Yes, Packed Binary Coded Decimal is two decimal digits per computer byte. Here is why:
A byte is eight bits. value could be 0 to 255, i.e. 256 possibilities. Half a byte is a Nybble, or four bits, i.e. 0 to 15 or 16 possible values. Binary coded Decimal uses 0000 to 10001 (0 to 9), Packed BCD uses two "nybbles" per 8 bit byte.
Thus from earliest days of computers in 1950s the Packed BCD was popular. Modern CPUs may have special hardware to speed up usage, historically it saved half storage but had to be "unpacked" to binary integers for arithmetic.
A Word is machine size / CPU type dependant and can be:
10 bits (1024 values)
12 bits (4096) Some 8 bit CPUs actually use 12 bit or 14 bit "words" in Program Memory
14 bits (16384 values)
16 bits (65536)
32 bits (4294967296 values)
64 bits (1.844674407×10¹⁹ values)
Signed Integers use half as many values per word as the most significant bit is the sign.
Note that physical addressing maybe unrelated to "word" size. Thus addressing can be 10 bits to 20 bits on an 8 bit CPU, and on 32Bit or 64bit CPUs can be 40 bits to 48 bits. Virtual and/or MMU internal CPU addressing is often more than the physical addressing possible.