dhis2-kenya team mailing list archive

[Ola Hodne Titlestad <olati@xxxxxxxxxx>]

Hi all,

*Some background*
Manya, Gikunda and I visited Kisumu District Hospital on Friday and had an
interesting discussion on the information needs for district hospitals.
What they need is to be able to monitor activities in each ward, e.g. look
at bed occupancy, turnover intervals, average length of stay etc.
The current data sets to not support this since they are designed to be
collected by the district hospital orgunit and not its children (the wards).

The 717 - Service workload has some very general data elements that are
based on the three groups of wards:  General Adults, General Paediatrics,
Maternity Mothers only, Amenity.
Here the data elements represent the activity (Admission, Discharge etc.)
and the category options are the types of ward (listed above).
This makes it possible for higher level people to see roughly which types
of wards that has the highest workload etc. but it is useless for district
hospital managers that want to manage individual wards.

The dataset Hospital Administrative Statistics is better, but still
difficult to use for hospital management, Here the data elements represent
the different types of wards and the category options the activities
(Admissions, Discharges, Available beds etc.). This has some design
drawbacks. First of all the hospitals call they wards differently and do
not have exactly the same type of wards. So the various hospitals will ask
for more data elements to be added, we have already seen requests in the
message board. This will be difficult to standardise across all hospitals.
Another issue here is the idea of putting Beds/Cots and
admissions/discharges/etc as options for the same data element ("Male
Medical"). This doesn't work well when we aggregate over time since beds
are not supposed to be summed together, but admissions are. A quarterly
report will show three times as many beds that are actually there..... We
need to discuss what to do with this dataset. If we want to keep it we at
least need to split it so that beds and cots are separate data elements.

*The proposed solution*
So we ended up with the following solution, which is a third way of doing
inpatient data for hospitals:
(I don't like that we have three different approaches, but think we at
least need this one. Let's talk about cleaning up this later.)

*- new orgunits under the district hospital to represent the individual
wards at the hospital, e.g. "Kisumu DH 1-Maternity", "Kisumu DH
2-Peadiatrics", "Kisumu DH 3-Female Medical" etc.*
*- some new orgunit groups to represent the different standard types of
wards at the district hospitals ("Medical", "Surgical" etc.)*

Representing the wards as orgunits and not as data elements or categories
gives more flexibility to the various hospitals to define their own wards
(give them the names that they call them), and to do analysis on ward
level. Kisumu DH has 10 wards. I suspect that the number of wards will vary
between 5-10 for most district hospitals.
This will add some more orgunits to the database of course, but I think it
is worth it considering the value it adds to local hospital management. And
these orgunits will only have 11 data values each month. While I think the
bigger hospitals like the national and provincial with many more wards
should use a separate hospital database, I don't think that is a good idea
for these small district hospitals. Better for the district hospitals to
have all their data in one place. Anyway, let's see how this scales.

*- a new data set called "Inpatient activities" with all the typical data
elements like "IP - Admissions", "IP - Discharges", "IP - Available Beds"
etc.*

*- a set of indicators for IP management based on the new data elements
described above*

*- for now, 1 chart and 1 standard report using the new indicators - more
needed here*

Kisumu District Hospital (in Kisumu East district) will start data entry
for this data set this week, so we can soon test the indicators and reports.

Ola
----------


----------------------------------
Ola Hodne Titlestad (Mr)
HISP
Department of Informatics
University of Oslo

Mobile: +47 48069736
Home address: Vetlandsvn. 95B, 0685 Oslo, Norway. Googlemaps
link<http://maps.google.com/maps?f=q&source=s_q&hl=en&geocode=&q=Vetlandsvn.+95B,+0685+Oslo,+Norway>