This is a very exciting development.
During the phases of development I am sure that a number of issues arise, which we can then clarify on the list.
Afterwards the collection of solved issues and ideas to solve them could be collected into a common wiki page, which will then explain how to create such a system and modify/extend it in the future. Since the idea of the wiki is to allow all to edit the contents this could be a learning experience. Obviously provided you think this is a good idea.
A comment would be: The "key" entity has a different meaning now in comparison with EpiData Classic 3.1 and before that Epi6. Now the key is used to define a unique key, which is based on one or more fields. On a given dataform one can only have ONE unique key.
So if you want to create a key for a given form, say person and date, which are stored in two different fields. Then you would define in Manager for that dataform a single unique key based on both the person and the date fields.
An example (hypothetical) might explain this better: You want to create a related system for family visits to a health centre, where a family are those living together on a given adress:
top level: start (key: "road name" "house number" "post code" "city name" ) related level: family (key: "role" "date of birth" ) where role is father/mother/grandmother/child/twin A/twin B" etc related level to family: visit (key: date)
To create such a system you follow this recipe - also shown when you create a new project in v2.0.... of Manager: For relational datasets: 1 Select "Dataform 1" on the left side a define fields: "road name" "house number" "post code" "city name" + other relevant ones b rename "Dataform 1" to "start" by clicking on the current name 2 Define Key variables on that top level (you do this in button "Dataform") add to key: "road name" "house number" "post code" "city name" 3 Add the family member dataform 3a Select "start" dataform and press the + on the left side, second level will be: family_members. 3b Notice that when you add this the key variables from top level are propagated to the second level 3c rename the dataform and add the relevant variables, including role and "date of birth" 3c Since we wish to have several records at this level, one for each family member: Extend the key at this level with "role" "date of birth" Notice that since twins could be born on the same day we must as a value label in role also have "twin a", and "twin b" 4 Add the visit dataform 3a Select "family" dataform and press the + on the left side, a third level dataform is created 3b The key variables from top level and second level are propagated to the third level 3c Rename the third dataform and add the relevant variables, including "date of visit" 3d Since each member can have more visits: extend the key on this form with "date of visit" 5 For each dataform now define 1:1 or 1:many in "dataform" properties. For one to many there must be at least one additional field in the key. 6 For each dataform define a variable, which will decide how to change dataform in focus during entry. (basically this is like a jump in entry sequence from one dataform to the next).
Continue as above with other dataforms. Notice that the sequence is always to first select the form which should get a child, on that form add the full key before you add the child, since otherwise the relevant fields will not be propagated to a given level. The system will warn you if you have not defined the appropriate number of key fields.
In the distribution of v2 test a working example "clinical example" is included. In this you can find further ideas.
regards Jens Lauritsen EpiData Association
On 02-11-2014 17:39, epidata-list@lists.umanitoba.ca wrote:
Hello List,
As a District Medical Officer, I will have to prepare for Contact tracing and risk assessment/managing the follow-up of contacts if an unsuspected Ebola case should surface in my community.
I have started to think of a call center / health care registration system in EpiData Manager with relate feature: One event/Index case, several contacts with contact details, and follow-up details of contacts. In the event that one of the contacts should develop symptoms, then the secondary case´s contacts will have to be traced. The design of the system certainly needs some consideration.
This is a global concern, and I imagine several organizations probably already have made or have plans to develop such a system. EpiData´s share-friendly approach should be ideal under current circumstance. ECDC have already done so for food/waterborne outbreak investigation.
- Anybody already made a system, wiling to share?
- Any major Organization seeing the benefit of same-format reporting, and
willing to support development of system/templates? 3. I assume that final release of ED 2.0 with relate feature is imminent. I therefore start to work on a functional but simple local system anyway, and will seek advice from the list as needed. Where I am right now, I struggle to carry the content of more than one unique key variable over from the parent to the child record.
Any comments?
Best regards, Vegard Hoegli District Medical Officer Skien, Norway _______________________________________________ EpiData-list mailing list EpiData-list@lists.umanitoba.ca http://lists.umanitoba.ca/mailman/listinfo/epidata-list