Ethics approval

This study was approved by the McMaster University Research Ethics Board in Hamilton, Ontario, Canada. To facilitate confidentiality, participants were assigned a unique identifier based only on country and sector; identifying details including job titles or affiliated organizations were dissociated from responses prior to analysis.

Data analysis

A thematic analysis approach was followed to extrapolate key themes from the qualitative responses. This involved an initial sorting of the responses into constructed descriptive categories and the subsequent definition of analytical themes. Themes were thus determined by the data itself and not in accordance with any initial hypotheses that might have influenced results. Quantitative descriptive statistics were extracted based on the coded responses, and representative quotations were identified.

Awareness of the Code

Eighteen respondents (41 %) indicated that they were largely unaware of the Code and its impact despite working in organizations for the licensure of international medical graduates (n = 2), health workforce modeling (n = 4), health workforce planning and research (n = 8), health policy and workforce development (n = 3), and international workforce equity (n = 1).

Ten respondents indicated that their colleagues lacked an awareness of the Code (23 %). One private sector respondent from Canada stated, “Neither the existence of such a global Code, nor the WHO role is known within our health services network” (CAN.PS.1). A civil society respondent from the USA emphasized, “[My colleagues] are unaware of the problem and any possible solutions” (US.CS.1). The two respondents reporting some level of awareness suggested that the Code was likely more familiar to colleagues in developing countries. One respondent mentioned, “My impression is that my Canadian colleagues have minimal awareness of the Code…while colleagues in developing countries where we work, such as Jamaica and Zambia, are much more aware of it” (CAN.CS.2).

Ten respondents reported that their colleagues were likely to be familiar with the Code (23 %). Of these, colleagues who were researchers and senior academics were reported as likely knowledgeable about the instrument (n = 5). However, colleagues whose work directly impacted health workforce planning, such as those working in health authorities and the health policy community were reported to be generally unaware of the Code (n = 4).

Among the three individuals reporting that their colleagues were “very aware” of the Code (7 %), two individuals were from Australia, while one Canadian respondent referred to the Code in conjunction with the Melbourne Manifesto, which, they surmised, collectively provided a basis for discussing issues affecting health worker migration. One of the Australian respondents stated, “I believe all senior academics would be aware of the WHO Code, including its global purpose” (AU.G.1).

Changes resulting from the Code

Of the four individuals reporting anticipated policy or regulatory changes as a result of the Code, a Canadian government respondent declared, “…we are consistently working with colleagues to ensure that our policies follow the Code – I would say that this would be true of all jurisdictions in Canada” (CAN.G.1). However, the vast majority of respondents indicated that they were not aware of any anticipated changes resulting from the Code (n = 21; 48 %). One UK respondent suggested that, as the UK Code of Practice on International Recruitment preceded the WHO Code, it was difficult to determine any specific impacts of the latter. Alternatively, a number of respondents attributed poor implementation to the primacy of domestic health workforce needs and competing priorities. One Canadian civil society respondent claimed, “…I expect there to be a greater emphasis on international recruitment of health workers to Canada as both the general population and key health cadres such as nurses continue to age” (CAN.CS.2). The prioritization of national deficits was echoed by American and Australian respondents, one of whom suggested, “My personal commitment is to produce a workforce in and for the regions in which I work. However, I’m heartened that this social justice issue is on the national and international agenda” (AUS.CS.3). A civil society respondent from the USA corroborated, saying “there is very little incentive for healthcare providers to adhere to the Code – there is actually more domestic incentive to not adhere to the Code, for example through increased funding for ‘diversity’ hires” (US.CS.5).

Key informants suggested that strong personal incentives for health workers to leave their countries of origin explained the lack of change the Code had hoped to achieve. These “pull” factors included the right of individuals to migrate for improved political and socioeconomic stability, as well as to achieve a higher standard of living (n = 3). One Canadian civil society respondent stated, “The real incentive to keep people in their country of origin is addressing the social inequalities in these countries and providing supportive environments…for them to build meaningful lives in safe conditions. People don’t leave unless they really have to” (CAN.CS.3).

