What is New in 2.7

Much has changed in Protean and GreEd from the demo version (2.5) available at http://proteme.org. Many of the changes have been beneath the surface. The most important, beneath-the-surface change is the removal of all non open-source code and libraries so that all code is unencumbered by commercial licensing requirements. We also removed many bugs.

Many other changes in Protean and GreEd, however, are those that will make knowledge author's job much easier. Here is the list of those changes:

Contents

1 What is New in Protean

1.1 Share executable knowledge with others
1.2 Unlimited undo and redo
1.3 Promotion and demotion
1.4 Move an item from one location to another
1.5 Search your library of components
1.6 Reordering of Data Elements
1.7 Things that have been removed

2 What is New in GreEd

2.1 Undo and Redo
2.2 Default Inference
2.3 Semantic Guidance and Constraints
2.4 N of M Operator
2.5 Between Operator
2.6 Date Fields and Operations
2.7 Things that have been removed

What is New in Protean

Share executable knowledge with others

Share executable knowledge you create with others by using the export toolbar button, and import knowledge created by others, by using the import button on the toolbar. Go ahead, show off your medical genius and knowledge authoring smarts to your colleagues.

Unlimited undo and redo

No need to worry about losing your existing work by one costly error. All actions now are undo-able. Simply press Ctrl-z. Press Ctrl-y to redo.

Promotion and demotion

You can promote a simple activity (AKC) to become a process (PKC) while preserving the semantics. When creating a clinical workflow you may think of an activity to be an atomic one but as your understanding of the clinical process evolves you want that activity to behave like a process. You can easily do so from the context (right-click) menu for the knowledge component. At other times, representing an activity as a process may seem like an overkill. You can as easily demote the process to be an atomic activity.

Move an item from one location to another

If you start authoring real world clinical processes, you are bound to run into this need. For example, as you begin authoring a workflow you think “Measure B.P.” belongs in the “Preliminary Examination” process. However, later you realize that it belongs more appropriately in another process, called “Risk Assessment”. You can simply control + drag the “Measure B.P.” from its previous location, the process in which currently exists to the one you now want it to be in.

Search your library of components

Over a period of time you will build up a large number of components, both data elements and knowledge components, in your library. From time to time, you will feel the need to reuse something that you had developed in the past for the new clinical workflow you are developing. To do so you will need to search through your existing corpus of components. Now you can do so.

To search components that you want to locate in the component repository tree, open up the branches that you want the component from. For example, if you know that what you are looking for is a Process KC, just open up the Process KC branch. Thus the opened branches are the scope of your search. To initiate the search press Ctrl-i. You will see a small tool-tip like search field displayed on top of the tree. Start typing any part of the name of the component. If any component matches the strings you type it will be selected in the component tree. If there is no match, the text you type turns red. At any stage, press Ctrl+Shift-A to select all the components that match the search string.

The Ctrl-i hotkey launches a case sensitive search. If you want a case-insensitive search, press Ctrl+Shift-i. In every other respect the case-insensitive search is similar to the case-sensitive one.

Reordering of Data Elements

You can change the position of a data element, relative to others, within an AKC by selecting the data element and pressing Ctrl-Up arrow or -Ctrl-Down arrow

Things that have been removed

The previous version of Protean was more for demo purposes. With the version 2.7 we have started moving towards making Proteus useful for real world clinical decision support. Being a demo, the previous version included some patient related demonstration features. These features have been now removed. For example, you can no longer select a patient from the sample patient records. Nor will you see the other tabs in Protean for seeing other clinical data.

What is New in GreEd

Undo and Redo

Like in the new version of Protean, in GreEd also you can undo and redo your editing actions. Press Ctrl-z to undo and press Ctrl-y to redo.

Default Inference

What happens if none of the conditions specified in the left hand side of your rule are met? In conditional (if-then) constructs of all programming and rule languages, you have 'Else' to allow to deal with such a situation. Now you can do so in GreEd too. Just double click an inference in any of the rules open for a KC. That inference is now shown as having bold magenta border, which means that if none of the other inferences are reached, this inference will be considered to be reached by the rule set for the KC. If you want to turn the default off simply double-click the inference again, or you may specify another inference to be default.

Semantic Guidance and Constraints

With this version, we have started moving towards developing power authoring features. The goal is to prevent authors from creating nonsense or stupid rules and to ensure that the author has not left any of his/her bases uncovered. This has been done by providing an alert at every stage that the user is trying to do something that is not possible or not rational. This alert is as an unobtrusive message that pops up when needed and vanishes automatically in a few seconds. Additionally, the status bar keeps the author informed of how many conditions or inferences he/she has yet to to take into consideration.

N of M Operator

Example: You want a rule which says, if the patient has any 2 of the possible features (e.g., 'headache nature - throbbing', 'headache unilateral', 'visual aura', 'nausea/vomiting', 'sensitivity to light and sound'), the inference should be 'potentially migraine'. You can create such a rule by using the 'N of M' operator. Just drop the operator, specify the N value and connect all the symptoms that are to be the part of the M set.

Between Operator

The between operator allows you to specify an expression with a range (e.g. 100-140) within which if a given value (e.g. Systolic Blood Pressure) should fall for it to be evaluated true

Date Fields and Operations

Limited date operations are now available in GreEd. For example, you can drop a date, or use a date field of a KC, and do some calculations on them. You can even use some numeric fields with date operands, for example, you can add 280 to Last 'Menstrual Period Date' to calculate 'Expected Date of Delivery'

Things that have been removed

Some of the operators that were available in the previous version had not been fully developed and we felt that they are rarely needed. These were removed.