I am talking about a real population number of persons receiving  lumbar surgery in 2007 compared to the number in 2014. I want to see if there was a significant increase.

The subgroups are inpatient and outpatient, however I want to calculate increase in each population independently. I am able to compare both groups using chi squared but need to see if prevalence/incidence increased significantly.

Thanks Jochen

If you have a reference population (for example, total surgeries in each period) you might calculate the difference between proportions of lumbar surgeries en each period. There are several ways to know whether the difference is significant or not. It depends upon the figures. With short series you might need Fisher's exact. Otherwise Chi squared is suitable. You even can use the 95% confidence limit of the difference to know if "p" is above or below 0.05.

I agree both with Jochen and Patrricio . The decision lies with the researcher , what he wants . He should proceed accordingly. 

Investigator has data count on lumbar surgeries only two points - 2007 and 2014. Simple subtraction of cases of 2014 from cases of 2007 and divide by the cases at the baseline (2007)  provide percent change. This percent is a good enough describe the change. In case you has data for every year from 2007 to 2014, depending on the shape of curve fit a regression model and test whether the slope (rate of change) is significant.

I can get the data points for all years and will try regression model.

Respectable researcher himself is not clear , what he wants to do ? 

I agree with Jochen W

Dear friend

If you have 8 pairs of values ( a value associated with each year : 2007, 2008, ... 2014) , the trend (increasing, decreasing, steady trend) can be assessed by the Spearman correlation coefficient ( use the free BiostaTgv ) . But if you only have two pairs of values ( for 2007 and 2014) , you can not use this coeffiicient (it takes at least five couples) . The best is actually using the chi-square test for a 2 * 2 table (year : 2007, 2014; lumbar surgery : yes, no).