Article Text

Original research
Developing an evidence assessment framework and appraising the academic literature on migrant health in Malaysia: a scoping review
  1. Allard Willem de Smalen1,2,3,
  2. Zhie X Chan3,
  3. Claudia Abreu Lopes3,
  4. Michaella Vanore1,2,
  5. Tharani Loganathan4,
  6. Nicola S Pocock3,5
  1. 1Maastricht Graduate School of Governance, Maastricht University, Maastricht, The Netherlands
  2. 2United Nations University—Maastricht Economic and Social Research Institute on Innovation and Technology, Maastricht, The Netherlands
  3. 3United Nations University—International Institute for Global Health, Kuala Lumpur, Malaysia
  4. 4Centre for Epidemiology and Evidence-based Practice, Department of Social and Preventive Medicine, University of Malaya, Kuala Lumpur, Malaysia
  5. 5Gender Violence & Health Centre, London School of Hygiene and Tropical Medicine, London, UK
  1. Correspondence to Mr Allard Willem de Smalen; allarddesmalen{at}gmail.com

Abstract

Background A large number of international migrants in Malaysia face challenges in obtaining good health, the extent of which is still relatively unknown. This study aims to map the existing academic literature on migrant health in Malaysia and to provide an overview of the topical coverage, quality and level of evidence of these scientific studies.

Methods A scoping review was conducted using six databases, including Econlit, Embase, Global Health, Medline, PsycINFO and Social Policy and Practice. Studies were eligible for inclusion if they were conducted in Malaysia, peer-reviewed, focused on a health dimension according to the Bay Area Regional Health Inequities Initiative (BARHII) framework, and targeted the vulnerable international migrant population. Data were extracted by using the BARHII framework and a newly developed decision tree to identify the type of study design and corresponding level of evidence. Modified Joanna Briggs Institute checklists were used to assess study quality, and a multiple-correspondence analysis (MCA) was conducted to identify associations between different variables.

Results 67 publications met the selection criteria and were included in the study. The majority (n=41) of studies included foreign workers. Over two-thirds (n=46) focused on disease and injury, and a similar number (n=46) had descriptive designs. The average quality of the papers was low, yet quality differed significantly among them. The MCA showed that high-quality studies were mostly qualitative designs that included refugees and focused on living conditions, while prevalence and analytical cross-sectional studies were mostly of low quality.

Conclusion This study provides an overview of the scientific literature on migrant health in Malaysia published between 1965 and 2019. In general, the quality of these studies is low, and various health dimensions have not been thoroughly researched. Therefore, researchers should address these issues to improve the evidence base to support policy-makers with high-quality evidence for decision-making.

  • statistics & research methods
  • public health
  • health services administration & management
http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Strengths and limitations of this study

  • This study provides a comprehensive overview of migrant health research in Malaysia, including a summary table, critical assessment tables and a multiple-correspondence analysis.

  • Methodological contributions by creating an evidence assessment framework, including a decision tree that identifies the type of study design and corresponding level of evidence, and modified Joanna Briggs Institute checklists.

  • Exclusive focus on vulnerable migrants within the non-citizen population in Malaysia.

  • Only English peer-reviewed academic articles were included in this study, and, therefore, much relevant information that could potentially be used to inform both policies and practice may have been excluded from this review.

Introduction

Worldwide, the international migrant population accounts for approximately 272 million people, with almost one-third within Asia.1 Due to its strategic geographical location and high labour demand, Malaysia is among the top destination countries for international migrants in the Asian region.2 According to the Department of Statistics Malaysia (DOSM), the documented non-citizen population represented 3.2 million people in 2019, which accounts for 10% of Malaysia’s total population.3 DOSM defines a non-citizen as a person that resides in Malaysia for 6 months or more in the reference year.4 However, no subcategories were included in this definition. According to the Office of the United Nations High Commissioner for Human Rights, a non-citizen is an individual that does not have an effective connection with the location where the person is situated according to the host nation, and includes various types of migrants, such as foreigners with permanent residency, refugees, asylum seekers, foreign labour, international students, stateless individuals and victims of human trafficking.5 Other definitions of migrant-related terms that are used in this paper are presented in table 1.

Table 1

Definitions of migrant-related terms

The vast majority of non-citizens in Malaysia are migrant workers, where foreign labour can be divided according to their visa status into regular and irregular migrant workers. According to the Ministry of Home Affairs, Malaysia issued 2 million work permits to documented migrant workers in 2019.6 However, the total number of migrant workers, both documented and undocumented, is estimated to fall between 4.2 and 6.2 million people.2 Another group that contributes significantly to the non-citizen population in Malaysia is refugees and asylum seekers. The terms refugees and asylum seekers are often used interchangeably, yet, these populations differ by their legal status in destination countries and subsequent vulnerabilities (see definitions in table 1). In 2019, an approximate 178 580 refugees and asylum seekers were registered with the United Nations High Commissioner for Refugees in Malaysia, where 153 770 (86%) came from Myanmar. The remaining number (14%) came from Yemen, Syria, Afghanistan, Iraq, Palestine, Pakistan, Sri Lanka, Somalia and other countries.7

