Background

The healthy migrant and salmon bias hypotheses stipulate that those who are healthier are more likely to migrate (a selection effect) and unhealthy migrants have a greater tendency to return to their origin communities than healthier migrants. It remains unclear whether the hypotheses are relevant for mental health, although they have robust evidence for physical health. The current study expands the theoretical foundation underlying these hypotheses by exploring the impact of migration status on depression among the middle-aged and older population in China.

Method

Longitudinal data from the China Health and Retirement Longitudinal Study (CHARLS) in 2011, 2013, 2015 and 2018 were used. Participants were divided into three groups, non-migrants, permanent migrants, and return-migrants according to their migration status. The Chinese version of 10-item Center for Epidemiologic Studies Depression Scale was used to assess clinically relevant depressive symptoms. Incidence were estimated by each migration status and compared across strata using log-rank tests. Cox regression models were fit to assess the relationship between migration status and depression incidence.

Results

A total of 9264 middle-aged and elderly participants were included, of whom 7047 (76.1 %) were non-migrants, 1065 (11.5 %) were permanent migrants, and 1152 (12.4 %) were return-migrants, respectively. Incidence rates of depression by different migration statuses were 76.6/10³ person-years (73.7/10³–79.7/10³) in non-migrants, 62.9/10³ person-years (55.9/10³–70.6/10³) in permanent migrants, and 67.3/10³ person-years (60.7/10³–74.5/10³) in return-migrants. After controlling for covariates, only permanent migrants had significantly lower incidence of depression than non-migrants (HR: 0.85 (0.75, 0.96)). After stratifying by current place of residence, urban-permanent-migrants had a lower risk of depression than both rural and urban non-migrants, and the HRs and 95%CIs were 0.62 (0.52–0.74) and 0.81 (0.68–0.97).

Limitations

Migration statuses of participants may have changed over time, but due to the lack of raw data on migration status in 2013 and 2015, we used migration status at baseline. We were also not able to contextualize the migration experiences among those who migrated, and these trajectories may be heterogeneous.

Conclusion

Using depression as an indicator of mental health, the healthy migrant hypothesis was supported among middle-aged and elderly permanent migrants living in urban areas while the salmon bias hypothesis was not. Significant urban-rural inequities in mental health existed among different types of migrants. Especially among permanent migrants who have migrated and converted Hukou from their original registration place to their current residence.