In an experiment, 500 families were given a medical self-help book

Posting the explanation that I found on Powerscore forum discussion-

The flaw in the stimulus's argument can be illustrated with the following diagram:

Has medical textbook → Less visits to doctor
Good health → Less visits to doctor
Has medical textbook → Good health (MISTAKEN INFERENCE)

or more simply:

A → C
B → C
A →B (mistaken inference)

(D) says that the flaw is "two different states of affairs could each causally contribute to the same effect even though neither causally contributes to the other." This would perfectly match the above.

The "different states of affairs" are having a medical textbook and good health and the effect is less visits to the doctor. While it's possible that having a medical textbook (state of affair 1) leads to less visits to the doctor (effect), and good health (state of affair 2) also leads to less visits to the doctor (effect), we can't conclude based on this information that having a medical textbook leads to improved health (improperly assuming causal relationship between two states of affairs).

Why option A is ruled out- It tells us nothing about the link between the books and improved health. Answer A might make us question any causal claims about the books and visits to the doctor (because if it were true that the group that wasn't given the book got one anyway, we would have a possible cause without the possible effect). That, however, is something of a side issue, while the main issue is what any of this has to do with improved family health. There could be no connection at all.