Existing national or bilateral agreements were also seen as preeminent to the Code, with respondents referencing tangible regulatory changes following the implementation of the Melbourne Manifesto and the UK Code of Practice, among other international codes (n = 4). The Code was often seen as complementary to these existing agreements, or providing further leverage for their support, albeit largely incapable of inciting change directly. As one civil society respondent from the UK mentioned, “…the UK already had a Code of Practice, [but] the WHO Code has proved useful in terms of encouraging cross departmental work in monitoring the UK Code” (UK.CS.1). One American civil society respondent also pointed out the “synergistic” impact of the WHO Code in legitimizing the work of NGOs in trying to reduce unethical international health worker recruitment (US.CS.8).

Suggestions to improve the Code’s impact

When asked whether the Code had a meaningful impact on health worker recruitment, fifteen respondents disagreed (34 %), with six and eight individuals indicating strong and moderate disagreement, respectively. However, thirteen respondents suggested that no specific amendments to the Code would improve its effectiveness in terms of producing change in health worker recruitment policy or regulation. In explanation, two respondents suggested that the Code’s “principles were good” (CAN.CS.1) and that it was “fairly comprehensive” (US.CS.5), while others cited current domestic changes to reduce health worker recruitment that were independent of efforts to implement the Code (n = 2). One American civil society respondent explained, “[The USA government has] strengthened their investments in health professional education domestically. This was not a result of the WHO Code but it has helped alleviate the problem” (US.CS.8).

Of the eleven respondents suggesting possible changes to the Code to strengthen its impact, four suggested the principles of the Code be more widely disseminated, while nine promoted the idea of international public reporting of individual nations’ adherence. As one Canadian civil society respondent pointed out, “I believe the Code itself is a fully developed tool, but regular international public reporting on developed countries with high levels of their health workforce educated in the second or third worlds would track the relevance or effectiveness of the Code” (CAN.CS.6). Five respondents urged clearer requirements for accountability and consequences for non-adherence, with one American civil society respondent favoring a tax for destination countries and corresponding financial compensation for source countries. However, despite supporting improved accountability mechanisms, one Canadian civil society respondent noted a possible barrier to doing so, citing “…those who are reporting are not those recruiting so there is a mismatch between activity and accountability” (CAN.CS.8). Two civil society respondents suggested promoting action on the social determinants of health to improve living conditions, while a UK government respondent urged, “the [Code] needs to be turned into a monitoring board based on indicators…intentions are nice but numbers are more tangible” (UK.G.1).

Effectiveness of non-binding Codes

Few respondents were of the opinion that voluntary, non-binding codes such as the WHO Code were effective instruments through which to influence national policy (n = 3). The majority of respondents indicated that the Code was only “somewhat” effective, with proponents offering the rationale that the Code raised the international profile of the issue (n = 4) and was an effective tool to support advocacy efforts or to justify implementation of existing bilateral agreements (n = 5). One respondent suggested a particular utility of the Code in enabling the “naming and shaming” of non-adherent countries, claiming, “wealthy countries such as Australia have ‘pride’ and would not wish to be exposed as an exploiter and so have made serious endeavors to adhere to the guidelines” (AUS.CS.3).

Participants reaffirmed that the Code was useful in principle, but that its utility was undermined by national health systems that involved multiple levels of leadership. As stated by one American civil society respondent, “They can be [useful], if they are created with the active participation of the leaders in the field. At least in the USA, that leadership is widely distributed, fluid, and definitely not of a single mind” (US.CS.4). These sentiments were echoed by another American respondent, who argued, “Given that the USA does not have a national health system, disseminating this information and getting multiple different systems and entities to follow this, I feel, will be virtually impossible” (US.CS.3).

Ten participants asserted that voluntary, non-binding Codes were not effective. Respondents cited a lack of necessary accountability mechanisms (n = 5), in addition to prevailing health inequities that remain the primary incentive to migrate (n = 3). Internal pressures to fill under-resourced health sectors, and a sense that the Code was more highly esteemed by developing countries were also listed as reasons for its ineffectiveness. One Canadian participant explained, “In developed countries, there is little sense that these types of WHO edicts apply to them, whereas in low-income countries, the WHO carries a lot of weight. In this case, the voluntary, non-binding nature of the Code makes it a low priority for all” (CAN.CS.5).

Comparative principal findings between current and initial study

The main findings between the initial impact evaluation of the Code and the current one (conducted at eleven months and four years post-adoption, respectively) are strikingly similar. In both studies, the majority of key informants reported that no significant policy or regulatory changes to health worker recruitment had occurred in their countries as a direct result of the Code. To explain this lack of impact four years after adoption, we turned to global health governance frameworks developed by Hoffman and Røttingen (2015) and Happaerts (2012) whose shared principles provide a theoretical basis in which much of the analysis of our responses can be grounded [14, 24].