Refugees, asylum seekers and both documented and undocumented low-skilled foreign workers can be classified as vulnerable migrants in Malaysia, as these populations may face significant hardships in their new country of residence.8 9 Vulnerable migrants are more prone to being exploited and abused, have an increased need to be protected by duty bearers and are not able to fully benefit from their human rights.10 Health is among these affected human rights, as migrant workers and refugees could encounter various challenges to maintain proper health and prevent poor health outcomes, including difficulties in accessing healthcare and obtaining quality health services.10–12 According to Sweileh et al,13 assessing the current status of scientific output and identifying research gaps could positively contribute towards improving the evidence base for advocating for migrant health needs. Scoping reviews can be helpful to map the academic literature and have been used by different researchers to present the available evidence on migrant health issues in other countries.14 15

Despite the burgeoning academic literature on migrant health in Malaysia, health information on migrant-related issues is still limited, and public data remain difficult to access. Aggravating the matter, there is no overall picture currently available of the evidence base on migrant health in Malaysia, including critical appraisal of the quality of research. Therefore, this study aims to map the existing academic literature on migrant health in Malaysia since 1965 to identify the trends and gaps in this field, as well as to present an overview of the topical coverage, quality and level of evidence of these scientific studies.

Methods

General methods

A scoping review was conducted, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses—Extension for Scoping Reviews guidelines16 (online supplemental file 1). A prereview protocol was developed to guide decisions for literature selection and structure of the review, and included the review question, aim, search strategy, selection criteria and risk of bias assessment. However, the protocol was not formally registered and changed to some extent over the course of this review. The prereview protocol can be accessed on request from the first author. Data were extracted and organised using the Bay Area Regional Health Inequities Initiative (BARHII) framework.17 In addition, a decision tree was developed to classify the type of study design and level of evidence of each journal article. Subsequently, a quality assessment of the included literature was conducted by using the Joanna Briggs Institute (JBI) critical appraisal toolkit.18 Lastly, the data were analysed, and a multiple-correspondence analysis (MCA) was applied to explore existing relationships between variables, including the type of migrant, main health dimension, quality of the study and research design.

Patient and public involvement

Patients and the public were not involved in this study.

Conceptual framework

The BARHII framework was used to organise the identified literature in this scoping review into specific factors that shape equitable health outcomes (figure 1). The BARHII framework was selected due to its comprehensive nature and inclusion of various health dimensions, whereas other models focused on specific public health elements or lacked clear explanation regarding the included health-related components of the model.19 20

Figure 1

Bay Area Regional Health Inequities Initiative framework.

The BARHII framework consists of six dimensions: (1) social inequities; (2) institutional inequities; (3) living conditions; (4) risk behaviour; (5) disease and injury; and (6) mortality. In addition, each health dimension contains various subdimensions (as presented in figure 1). Except for ‘social inequities,’ the other five categories were used to describe which health dimension the particular articles focused on. The social inequities element was incorporated by describing the population of interest, which was divided into three categories: foreign workers, asylum seekers and refugees, and unclassified migrants. The lattermost category was applied if a paper used the term ‘migrants’ or ‘immigrants’ but lacked specific information to classify the study population as foreign workers or asylum seekers/refugees.

Institutional inequities include the practices of corporations, businesses, government agencies, schools, not-for-profit organisations as well as laws, regulations and policies that could influence health outcomes (eg, a regulation that obligates companies to financially compensate an individual in case of a work incident).

Living conditions consist of the physical environment (eg, indoor air pollution), economic and work environment (eg, unemployment), social environment (eg, discrimination in the neighbourhood) and service environment (eg, healthcare) that people live in, and that play a role in determining their health outcomes (eg, denied healthcare access due to visa status).

Risk behaviour includes smoking, poor nutrition, low physical activity, violence, alcohol and other drugs and sexual behaviour. This dimension reflects the way someone acts and how that increases or decreases the risk of obtaining a particular health outcome (eg, the attitude and related behaviour towards smoking could influence an individual’s level of risk of developing lung cancer).

Disease and injury consist of communicable diseases (also known as infectious diseases; eg, chlamydia), chronic diseases (also known as non-communicable diseases; eg, cancer) and injuries (eg, fractured bone). This dimension describes the number of people or individual cases with a particular health outcome (eg, 10 out of the 100 people suffered from cancer).

Mortality was changed to ‘mortality and morbidity’ and focused on death and disease rates of the study population (eg, 10 out of 1000 live births of children under the age of 1 died) to distinguish epidemiological studies with larger samples from descriptive studies with smaller samples, where the latter were categorised as disease and injury studies.

Furthermore, some additional subdimensions were created during the data extraction stage, as these were lacking in the original BARHII framework (eg, the subdimension ‘mental health’ was added to the disease and injury dimension).

Search strategy

Based on the guidelines of the London School of Hygiene and Tropical Medicine21 and Bramer et al22 on selecting the number and types of databases that should be included in biomedical systematic searches, six databases were selected for this study: Econlit, Embase, Global Health, Medline, PsycINFO and Social Policy and Practice. This scoping review includes studies from 1965 onwards until 2019. However, all identified records were retrieved from the six databases to manually screen the data for publication date-related issues. The search process was conducted by AWdS and included a two-stage procedure to ensure that the search was exhaustive and to minimise the risk of missing potentially eligible studies. The first stage focused on identifying English-language key words and Medical Subject Headings terms for migrants (eg, immigrants, foreign workers, refugees), health (eg, disease, infection, disorder) and Malaysia (eg, Sabah, Kuala Lumpur) through reading search strategies of other review studies on migrant health as well as using medical terminology of renowned medical institutions, such as the Mayo Clinic. Subsequently, these items were combined by using Boolean operators (eg, migrant AND health AND Malaysia) in the search platform of each database (online supplemental file 2).

