This is possible without having all of the cards to be the same. I assumed that the audience, participant, and assistant (not the magician) could see all five cards after they had been blindly picked.

There are 4*3*2 = 24 ways for the remaining four cards to be arranged. The magician can calculate the correct order of the cards, and can then figure out which permutation it is.

Since the deck has 52 cards with four that are being used for the permutation, the magician has to choose one of 48 cards, or 2*24. This means that there is a way to encode a simple yes-no answer by card choice alone.

Now, we need to see how many ways that 5 cards can be partitioned into at most 4 suits.

5

4:1

3:2

3:1:1

2:2:1

2:1:1:1

There are six ways.

Now, how many ways are there to partition 4 into 4?

4

3:1

2:2

2:1:1

1:1:1:1

There are five ways. Somehow, by choosing how to partition the four cards by removing one card, the assistant conveys some information as to which suit the cards are in.

How about this encoding scheme. This shows which four card partitions come from which five card ones.

2:1:1 <- 2:1:1:1, 3:1:1 (2)

4 <- 5 (1)

2:2 <- 2:2:1 (2)

3:1 <- 4:1, 3:2 (2)

Perfect.

This manages to encode all of the possible partitions of five cards into a partition of four cards while limiting the uncertainty to be no more than two.

If the cards are divided 2:1:1, then the missing card is either in the same suit as the two cards or the odd one out.

If the cards are all of the same suit, then the missing card is of the same suit.

If there are two cards of one suit and two of another, then the missing card is in one of the two other suits.

If there are three cards of one suit and one of another, then the card is in one of those two suits.

If the cards are all of different suits, then your assistant messed up because that should never happen.

This will always limit the cards to be within two suits, so there are 24 possible cards (remember that four are already known). This is enough so that knowing which of 24 permutations were used can tell which is the right card. I have not decided on an encoding scheme for 24 permutations into 24 cards, but it should be much easier to come up with.