International medical graduates' experiences before and after migration: A meta‐ethnography of qualitative studies

International medical graduates (IMGs) represent a large portion of practising doctors in many countries. Many experience difficulties, including higher rates of complaints against them and lower exam pass rates. The UK's General Medical Council (GMC) recently set targets to ‘eliminate disproportionate complaints’ and ‘eradicate disadvantage and discrimination in medical education’. Our timely meta‐ethnography aimed to synthesise existing qualitative literature on the wider personal and professional experiences of IMGs to identify factors affecting IMGs' professional practice (either directly or indirectly).


| INTRODUCTION
An international medical graduate (IMG) is a doctor who practises medicine in a country different to the country where they received their primary medical qualification (PMQ). 1 Around 40% of practising doctors in the United Kingdom are IMGs. 1 This figure is over 25% in the United States and Canada and over 40% in Australia, Ireland, Israel, New Zealand and Norway. [2][3][4] Compared to domestic medical graduates (DMGs), IMGs are more likely to receive patient complaints, 5,6 have lower pass rates at postgraduate exams [7][8][9][10] and are less likely to achieve board certification. 11 There is a similar pattern of higher complaints rates and differential attainment for doctors from ethnic minorities (some of whom are IMGs). 5,12 In recognition of the urgency in addressing these issues, the General Medical Council (GMC) recently set a target to 'eliminate disproportionate complaints from employers about ethnic minority doctors' by 2026 and 'eradicate disadvantage and discrimination in medical education and training' by 2031. 8 A growing body of qualitative literature has sought to explore IMGs' migration motives and integration experiences. 13 Although offering valuable insights into the range of challenges IMGs experience in the country of practice (host country), published studies are typically confined to a single setting (e.g. the local context of a hospital 14 or (a) training programme(s) 15,16 ) or region while mostly relying on small study samples. 17 Studies recruiting participants across a whole country are uncommon, [18][19][20] and those recruiting across more than one country are rare. 21 Yet, identifying commonalities in the experiences of IMGs, regardless of country of PMQ, host country or medical specialty, might be valuable in providing some explanation for the disproportionate complaints and differential attainment. Identifying these commonalities might also be valuable for shared learning and joint efforts in designing educational interventions and policies to achieve the GMC targets.
Approaching our research question from a constructivist worldview and considering our aim to provide an in-depth understanding of the experiences of IMGs worldwide, we chose to conduct a systematic review and qualitative evidence synthesis (QES) and, specifically, used the technique of meta-ethnography.
Meta-ethnography 22,23 is one of the most commonly used methods for QES. 24 Rather than simply aggregating existing findings, it involves developing new conceptual understandings and 'making a whole into something more than the parts alone imply'. 22 Our review question was: What are the personal and professional experiences of IMGs worldwide before and after they migrate to the country in which they practice?

| METHODS
We followed Noblit and Hare's seven steps of conducting a meta-ethnography 22,23 and report our findings in accordance with the eMERGe Meta-Ethnography Reporting Guidelines. 25 Online Supplement 1 provides a summary of the guidelines with the corresponding page in this manuscript. The review protocol was registered with the PROSPERO International Prospective Register of Systematic Reviews (CRD42020176053). To identify potential additional articles, we also manually screened the reference lists of relevant reviews. Articles were eligible for inclusion if they (a) were published as original research articles in peer-reviewed journals; (b) were written in English;

| Search strategy and selection criteria
(c) explored the experiences of IMGs in any context; and (d) employed a qualitative design, with interviews or focus groups as the primary method of data collection. Mixed-methods studies were also eligible for inclusion, provided that the qualitative component was presented in sufficient detail. We excluded studies that recruited IMGs as part of a larger group of participants (e.g. doctors in training) and did not separately report findings from IMGs. We also excluded articles reporting experiences of IMGs who were on short-term placements abroad or had returned to their country of PMQ.

