Is it a capacitive coupled amplifier? If so, is this 15pF input capacitance or the feedback capacitance?

Extracted netlist generally does not have mismatch enabled and needs to be explicitly added (in most cases). Therefore the issue might not be related to mismatch but rather due to parasitics. Moreover good matching practices include laying both feedback capacitance and input capacitance in a common centroid fashion and adding dummies to create similar looking environment for all caps.

Please post the schematics, pre and post layout sims, including the extraction type (C+CC or R+C+CC) to better understand the problem.

There are several documents available on the web detailing good layout practices for matching passives. One of those is http://www.eecg.toronto.edu/~johns/ece1371/slides/ECE1371_Mismatch-2.pdf

Since you have a differential amplifier at hand, you need to match sensing electrode and reference electrode paths to avoid cm-dm conversion.

Use of mos devices as pseudo resistors, while is a common practice in biopotential amplifiers, their modelling is rather poor under such operating conditions.

Good to know that you got it working. If you could post the problem and its solution, it would help others who might run into similar issue in future.

Best regards