Regex Search Criteria

You can use also use regular expressions (regex) in the Log Text Query search field, simply enclose the expression in slash characters /<regex>/.
Note: The closing slash is optional unless you wish to combine multiple regex expressions.

Here are a few ideas for (hopefully) useful things you can do with regex queries:

1. Match all HL7 messages for all channels
   • Use an anchor “^” to match all channels beginning with “MSH”:
     

     /^MSH/

     Alternatively you can omit the second slash:

     /^MSH

   • This query will return all log entries starting with MSH, to restrict it to messages select “Messages” from the Type list in the Search Criteria panel.
2. Match non-HL7 messages for all channels:
   • First select “Messages” from the Type list in the Search Criteria panel.
   • Then negate the regex (from above) that matches all HL7 messages:
     

     NOT /^MSH/

   • Alternatively use a regex with look-ahead not equals “?!”:
     

     /^(?!MSH)/

3. Match all messages for all patients with the same surname, for a specified channel.
   • Select “Messages” from the Type list in the Search Criteria panel.
   • Select the the desired channel name from Channel drop down (we used Sender).
   • This regex returns all records containing the name, unfortunately it matches relatives (NK1) and doctors (PV1) named “Smith” as well as Patients (PID):
     Note: A simple (non-regex) search (without the slashes) will give the same result.

     /Smith/

   • This regex checks the Name field in the PID segment so it only matches only matches patients named “Smith”:
     

     /PID(\|[^|]*){5}Smith/

   • Check for alternative spelling of Smith, Smithe, Smyth, or Smythe:
     

     /PID(\|[^|]*){5}(Sm[iy]the?)\b/

   • You can also check for multiple names:
     

     /PID(\|[^|]*){5}(Smith|Jones)/

4. Exclude all messages for all patients with the same surname, for a specified channel.
   • Select “Messages” from the Type list in the Search Criteria panel.
   • Select the the desired channel name from Channel drop down (we used Sender).
   • Negating the matching regex excludes all records containing the name, unfortunately it excludes relatives (NK1) and doctors (PV1) named “Smith” as well as Patients (PID):
     Note: A simple (non-regex) search (without the slashes) will give the same result.

     NOT /Smith/

     Alternatively use regex with look-ahead not equals “?!”, unfortunately this also excludes all messages containing “Smith” not just patients:

     /^(?!.*Smith)/

   • Negating a regex that checks the Name field in the PID segment excludes patients named “Smith”:
     

     NOT /PID(\|[^|]*){5}Smith/

     Alternatively use regex with look-ahead not equals “?!”:

     /PID(\|[^|]*){4}\|(?!Smith\b)/

   • Negate checking for alternative spelling of Smith, Smyth, or Smythe:
     

     NOT /PID(\|[^|]*){5}(Sm[iy]the?)/

     Alternatively use regex with look-ahead not equals “?!”:

     /PID(\|[^|]*){4}\|(?!Sm[iy]the?)/

     Note: This regex relies checking the first (surname) sub-field in the name field, by using “\|” (/PID(\|[^|]*){4}\|(?!Sm[i|y]th[e]?)/) to match the start of the name field.

   • You can also negate checking for multiple names:names:
     

     NOT /PID(\|[^|]*){5}(Smith\b|Jones)/

     Alternatively use regex with look-ahead not equals “?!”:

     /PID(\|[^|]*){4}\|(?!Smith\b|Jones)/

5. Match a specified value in a specified field in an HL7 message.
   • Use this regex format:
     

     /<segment>(\|[^|]*){<field>}<value>/

   • To find Toronto in the address field of the PID segment:
     

     /PID(\|[^|]*){11}Toronto/

   • To find LA or Los Angeles in the address field of the PID segment:
     Note: The “\b” (word boundary) to only match “LA” as a whole word, to prevent matching things like “Miller Lane” or “Lake Rd”

     /PID(\|[^|]*){11}(Los Angeles|LA\b)/

6. Exclude a specified value in a specified field in an HL7 message.
   • Negate this regex format:
     

     NOT /<segment>(\|[^|]*){<field>}<value>/

   • To exclude Toronto in the address field of the PID segment:
     