| Data screening and quality assessment
We imported all identified references into EndNote (Clarivate Analytics, Philadelphia, PA) and removed duplicates. We then uploaded the results to Covidence (Veritas Health Innovation, Melbourne, Australia) and removed more duplicates. Titles and abstracts of all articles were screened independently by two reviewers: MAH screened all articles, whereas SJ and EG each screened half of the articles. During the second stage of screening, full texts of potentially eligible articles were retrieved and assessed independently, as described above. We resolved disagreements by discussion at monthly meetings and kept a journal to record all decisions. MAH read all included articles again and extracted key study characteristics (e.g. aims, country and sampling) in a spreadsheet using Microsoft ® Excel for Mac (Microsoft Corporation, Redmond, WA).
We assessed the quality of included studies using the Critical Appraisal Skills Programme (CASP) tool for qualitative research. 26 The use of purposeful sampling of articles has been advocated as a way of making a qualitative synthesis more manageable. 27 Yet, we opted to include all articles that met our inclusion criteria. The large number of studies meant that there was sufficient material for us to synthesise within each of the subtopics, for example, migration and acculturation.
It also enhanced the confidence in our findings using the GRADE-CERQual framework. 28 In meta-ethnography, the order in which articles are synthesised can be important and remains a matter of debate. 23,29,30 It is likely that articles synthesised earlier have a stronger influence on the development of ideas. 29,30 While synthesising articles in chronological order is common practice, some have advocated starting the synthesis with articles based on quality appraisal 'index papers' 23,31 or combining the two approaches. 30 To identify 'index papers', we assessed aim congruence, conceptual clarity and interpretive rigour. Aim congruence was the extent to which the aim of the article was aligned to our research question. Conceptual clarity and interpretive rigour have been previously described by Toye and colleagues. 23 They pertain to the richness and clarity of findings and rigour of data analysis, respectively. MAH evaluated all articles in each of the three domains and scored them as 'high', 'moderate' or 'low'. First, we ordered articles from high aim congruence to moderate then low. Within each of these three groups, we ordered articles from high conceptual clarity to moderate then low. Within each of the resultant groups, we ordered articles from high interpretive rigour to moderate then low. Finally, within each of the resultant groups, we ordered articles chronologically. Our 'index paper' 32 was therefore the oldest article that received a high score in all the domains of aim congruence, conceptual clarity and interpretive rigour.

| Data analysis
We imported all articles into NVivo 12.6.1 (QSR International, Doncaster, Australia) and extracted second order constructs. A second-order construct, as defined by Shütz, is a concept described by the original authors and is often presented with first-order constructs (participant quotes). 33 Third-order constructs are concepts arising from our (the reviewers') interpretations of identified second-order constructs and one of the outcomes of this metaethnography. We coded each second order construct as a 'node' in a way similar to that described by Toye and colleagues. 34 We kept a 'memo' for each node and updated it each time a new article contributed to it. This meant that we ended up with fewer nodes than the sum of second-order constructs listed in the articles.
Starting with Wong and Lohfeld's article, 32 we translated (compared) each second-order construct into the same second-order construct from the next article that it appeared in. The translation was either reciprocal or refutational depending on whether it corroborated or refuted the argument already presented in the previous article(s).
We therefore carried out reciprocal and refutational translations by second-order construct rather than by article. For instance, many articles reported on experiences of IMGs in the initial period in the host country as being stressful (reciprocal), 18,20,[35][36][37][38][39][40] but Skjeggestad and colleagues reported that some IMGs trivialised these experiences (refutational). 18 The resultant translation was that although stress was widely experienced by IMGs during this period, the experience was not universal. This contributed to our line of argument that experiences were varied and personal.
We carried out the translation until all articles with the same second-order construct were translated into each other, and a line of argument was formed. We repeated this process for all secondorder constructs and considered alternative interpretations at monthly team meetings. Reading and rereading the articles within the context they were presented allowed immersion in the data. This allowed us to develop an overall line of argument and a conceptual model that encompassed, described and explained experiences of IMGs worldwide. The resultant synthesis was thus grounded in the published data and the accounts of the hundreds of IMGs who contributed to them.

| Reflexivity
The lead author (MAH) is an IMG practising in the United Kingdom.
He has significant involvement in supporting IMGs in Scotland. MAH recognised many of the experiences described in this article. To mitigate potential bias, all findings, interpretations and the conceptual model were challenged by SJ and EG at monthly team meetings, and alternative interpretations were explored. SJ is a professor of health professions education with experience in biomedical and education research. EG is an experienced qualitative researcher, with a social sciences background. Both SJ and EG were involved from conception of the project and through all stages including reviews, checks and challenges of translations, third-order constructs, the conceptual model and drafts of this article.

| RESULTS
The flow diagram of study selection is presented in Figure 1. 41 No article was excluded based on the language it was written in. Online Supplement 3 provides a summary of the critical appraisal.

