Blood donation barriers and facilitators of Sub‐Saharan African migrants and minorities in Western high‐income countries: a systematic review of the literature

SUMMARY Objectives The present study aimed to gain more insight into, and summarise, blood donation determinants among migrants or minorities of Sub‐Saharan heritage by systematically reviewing the current literature. Background Sub‐Saharan Africans are under‐represented in the blood donor population in Western high‐income countries. This causes a lack of specific blood types for transfusions and prevention of alloimmunisation among Sub‐Saharan African patients. Methods/materials Medline, EMBASE, PsycINFO and BIOSIS were searched for relevant empirical studies that focused on barriers and facilitators of blood donation among Sub‐Saharan Africans in Western countries until 22 June 2017. Of the 679 articles screened by title and abstract, 152 were subsequently screened by full text. Paired reviewers independently assessed the studies based on predefined eligibility and quality criteria. Results Of the 31 included studies, 24 used quantitative and 7 used qualitative research methods. Target cohorts varied from Black African Americans and refugees from Sub‐Sahara Africa to specific Sub‐Saharan migrant groups such as Comorians or Ethiopians. Main recurring barriers for Sub‐Saharan Africans were haemoglobin deferral, fear of needles and pain, social exclusion, lack of awareness, negative attitudes and accessibility problems. Important recurring facilitators for Sub‐Saharan Africans were altruism, free health checks and specific recruitment and awareness‐raising campaigns. Conclusion The findings of this review can be used as a starting point to develop recruitment and retention strategies for Sub‐Saharan African persons. Further research is needed to gain more insight in the role of these determinants in specific contexts as socioeconomic features, personal histories and host country regulations may differ per country.

In many Western countries, minority populations (such as immigrants and refugees but also individuals with total or partial ancestry from non-White racial groups) are under-represented in the blood donor population (Murphy et al., 2009;Rastogi et al., 2011). Certain specific blood types are more common in certain ethnic groups than others, especially among those of Sub-Sahara African (SSA) background (Reid et al., 2002). For instance, the Duffy negative phenotype (Fy(a-b-)) is frequently found in the Sub-Saharan region of Africa but is rarely present among individuals in countries consisting largely of White European-origin people (Howes et al., 2011).This discrepancy in blood types poses a problem because, if donor blood and patient blood do not match well, serious complications can occur (Yazdanbakhsh et al., 2012), such as haemolytic transfusion reactions caused by the development of antibodies in response to antigens in donor blood (Miller et al., 2013). Patients in need of repeated blood transfusions are especially at a high risk of alloimmunisation. One example is sickle cell disease (SCD), a relatively common inheritable blood disorder among SSA individuals (Rees et al., 2010). Many patients with SCD who receive red blood cells produce antibodies and are thus alloimmunised (Miller et al., 2013;Alkindi et al., 2017). An adequate supply of well-matched, antigen-negative red blood cells is needed to improve the blood supply and to enable helping patients with an SSA background. This makes SSA individuals an important target group for blood donation agencies (van Dongen et al., 2016).
Unfortunately, blood agencies all over the world have problems recruiting SSA blood donors (Grassineau et al., 2007;Shaz & Hillyer, 2010a). In part, this is attributable to existing regulations in some countries, such as the exclusion of individuals with language barriers and SCD and Thalassemia carriers (van Dongen et al., 2016). On the other hand, attempts to recruit healthier SSA donors have fallen short, or some recruitment programmes seem to appeal to the majority population only (Frye et al., 2014;Muthivhi et al., 2015). To optimise recruitment and retention strategies, more insight is needed on what prevents and motivates people of an SSA background to donate blood.
Recent systematic reviews of the literature have focused on SSAs in their birth countries rather than on those living as ethnic minorities or migrants in Western countries (Tagny et al., 2010;Burzynski et al., 2016). According to the qualitative syntheses in these systematic reviews, health-and knowledge-related barriers are commonly cited by SSAs. More specifically, there is a fear of being exposed to various infectious diseases (Burzynski et al., 2016), but there is also a high prevalence of transmissible infections among blood donors, which impacts blood safety (Tagny et al., 2010). Replacement/family donations are also predominant in SSA countries instead of voluntary non-remunerated donations. Due to the different blood donation and supply systems between SSA countries and Western countries, the barriers and facilitators experienced may differ. Earlier studies regarding barriers and motivators of SSAs in non-African countries were summarised but have not been systematically reviewed before (Shaz et al., 2008;Shaz & Hillyer, 2010a). In addition, these summaries focused only on African Americans (AAs) in the United States but not on other countries where their blood is needed, such as Australia or European countries.
A better understanding on what prevents and motivates potential SSA blood donors in different Western countries to donate blood would allow the development of more effective recruitment and retention strategies. The present study aimed to gain insight into the barriers/facilitators of blood donation among SSAs in high-income countries where the majority were White or Caucasian and into differences between SSA and White individuals by systematically searching and analysing the current literature.

