Agreed. Also, the crude CI is almost crossing 1 and the width (and PR) seems to increase when controlling for other things. This may suggest that those other th

Agreed. Also, the crude CI is almost crossing 1 and the width (and PR) seems to increase when controlling for other things. This may suggest that those other things may be of more significant interest than the main effect the authors are modeling.