Selection criteria

Studies were eligible for inclusion if they met the following inclusion criteria: (1) conducted in Malaysia, including cross-national studies in which Malaysia was included; (2) published in peer-reviewed academic journals; (3) primary outcomes of the study included a health-related variable from at least one of the five health dimensions of the BARHII framework; (4) employment of one of the following study designs: literature synthesis (systematic review, meta-analysis, other scientific review designs), qualitative (interviews, focus group discussions) and/or quantitative (randomised controlled trial (RCT), cohort, case-control, cross-sectional, case series, case report) study design; (5) written in English; (6) inclusion of international (im)migrants, foreign workers, asylum seekers and refugees, as these groups were considered as vulnerable migrant populations in Malaysia. Articles that included both migrants and the general population were included in this study if sufficient information concerning the migrant population was available.

Studies were excluded if they were: (1) conducted or included data from 1965 or earlier, as Singapore was part of Malaysia until 1965, and this study is careful to only include Malaysia studies without Singapore; (2) grey literature; (3) opinion papers, editorials, fieldnotes of symposia, conferences and workshop abstracts; (4) focused on non-citizens and foreigners, where it was unclear whether a vulnerable migrant population was included (such as permanent residents, naturalised persons, expatriates, temporary visitors, tourists, Malaysian returnees and international students); (5) only presented migrants as a control variable and no other information regarding migrants was available.

Data extraction

Three reviewers (AWdS, ZXC and NSP) were involved in the screening process, where all had experience in the domain of public health and AWdS and NSP had practical knowledge with respect to conducting systematic reviews due to previous research work. Titles and abstracts were exported by AWdS and subsequently moved into Rayyan, an open-source software designed to support systematic reviews. AWdS and ZXC were the main reviewers, where AWdS conducted an entire screening of titles and abstracts and ZXC assessed a randomly selected 20% sample. Independent screening was carried out by using the ‘blind’ function of Rayyan, with both researchers working separately. The first stage involved screening titles and abstracts according to the inclusion criteria. Subsequently, AWdS and ZXC conducted an independent full-text screening of all potential articles and attached comments to each article on why the paper was included or excluded. After each screening stage, AWdS and ZXC compared their findings and discussed the discrepancies. In both stages, the discrepancies were about 13%–14% of the papers and were mostly around the study design and target populations. Conflicts were examined and resolved by NSP.

Following the full-text screening stage, the data were extracted by one reviewer (AWdS) and disaggregated by the different dimensions of the BARHII framework, including the type of migrant (social inequities), main health dimension (institutional inequities, living conditions, risk behaviour, disease and injury and mortality and morbidity) and health subdimensions.

For the next stage, a decision tree was developed to ensure that the correct quality appraisal tool by study design was selected and to identify the level of evidence of the included literature (figure 2). Although various research designs were included in the decision tree, some study designs did not fit in this model, such as the mixed-method design.

Figure 2

Decision tree to identify the type of study design and corresponding level of evidence.

The decision tree built on the study design tree from the Centre for Evidence-Based Medicine23 and essentially allowed research of varying designs to be consistently, reliably classified into one of several design families. The newly developed decision tree was created through a two-step process. First, a table was created that included definitions of various research designs, and, subsequently, specific traits of these definitions were used to develop guiding questions for the decision tree (table 2).

Table 2

Definitions of included study designs

Second, Tomlin and Borgetto’s24 model was used to identify the level of evidence of the included literature, as the study designs that were included in their model were in line with the research designs in the definitions table. In addition, it was one of the few models that deconstructed the single-hierarchy framework and assigned study designs to different categories depending on the study objective (eg, if the study design did not aim to provide a causal relationship, but simply describe a particular outcome, the study design would be classified as descriptive research), and, therefore, valued studies with different objectives equally. Tomlin and Borgetto’s model consists of four dimensions, including descriptive research, experimental research, outcome research and qualitative research. Each of these dimensions contains four subclasses to show the level of evidence within each class, where level 1 is the highest level of evidence and level 4 the lowest. The assignment of these levels to the different study designs is based on the degree of internal validity/authenticity and external validity/transferability, where level 1 is regarded with the highest level of these two measures and level 4 ranks the lowest. Table 3 shows the different research dimensions that correspond with the included study designs and level of evidence.

Table 3

Level of evidence for each study design

After incorporating feedback on the questions used to identify the research design and multiple testing rounds to assess if the questions were specific enough to distinguish these designs within the full set of articles, the final version of the decision tree—as seen in figure 2—was used to extract the data.

