Okay. The big picture. If there’s only two concepts you
remember from this entry, I’d hope that they are “malaria prevention” and “long
lasting insecticide-treated nets (LLIN)”.
Those are the end and the means of the Tanzania National Voucher Scheme
(TNVS), which I am involved in.
TNVS, as the name suggests, implements a voucher
system. When a pregnant woman or infant (the
ones at highest risk of contracting malaria) visit a health clinic, the health
worker issues them a voucher. This
voucher can be taken to a retailer to obtain a LLIN at a subsidized cost (funding
by USAID and DFID). The price paid by
the beneficiary is very affordable, and I think it’s more of a token amount just
so that they have some skin in the game.
Besides malaria prevention, the program also aims to
jumpstart the market for mosquito nets. This
includes creating both an awareness of and the demand for mosquito nets. In addition to driving down net prices
through economies of scale, there is a second strategy. Initially, the program partnered with a
single net manufacturer. However, we are
currently working on introducing a second supplier into the program. The hope
is that providing the beneficiary with a choice in net type will create a
little competition between the manufacturers.
This will motivate manufacturers to make better nets at lower
costs. Don’t worry, LLINs must be
inspected and certified which ensures no sacrifice in quality. The desire to drive down net prices stems
from a desire to make nets more affordable to low income families after the
voucher system is removed. The hope is
that the LLIN market will remain even after TNVS shuts down. If it all works according to plan…net
manufacturers will be creating more jobs and income, retailers will also have an
increased income from selling nets, and the general public will have an
opportunity to better protect themselves against malaria.
Okay, now for some of the challenges. I’ll give just two examples of the types of
projects I have been working on. The
first one is automated reporting. The second
is market actor profiling. Grab a
coffee!!
TNVS is a nation-wide program with around 5,000 clinics and 6,000 retailers redeeming just over 370,000 vouchers every month. MEDA TZ handles
the logistics of the entire program. MEDA
TZ only has about 25 employees in the office (there is also another 10 field
officers with drivers scattered throughout Tanzania conducting training). It’s a great opportunity for technology to
help ease the workload!
In order for a program like TNVS to thrive, it’s important
to know which locations are succeeding and which ones are failing. This feedback is extremely useful to learn
from success and nurse weaknesses. It’s
kind of like a strategy game. MEDA TZ
keeps a pulse on the health of the program through weekly reports. Weekly reports guide field officers to the
locations which require attention.
Performance indicators in the weekly report include figures
like the number of issued vouchers, the number of redeemed vouchers, and the
percentage of issued vouchers which were redeemed. Manually gathering and summarizing this information
for all the clinics and retailers can easily take a half day of work. Every
week. Similarly, the payment report,
which tallies the number of nets distributed and the money owed to the net supplier,
will take a half day to compile. It’s
done every other week. To put it in
perspective, one employee (I think it was supposed to be me) can spend well
over a week of every month compiling reports.
My first project was to automate the process so that the reports could
be generated by a button click and save a lot of time and manual work.
But IT skills can do much more than just improve the
efficiency of report generation.
The paper voucher system suffers from a problem. The problem is its limited visibility of what’s
actually taking place in the field. For
instance, we are unaware when a clinic runs out of voucher stock and stops issuing
vouchers to patients. We are unaware
when a retailer stops redeeming vouchers because they have run out of
nets. Furthermore, we can’t detect if a
clinic worker and retailer are colluding together to steal nets – they could get
together and make up fictitious beneficiaries to issue vouchers to and redeem
vouchers from…and then keep the nets for themselves. The voucher system needs a way to extend its
sense organs into market actor transactions.
This was the motivation behind the eVoucher system. It is an sms based tracking system which documents
market transactions. Clinic workers must
use a cellphone (everyone’s got one!) to text MEDA’s shortcode when issuing a
voucher. Retailers do the same to inform
MEDA of a voucher redemption. A retailer
will only be restocked with nets for voucher redemptions which were reported. A voucher redemption will only be successfully
reported if its issuance was also reported.
Basically, the system works…and we have information about the time and
location of every issuance and redemption.
Information which allows us to profile clinic/retailer behaviour. This is known as data mining, and it’s
another project that I’ve been involved in.
I create algorithms which try to determine when a retailer is out of net
stock. I also create algorithms which
try to determine when a retailer or clinic is engaging in fraud.
Okay okay, I’m beginning to realize how incredibly long this
story is. I’m curious how many people
made it to the end. Anyway, I hope I’ve given
you a taste of my work and satisfied some of the curiosities floating around. Let it be known that it’s not ONLY exploratory
adventures for me. I work hard too!