Relationship between Sleep and Mental Health among Pregnant Women in Saudi Arabia

Endogeneity bias in the relationship between sleep and mental health

“…A reciprocal relationship invalidates results of linear regression achieved by [1] because of the endogeneity issue. A reciprocal relationship between the dependent and an independent variable is a form of endogeneity called simultaneity.…”

“…In a recent published paper, Kahn (2023) [1] studied the relationship between sleep and mental health among pregnant women in Saudi Arabia. Using a sample of 110 individuals, and after applying a stepwise regression model, they found that insomnia and fatigue were significant predictors of depression and anxiety.…”

“…However, not considering the reciprocal relationships between some of the concepts analyzed is a major concern for acknowledging the validity of results. Kahn (2023) [1] predicted depression and anxiety from some variables related to sleep, considering both insomnia and fatigue as determinants of mental health (depression and anxiety), although they recognized as a limitation the inability of determining a causal relationship between poor sleep and mental health issues.…”

“…However, [1] did not take into consideration an important feature of these types of variables which is essential to achieving statistical analysis; the reciprocal relationships among them. There is ample support in the literature regarding the reciprocal relationships between insomnia and depression [2,3,4,5,6,7,8], and regarding the bidirectional relationship between insomnia and anxiety [7,9,10,11].…”

“…The importance of addressing endogeneity is well explained in [14], being a part of the editorial policy in the most rigorous academic journals. Therefore, the main conclusion reached by [1]: "pregnant women who have poor sleep quality are more likely to have mental health issues, including anxiety and despair", could be reversed by, for example: pregnant women who have mental health issues are more likely to have poor sleep quality. To deepen into these relationships, it would be necessary to re-analyze data, using instrumental variables to address endogeneity.…”