It is commonly assumed that reaction measurements for astrophysics should be preferably performed in the direction of a positive Q value to minimize the impact of the stellar enhancement factor, i.e., the difference between the laboratory rate and the actual stellar rate. We show that the stellar effects can be minimized in the charged particle channel, even when the reaction Q value is negative. As a demonstration, the cross section of the astrophysically relevant 85Rb(p,n)85Sr reaction has been measured by activation between 2.16 < or = Ec.m. < or = 3.96 MeV and the astrophysical reaction rate for (p, n) as well as (n, p) is directly inferred from the data. The presented arguments are also relevant for other alpha- and proton-induced reactions in the p and rp processes. Additionally, our results confirm a previously derived modification of a global optical proton potential.