All facilities included in the study were found to serve a diverse range of patients including long-term residents, recent arrivals, internal and cross-border migrants. Women account for the majority of public healthcare users surveyed, illustrating the known gendered dimension of healthcare seeking—and associated gendered burden of healthcare-associated challenges—documented elsewhere [23–25]. Recent arrivals were significantly younger than longer-term residents, in line with existing literature—including in South Africa [21]. The demographics of our sample population—generated through a convenience sampling approach and chosen to reflect areas of relatively high as well as relatively low levels of cross-border migration—are also in line with South African migrant profiles that have been documented elsewhere, although not in healthcare settings [for example, 3]. Documentation status amongst our sample was associated with length of stay. Compared to recent migrants, new arrivals and long-term residents were most likely to be documented. This suggests that, on arrival, non-nationals have documents but that these expire and non-nationals ‘overstay’ the duration of their visa, and it takes several years for individuals to regain a documented status—often through a temporary residence permit.
The majority of public healthcare users we surveyed were born in a province (or country) different to where they were interviewed, indicating the prevalence of a migration history within the study sample. There were significant differences between clinic location and classification of migrant groups: a higher proportion of cross-border migrants were found at urban clinics; internal South African migrants at peri-urban clinics; and, non-migrants (born in the province of interview) at the rural facility. Regardless of migration status or length of stay, the majority of respondents reported using a clinic due to its proximity, followed by a smaller number reporting that their choice of clinic is because ‘staff are nice’. This indicates that, when seeking health care, convenience and- to some extent- staff attitudes are paramount. When comparing nationality or migration status, although not statistically significant, a greater proportion of migrants (non-nationals vs nationals, Table 1; internal and cross-border migrants vs non-migrants, Table 2) reported having experienced problems at the clinic at which they were interviewed. When exploring length of stay, however, long-term residents were significantly more likely than recent arrivals to report problems. This suggests three possibilities: (1) that long-term residents are more confident in expressing their dissatisfaction; (2) that recent arrivals have either not yet accrued sufficient PHC user experience to have experienced problems at the clinic; or, (3) have not yet accrued sufficient PHC user experience to consider reporting their experience as problematic. The latter two suggestions are supported by the data showing that length of stay in current residence is associated with length of time accessing PHC services at the clinic of interview (discussed below).
Many users (recent arrivals, migrants and long-term residents) had been resident at their current address for a longer period of time than they reported accessing healthcare at the particular clinic in which they were surveyed (Table 3). These findings show that these respondents did not start accessing healthcare, at least at that particular clinic, at the same time that they moved into the area of interview, suggesting that they did not move in order to access PHC services.
The survey asked users whether they were accessing other healthcare facilities simultaneously; this was almost exclusively reported not to be the case with the majority of users indicating they only make use of the PHC facility at which they were interviewed (data not shown). Together, these findings suggest that whilst most PHC users surveyed have migration histories, they are not moving in order to access PHC. In the survey, non-national PHC users reported moving for other reasons—mostly to seek employment or to join family (data not shown). This suggests that, in line with existing research [26], the majority of PHC users surveyed have not moved in order to access health services.
Respondents who had been resident at their current address for less than 1 year were more likely to be found at the rural clinic. This group—mostly South African nationals—reported using the rural clinic for a longer period of time than they had been living at their current residence, suggesting high levels of local (intra-provincial) mobility of South African nationals within the rural clinic catchment area.
In our sample, recent arrivals and migrants are younger than long-term residents (no difference in age between new arrivals and migrants). Long-term residents have been accessing PHC services for a longer time (>3 years) at the clinic of interview when compared to new arrivals and migrants (1–5 years). An obvious confounder here is the length of time that the person was living in the area of interview. However, our findings do suggest that the PHC users we surveyed have not moved in order to access healthcare. Our findings support existing research that shows that individuals who move are positively selected: they are younger and therefore likely to be healthier than both the population they leave and the population they join, suggesting the presence of a healthy migrant effect [2, 26–28]. Whilst further research is required, these findings are important in challenging the prevalent assumptions that associate migration with poor health and healthcare-seeking [5].
