrast, a 2019 survey of all the specialities and subspecialities offered in Canada showed that 34% of the total physicians surveyed were unable to secure employment at 12-17 months following graduation [18].

Health human resource (HHR) planning is a necessary endeavour to optimize the quantity and skillset of our healthcare workforce, which in turn ensures that population health needs are adequately serviced. It is the most sustainable approach to attaining long-term stability of the Canadian universal healthcare system [2]. Physician resource planning (PRP), a subset of HHR strategy, aims to optimize physicians’ available supply to meet population demands. Many nations, including Canada, are predicted to face critical deficits in physician supply over the next 30 years [19]. The goal is not to simply train a set number of physicians, but rather to train the right types of physicians based on population needs. This avoids the present circumstance of simultaneously having too many physicians, and yet, not enough. These shortages may be effectively overcome by improving PRP strategies for domestically trained physicians to close gaps in attrition and supply-demand mismatch [11]. However, due to the paucity of concerted national PRP efforts, Canadian medical trainees are not equipped to make evidence-informed decisions regarding specialty choice, and thus, the cycle persists [20]. Medical training is an arduous, lengthy, and costly process that can—and should—be optimized by a national PRP strategy.

Currently, there are two main checkpoints in the medical training process that affect the number, mix, and distribution of future physicians, thus acting as a component of national PRP strategy. First, undergraduate medicine admissions; second, postgraduate residency admissions.

The HHR Platform was created by FireNet Designs (FireNet Designs, Winnipeg, MB, CA) on a DigitalOcean (DigitalOcean Inc., New York City, NY, US) server to host the backend database using MySQL (Oracle Corp., Redwood Shores, CA, US). Construction of the platform spanned from November 2019 to September 2020. The backend is composed of a comma separated value (CSV) parser application used for adding data to the database, and flexible application programming interface (API) endpoints for querying data. The frontend is constructed using React (Facebook Inc., Menlo Park, CA, US) to manage user functionality and Material UI (Material-UI SAS, Paris, FR) to simplify data processing. Lastly, Leaflet (Vladimir Agafonkin, Kyiv, UA) and Google Charts (Google LLC, Menlo Park, CA, US) are utilized for the map and graph views, respectively. Geographic representation was aided by a public API acquired from the Environmental Systems Research Institute (ESRI) (ESRI, Redlands, CA, USA). The HHR Platform is hosted publicly and free-of-cost on the CFMS website [41].

The HHR Platform coalesces data from the following Canadian agencies: Canadian Institute for Health Information (CIHI), Canadian Medical Association (CMA), Canadian Post-MD Education Registry (CAPER), CaRMS, Ontario Medical Students’ Association (OMSA), and ESRI. Appropriate permissions and data-sharing contracts were obtained prior to construction of the platform. The unique datasets obtained from these agencies are categorized in Table 1. The delimiters included in the HHR Platform are categorized in Appendix 1. Program directors—who are responsible for guiding the selection and training of postgraduate resident physicians—were contacted to offer their expert opinion on PRP in their specialty and region given their unique position to comment on medical students’ interest in their program, postgraduate training logistics, and job availability in their specialty.. Aside from Ontario, program directors of all Canadian residency programs were identified via the CaRMS and Royal College websites and contacted via a standardized email template (Appendix 2) for their feedback regarding physician needs pertaining to their specialty and region of practice. Program director responses from Ontario in 2019 had already been collected by OMSA and were accordingly shared with the CFMS. Consent was obtained to share their de-identified responses publicly. No altering of responses were conducted and there was no text analysis; all data were subsequently posted as received onto the HHR platform. Members of the HHR Task Force—comprising nine medical students from across Canada—also performed a standardized search on the public domain for specialty-specific resources, which were amalgamated and included as part of the HHR Platform. Google analytics (Alphabet Inc., Mountain View, CA, USA) was integrated to demonstrate the utilization of the HHR platform in the first 4 months following the day of its official launch. An issue tracker using GitHub (Microsoft Corp., San Francisco, CA, USA) is available for user feedback on the CFMS website.

Sustainability was a key consideration in determining the backend database architecture. Thus, we prioritized construction of the database to be flexible such that new datasets may be added on an annual basis and that complex datasets may be promptly queried. Data presented in the platform was unmanipulated with one exception: data were averaged if the same metric was provided by multiple separate sources; averaged data were indicated by parentheses next to the values to indicate the number of sources contributing to the presented data. The range of the data is available in the graph view. Additionally, per CMA privacy policy, cells with fewer than five data points were suppressed. Specialties on the HHR Platform are defined per CIHI’s catalogue. Specialities represented in the platform are shown in Appendix 1. Jurisdictions are computed based on provincial boundaries and further subclassified by regional health authorities. Further information regarding data construction and considerations may be found in the HHR Platform user guide, which supplements the web tool [42].