In assessing the effectiveness of global health treaties, Hoffman and Røttingen (2015) derived three characteristics necessary for such agreements to achieve intended social impacts (defined as those without principally economic aims). These included: 1) incentives for participating actors in alignment with their self-interest and priorities 2) institutional mechanisms to promote compliance or penalize divergence from treaty principles and 3) interest mobilizers, or strong coalitions and advocacy groups with sufficient resources to promote implementation [14]. Similarly, Happaert (2012) posits that successful uptake of ‘outside-in’ policies (those that are internationally defined but rely on national and sub-national leadership for implementation, such as the Code) requires both 1) the compliance of national governments with binding agreements, or ‘international harmonization’ and 2) the promotion of policy models, or ‘norm-setting’ by international organizations. [24] The latter is further explained by Bernstein and Cashore’s discussion of international normative discourse (2002), where international organizations like the WHO are able to shape domestic policy by first defining principles of appropriate behaviour and then driving their international institutionalization by developing a pervasive discourse around them (through mechanisms such as global fora and inter-state conventions [24].

Bilateral or regional Codes in incentivizing compliance

To explain the Code’s lack of perceived uptake, many participants emphasized a lack of incentives for compliance (due to its voluntary, non-binding nature) as well as the existence of disincentives for adherence (including the primacy of fulfilling domestic health worker needs. In both studies, the Code’s impact was perceived to be secondary to that of the multiple existing bilateral or region-specific Codes. The precise nature of these agreements better approximate each party’s self-interest (thereby providing adequate incentive), and include institutional mechanisms such as penalties for non-compliance, thus also facilitating international policy harmonization. For example, during the 1990s the proportion of foreign-trained physicians working in the UK had increased steadily as a result of recruitment efforts directed at low-income countries [25–29]. In 2001, the UK Department of Health introduced the UK’s Code of Practice on International Recruitment to mandate that no National Health Service (NHS) employers would engage in active recruitment of health workers from low-income countries in recognition of its detrimental impacts on the latter’s health workforce capacity [27–29]. Penalties for offending agencies include revocation of ability to supply health workers to the NHS, which was particularly significant given that the NHS remains the UK’s largest employer and most important deliverer of healthcare [27]. An evaluation of the UK Code by Buchan et al. (2009) demonstrated a “considerable” reduction in the influx of health workers to the UK following implementation of the Code: by 2007, only 45.2 % of all new physicians were international medical graduates, compared to 74.7 % in 2003, with a concurrent decrease in the proportion of international medical graduates from low income countries as well (although it was aptly noted that these findings could not be attributed solely, or perhaps even principally to the Code subsequent to the broad range of factors affecting health worker migration) [29–31].

Bilateral agreements also facilitate the development of explicit incentive-institution mechanisms that are agreeable to both nations and thus enable policy harmonization. For example, the Philippines had become the lead source country of nurses worldwide, and had the largest proportion of registered nurses (RNs) working in foreign countries, comprising one-third of all migrants leaving the country with Canada as one of the main destination countries [32–36].

Conversely, in 2008, 50 % of Canadian RNs were expected to retire within 15 years, contributing significantly to the nursing shortage expected: the healthcare needs of its aging population were expected to demand 361 000 RNs by 2016 in the context of a projected shortage expected to reach 113 000 by the same year [37]. The development of the Philippines-Canada memorandum (2008) facilitated regulated labour movement between certain Canadian provinces and the Philippines on the terms that Canada committed to supporting human resources development in the source country, including specifically by describing Canadian funding mechanisms for recruitment [38]. These explicit provisions thus enabled both the source and destination country to achieve better regulation of health workforce flows amongst them, while being able to negotiate terms acceptable to their self-interest.