Search strategy
Medline, EMBASE and PsycINFO were systematically searched for articles or abstracts published from inception until the 22nd of June 2017. BIOSIS was searched until the 19th of October, 2015, due to the discontinued licence of the database. The search resulted in a total of 4672 articles (Medline, N = 776;EMBASE, N = 1853;PsycINFO, N = 1596;BIOSIS, N = 447). No additional relevant articles were identified through manual searching of other sources (n = 0). After removing duplicates, 3859 articles were screened on initial relevance based on the title, and the resulting 679 articles were screened by title and abstract. Of the resulting 152 articles screened by full text, 121 were excluded based on the eligibility criteria, thus leaving 31 articles for the present quality assessment ( Fig. 1) (Moher et al., 2015).
An initial scoping of the literature led to the identification of three relevant search concepts: [blood donation] AND [race, minorities and ethnicity] AND [factors -barriers & facilitators]. For each concept, relevant (controlled) terms were employed. Animal studies were excluded. Appendix A presents details for each database.

Eligibility criteria
We included studies if they explicitly focused on possible barriers and facilitators that may influence blood donation behaviour and intention among adults (about 18-65 years) of SSA origin or background living in a high-income country with a White European or Caucasian majority. The possible barriers and facilitators could be either experienced or self-reported and could refer to factors either negatively or positively associated with blood donation behaviour, blood donor status or intention to donate or become a blood donor. Both descriptive studies on SSA minorities or migrants only and comparative studies with White or other subgroups were included.
SSAs were defined as individuals who originated from countries lying south of the Sahara Desert in Africa. In American studies, those of African ancestry are commonly referred to as Blacks or AAs. Although the precise definition of these labels is unclear, most AAs came to the United States during the Colonial era. We decided to include these latter studies as the terms are commonly used for persons who originate from West or Central Africa and are, thus, carriers of blood types not common in the White European or Caucasian population and are an important target population for blood donor recruitment and retention (Reiner et al., 2011).
Only empirical studies were included: quantitative questionnaire or database results and qualitative interview or focus group results. We excluded case reports, reviews and viewpoints. Studies in countries where whole blood donors are remunerated in cash for their whole blood donations are excluded, as well as studies that are solely on other types of donation (e.g. organs, platelets).

Quality assessment
We created two quality criteria lists for quality assessment of the quantitative and qualitative studies (Appendices B and C). They included items from different quality assessment tools, thus creating comprehensive lists to assess the risk of bias in the varying designs of the studies. The Critical Appraisal Skills Programme (CASP) (Singh, 2013), the STROBE statement (Von Elm et al., 2007), the QualSyst tool (Kmet et al., 2004)  quality criteria for the quantitative studies. The CASP (Singh, 2013), the QualSyst tool (Kmet et al., 2004), the Consolidated Criteria for Reporting Qualitative Studies (COREQ) (Tong et al., 2007), the modified quality checklist used by Mills et al. (2005) and the Cochrane risk of bias tool (Offringa et al., 2003) provided quality criteria for the qualitative studies. For each quality criteria list, two authors scored each article and compared each other's assessment and resolved differences. All items were weighed equally for the overall quality score. Similar methods and score systems were used in previous systematic reviews of the literature (Hoogerwerf et al., 2015;Piersma et al., 2017).

Characteristics of the included studies
The characteristics of the quantitative studies are presented in Table 1 and the characteristics of the qualitative studies in Table 2. All included studies were published between 2002 and 2016. Most were conducted in the United States (n = 21), followed by Australia (n = 5) and Canada (n = 2). The remaining three studies were conducted in Israel (n = 1), the UK (n = 1) and France (n = 1). All Australian studies, as well as the two Canadian studies, were conducted by the same research group in each country with recurring authors. The Australian quantitative studies used the same data (425 migrants and refugees from Africa), as did the Australian qualitative studies (88 migrants and refugees from Africa). In the United States, 16 of the 21 studies were conducted by recurring (groups of) authors. Both the studies by Boulware et al. used the same data (385 individuals from households in Maryland, USA) (Boulware et al., 2002a,b).

