Restrict indicator calculation to data dimension of type orgunitgroupset


When creating indicators, it is possible to select in the formula either a DE (total) or any of its disaggregation (categoryoptioncombos). For instance:


My question:

Is there any way to restrict the calculation according to the organizationUnitGroups where the datavalues are stored (i.e. pre-defining an option of an existing data dimension of type orgunitgroupset at indicator level or even by DE in the formula)? Would it be something feasible in the future?


Our use case is the following. In the design phase we decided to harmonize DEs among data sets as much as possible. E.g. Using a single DE for “Admissions” across different IPD datasets (Hospitalization Ward, Paediatric Ward, ITFC…). Datasets are assigned then in OUs of level 6, that for us are “Services” within a Health Facility (our OU level 5).

This approach has the benefit that we have to maintain fewer DEs (name, description…). During analysis, we can have the total number of admissions at Health Facility level (if we analyze at OU level 5) or number of admissions by service (either analyzing at OU level 6, or at OU level 5 in combination with the use of OrgUnitGroups in a data dimension).

However, the limitation comes when we have a service specific key indicator. E.g. If for a certain context it is especially relevant to calculate “Admissions in ITFC” and they want to present it in the same table as “Exits in Paed Ward”, we cannot do it (we are forced to analyze at Health Facility level and we have incompatibilities in the orgunitset data dimension). It would be also helpful for specific relevant indicators where we want to avoid the human error of not setting the data dimension of type orgunitset.

For the time being the solution is dividing the analysis in different charts/tables.



Hi @acasrod

Have you tried to assign the different wards/levels of service to an organisation unit group? Each of the services would be assigned to a single group (Hospital, Paediatric, IFTC…) and these groups would be part of a group set. Using the pivot tables, you could then restrict your analysis to a single group using a filter, or see the data aggregated across each of the different groups.

Analysis using orgunit group sets I think would be the preferred approach here if possible.

Best regards,