Pattern Attribute Keywords

Each pattern attribute has a set of keywords, and each keyword has a name and a collection of matching phrases:

Suppose that a pattern attribute has keywords as follows:

Here are some clinical notes converted to lists of these keywords:

When a pattern attribute appears in a case, its value is that of the associated attribute converted to a list of keywords. This is done for every episode. For example, here is a pattern attribute called “NotesSummary” that analyses values of “ClinicalNotes”:

It is possible to write rule conditions that directly compare a required string with these values. For example, for this case we could use a condition

NotesSummary contains "ANAEMIA"

To add a comment about the patient having a known disease.

However, this is not the recommended approach to using pattern attributes. One reason for not using keywords directly in a condition is that we may need to change the name of a keyword, and if we do, rules that rely on the presence or absence of the original keyword name will fail.

Keywords are regular expressions

A regular expression is a string of text in which certain characters take on a special meaning, so that the expression itself represents a whole class of texts. A familiar  example is that “*.jpg” represents any file with the extension “.jpg”.

In Pattern Attributes, the matching phrase of keywords are interpreted by RippleDown as regular expressions. This is mostly of use in analysing lengthy documents. For the  analysis of clinical notes and other short texts, regular expressions can pretty much be ignored.

Matching punctuation

Because a matching phrase for a keyword is a regular expressions, care must be taken when adding a matching phrase for punctuation such as a full-stop (period). These characters need to be ‘escaped’ with a backslash symbol ”.

Another character that must be escaped is the question mark.

• Pattern Attribute
• Derived Pattern Attribute
• Pattern Attribute Maintenance