Quality descriptives and issues
Tables 3 and 4 present an overview of the quality criteria and the scores for the quantitative studies and the qualitative studies, respectively. A total score of 100% means that the study meets all criteria, whereas a score of 0% would mean that the study meets USA Evaluate whether mistrust for the healthcare system among African Americans affects attitudes towards blood donation. Self-administered questionnaire 930 individuals from African American religious institutions in Atlanta.
Barriers → Rarely think about it, afraid to give blood, afraid of needles, pain or discomfort, afraid of feeling faint, dizzy, or unwell and mistrust in hospitals. Facilitators → Help save a life, it is the right thing to do, help the community and because blood is needed.
USA Identify barriers and factors that can be effectively addressed by blood centres.
Self-administered survey in 6 American blood centres 4142 lapsed whole blood donors. none of the criteria. Almost all quantitative studies addressed a clearly focused issue and described specific objectives, and all qualitative studies provided a clear aim of the study. However, we encountered many methodological issues for both the quantitative and qualitative studies. For the quantitative studies, the study sample was often not representative of a defined population, or it was not sufficiently explained why this particular sample was chosen or necessary to study. In addition, the response rate and characteristics of the study sample were often not mentioned, and many studies did not control for possible confounders, which are partly due to the descriptive, rather that analytical, approach of many studies. Regarding the methodological issues of the qualitative studies, the role of the researcher was only discussed in two of the seven studies. The researchers' own ethnic and cultural background may be a potential bias, especially in studies on ethnic communities. Besides, the locations of the interview/focus groups were often not described, and for almost half of the studies, it remained unknown whether the researchers had taken ethical issues into consideration.

Lack of knowledge and awareness
McQuilten et al. (2014) found African migrants and refugees with moderate blood donation knowledge to have an almost 4·5 times higher odds on having donated previously compared to those with poor knowledge (adjusted odds ratio, AOR [95% confidence interval, CI] = 4·46 [1·57-12·67]; P < 0·01). For those with a high level of knowledge, the odds were more than 10 times higher compared with those who had poor knowledge [AOR (95% CI) = 11·30 (3·79-33·70); P < 0·001]. In addition, Polonsky et al. (2013) found that adding knowledge to the original Theory of Planned Behaviour (TPB) model increased the model fit for SSAs. The TPB is a commonly used theory in blood donor studies, whereas attitudes, social norms and self-efficacy predict the intention and behaviour to donate blood (Ajzen, 1991;Lemmens et al., 2005). However, James et al. (2011) found AAs to have a fairly good knowledge of blood donation and that there were no differences in the scores between AA donors and AA non-donors. In addition,  found marginalisation to be negatively related to blood donation knowledge, but there was no evidence that marginalisation was related to actual blood donation. Concerning the lack of awareness, for both AA donors and AA non-donors, not knowing that donating blood is important (23·1% donors; 21·8% non-donors) and not knowing where to donate (23·9% both donors and non-donors) were important self-reported barriers (Shaz et al., 2010b). There was evidence that AAs from the general population in Atlanta, Georgia, more often did not know where to donate compared with White individuals (AA 31%, White 19%) (James et al., 2013). In the qualitative study by Polonsky et al. (2011b), respondents from Australian-based African communities reported that they had never discussed blood donation or had never been approached about blood donation before their research.
Barriers → Perceived discrimination and social exclusion. Facilitators → increased awareness about sickle cell anaemia and the importance of their contribution. Schreiber et al. (2006) found AA first-time donors being more likely to report poor staff skills (P < 0·01) and experiencing bad treatment (P < 0·01) compared with White first-time donors. The African migrant respondents in Australia in Polonsky et al. (2011a) also stated, in interviews, that they experienced poorer treatment and longer waiting times compared with other patients. Accordingly, Ethiopians, compared with native Israelis, had a more negative behavioural attitude towards blood donation [t(124) = 4·0, P < 0·01] (Merav & Lena, 2011).
Regarding mistrusting the blood supply or donation agencies, Steele et al. (2012) found that AAs had more concerns about the safety of blood donation than White individuals, e.g., that not all blood donations were tested for AIDS (acquired immunodeficiency syndrome) [OR (95% CI) = 0·7 (0·6-0·8); P < 0·001] and that they could get AIDS from donating blood (43·1% AAs; 15·9% White; P < 0·001). AAs were more distrustful towards shortage claims and were more likely to believe that their blood was not wanted and would not be used (Mathew et al., 2007;Merav & Lena, 2011;Tran et al., 2013). In contrast, James et al. (2013) found that only 6% of the AAs reported mistrust for blood centres as a barrier.
Even experiences of discrimination outside the blood donation setting had a negative impact on AAs' views towards blood donation (Polonsky et al., 2011b). Discrimination was also experienced in healthcare settings where SSAs felt that they were treated worse than others by medical staff (Polonsky et al., 2011a).

