The paper in the link includes an instrumental variable model to look at exogenously caused birth spacing (caused by miscarriages). So if the model does what the abstract says it does, then that pins down the direction of causation to some extent.

Of course to be really sure you'd need to take a closer look and probably do more research.