The HHR Platform includes information and resources for 62 specialties with over 500,000 unique numerical data points obtained from 37 datasets. Data summarized in the HHR Platform is from 2015 onwards, with succeeding data added on an annual basis. There were 96 program director comments in 2019, representing 24.6% of the total individuals contacted. In the first 4 months (September 20, 2020, to January 20, 2021) of the HHR Platform being hosted publicly on the CFMS website, there were an average of 20.0 users per day with a total of 2434 different users, of whom 985 returned to the platform. This comprised 24.8% of all CFMS homepage views and 2.9% of all total site pageviews. 95.5% of users visited the platform from within Canada, with the majority visiting from Edmonton (10.3%), Toronto (8.7%), Ottawa (8.3%) and Calgary (6.1%). Assuming that these Canadian users were all medical students, this represents approximately 28% of CFMS’ membership, which includes 15 out of the 17 medical faculties in Canada. The average time spent on the platform was 3 min. The specific actions of users while using the platform, known as event tracking, was not available and this is further described in the limitations section. The HHR Platform was designed to present data in a customized fashion, whereby users select their preferred view between a map, table, or graph view. Each view utilises the same database but provides a unique purpose based on user needs. No feedback or issues were submitted on the issue tracker regarding the HHR platform since its implementation.

The following results section presents the different parts of the HHR platform, which was the primary outcome of this project. The software code can be found in the supplementary materials, with copyright details listed.

In Canada, the majority of national PRP strategies aimed at the medical training process rely on two checkpoints: admissions into medical school, and admissions into a residency program. While these approaches have shown to be effective in both planning for the number of future physicians entering the workforce and attaining a set specialty mix, the distribution of physician to areas of high needs remains an issue in Canada. Additionally, information and data specifically intended for undergraduate trainees regarding the current and projected demand for physicians in specialties and locations of high population need are lacking [43]. Trainees advance in their medical education by basing their specialty choice on factors such as personal interest, academic opportunities, gross income, and prestige while often not considering social accountability [16, 17]. New strategies that go beyond the admissions checkpoints are warranted to inform trainees of societal needs so that they are inclined to work in areas of high need, ultimately providing better access of care to the Canadian population.

One example of an approach outside of the admissions checkpoint to inform trainees of the population needs was developed by the Fédération Médicale Étudiante du Québec (FMEQ) in 2018. In order to balance physician distribution, the Québec Health Ministry’s rearranged their available residency positions to reflect 55% of seats reserved for the family medicine specialty, and 45% to all other specialties [44]. In that year alone, there were 65 vacant positions following both iterations of the CaRMS in Québec family medicine programs [45, 46]. In response, the FMEQ promoted family medicine education under the pretense that disinterest in family medicine stems from the medical students’ lack of knowledge about this specialty [47]. In the following year in Québec, there was a greater interest in family medicine as demonstrated by an increase from 373 to 440 applicants and only 23 vacancies in family medicine after the first iteration (of two) of the CaRMS match [45, 48]. The FMEQ approached PRP by targeting and educating medical students about the mix and demand of specialties, particularly with respect to the need for family physicians in the Québec population.

Additionally, there are many public resources that play a role in informing Canadian medical trainees on job opportunities and of the different specialties (many of which are outlined in Table 1 and accessible from the HHR Platform) to guide trainees in their career orientation. One limitation of these resources is their focus on a particular dataset or aspect of PRP. Thus, it is difficult for the user to compare these in relation to the data and information they may receive from career advisors, web resources, media reports, or health ministry statements. Another limitation is the frequency at which these various sources share their datasets; most being shared annually with a delay of many months between collection and public availability. In a healthcare system that is vulnerable to change in response to many types of factors (e.g. a pandemic), data must be updated frequently – if not in real-time – for it to remain pertinent. Finally, despite these efforts, strategies aimed at informing medical students’ specialty choice based on predicted population needs remains largely unexploited from a PRP strategy perspective.

The HHR Platform represents a national PRP approach towards improving transparency and ease of access to data regarding population health needs and associated practice opportunities to guide medical student specialty choice to be concordant with population health needs. The CFMS has approximately 8300 active medical student members across Canada. With over 2400 different users in the first 4 months of the HHR Platform launch, and over 20 daily users on average, there is significant outreach and interest shown by aspiring physicians. Designed with a flexible back-end, the HHR Platform is able to integrate new, changing or even incomplete datasets. This will enable it to become increasingly comprehensive over time, therefore functioning as a reliable and up-to-date – potentially even real-time – Canadian resource for career planning among medical trainees. Future directions include incorporating subspecialty data, additional information about workforce composition and demographics, predictive modelling of physician supply-demand dynamics, and inclusion of allied health professionals into the HHR Platform. Future analyses may be performed to explore how the HHR Platform and population-needs data specifically inform medical student specialty decision-making. The HHR Platform is currently equipped for and awaiting datasets for the projected demand of specialties in 5 and/or 10 years. This is likely the most anticipated and relevant dataset to inform medical students of societal needs prior to choosing a specialty, and continued advocacy for its public availability is warranted [49].

There is mounting evidence that Canadian medical students have limited information and data on the determinants of PRP, thereby lacking the tools to inform their specialty decision to be socially accountable based on population needs [53, 54]. The purpose of the HHR Platform is to provide a national approach to PRP which informs undergraduate medical trainees on the number, mix, and distribution of physicians to meet population health needs of Canadians. Modernising health data delivery by depicting the workforce needs of the population to medical students prior to making their specialty choice may function to improve the utilisation of current human resources, thus alleviating the impending shortage of health workers. Given the worldwide physician shortage, such a platform may be modelled in other countries with the goal of optimizing their physician workforce composition to better align with the needs of their population.