| Characteristics of included studies
Characteristics of all 57 articles included in the meta-ethnography are presented in the order they were synthesised in Online Supplement 4. Among included articles, only five used mixed methods.
Articles described 46 studies that included 1142 IMGs from low-, medium-and high-income countries. Twenty-one articles were conducted in Europe, 21 in the United States and Canada, 13 in Australia and New Zealand, one in Asia and one in the Middle East.
A study reported in two articles included participants from two European countries 21,42 ; one article included participants from the United States, Canada and Israel 43 ; and the rest included participants from only one country. Sixteen articles were set in hospitals, 13 in primary care, six in the rural setting and 17 in mixed settings, and five did not state the setting the study was conducted in. Fifteen articles involved IMGs in training, 13 involved IMGs in permanent posts, 13 involved a mixture of grades, 15 did not state the grade of IMG, and one article involved IMGs after migration but before employment.

| The synthesis
Ultimately, 56 articles contributed to the synthesis; one article was excluded, as it addressed a specific issue related to the use of translators by IMGs, 44 which was not reported in any other article. We identified five third-order constructs related to 14 second-order constructs ( Table 1).

| Line of argument
Overall, IMGs made comparisons across several migration dimensions (safety; professional education and development; work conditions; connectedness; belonging; status and lifestyle) to make their own individual decision to migrate and to which country they migrate to. In the host country, they faced a challenging start to varying degrees.
There were four main barriers (language, culture, medical education and belonging) that they had to overcome. The impact of these barriers was different for each IMG depending on the degree of dissonance in relation to each of the barriers. Support to reduce the impact of the four barriers could level the playing field. IMGs aimed to survive initially, then adjust to overcome these barriers and thrive ( Figure 2). We will now describe each of these third-order constructs.

Migration dimensions
Migration and the decision to migrate were governed by the balance of a variety of push, pull and plant factors. 18,19,21,42,[45][46][47][48][49][50][51][52][53][54] Push factors are those that push people away from a country; pull factors are those that pull them to a certain country; and plant factors are those that keep them planted in the country they reside in. 45 These factors might be better considered along certain dimensions. For example, war, political instability, violence and crime 18,19,21,42,45,47,48,51  where the IMG was when they considered migration and between which countries they made the comparisons. Comparisons were repeatedly made, typically-but not exclusively 52 -between country of origin and host country. These comparisons were relative and subjective. In the case of migration from a war-torn country to a politically stable country, the difference was stark, but differences in other dimensions were not always as clear-cut.
Another dimension was opportunities for professional education and development, which was both a push and pull factor. 19,21,42,[46][47][48][49]54 This dimension was also considered when contemplating whether to remain in the host country, go back to the original country or migrate elsewhere. An example of the latter was seen in the accounts of IMGs, who did not have opportunities to develop professionally in Ireland and were considering moving elsewhere, for example, to the United States. 52 Work conditions and pay were other dimensions that IMGs considered. 19,21,42,[47][48][49] The difference between countries could be stark, for example, when comparing these in some low-income countries with higher-income countries. 19,21,42,[47][48][49] It could also be more subtle, for example, migration of UK doctors to Australia 47 or between Nordic countries. 18 Connectedness, belonging and status were other dimensions that IMGs considered. 19,20,[54][55][56][57] On these dimensions, IMGs generally moved towards a country where they felt less connected and had a lower sense of belonging and status compared with their own countries. As such, connectedness, belonging and status could be seen as factors that kept IMGs planted in their original countries. As the levels on these dimensions improved for IMGs, they acted as plant factors that tied them and their families to the host country. 46,48,50,53 IMGs made these evaluations in the different dimensions based on their values, perceptions and judgements. There were no clearer examples of the subjectivity of these dimensions than in the T A B L E 1 Third-order constructs with definitions, their relation to the second-order constructs and articles from which the second-order constructs were extracted Third-order construct Definition Related second-order constructs Articles contributing to second-order constructs

