Determining cutoff point after -lsens-

[Musa Waziri]

Dear all,
The accuraccy rate of my prediction model is very high that can be explained by imbalanced number of failure and non-failureobservations. I decided to use -lsens- command to find a more optimal cutoff point. What Stata command or option would allow to find the intersection point between sensitivity and specificity? Visually it is hard to figure it out.

[Andrew Musau]

*OPEN DATA SET AND RUN LOGIT REGRESSION
webuse lbw, clear
logit low age i.race
*GENERATE CUTOFF, SENSITIVITY AND SPECIFICITY USING -LSENS- COMMAND 
qui lsens, genprob(cutoff) gensens(sens) genspec(spec) nograph
gen difference= sens-spec
levelsof sens if abs(difference)<0.01, local(yval)
levelsof cutoff if sens==`yval', local(xval)
lsens, yline(`yval') xline(`xval') scheme(s1color) ylab(0 0.2 `yval' 0.8 1) xlab(0 0.2 `xval' 0.8 1)