Quality appraisal and level of evidence assessment

The quality assessment of the included studies was conducted by one reviewer (AWdS) based on the JBI critical appraisal tools, as this toolkit includes checklists for a wide variety of study designs that are most in line with the research designs included in this study. Additional objective criteria specific to migrant health studies were developed for each question of the JBI checklists to increase the reliability of the quality assessment. An example is provided in table 4.

Table 4

Example of additional objective criteria for the Joanna Briggs Institute toolkit

After discussing the additional criteria and piloting the tools, slight modifications were made for the JBI tools, and these final versions were used to assess the quality of the papers. The modified checklists can be accessed on request from the first author.

Questions were answered with ‘Yes (V)’ if the study met the criteria according to descriptions provided in the final version of the JBI toolkit. ‘No/Unclear (X)’ was selected if the study did not address the question or if information to assess the given criteria was lacking. The score concerning the quality of the study was determined by summing up all ‘Yes’ answers and dividing this number by the total number of answered questions, which differed by study design in the JBI tools. Questions that were answered with ‘Not applicable (N/A)’ were excluded from the calculation. As the JBI toolkit has no standard scoring index, the following scoring system was applied: (1) low quality=0%–50%; (2) moderate quality=above 50% and below 75%; (3) high quality=75% or higher. Although a four-band scoring system—where each category would include a 25% scoring range—was considered, a three-band scoring system was selected because the three given categories—low, moderate and high—would simplify the interpretation concerning the quality of the study. In a four-band system, the distinction and classification of the two middle categories are less straightforward compared with the three-band scoring system. Further, the first two categories in a four-band scoring system would still represent a poor-quality study, and, hence, should be used to signal more cautious interpretation of the study results among readers. The cut-off score was based on the idea that if a study could answer yes to only half or less of the questions, it would not be sufficient to transmit a reliable message to the audience. Therefore, at least more than half of the questions should be answered with yes to obtain a moderate score. The 75% cut-off was still based on the idea of having four equal scoring categories, where 75% and above would be classified as a high-quality study and would inform the audience with a more credible message.

Data analysis

Data concerning the type of migrant, health dimension, health subdimension, research design, level of evidence and quality assessment score were imported into Microsoft Excel for Mac (V.16.28). Mean quality scores were calculated for the different variables by using Microsoft Excel, including the type of migrant, health dimension, health subdimension, research design and level of evidence. RStudio (V.1.0.136; Macintosh; Intel Mac OS X 10_15) was used to conduct χ2 tests and an MCA. An MCA is a descriptive technique that can be used to visually demonstrate relationships among the levels of several categorical variables—here, these include the type of migrant, main health dimension, quality of the study and research design—in a two-dimensional (2D) space. The MCA projects categories in a 2D space with axes defined by latent dimensions (and, therefore, it is not possible to label the axes), based on weighted Euclidean distances.25 The MCA allows categories with similar profiles to be grouped together, where a closer distance of categories within the same quadrant demonstrates a stronger relationship, whereas categories that are further apart and in opposite quadrants present weaker associations.26 In addition to the MCA, χ2 tests were conducted to assess whether categorical variables were independent (eg, not associated). It should be noted that a few studies included two BARHII dimensions, yet, the analysis only allowed one dimension to be included. Therefore, only the most prominent dimension, based on the amount of attention given to the specific dimension in the article, was selected and used for the analysis.

Results

The study selection process is presented in figure 3. After removing the duplicates, 1282 original records were identified. A total of 1136 papers were excluded after the title and abstract screening stage due to focusing on another population of interest, lacking focus on a BARHII health dimension, not being a peer-reviewed academic article, and including data before 1965. As a result, 146 articles were eligible for the full-text screening stage. Subsequently, full-text articles were retrieved from these 146 records, and eventually, 67 papers met the inclusion criteria and were included in this review.

Figure 3

Flow chart of the data selection process.

Characteristics of included papers

This section first demonstrates the findings of each BARHII dimension, followed by the results on the quality and level of evidence of the included studies. Lastly, existing relationships between the type of study design, study quality of the study, type of migrant and main health dimension are shown. Table 5 presents a descriptive summary of all included articles, including the study design and corresponding level of evidence, study period, type of migrant, sample population, main health dimension, health subdimension, quality assessment score and a short description of the study.

Table 5

Summary table of included articles

Health dimension and type of migrant

The literature was first assessed to understand the trends and topical coverage of research against the six dimensions of the BARHII public health framework. The first dimension, social inequities, was used to describe the population of interest and refers to the type of migrant (eg, foreign workers, asylum seekers and refugees, or unclassified migrants). The other five dimensions focused on elements that influence the health status of the population of interest, including institutional inequities, living conditions, risk behaviour, disease and injury and mortality and morbidity. These latter five categories are outlined below and include results on the types of migrants researched within these dimensions. Figures 4 and 5 present overviews of the number of studies disaggregated by health dimension and type of migrant, respectively.

Figure 4

Number of studies disaggregated by health dimension.

Figure 5

Number of studies disaggregated by type of migrant.

Institutional inequities

One paper addressed the institutional inequities dimension27 by exploring the inclusion of migrant workers into national universal health coverage (UHC) policies in five countries of the Association of Southeast Asian Nations: Indonesia, Philippines, Malaysia, Thailand and Singapore. The researchers stated that Malaysia has implemented a medical insurance policy for foreign labour by obligating documented migrant workers to be enrolled in private insurance schemes, as non-citizens have no access to UHC at public facilities.