Migration Dimensions
When considering migration, IMGs compared countries they were going to emigrate from and countries they were immigrating to. These individual and subjective comparisons were made along a few dimensions before a final decision was made regarding migration Immigration push, pull and plant factors 18,19,21,42,[45][46][47][48][49][50][51][52][53][54] Immigration is individual and often pragmatic 18,19,21,42,45,46,49,54,73 A challenging start Landing in the host country was challenging. It was a stressful period that was characterised by loss, shock, disorientation and difficulties in entering the medical profession Loss 14,32,37,47,52,58,59 Shock and disorientation 14,17,32,36,40,47,55,56,58,60,61,65,68,69 The barrier of entry into the profession 18,19,32,35,36,40,42,43,46,47,55,[57][58][59]62,63 Degree of dissonance There were four specific and significant barriers faced by IMGs. These were the differences in language, culture, medical education and belonging. Information about these exams could be hard to find, and the assessment methods could be alien, all leading to a heightened state of anxiety regarding the licensing process. 18,[35][36][37]40,47,58,62 Added to that, IMGs often had to find other jobs to sustain them and their families until they passed exams. 18,19,36,43,57 The pressure led some IMGs to abandon a career in medicine, especially women who often felt that it was their responsibility to look after their families and children. 42,43,57 For some IMGs, for example, refugees, there was the added stress related to the uncertainty about visas and residency status. 37 It was also during this period that IMGs had to secure their family's basic human needs 64 from accommodation to schooling and banking, adding to the anxiety and stress. 37,47,55,65 Once IMGs obtained their licences, they competed for posts with DMGs and often lost out to the competition for popular posts and had to settle for less popular ones, often in areas remote to where they landed. 15,18,66 This could lead to separation from their families, which negatively affected their mental well-being as they lost a vital source of comfort and support. 20,67 IMGs suffered culture shock and disorientation at the workplace once they commenced their posts. 14 idioms. 15,74 Although IMGs attended courses if available, these did not prepare them for the necessary language skills required at the workplace, where abbreviations and special terms were used. 17,18,63,72 Some IMGs found it difficult to express themselves, especially when a response was expected quickly. 75 That resulted in avoiding speaking and discussions, 36,70 especially in groups, which IMGs feared came F I G U R E 2 A conceptual model of the findings of the meta-ethnography. After considering migration dimensions, IMGs migrated to the host country and immediately faced a challenging start. IMGs adjusted to survive and ultimately thrive. However, they faced four main barriers that were experienced by each IMG uniquely depending on the degree of dissonance (between the IMG and host country) in relation to each of these barriers. Support and interventions designed to reduce the magnitude of these barriers could level the playing field across as lacking knowledge. 76 This also made it difficult for IMGs to socialise at work, which made it more difficult to fit in and made some IMGs feel like 'outsiders'. 36 The cultural differences that IMGs faced included culture in the workplace and in society at large. 15,16,20,36,39,46,[54][55][56]58,59,65,69,[72][73][74][75][76][77][78][79][80][81][82][83] An IMG who was a native of the host country but qualified in another country did not suffer culture shock in society compared with an IMG who immigrated from another country. 14,47,71,73 However, the former was likely to suffer a workplace culture shock, especially if the country of PMQ had a very different healthcare system and workplace culture. 68 The same principle applied to differences in medical education between country of PMQ and host countries. Some IMGs immigrating from a medical education system where the emphasis was more on knowledge and science found it challenging to work in a healthcare and medical education system where there was more emphasis on communication skills and placed patients-rather than science-at the centre of care. 14 Some IMGs felt alienated and that they did not fit in or belong in their new workplace and communities. 19,20,55,57,72 They did not feel they were trusted as their practice was constantly being observed and felt unfairly judged if they made a mistake similar to one made by a DMG. 17,20,35,61,76 Some were made to feel unwelcome and were subjected to overt or subtle racism, marginalisation and/or discrimination. 15,20,35,46,54,59,68,81 They could experience institutional and systematic discrimination 15,[18][19][20]36,46,52,[55][56][57][58]66,68,81 that put them at a disadvantage when entering the profession, training programmes and progressing within these. 19,20,36,46,52,66 Levelling the playing field Interventions that targeted these barriers could reduce their impact on IMGs' progress and mental health. Orientation programmes were not universally effective. 17,18,36,46,58,62,67,69 When comprehensive, they orientated IMGs to their new workplace and sociopolitical environment, helping reduce the impact of the cultural barrier. 62,67 IMGs were expected to perform at a high level from the start despite their disorientation and stress. 17,18,40,58 Allowing a longer period to settle in before assuming their full duties would give IMGs time to work on reducing the impact of language, cultural and medical education barriers. 17,32,36,46,67 Language courses appeared to reduce the impact of the language barrier to some extent. 17,18,63,72 Well-designed and context-focused language 17 and communication training as well as targeted medical education training could also reduce the impact of the language and medical education barriers respectively, although empirical evidence is required to support that.
Mentors or buddies were invaluable in reducing the impact of some of the barriers and alleviating some of the stress for IMGs, 32,35,54,62 as did immigrating to a welcoming community 53,55 and workplace with supervisors who were familiar with the needs of IMGs and were able to support them. 20,36,40,54,63 Survive then thrive Ultimately, IMGs migrated to improve their personal and professional lives. 18,19,21,42,[45][46][47][48][49]54 During the initial period, they just wanted to survive. 15,17,18,32,35,36,40,46,47,54,56,58,61,62,65,67,68,70,72,80 They had to quickly adjust and adapt to fit in. Initially-especially when language, culture and medical education were very different-they tried not to stand out 17,32,70 and contributed as little as possible to discussions. 70 All the while, their senses were on high alert. IMGs adjusted to a lower status 61 that could be frustrating 35,56,61 and extended this adjustment to within the community. 56 At times, the adjustment was to downgrade career plans and level of ambition. 18 Nevertheless, IMGs forged ahead by using what they learnt in their initial heightened state of awareness and continued making adjustments. 61 They adjusted the way they spoke, 15,16,65,74,78,80 wrote 36,61,68 and acted 54,61,74,79 to blend in. 32 Table 2 summarises the confidence in our findings using the GRADE-CERQual method. 28