Study Limitations
The survey described follows a methodology type that aimed to produce a cross-sectional sample that is sufficiently representative of the diverse migration context of South Africa [15, 16, 20], for example, see [29–33]. We undertook statistical analyses similar to those of previous studies that involved a comparable, non-traditional sampling approach, including studies that were also undertaken in the absence of a probability sampling frame and involved small sample sizes [15, 16, 30]. Importantly—and key to this study—such an approach allows for exploratory cross-sectional research in a situation where it was not feasible to develop a probability sampling frame [15, 16, 30], and thus allows for exploratory analysis in a context of resource and time constraints. Previous studies exploring migration have raised concerns relating to possible bias, particularly that associated with nationality, documentation, gender and vulnerability [15]. Whilst we did not collect data on non-responders, the demographic data presented here in our non-probability sampling approach reflect results from other surveys of migrants and non-migrants in SA in each location (Census, 2011), suggesting we obtained a sufficiently representative sample. For example, our study obtained a range of responses describing documentation types similar to that which would be expected based on previous estimates [26, 29]. Importantly, this included reporting of undocumented status—often assumed to be a reason for non-response. In line with global trends relating to gender bias in healthcare-seeking behaviour, women in our sample accounted for more respondents than men. Vulnerability bias, in this case, would refer to individuals who had previously had bad experiences at PHC facilities and, as a result, would not return, and therefore be excluded from our sample. Whilst we recognise this limitation, our sample also included many respondents who reported significant challenges with accessing PHC but still chose to return.
Conclusion
The findings support and strengthen existing knowledge on migration profiles in South Africa, and provide new insights both empirically in terms of migration profiles and experiences in PHC settings in SA, and also methodologically in terms of conducting exploratory research in resource-constrained settings. The findings also raise further research questions relating to how the dynamics of migration and mobility affect the experiences of PHC users in South Africa. The results highlight the high prevalence of a migratory status amongst PHC users—regardless of nationality—and emphasise the importance of gaining improved understanding of local (intra-provincial), internal (inter-provincial) as well as external (cross-border) mobility. The findings suggest that the population making up the rural sample is associated with higher levels of local (intra-provincial) mobility than the urban and peri-urban samples; further research is required here.
In line with prior research comparing the experiences of different South African and cross-border migrants [19, 21, 26], our findings suggest that nationality alone does not explain the different experiences of PHC users. Our study indicates that both length of time accessing a particular PHC facility and length of stay in current residence are important in explaining the differences in experience of PHC users. This does not mean that healthcare responses should ignore nationality and the ways in which this mediates the experiences of PHC users—including through language, culture, documentation status and the attitudes of staff. To the contrary: our findings highlight that the diverse migration profiles of all healthcare users must be considered within the strengthening PHC provision in South Africa. In support of this, we suggest that future research and associated health systems strengthening should move away from “methodological nationalism”—a focus solely on nationality when exploring population movements—in the ways that migration and mobility are conceptualised, measured and responded to [19, 34, 35]. Instead, engagement with population movement for strengthening health systems needs to consider the complexities of migration histories, including where these relate to internal migration as well as cross-border migration.
Critically, this paper highlights the need for further research to improve our understanding of the dynamics of population movement amongst users of the South African PHC system, with implications for the design and implementation of improved public health system responses in both South Africa and the southern African region. Key here are concerns relating to the ways in which migration and mobility influence access to MCH services, continuity of access to chronic treatment provided through PHC services, and associated concerns surrounding the successful control of communicable and non-communicable diseases in the region. We hope that the findings from this study will go some way to supporting efforts towards strengthening PHC services in South Africa, including through the current PHC re-engineering and National Health Insurance (NHI) processes.