Living conditions

Eleven papers were classified under the living conditions dimension, where most articles (n=9/11) addressed the service environment subdimension.8 28–35All of these papers studied the asylum seeker and refugee population, except for one article that focused on migrant workers.33 Half the studies used qualitative methods to explore barriers to healthcare utilisation and showed that language difficulties, discrimination, insufficient health literacy and cultural differences were common issues. One study focused on the social environment subdimension and showed that refugee children experienced discrimination by locals and other refugees of different ethnicities and national origins, such as stereotyping them as criminals.36 Santos et al37 assessed elements related to the work environment subdimension by investigating perceived environmental hazards among foreign workers, demonstrating that noise and dust were perceived as the greatest occupational health threats.

Risk behaviour

Ten studies researched the risk behaviour dimension, with most articles (n=8/10) conducted on general migrant populations without clear identification of which migrant categories were included in their study.38–45 Three of these articles focused on the sexual behaviour subdimension, exploring risk behaviour related to human papillomavirus (HPV). The studies showed that a significant number of migrant women have high HPV risk behaviour due to lack of understanding with respect to cervical cancer, the screening process, and poor knowledge concerning HPV vaccination.41 42 45 Two papers, classified within the poor nutrition subdimension, showed poor health outcomes among detained migrants due to nutrition deficiencies.39 44 The other articles among unclassified migrants included two studies on violence and abuse, exploring maternal filicide43 and neglecting children38; and one study on alcohol and other drugs, pertaining to inhalants’ usage.40 These three studies simply showed that migrants represent a certain proportion of the identified cases. Only the study on the use of inhalants presented more cases among migrants than locals. Two final studies included foreign workers and explored the hygiene and sanitation and hazard and safety awareness subdimensions.46 47 Kamaludin and How46 stated that migrant workers had significantly less knowledge regarding environmental health, such as air quality, natural hazards, sanitation and industrial hazards, compared with local workers. Woh et al47 investigated the level of hygiene among migrant food handlers and argued that personal hygiene and sanitation measures should be improved among this population.

Disease and injury

With a total of 46 studies, the disease and injury dimension presented the largest study field of interest related to the BARHII framework. Most articles (n=36/46) studied foreign workers,34 37 48–81 while only six and four articles included unclassified migrants82–87 and refugee populations,88–91 respectively. The majority (n=27/46) of the articles studied communicable diseases, where 18 of these studies focused on parasites,48 50 51 54–58 64 68 69 74 75 77 78 81 82 89 eight on bacteria,53 65 71 72 83 85–87 and two on viruses.61 87 Most of the studies were descriptive and presented that migrants, irrespective of the defined type, represented a significant share among the study populations. Non-communicable diseases were studied far less compared with communicable diseases and were only specifically addressed in three articles.59 67 88 Scheutz et al88 found high numbers of different non-communicable oral complications among Vietnamese refugees, such as tooth decay and missing teeth. Kugan et al67 compared the difference in characteristics between foreign workers and Malaysian patients with perforated peptic ulcers, showing that the treated foreign labour population were younger, experienced fewer postoperative complications, and had smaller-sized ulcers compared with locals. Murty59 reported a case study, presenting a deceased migrant worker due to a cystic tumour in the heart region.

In addition to the studies that focused on single disease outcomes, two studies were conducted that presented distributions of various diseases among foreign workers, including communicable and non-communicable disorders.34 52 Five studies focused on the mental health subdimension, where these studies concentrated on describing psychiatric disorders,49 determining quality of life-related risk factors,84 90 and testing the effect of different coping mechanisms and therapy sessions on the level of stress.73 91 Nine studies explored the injury subdimension, where nearly all (n=8/9) studies focused on work-related injuries. Most of these studies examined the prevalence of particular injuries and traumas, including fatal lightning strikes,60 ocular traumas,70 and musculoskeletal pain.37 66 76 Ratnasinga et al63 compared the number of occupational incidents between local workers and migrant workers, where foreign workers had less accidents. In addition, two papers described risk factors for work-related injuries, such as high machine-related vibration exposure62 and low levels of the company’s safety commitment (as assessed by foreign workers themselves).79 Ya’acob et al80 conducted an RCT to evaluate the impact of a specific workplace intervention on musculoskeletal symptoms (MMS) among foreign labour and showed that the intervention reduced MMS in the foot and ankle regions significantly compared with the control group.

Mortality and morbidity

Two papers addressed the mortality and morbidity dimension by showing incidence rates among general cohorts of migrants. Zulkifli et al28 conducted a study on maternal and child health in Sabah and identified that infant mortality rates were significantly higher for migrants compared with locals. Dony et al92 also conducted a study in Sabah and showed that at least 24% of new tuberculosis cases detected since 1990 were among migrants and that leprosy incidence rates among migrants were on average 3.7 times higher than incidence rates among Malaysians.

Level of evidence and quality of the study

In total, 65 articles were included in the quality assessment; tables 6 and 7 show the mean quality scores of the papers disaggregated by BARHII dimension and level of evidence, respectively. Two articles—representing a scoping review27 and mixed-method design32—were excluded from this assessment, as the JBI toolkit does not accommodate these study designs. The quality assessment scores can be found in online supplemental file 3. In addition, figure 6 shows an overview of the number of studies disaggregated by research design.

