AFRO-NETS> HMIS reform (2)

HMIS reform (2)
---------------

I have been intrigued by the fundamental questions asked by Bruce Campbell
and have been struggling to formulate a response. There are a number of
serious questions that have been asked and I will give my comments.

But before I do that I want to mention a myth that affects this discussion.
The myth is that data or information generated from data is used for
decision making. There is even a USAID project called Data for decision
making. What I have observed is that apart from financial data the other
morbidity, activity and preventive data dutifully collected is rarely if
ever used at national level. There are many reasons for this particularly
related to the incompleteness of the data and the delays in consolidating
it. But most decisions are made for political, personal or convenience
reasons.

So where does data get used? I have seen routine data used at facility
level and at district level when there is not too much. The problem however
is the amount of data that is demanded that sinks the system. The KIT
publication is good but there is still too much morbidity data being
collected and as Bruce said in a previous posting this was a compromise.

What would be my minimal monthly data set for a facility? I would like to
know basic activity data such as Number of: outpatients seen, first
prenatal visits, deliveries, BCG and measles vaccinations, percent
undernourished if a master card is used and new FP acceptors. I would also
like to know about stock outs of vital drugs and vaccines and any mechanical
failures. I would also need some basic indicators of environmental health
activity work e.g. inspections done, wells protected etc. There should also
be a notification system for diseases to which there will be a response
such as cholera, polio and possibly measles. Also for maternal deaths. I
don't want to be notified every case of jaundice if I am not going to do
anything about it.

I do not believe that it is necessary to collect a continuous tally of
diagnoses. A one day survey per month would give the distribution of
diseases seen that could then be combined with the total attendances to
give the needed results. I also do not need to know all of the different
vaccines given DPT1, DPT2, Polio1, Polio2 etc. just the number of kids
starting and leaving the primary course or all the different FP methods
used. If each facility has catchment populations it is then very easy to
calculate utilization rates, coverage rates etc.

If this minimal data was collected and used for facility and district
planning the quality of information could only improve.

At the national level I believe in regular well performed sample surveys
combined with sentinel systems. The Essential drugs program in Zimbabwe has
used this method very successfully to monitor and evaluate the effects of
their programme.

So my response to your questions are:

1. What steps to review and reform? I would suggest start at the bottom
with what is working and build on that.

2. Relationship between HMIS reform and HS Reform. It depends on the model
   of HS reform but it is important to have a simple robust system that is
   useful to the decentralized management units. What I have seen in the
   Philippines is not encouraging.

3. Participatory versus task force approach to development. I believe in a
   participatory process based on low level consultation and iterative
   field testing.

4. Site for the HMIS home. My preference is for the planning or management
   support unit and not the EPI department.

5. Can HMIS be kept simple in the face of growing needs for information? I
   am not convinced of the need just the want. If information is used
   effectively the HMIS system will grow to meet the real need from a
   simple system.

6. Is routine data sufficient? Absolutely not, particularly for national
   level planners. They need to do regular surveys using good sampling
   methods combined with sentinel systems.

7. Accuracy of diagnoses for morbidity reports. This is a concern but in
   the drug use surveys done in over 30 countries we have found that a few
   easily diagnosed conditions predominate. The key is to ensure that the
   minimal history and examination criteria are fulfilled. And even if the
   diagnosis is incorrect the patient is being treated for that condition
   and the drugs are being used.

8. How can data accuracy and consistency be ensured? By rapid analysis and
   feed back. When we running the sentinel system for the EDP in Zimbabwe
   staff at the 40 randomly selected facilities were told to submit the one
   page forms by the 5th, we received them by the 10th, analyzed by 15th,
   sent out feedback reports by the 20th and the expected to receive the
   reports and the next months forms by the 25th. With this system we had
   about 85% response rate from facilities about 75% for districts and
   lower rated for the provinces. When we did a national survey after
   18 months the sentinel sites were within 5% of the national figures!

9. Denominator data is tough but you just have to work at it at the lower
   levels. Get an agreement on the best data available and then review the
   results each year.

10. What happens to vertical programs? They have to integrate and simplify.

11. Evidence of effect. Not that I know of from a large routine system. For
    simple focused systems I have seen some good results.

12. How should data processing be done? At facility level pencil and paper
    and calculator. At district level the same though a computer can be
    used effectively particularly for preparing attractive feedback
    newsletters.

13. Which come first strengthened management capacity or strengthened
    information systems? I would suggest that the two go together.

14. Appropriate reporting and feedback frequency. Facility to and from
    district monthly. District to and from province quarterly, province to
    and from national level quarterly or 6 monthly would probably be
    adequate.

Those are my comments for what they are worth. I do think this is an
important topic that should not be left to computer or data specialists
alone.

Richard Laing

Associate Professor of International Public Health
Boston University School of Public Health
53 Bay State Rd,
Boston MA 02215
Tel +1-617-353-6630
Fax +1-617-353-6330
mailto:richardl@bu.edu

--
Send mail for the `AFRO-NETS' conference to `afro-nets@usa.healthnet.org'.
Mail administrative requests to `majordomo@usa.healthnet.org'.
For additional assistance, send mail to: `owner-afro-nets@usa.healthnet.org'.