Strengths and limitations
The main strength of the RODAM study is the use of well-standardised approaches across the various study sites. Another unique strength of this study is the homogenous study population of Ghanaians living in different settings in Africa and Europe. So far, only a few studies have attempted to assess the potential role of migration on obesity and type 2 diabetes among African populations by comparing native Africans with people of African ancestry living in the Caribbean, UK and USA [
11,
12]. However, these studies were limited due to the heterogeneous ancestry of populations who were transported out of Africa several centuries ago. This factor, as well as genetic admixture primarily with European ancestry population groups, make it difficult to assess the potential role of migration and its impact on health in African populations [
20]. Furthermore, these studies were based on secondary data with different measurement protocols. The RODAM study overcomes these previous limitations by focusing on one population using the same measurement procedures in all sites.
Our study also has limitations. First, as in most epidemiological studies, type 2 diabetes was defined by a single blood glucose measurement, which traditionally would have to be confirmed. Second, although the same methods were applied in all sites, the recruitment strategies had to be adapted to suit the local circumstances due to differences in registration systems. Ghanaian participants in Amsterdam, for example, were drawn from the Amsterdam Municipal Population register, whereas London participants were drawn mainly from Ghanaian organisations lists. It is possible that individuals who were not on the lists of these organisations differ in terms of sociodemographics, which might somewhat affect the representativeness of Ghanaian migrants in London and Berlin. In a non-response analysis, men more often were non-respondents than women in all sites except for Berlin. Non-respondents were younger than respondents in all sites. Further, the non-response analysis in Berlin revealed that the distribution of respondents and non-respondents across Berlin city districts was fairly similar. Additionally, evidence suggests that most Ghanaians in Europe are affiliated with Ghanaian organisations [
13,
14], suggesting that members within these organisations may be representative of the Ghanaian population living in various European countries. Therefore, although a certain level of bias is likely, as in all population-based surveys, we consider it unlikely that the differences in prevalence rates between European sites are substantially biased by the variations in sampling strategy. Finally, only fasting plasma glucose was used to diagnose diabetes, which may underestimate the prevalence of diabetes. Evidence suggests that the 2-h plasma glucose value after a 75-g oral glucose tolerance test diagnoses more people with diabetes fasting plasma glucose.
Discussion of the key findings
Our current findings show that obesity is extremely common among women at all study sites, including a notable prevalence in rural settings. The prevalence rates in men were less than half of those among women. Despite the higher burden among urban populations, overweight/obesity is rapidly increasing also in rural communities in low- and middle-income countries, especially among women, as our study clearly shows. Therefore, the notion that overweight/obesity is affecting typically the urban populations can no longer be substantiated [
7,
8]. In fact, over a third of women in rural Ghana were either overweight or obese. This corroborates recent findings in rural South African youth [
21]. Rapid urbanisation and improved contact between rural and urban settings due to infrastructure improvements may be facilitating the transfer and introduction of urban practices to rural settings with consequent changes in diet, resulting in consumption of energy-dense traditional or processed foods as seen in urban Ghana and some settings in SSA [
22,
23]. Of note, the present study shows that the obesity rate among women in urban Ghana is nearly as high as those reported among women in the USA [
24], and far higher than the prevalence rates reported among women in many European high-income countries [
25‐
27]. We show that Ghanaian migrants in Europe are particularly affected by obesity, the rate being up to 15 times higher than among their rural counterparts in Africa. Among migrant Ghanaian women, the obesity rate greatly exceeds the figures of the host European populations in all three European countries. In the 2013 Health Survey for England, the prevalence of obesity among English general population women was 24 % [
25] compared with 54 % observed in the present study among Ghanaian migrant women in London. Similarly, the prevalence of obesity among Dutch women is 13 % [
26] compared with 49 % in Ghanaian migrant women in Amsterdam, and 24 % in German women [
27] compared with 39 % among Ghanaian migrant women in Berlin.