| DISCUSSION
This meta-ethnography is a comprehensive synthesis of the published qualitative literature, spanning from 1997 to 2019 and reporting the experiences of more than 1000 IMGs. We aimed to enhance understanding of experiences that were common to all IMGs, regardless of host country and country of PMQ. We did this by developing five third-order constructs and a line of argument that helped explain these experiences in a way that is both comprehensive and accessible.
We have presented a conceptual model that explains how these constructs are related. We found that there were commonalities in the experiences of IMGs worldwide, but each IMG was unique. Our findings provide a simple and pragmatic framework to guide the understanding and assessment of each IMG's circumstances, experiences and needs.
To our knowledge, our meta-ethnography is the only QES covering the broad range of experiences of IMGs before and after migration. Other available reviews have used different methods and have focused on specific issues or time periods in relation to IMGs. 13,84,85 Jalal and colleagues, 84 as well as Michalski and colleagues, 13 conducted systematic reviews and aggregated themes from empirical studies into categories. The reviews focused on intercultural issues and transitioning. In relation to these two topics, both reviews highlighted similar second-order constructs to the ones we identified.
T A B L E 2 A summary of the confidence in our findings using the GRADE-CERQual method 28 Review finding Articles contributing to the review finding

Methodological limitations
Coherence Adequacy Migration dimensions 18,19,21,42,[45][46][47][48][49][50][51][52][53][54]73 Minor concerns Levelling the playing field [14][15][16][17][18]20,32,[35][36][37][38][39][40]46,47,54,55,57,58,62,63,65,[67][68][69]72,[75][76][77]80,83,88 Minor concerns (11 articles with no concern or very minor concerns, 9 with minor concerns, 12 with moderate concerns) This study is not without limitations; however, some of these may indicate important avenues for future research. First, all studies included in this review were conducted in high-income countries. This might be because these countries rely heavily on IMGs to deliver healthcare. 4 Moreover, the publication dates of the articles in this synthesis spanned from 1997 to 2019. Although this meant that we provided a comprehensive picture of all available published qualitative literature on the topic, it is also possible that some of the specific issues explored in these articles might have changed or have been resolved over the years. A further limitation that applies to most QES is that qualitative research is often not well indexed. The qualitative filters that we used, however, had a 95% sensitivity in Medline and 94% sensitivity in Embase and PsycINFO. 89 Last, the quality of any evidence synthesis depends on the quality of identified primary research, and, in our case, the quality of included studies was variable, with the areas of well-being and colleague relationships with IMGs being significantly understudied.
Our findings and conceptual model provide a useful framework to view the experiences of IMGs worldwide and provide insights into potential causes of disproportionate complaints and differential attainment. It might help IMGs to know that their experiences are shared, to look out for stressors and act to mitigate them and seek support at specific points in time. Our model is comprehensive, yet accessible and easy to use in a pragmatic way for individual IMGs' needs assessment or to help those who are in immediate contact with IMGs (e.g. from colleagues and supervisors to members of the multidisciplinary team) to understand IMGs' experiences. This will hopefully encourage these colleagues to nurture and support IMGs to enable them to thrive for the benefit of IMGs, colleagues, patients and the healthcare systems within which IMGs work. Moreover, the model and concepts we present are of practical use to guide policymakers in countries that rely heavily on the IMG workforce to design policies that are IMG-friendly to help 'eliminate disproportionate complaints from employers' and 'eradicate disadvantage and discrimination in medical education and training'. 8 To achieve these goals, we recommend that our findings are used to raise awareness of experiences and challenges IMGs might face, especially to their immediate supervisors; those working alongside IMGs, supervising them or employing them should strive to level the playing field for IMGs to improve their experiences and facilitate integration; policymakers should take our findings into account and strive to create IMG-friendly policies, especially if the host country relies heavily on IMGs in delivering healthcare; specific interventions at specific times should be designed to address the individual needs of IMGs; and attention should be given to under-researched aspects of the IMG experience, including well-being and colleague relationships.