Table 6

Number and average quality of included articles disaggregated by type of migrant and Bay Area Regional Health Inequities Initiative (BARHII) dimensions

Table 7

Number and average quality of included articles disaggregated by research design category

Figure 6

Number of studies disaggregated by research design.

In general, the quality of the evidence base on migrant health in Malaysia is low (49.2%) and consists mostly of level 3 evidence papers (n=27/65). Level 2 evidence represents 38.5% of the evidence base (n=25/65), followed by level 4 evidence papers (n=13/65). No level 1 evidence studies (systematic reviews or meta-analyses) were identified. The majority of the papers (n=41/65) focused on foreign workers; however, studies that included asylum seekers and refugees have the highest mean quality (58.4%). Furthermore, only four out of five BARHII health dimensions were included in the quality assessment. The living conditions dimension has the highest average score (59.7%), followed by the risk behaviour dimension (48.7%), mortality and morbidity dimension (47.9%), and the disease and injury dimension (46.3%). Moreover, the descriptive research category represents the majority (70.8%) of the evidence base with a mean quality of 47.7%. The qualitative research category has the highest mean quality and is the only research category with a high-quality score (76%).

Associations between different variables

Figure 7 presents the results of the MCA, showing different associations between four dimensions: (1) type of study design; (2) quality of the study; (3) type of migrant; and (4) main health dimension. χ2 test results were used to assess whether categorical variables were independent.

Figure 7

Results of the multiple-correspondence analysis.

High-quality studies tend to include refugees and asylum seekers (χ2=17.005, df=4, p=0.001928), focus on living conditions (χ2=131.94, df=6, p<0.001), and have a qualitative research design (χ2=656.35, df=12, p<0.001). Moreover, studies that included foreign workers tend to focus on diseases and injuries (χ2=374.52, df=6, p<0.001) and contain a case report study design (χ2=576.87, df=12, p<0.001). Furthermore, research that included the unclassified migrant population tend to study the risk behaviour, and mortality and morbidity dimensions (χ2=374.52, df=6, p<0.001). Lastly, prevalence studies, and, to a lesser extent, analytical cross-sectional studies, tend to have a low-quality score (χ2=656.35, df=12, p<0.001).

Discussion

Key findings

This study mapped the existing academic literature on migrant health in Malaysia and assessed the quality and level of evidence of these scientific studies. The majority of these studies focus on the ‘disease and injury’ dimension, especially infectious diseases, and includes mostly foreign workers. Two health dimensions (institutional inequality, and morbidity and mortality) as well as various subdimensions of each health dimension are lacking substantial research. In addition, only a few papers include the asylum seeker and refugee population, and a vast amount do not provide any details to classify the type of migrant. The average quality of the papers was low, yet quality differed significantly among the studies. High-quality studies were mostly qualitative designs that included refugees and focused on living conditions, while prevalence and analytical cross-sectional studies were mostly low quality. In terms of research trends, no specific changes in type of migrant, health dimension or quality of the study have been observed over the last six decades. However, it should be noted that qualitative research made its entry in the early 2010s and made up a vast amount of the papers published in recent years. Future research priorities based on the existing evidence and identified gaps are summarised in box 1.

Box 1

Main recommendations to improve future research on migrant health

Recommendation

  • Improve the description of the target migrant population by including information regarding the type of migrant (eg, foreign worker, refugee), visa status (eg, regular, irregular), country of origin, socioeconomic variables (eg, level of education, income), mode of transport during migration journey (eg, boat, car) and the existence of forced entry (eg, human trafficking, forced marriage).

  • Create associations between different stages of migration (predeparture, travel, destination, interception and return phase) and health outcomes.

  • More research output concerning governance and institutional inequities and mortality and morbidity, and, consequently, conduct a time series analysis between these two dimensions to identify and possible relationships.

  • More research output regarding non-communicable diseases, especially on the main causes of death in Malaysia; cardiovascular diseases, chronic respiratory diseases and diabetes.

  • More research output concerning several subdimension of risk behaviour, especially on smoking, physical inactivity, and alcohol abuse.

  • Evaluate the impact of health and non-health policies on migrant health.

  • Explore living conditions regarding the physical environment, such as housing and environmental conditions, and the impact on migrant health outcomes.

  • Promotion of guidelines on study conduct and reporting among researchers.

Among the five BARHII health dimensions, institutional inequities, and mortality and morbidity were the least represented. Yet, studies concerning the influence of governance on migrant health are of utmost importance, as overarching governance can affect health outcomes of the other BARHII dimensions.93 94 Similarly, epidemiological research on mortality and morbidity rates is necessary for population health statistics, to identify disease patterns, document changes over time and inform plans of action to tackle these health issues.95 Further research should focus on migrant health governance, as well as epidemiological research on morbidity and mortality among both migrants and non-migrants, to better understand the effects of policies on migrant health, which is particularly relevant in low-income and middle-income countries where the evidence gap is so acute.96 Furthermore, a recent systematic review on the effects of non-health-targeted policies on migrant health in high-income countries showed that non-health policies (eg, restrictive immigration policies) were associated with poor health outcomes.97 It is therefore important that policies in other sectors (potentially including, eg, immigration, labour, education) are also assessed for their potential consequences for migrant health.

