I am running a regression : hours worked per year= days ill per year+ age+ gender+ region + u. Would number of doctors visits be a good instrumental variable for days ill per year? Thank you.