Worryingly, type 2 diabetes occurred at a similar prevalence among individuals in urban Ghana and in Europe. Previous studies among SSA populations found a rising gradient of type 2 diabetes from SSA through the Caribbean to the UK and USA [
11,
12]. Mbanya et al. [
12], for example, reported an age-standardised prevalence of diabetes of 1 % among urban Cameroonian men compared with 15 % in African Caribbeans in the UK. This gradient was due to extremely low prevalence of type 2 diabetes in SSA, which has been documented from the earliest studies that were conducted more than five decades ago. For example, in a 1958 study, Dodu et al. [
28] observed a diabetes prevalence of 0.4 % in an urban population in Accra, Ghana. Likewise, a community-based study in the Volta region of Ghana in 1964 found a diabetes prevalence of 0.2 % [
29]. In contrast, the results of the present study suggest that the gradient between urban Africans and diaspora African living in high-income European countries is fading rapidly. In fact, the prevalence of type 2 diabetes among women was marginally higher in urban Ghana (9.2 %) than in London (8.4 %). Thus, the increasing risk of type 2 diabetes is no longer an issue of only migrant populations, but appears to have reached urban communities in SSA. This implies increased risks for rural African communities, especially given the rapid changing lifestyles in these settings. The rise of obesity and type 2 diabetes among SSA populations can be partly attributed to modernisation with consequent adoption of unhealthy aspects of globalised lifestyles such as physical inactivity and poor dietary behaviour [
22]. The key specific drivers within these broad categories, however, still need to be identified. Interestingly, the prevalence of type 2 diabetes was higher in men than in women despite the higher levels of obesity in women in all sites except rural Ghana. The explanations for these differences are unclear, but may be partly due to a more favourable body fat distribution in women [
30]. Alternatively, it is possible that body weight has a larger impact on type 2 diabetes risk among men than among women, as is suggested by the current study.
Another important finding from this study is the high prevalence of IFG in all sites. The IFG rates in both rural (12 %) and urban (13 %) Ghana are far higher than those in most urban populations in Africa [
31,
32]. In a community-based study conducted more than a decade ago in urban Accra, the IFG prevalence was 6.2 % [
33], indicating a nearly 110 % percentage increase in IFG in urban Ghana in a decade. In the present study, IFG was exceptionally common in Amsterdam Ghanaian migrants, which is consistent with our earlier findings [
10,
34,
35]. In a previous study, the prevalence of IFG was 35 % and 14 % among African Caribbeans in the Netherlands and in England, respectively [
34]. The high rate of IFG is worrying given the increased risk of developing type 2 diabetes and related complications [
36]. The reasons for the abundance of IFG among Ghanaians in the Netherlands is unclear but might be due to contextual factors such as differences in treatment of diabetes and/or unknown aetiological factors; this requires further study.
Despite varying prevalence rates among the host populations in the three European countries, the respective differences among migrants residing in these countries were rather small. Still, although higher, the type 2 diabetes prevalence among the migrant populations mimics their respective host European populations. Recent International Diabetes Federation age-standardised estimates indicate a prevalence of type 2 diabetes of 4.7 % in the UK, 5.5 % in the Netherlands, and 7.4 % in Germany [
1]. Despite the lower prevalence of type 2 diabetes in the UK [
25], obesity is more common in the UK than in most European countries [
26,
27]. Interestingly, a similar pattern was observed among our study populations with Ghanaians in London having a lower prevalence of type 2 diabetes but a higher prevalence of obesity compared with Amsterdam and Berlin. This observation seems to suggest that the national contextual factors, such as prevailing health behaviour, health-related policies and access to preventive services, may influence metabolic risk factors in different ways in various countries [
6,
37].
Our findings have important public health implications for health planners in Europe and Africa. The prevalence rates of obesity and type 2 diabetes among African migrants exceed those of the European host populations. Ghana is a lower middle-income country with a substantial burden of communicable diseases. The high levels of overweight and type 2 diabetes will undoubtedly put more pressure on the already overburdened health system suggesting an urgent need for action with strong support by government and civil societies in Ghana. This requires a health policy shift towards prevention and control of obesity and diabetes and other non-communicable diseases [
38].