dhis2-devs team mailing list archive

[Ola Hodne Titlestad <olatitle@xxxxxxxxx>]

Hi Thuy,

I think I understand your problem and I had a look at your report. Could you
please translate the columnn headings for me, it's been to long since I was
in Vietnam. Still, from what I see in the columns and the few words I
remember it seems the table represents one of those cases where there is no
total for the row that sums up all the fields, I mean for every data element
(disease) there is no total that is the sum of all the categoryoptioncombos
(fields from both green and red).

As you say the new fields (green in your report) are not mixed with the old
(the red), so in fact they should be two different categorycombos.
CategoryCombos should represent a hierarchy where the total at the top/root
should make sense. In your case it seems the total of green and red is not a
real total that makes sense to use. The built-in logic of indicators,
validation rules, report tables etc. in DHIS is very much built around data
element totals, so having a data element total that does not make sense
makes it difficult for you to make good use of the rest of the system. Then
you end up with custom forms + tailored (developer) made reports in excel
and that's it, which might work of course, but a pitty thinking of all the
other functionality you miss out on.

Any chance you could modify either the green or red so that they actually
became the same? As far as I can see your are collecting both new cases and
deaths for both clinics (green) and hospitals (red), but the agre breakups
are different for the two. You collect children under 15 years total and <5
for red, but only children (no age specified) for green, is that correct? If
it would be possible to collect <15 and <5 also in the green part it would
be a lot easier for you.

The fact that you already have data for the red part and now want to start
collecting for the green part as well, makes it more complicated. If you can
make green and red the same then you could keep all the data with some
tweaking and continue with another categorycomo with green+red, but if you
need to keep them as they are in your report I would consider splitting up
into 2 categprycombos and two data elements in stead.

If I was to design datasets based on this report without having to think
about the data already in the system, a fresh start if you like, I would
create two data element per disease (ICD code), 1) "<disease name> + Phong
Kham)", and 2) "<disease name> + Benh Nhan dieu tri". And then I would
create two categorycombos, one for the green and one for the red, and then
assign data elements with the Phong Kham name to the green and data elements
with the Benh Nhan name to the red. That would give you two datasets.

You could later design reports that would give you the exact same table, and
with some changes to data entry forms (as in my previous suggestion to allow
multiple datasets to 1 form) you could also create a custom form that would
look like your report.

That probably didn't help much...

Just note that having categorycombos that doesn't make sense to add up as a
total will not be very useful, except for allowing you to design the data
entry form you need.

Another general comment would be to try to think of your paper forms as your
outputs (reports) and not necessarily your inputs. Your DHIS data entry
forms are your inputs and that they do not have to be exactly the same
(although that is what the user will ask you for). If you need some of these
data in other outputs, like an indicator to view on a map or in a chart, or
another summary report, or in a pivot table,  then you might have problems
getting what you need because of the way you have designed your datasets,
data elements, and dimensions.

If we can improve the data set + data entry forms model in DHIS 2 a bit to
allow for multiple data sets (categorycombos) in one form, then we are much
more free to design appropriate datasets (data elements with dimensions)
that makes sense for data analysis and still be able to meet the users needs
of having to make data entry forms look exactly like the paper forms.

With the current constraints of the 1-1 relationship between form and
dataset we are struggling to make good use of the multidimensional model
(because we put everything in one categorycombo to mimic the forms) for
anything but data entry, unfortunately.

Ola
-------

2009/12/4 Thuy Nguyen <thuy.hispvietnam@xxxxxxxxx>

> I got your point Abyot that we need to show the old data with old category
> options
> But if there is another sistuation like I used to presented some weeks ago.
> In the ICD 10 report
> There are many data elements (about 300).
> After collecting data some time, then the report add more column (which we
> will define new category options), these category options are not nested or
> nest any other category option (this can be understand as language and
> meaning) and we will continue collect data values with the old category
> options and new category options in the same data elements.
>
> If there is not addition for category options, then how can I solve this
> problem?
> It's would be painful to
> - create new categories with new category options and new category
> combination and assign those to many data elements
> If there is the solution that I can create new category with completely new
> categoy options, and then combine the old category with the new category to
> new category combination and assign to those 300 data elements again. But I
> don't think this combination won't mess up the structure of colums which we
> need to display in the report.
>
> Please find enclosed the attached file to see the report.
> --
> --
> Thuy
> HISP Vietnam
> +84902079126
>
>
hI THUY