More than looking inefficient, it looks unnecessarily cluttered, which can contribute to bugs:

• [\D]? is the same as \D?. This is repeated four times.
• [0|1] is probably a mistake; character classes don't use alternation, and I don't see any | characters in your sample input. You probably mean [01]
• \D?\d{4} is repeated twice in a row. How about (?:\D?\d{4}){2} ?

Making those changes would yield:

9818\D?9[01]\D?\d{2}(?:\D?\d{4}){2}
[download]

Now with the /x modifier, you can further clarify things like this:

m/
  9818        # A literal.
  \D?         # Optional non-digit.
  [01]        # Require a zero or a one.
  \D?         # Another optional non-digit.
  \d{2}       # Require two digits.
  (?:         # Group but don't capture.
    \D?       # Another optional non-digit.
    \d{4}     # Followed by four digits.
  ){2}        # Repeated twice.
/x
[download]

As for efficiency, what problems are you encountering? If you're dealing with huge input you're probably IO bound anyway.