Living conditions were represented in 11 studies and focused mainly (n=9/11) on the service environment by addressing the healthcare setting. However, there is scarce information on the social and economic environments that different categories of migrants must navigate and no data on the physical environment at all. Research conducted in other countries demonstrates the importance of these three subdimensions on migrant health.98–100 Shao et al98 argued that inequalities regarding the level of income (economic environment) influenced health outcomes among internal migrant workers in China. He and Wong99 stated that poor mental health among female migrant workers in China was related to gender-specific stressors (social environment). Al-Khatib et al100 demonstrated that poor housing conditions (physical environment) in a refugee camp were directly associated with various upper respiratory tract diseases. These studies underscore the importance of different environments on migrant health, motivating a focus of future research on the health impact of living conditions other than healthcare utilisation.

Ten studies were conducted on risk behaviour with different subdimensions, from hygiene and sanitation to violent and abusive behaviour. However, all of these subdimensions were under-researched, as only limited elements of each subdimension were discussed. For instance, three studies focused on sexual behaviour by addressing HPV knowledge.41 42 45 Yet, no attention was given to other sexual behaviour-related topics, such as condom use, HIV knowledge and birth control. Although these studies have been conducted in Malaysia, this research is lacking in the migration context.101–103 Therefore, future research should focus on broader aspects of each subdimension, as demonstrated in research elsewhere. For example, Renzaho and Burns104 addressed the poor nutrition subdimension by showing that dietary patterns among African migrants changed negatively after arriving in Australia due to the increased intake of fast food and processed food. Ganle et al105 concentrated on the sexual risk behaviour subdimension and stated that 71% of the sampled refugees in Ghana had transactional sex, and only 12% used contraceptives. Bosdriesz et al106 compared smoking between migrants and non-migrants in the USA and showed that migrants smoked less than US citizens. As a significant number of migrants in Malaysia come from Indonesia, a population that smokes almost twice as much as Malaysians, smoking behaviour among this migrant group may differ from locals.107 Therefore, future research should further explore the differences in other risk behaviours, such as smoking, between Malaysians and migrants in Malaysia.

Disease and injury was the most researched dimension, representing more than two-thirds of the evidence base on migrant health in Malaysia. Despite the strong representation, over half the research papers concentrated on communicable diseases, while only a few examined non-communicable diseases, consistent with global research output on international migrant workers.108 As the WHO109 states that approximately 74% of all deaths in Malaysia are attributable to non-communicable disease, in particular cardiovascular disease, chronic respiratory disease and diabetes, there is a need to expand research on non-communicable disease trends and outcomes among the migrant population in Malaysia.

We found that the majority of studies involved foreign workers (n=41/67), and only 10 studies examined asylum seekers and refugees as the primary population of interest. Our findings, therefore, offer useful synthesis on migrant worker’s health specifically, which is lacking relative to studies on asylum seekers and refugees in global migration health research.13 Furthermore, 11 studies did not specify the included migrant population. This issue could have occurred due to missing information on the type of migrant in the dataset that the researchers used for their studies. For example, the Ministry of Health (MOH) will not report anything more detailed than ‘non-Malaysian,’ as no further information on non-citizens are collected during patient registration at MOH facilities. Ideally, all research on migrants should clearly specify the type of migrants being studied and not omit crucial details, such as gender, visa status and country of origin. Also, human trafficking could significantly affect a person’s health and vulnerability, yet, there is very little known about the health issues experienced by trafficked persons in Malaysia.110 While the vulnerabilities experienced by trafficked persons intersect with other migration-related vulnerabilities like gender, ethnicity or documentation status, victims of human trafficking should be categorised separately, to reflect their own unique status and vulnerability. The travel routes or modes of transportation used by migrants to come to Malaysia may influence migrant health in different ways a well, as different routes or modes of transportation may be linked with specific hazards. Related to this issue is the lack of evidence on migrant health with specific stages of migration, including predeparture, travel, destination interception and return, where health outcomes might differ between these stages.111

Lastly, this scoping review revealed that the average quality of studies on migrant health in Malaysia is poor (49.2%) and that most of these studies have level 3 (n=27/65) or level 2 (n=25/65) evidence. Only qualitative studies with more rigour (level 2 evidence) and those that focus on living conditions and include the refugee and asylum seeker populations, tend to have a high-quality score. Therefore, there is a clear need to conduct research that will provide strong evidence to support practices and policies that will positively impact migrant health. Creating standard research design-specific guidelines, if not existing already, and, subsequently, promoting these materials among academics and research institutions, could increase the quality of future research work. Furthermore, researchers should follow study design specific reporting guidelines, to ensure that all relevant information is captured in publications for further evidence synthesis, such as this review.

