Use Lambert-Beer's law.

Before that you need to know the absorption coefficient of the drug in the oil at a particular wavelength. To calculate absorption coefficient take a known weight of sample and measure its absorbance in the oil. Dilute your sample so that the absorbance is less than 1. Do a series of sample to make a calibration curve (y=mx type) for precious calculation.

Thanks @ Mr. T. Khan, can I take absorption coefficient value from literatures. 

Prepare four/five solutions of known concentration of the compound you are trying to measure the concentration and make sure that the UV spectra is not showing saturation. Measure the absorbance and plot the calibration curve. Make a graph of absorbance versus concentration, and use the “method of least squares” procedure to find the best straight line through the linear  portion of the data.  

You have an equation y= mx+b where y is absorbance, x is concentration, m the slope and b the intercept. From the method of least squares you have m and b, so if you analyze the unknow sample you can have y.

So, to determine the concentration of an unknown sample , analyze it  and use the equation  to determine the concentration based on your calibration curve.

Yes, of course you can use literature value. Since absorption coefficient is wavelength dependent make sure to use same wavelength during calculation and if possible use same solvent.

Dear Raeesh,

I'm in complete agreement with you regarding UV absorption coefficients of unknown sample. To perform experiments use two beam and place solvent in one chamber and use it as base.  Further only want to add that though we're following least square method as best option for calibration but there are very few example where as per requirement we've to apply non-linear curve fitting under special situations. But it's acceptable otherwise. Don't rely on literature values for calibration try to perform calibration by yourself only.

Cheers.