     NOT /PID(\|[^|]*){11}Toronto/

   • To exclude LA or Los Angeles in the address field of the PID segment:
     Note: The “\b” (word boundary) to only match “LA” as a whole word, to prevent matching things like “Miller Lane” or “Lake Rd”

     NOT /PID(\|[^|]*){11}(Los Angeles|LA\b)/

   • We didn’t come up with a regex to exclude values in a field (let us know if you figure it out). However we were able to exclude values in a sub-field (see point 8 below).
7. Match a specified value in a specified sub-field in an HL7 message.
   • Use this regex format:
     

     /<segment>((\|[^|]*){<previous-field>}\|([^^]*\^){<sub-field>})<value>/

   • To find Toronto in the address field of the PID segment:
     

     /PID((\|[^|]*){10}\|([^^]*\^){2})Toronto/

   • To find LA or Los Angeles in the address field of the PID segment:
     Note: A “\b” (word boundary) to match “LA” only is not required, because we are now only searching the specific sub-field.

     /PID((\|[^|]*){10}\|([^^]*\^){2})(Los Angeles|LA)/

   • To find the name Smith, Smyth, or Smythe in the surname sub-field:
     Note: A “\b” (word boundary) is not required because we are only searching the specific sub-field.

     /PID((\|[^|]*){4}\|([^^]*\^){0})(Sm[iy]the?)/

8. Exclude a specified value in a specified sub-field in an HL7 message.
   • Use this regex format:
     

     /<segment>((\|[^|]*){<previous-field>}\|([^^]*\^){<sub-field>})[^(<value>)]/

   • To exclude Toronto in the address field of the PID segment:
     

     /PID((\|[^|]*){10}\|([^^]*\^){2})[^(Toronto)]/

   • To exclude LA or Los Angeles in the address field of the PID segment:
     Note: A “\b” (word boundary) to match “LA” only is not required, because we are now only searching the specific sub-field.

     /PID((\|[^|]*){10}\|([^^]*\^){2})[^(Los Angeles|LA)]/

   • To exclude the name Smith, Smyth, or Smythe in the surname sub-field:
     Note: A “\b” (word boundary) is not required because we are only searching the specific sub-field, also we used “(i|y)” instead of the bracketed character class “[iy]” we used earlier (because it raises an error when used with character class negation [^…]).

     /PID((\|[^|]*){4}\|([^^]*\^){0})[^(Sm(i|y)the?)]/

9. Match or exclude messages where the patient surname matches the surname of a relative.
   1. Use this regex to match:
      Note: “(?s)” is a PCRE option to match newlines (\n), “\d” matches a single numerical digit, “\3” represents the 3rd capture group (that captures the surname).

      /(?s)PID((\|[^|]*){4}\|([^^]*[\^|\|]){1})(.*NK1\|\d\|\3)/

      Alternatively using a named group can be easier to understand:
      Note: “?P<surname>” names the group, “?P=surname” references it, also we used “?:” to make other groups non-capturing (more efficient).

      (?s)PID(?:(?:\|[^|]*){4}\|(?P<surname>[^^]*[\^|\|]){1})(.*NK1\|\d\|(?P=surname))

   2. Use a regex with look-ahead not equals “?!” to exclude:
      

      /(?s)PID(?:(?:\|[^|]*){4}\|(?P<surname>[^^]*[\^|\|]){1})(?!.*NK1\|\d\|(?P=surname))/

10. Match or exclude messages where the patient surname matches the surname of a (referring) doctor.
    • Use this regex to match:
      Note: “(?s)” is a PCRE option to match newlines (\n), “\d” matches a single numerical digit, “\3” represents the 3rd capture group (that captures the surname).

      /(?s)PID(?:(?:\|[^|]*){4}\|(?P<surname>[^^]*[\^|\|]){1})(.*PV1.*[\d]*\^(?P=surname))/

      Note: This regex assumes that the Person Identifier code for the Doctor contains a numeric code, or if the code is alpha, the regex will not match.

    • Use a regex with look-ahead not equals “?!” to exclude:
      

      /(?s)PID(?:(?:\|[^|]*){4}\|(?P<surname>[^^]*[\^|\|]){1})(?!.*PV1.*[\d]*\^(?P=surname))/