How to organise data in DHIS2 to capture project "context"

Hi lars ,

Did you get a chance to look at this issue ?

···

On Thu, Aug 6, 2015 at 12:20 PM, David Siang Fong Oh doh@thoughtworks.com wrote:

Hi Lars,

We are currently working with MSF to attach project context to the data entries and would like some input on how best to structure this in DHIS2.

A bit of background to get started

Our org unit structure:

  • Company (i.e. MSF)
  • Operational Centre (i.e. OCP)
  • Country (i.e. Afghanistan)
  • Project (i.e. Kabul)
  • Operational Unit (i.e. Kabul Hospital)
  • Modules (i.e. Neonatal ward)

At the project level we capture additional information to describe the “context”:

  • Mode of operation (i.e. Direct, Remote etc.)
  • Reason for intervention (i.e. Armed conflict, epidemic etc.)

We currently assign this information as attributes for the Project org units. To assist with data analysis we also organise the projects into org unit groups based on this information, for example an org unit group for “Direct Mode of Operation” and another group for “Remote Mode of Operation”.

The problem

The problem we are trying to solve is what happens to the data when a project “context” changes. For example if Kabul changes mode of operation from ‘Direct’ to ‘Remote’ on Week 22, we need to keep track that Weeks 1-21 have data where the mode of operation was ‘Direct’, whereas Week 22 onwards will be data where the mode of operation was ‘Remote’.

Our proposed approach

We intend to store the project “context” information as data elements. For example using the example above, for Week 10 in Kabul the data entered in the Neonatal Ward will have the actual numbers (i.e. admissions) as well as data elements which represent context (i.e. data element ‘Mode of Operation’ will be set to the option ‘Direct’, and data element ‘Reason for Intervention’ will be set to the option ‘Armed conflict’).

This approach works well as far as recording the data. When we use a pivot table the data is represented as expected (see below). The problem is that when we represent this as a chart in the visualiser we are not able to use the data element ‘context’ as a filter.

Is there a way to use the value of a data element as a filter in the visualiser charts? Can you think of a better way to organise our data to capture this ‘context’ information?

Thanks for your assistance.

Cheers,

David Oh

doh@thoughtworks.com

+91 9900246052

Sandesh DoolipetaApplication Developer
Email
sandeshd@thoughtworks.com

Telephone
9686062727

ThoughtWorks