Similarly, China and the UK entered into a memorandum of understanding regarding health workforce flows in 2006 with unique criteria negotiated to optimize the UK’s health workforce shortages while preventing further health workforce inequalities between urban and rural China. Since 1998, China has produced an excess of health workers relative to their capacity for absorption into the health workforce [39]. Comparatively, the UK had 57 000 fewer nurses than required to fulfill staffing requirements of the NHS in 2001, with already a large proportion of the health workforce coming from low-income countries overseas [26, 28, 29, 40]. However, Chinese physician-density in urban areas was and continues to be twice that in rural areas, whereas nurse-density is tripled in urban vs. rural centres, resulting in within-province inequalities in health worker density of up to 82 % [41]. This rural to urban displacement combined with overall excess health worker production results in a unique set of challenges that China attempted to mitigate in part through bilateral agreements regulating where destination countries including the UK could recruit workers from, resulting in a UK-China memorandum (2006) that mandated that the UK recruit from China through regulated agencies only, and prohibiting any direct recruiting of physicians or nurses and any recruiting from rural areas [38].

Consequently, country-/region-specific codes such as the bilateral agreements and the UK Code have been comparatively more successful in managing health workforce flows than the WHO Code itself, and indeed are a precedent encouraged by the WHO Code. Thus, a focus on developing such targeted agreements would facilitate nations’ compliance with the international norm, while doing so in a way that may more effectively achieve desired outcomes as a result of their binding nature and clearer incentives for adherence.

Institutional mechanisms for compliance and accountability

The impact of a lack of incentives on the Code’s effectiveness is compounded by a lack of ‘institutional mechanisms’, including penalties to discourage non-compliance or strict accountability mechanisms to monitor progress. Many participants proposed a greater role for the WHO in imposing consequences on deviant countries to better facilitate compliance: participants touted the benefits of levying a “solidarity tax” on recipient countries for each health worker gained, or through “naming and shaming” non-compliant countries through public fora. While substantive penalties are absent from the voluntary Code by definition, the WHO could remedy this deficiency by instead increasing accountability and reinforcing responsibility of destination countries through a renewed focus on gathering more specific, measurable indicators. These could include assessing mechanisms used to increase the size of the domestic workforce, including increasing health professional school enrollment, subsidies for health professional students, and incentives for rural and remote practice. These additional indicators could be used to quantify the health human resources gap more precisely, while also serving as benchmarks of national progress in achieving health workforce sustainability.

Shaping normative discourse

As a voluntary instrument ratified by the WHO, where the Code was poised and indeed perceived to have the most impact is through shaping the normative discourse around the international health workforce [24]. Indeed, despite the majority of participants believing the Code had not produced tangible changes, a large number of participants reported also their belief that no revisions could be made to improve its impact. Where the first study ascribed the resulting lack of impact in part to insufficient elapsed time, publicity, and support for implementation, little has changed three years later in the way of the Code’s promotion and subsequent uptake. And when the WHO was asked to develop a Global Strategy on Human Resources for Health, the WHO Code was seen more prominently as a watershed moment than as a catalyst for change in global health human resources [18]. Yet participants in this study still described the Code as a “fully developed tool” with sound principles, despite few member states adhering to its norms in practice. Clearly, despite having few implementation or accountability mechanisms, voluntary instruments such as the Code can serve an important role in galvanizing stakeholder attention and subsequently ensuring the issue is prioritized on both global and national agendas [24, 42]. The ability of non-binding codes to raise an issue’s global profile was recognized even by the respondents who suggested that voluntary codes were generally of low- to mixed-effectiveness. These merits of the WHO Code became evident when it received renewed commitment from attendees of the Third Global Forum on Human Resources for Health, which cited the Code as an important global commitment whose principles should be adopted by all countries [18].

WHO as interest mobilizer

Instead of focusing on the need for external advocacy efforts, the WHO’s inherent capacity as an international norm-setting agency can be capitalized on to serve as its own ‘interest mobilizer’. For example, the Third Global Forum framed the Code and its focus on human resources for health as an integral part of achieving universal health coverage (UHC), which has emerged as a topical global priority [18]. By purposefully linking the Code with current global health priorities, the Third Global Forum highlighted an opportunity for renewing attention to the principles of the Code and encouraging its implementation without expending additional resources on its advertisement. WHO could further capitalize on this momentum by linking the Code with other priorities, such as the Sustainable Development Goals (SDGs), accomplished by referring to the Code in further discussions on UHC and the SDGs, or in global fora or technical consultations concerning the ongoing development of the WHO’s global strategy on human resources for health, most recently presented at the 69^th World Health Assembly for draft consideration [43]. These efforts could serve the dual purpose of ensuring that the Code continues to be publicized among member states, and that national and sub-national discussions concerning UHC, the SDGs and Human Resources for Health would inevitably involve a deliberate focus on the state of implementation of the Code.

Framing and naming to optimize relevance