Fear
Different types of fear were distinguished in the included studies. For instance, AA first-time donors were significantly more afraid of needles (P < 0·05) and were more afraid that donating is painful (P < 0·01) compared with White first-time donors . The overall prevalence of needle fear ranged from 14 to 38% (Shaz et al., 2009a,b;James et al., 2013). Another type of fear identified in the studies was for fainting. James et al. (2013) found White individuals to have a higher prevalence of fear of fainting than AAs (AA 18%, White 29%). Still, fear of fainting is a major barrier for AAs, with a prevalence of 34% among AA non-donors (Shaz et al., 2010b). Fear of hospitals was also found to be a donation barrier. Those afraid of hospitals had 70% lower odds of prior blood donation compared with those who were not [OR (95% CI) = 0·3 (0·1-0·9)] (Boulware et al., 2002b). Lastly, fear of contracting a disease was mentioned by 12% of the AA respondents in the study of Grossman et al. (2005) and 22% of the AA respondents in the study of Shaz et al. (2010b) but was also commonly mentioned among other ethnic groups (Mathew et al., 2007).

Deferral and exclusion factors
SSAs had the highest chance of haemoglobin (Hb) deferral compared with other ethnic groups (Cable et al., 2011;Custer et al., 2012). While 1·6% of the White men and 16·6% of the White women were deferred for low Hb on their donation attempt, for SSA donors, these rates were 2·4 and 29·2%, respectively (Mast et al., 2010). James et al. (2012) found the Hb deferral rate for White persons to be 3·6%, compared with 12% for AA donors. Other commonly reported deferral or exclusion factors for donating blood for SSA donors were: difficulty to find or palpate the veins, high blood pressure or pulse deferral, hepatitis C infections, hepatitis B infections, minor infections (e.g., a cold), tattoos, institutionalisation, pregnancy, cancer, syphilis, malaria, diabetes and cardiovascular problems Grassineau et al., 2007;Shaz et al., 2010c;Custer et al., 2012;James et al., 2012). These factors cause SSAs to be more often temporarily or permanently deferred for blood donation.

Inconvenience
Six studies found evidence inconvenience to be an important barrier to donate among SSAs. Although most studies focused on an inconvenient location of the donation centre only (n = 5) (Grossman et al., 2005;Schreiber et al., 2006;Mathew et al., 2007;James et al., 2013James et al., , 2014, one study also took inconvenient opening times into account (Shaz et al., 2010b). From focus group interviews, Mathew et al. (2007) found that most individuals felt the opportunities to donate to be limited and that blood centres were not easily accessible. Grossman et al. (2005) also found inconvenience to be a common barrier among AA women (19%). AA repeat donors reported inconvenience more frequently (31·4%) compared with White repeat donors (26·3%) . Shaz et al. (2010b) found a high prevalence of inconvenience as a barrier, which was 47% for AA current donors and 87% for AA non-donors. James et al. (2014) found that minority communities lacked mobile sites and that these people were thus less likely to donate within their own living area.

Synthesis of results
This systematic review indicates that most specific barriers for blood donation in African minority and migrant groups in White/Western majority high-income countries are: fear of needles, social exclusion, Hb deferral, not being aware of the need, having a negative attitude towards the blood bank policy or organisation and not having a convenient place to donate blood. Fear and a lack of awareness about blood donation are also important and commonly reported barriers for White individuals. White individuals in the included studies also frequently experience Hb deferral and no convenient place to donate blood as important barriers, but there is evidence that these barriers have a bigger impact on SSAs and AAs. For instance, the overall Hb is lower for individuals with an African background (Cable et al., 2011), and blood drives more often visit places with a relatively low proportion of African individuals (James et al., 2014). Lastly, the (perceived) experiences of social exclusion and discrimination are factors that have a large impact on SSA minority groups' intention to donate blood .
Among the possible facilitators to donate blood in the included studies, we found altruism, health checks and community involvement and campaigns to present promising factors to target in order to facilitate blood donation among SSAs. Altruism was also an important facilitator for White individuals in these studies. There is evidence that SSAs would be more motivated by campaigns focused specifically on (the needs of) their ethnic group and by creating awareness inside their communities.
The barriers and facilitators we found in this review do partly resemble findings from the systematic review by Burzynski et al. (2016), which focused on SSAs living in their countries of birth. They too found a lack of knowledge to be a main barrier and helping others to be a main motivating factor. However, although they found health concerns to be an important barrier, in the studies reviewed here, this barrier was not as prevalent. Likewise, although we did find some evidence of SSAs being more concerned with the safety aspect of donating blood, we did not find evidence that a large proportion of the SSAs in Western countries is concerned with a shortage of blood after donating or with adverse health effects to themselves.