Limitations

This study is the first systematic literature synthesis and scoping review on migrant health in Malaysia and presents a comprehensive overview of all identified peer-reviewed articles that met the inclusion criteria. Specific recommendations based on this research are provided to improve the evidence base on migrant health in Malaysia. Furthermore, we used a self-developed decision tree and modified JBI checklists to help identify the type of study design and corresponding level of evidence of the included studies. We found this evidence assessment framework to be useful for the quality assessment of migrant health-related studies, and it might be useful for other research fields as well. Yet, our review has several limitations. As this paper focuses exclusively on vulnerable migrants within the non-citizen population in Malaysia, we excluded other non-citizen groups, such as expatriates and international students, based on the assumption that these groups are less vulnerable (eg, expatriates in Malaysia have more privileges in terms of recognition regarding their roles in society, receive better financial compensation and tend to have access to many other benefits compared with foreign workers). However, we acknowledge that other non-citizen groups may face challenges in obtaining proper healthcare in Malaysia as well, such as issues related to cultural competency among foreign students and retirees.112 113 In addition, papers including non-citizens without further description were excluded, although these studies may have included the vulnerable migrant population.

Only academic peer-reviewed studies were included, thus excluding grey literature, editorials and opinion papers. Also, only English language articles were included, resulting in the exclusion of one identified paper in Bahasa Malaysia (the Malay language).114 Aggravating the issue, other Malaysian articles might not have been identified due to the lack of Malaysian keywords in the search strategy. As a result, much relevant information that could potentially be used to inform both policies and practice, as well as to make this review more comprehensive, may have been excluded from this review.

Inter-rater reliability was limited to a 20% sample of the records in the first (abstract and title) screening stage, and no data extraction nor quality assessment was verified by a second reviewer due to time and resource constraints. Yet, we anticipate low selection bias as the health dimensions in the BARHII framework present clear distinctions between each other, and most of the included papers used objective indictors. For example, when a paper was measuring the knowledge and awareness regarding the pap smear test among female migrants, it would be classified as a ‘risk behaviour’ study. Furthermore, we believe that the development of the decision tree and additional objective criteria for the JBI tools—an example was given earlier in table 4—reduced the subjectivity of this study, and, hence, increases the reliability. Yet, future research is needed to validate both the decision tree and modified JBI toolkit.

Besides the BARHII framework, various conceptual public health models are available, and these models may include different (sub)dimensions. For instance, the WHO Commission on Social Determinants of Health framework includes material circumstances, such as food availability, whereas this dimension is not included in the BARHII framework.115 Similarly, critical appraisal tools other than the JBI toolkit are available, which could address different points to determine the quality of the study. Therefore, it would be helpful to assess other public health models and critical appraisal tools to see if they include additional elements (eg, food availability) that would be beneficial for future studies.

Likewise, a decision tree was developed by using the characteristics of the used definitions of different research designs as well as the specific traits of Tomlin and Borgetto’s24 level of evidence model. Using other definitions and level of evidence models could result in a different level of evidence categorisation. However, we believe this review makes a strong methodological contribution by combining study designs and level of evidence in a unified decision tree, which can be used by researchers conducting systematic or scoping reviews where accurate classification of the study design and associated evidence level, is important.

In order to conduct the MCA, the dataset could only include one unit per dimension for each paper. As some studies included multiple BARHII dimensions, only the most prominent dimension was included in the analysis. As a result, the analysis may suffer from some selection bias and present slightly different outcomes compared with an analysis that includes the other BARHII dimensions.

Lastly, no adjustments were made for outliers in the quality assessment. Therefore, some papers with extremely high or low scores could have influenced specific dimensions and might not reflect the quality of those dimensions perfectly.

Conclusion

Migrant health remains an issue in Malaysia, yet, the quality of the evidence needed to inform policies is currently lacking. Research-specific reporting guidelines should be followed to improve the credibility and quality of the evidence base. Furthermore, future research should focus more on evidence gaps in the mortality and morbidity, and institutional inequities dimensions, and certain subdimensions, such as non-communicable diseases, housing conditions and physical inactivity, to provide a comprehensive picture of migrant health in Malaysia. Apart from demonstrating the research gaps, this paper also makes methodological contributions to migrant health research by providing a modified JBI toolkit and a decision tree that identifies the type of study design and corresponding level of evidence, both of which can be used in other research fields as well.

Acknowledgments

The authors would like to thank Andrew Seko for granting them permission on behalf of the Bay Area Regional Health Inequities Initiative (BARHII) for the use of their public health framework in this scoping review. They would also like to thank Norhayati Binti Yahya for providing them with the definition for non-citizens on behalf of the Department of Statistics Malaysia (DOSM). Furthermore, they would like to thank the editor and four independent reviewers for their helpful feedback. Lastly, they would like to thank the United Nations University – International Institute for Global Health for its continuous support throughout this project.

References

Supplementary materials

Footnotes

  • Contributors AWdS, NSP, MV and ZXC created the study protocol. AWdS and ZXC conducted abstract and full-text screening and NSP solved undisputed conflicts. AWdS extracted the data, drafted the decision tree and modified the JBI tools, and NSP, MV, ZXC and CAL provided feedback during these processes. CAL and AWdS conducted the data analysis. AWdS drafted the initial version of the manuscript and AWdS, NSP, MV, CAL, TL and ZXC critically revised and approved the final version.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests All authors have completed the Unified Competing Interest form (available on request from the corresponding author) and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information. The prereview protocol and modified JBI checklists can be accessed on request from the first author.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.