Limitations
While most studies reported similar results, some factors yielded mixed results, making the results we found less certain. For instance, the prevalence of medical mistrust differed considerably between the studies: ranging from 14% for AA donors according to James et al. (2011) to 72% for AA men according to Boulware et al. (2002a). Large differences between studies in percentages for barriers/facilitators were also found for fear, inconvenience and incentives. We speculate that these differences could be attributable to differences in measurements, sample size, sample characteristics of study populations (e.g. students, immigrants, refugees and church members), varying healthcare systems or cultural differences between countries. Most studies originate from the United States and Australia, where the economic and social differences between their racial/ethnic groups are different compared with European countries (OECD, 2015). It remains unknown whether the barriers/facilitators AAs experienced in the United States also apply to SSAs in different continents, especially for the European context. AAs are often descendants of African slaves during the Colonial era and are thus born and raised in the United States, whereas SSAs in European countries are often first-or second-generation immigrants. Arguably, these groups may have different barriers and motivators for donating blood, which we were not able to distinguish, partly due to an under-representation of studies conducted in Europe. Moreover, some statistically significant differences between SSAs/AAs and White individuals are relatively small in effect sizes or proportions (e.g. for a negative attitude or being motivated by incentives). Therefore, we argue that adjusting recruitment or retention strategies in SSAs regarding these factors -wherever they live -has limited added value.
In addition, only a few quantitative studies used advanced statistical methods, whereas other studies limited themselves to descriptive analyses only. Creating a funnel plot or discussing different effect sizes was deemed impossible because the studies used various research designs. For a more coherent review, it would have been practical to limit the focus to a specific type of design. However, because the main goal of the present study was to explore the barriers/facilitators that are currently studied, we decided to include descriptive studies as well.

Implications for practice and research
We would encourage the development of strategies, in collaboration with African communities, to create more awareness of the need of blood (especially for SCD patients and other patients requiring repeated transfusions, such as patients with haematopoietic disorders). There is evidence that interventions developed for and together with the community are more effective, and this may improve trust in the blood bank organisations (van Dongen et al., 2016). Strategies to reduce barriers for blood donation in this group should focus on investigations on Hb deferral, such as examining possibilities for implementing different reference standards that are still safe for the donor but may reduce deferral rates (Beutler & West, 2005). Finally, the blood bank organisations should contribute to a comfortable environment for SSAs, e.g. by reassuring the blood donors, but also demonstrating what happens with the blood once it is donated. This may contribute to less experienced fear and less mistrust towards the blood bank organisations or their staff.
More research is needed to gain a deepened insight into underlying mechanisms of blood donation among SSAs/AAs. For instance, it would be valuable to more extensively study how specific barriers and facilitators for blood donation actually influence blood donation intention and behaviour. This approach may enable more careful and context-specific intervention development to increase the chances of implementing more effective recruitment methods. We particularly encourage studies in European countries as most studies are performed in the United States, whereas there is an under-representation of SSAs in the European blood donor population as well. Although we managed to distinguish important determinants that seem to play a role for Sub-Saharan minorities in Western high-income countries, especially the United States and Australia due to the larger amount of studies performed there, the social and personal contexts vary between countries, which may relate to more specific determinants. Future quantitative studies should carefully report the methodology and use statistical hypothesis testing for better generalisability and comparison of results between studies. Measuring the relation between the barriers/facilitators and the donor intention/behaviour would provide more evidence of what kind of interventions may work instead of giving a descriptive overview of the most reported determinants only. In addition, as most results are based on self-reported barriers and motivators, it may be interesting to look more into the underlying mechanisms of these determinants. For instance, as fear is often reported as an important barrier among SSAs, it would be valuable to monitor whether there are actual differences in levels of stress or anxiety between SSAs and Whites before and after initiating blood donation or seeing a needle. A general overview of possible future research questions based on this systematic literature review can be found in Table 5.