WEBVTT

00:00:00.000 --> 00:00:02.400 align:middle line:90%
[MUSIC PLAYING]

00:00:02.400 --> 00:00:07.318 align:middle line:90%


00:00:07.318 --> 00:00:08.860 align:middle line:84%
EME OWOAJE: So in
this section, we'll

00:00:08.860 --> 00:00:15.243 align:middle line:84%
be looking at the importance
of incentives and data quality.

00:00:15.243 --> 00:00:16.660 align:middle line:84%
In this slide,
we're going to look

00:00:16.660 --> 00:00:21.120 align:middle line:84%
at elements of data
quality and these elements

00:00:21.120 --> 00:00:25.770 align:middle line:84%
include validity, reliability,
completeness, precision,

00:00:25.770 --> 00:00:28.260 align:middle line:90%
timeliness, and integrity.

00:00:28.260 --> 00:00:30.960 align:middle line:84%
The validity of the data
is extremely important

00:00:30.960 --> 00:00:33.150 align:middle line:84%
because it shows
whether it's been

00:00:33.150 --> 00:00:36.570 align:middle line:84%
able to capture that which
is supposed to capture

00:00:36.570 --> 00:00:39.510 align:middle line:84%
and with the tools that--
the indicators that we use

00:00:39.510 --> 00:00:41.130 align:middle line:90%
were actually appropriate.

00:00:41.130 --> 00:00:43.650 align:middle line:84%
Then reliability
looks at the ability

00:00:43.650 --> 00:00:48.000 align:middle line:84%
to go back and measure the
same thing by other people

00:00:48.000 --> 00:00:49.970 align:middle line:90%
and get the same results.

00:00:49.970 --> 00:00:51.510 align:middle line:90%
Completeness.

00:00:51.510 --> 00:00:54.400 align:middle line:84%
There are various aspects of
the data that are looked at

00:00:54.400 --> 00:00:58.020 align:middle line:84%
and it's important to get
those various aspects and not

00:00:58.020 --> 00:00:59.190 align:middle line:90%
just part of it.

00:00:59.190 --> 00:01:02.220 align:middle line:84%
So it's-- data complete
this is important.

00:01:02.220 --> 00:01:07.470 align:middle line:84%
Precision looks at the ability
to measure a specific aspect

00:01:07.470 --> 00:01:11.130 align:middle line:84%
that we actually want to
measure for that data.

00:01:11.130 --> 00:01:13.890 align:middle line:84%
Then timeliness looks
at the properties

00:01:13.890 --> 00:01:15.720 align:middle line:84%
that which that data
is being measured.

00:01:15.720 --> 00:01:18.310 align:middle line:84%
And in a campaign
like polio, it's

00:01:18.310 --> 00:01:22.140 align:middle line:84%
important to get the right
data at the right time

00:01:22.140 --> 00:01:23.220 align:middle line:90%
because people--

00:01:23.220 --> 00:01:26.790 align:middle line:84%
children can move around and
it would affect the dominators,

00:01:26.790 --> 00:01:29.730 align:middle line:84%
it would affect the
percentage of coverage.

00:01:29.730 --> 00:01:33.600 align:middle line:84%
Then integrity-- data integrity
is very important because it

00:01:33.600 --> 00:01:35.610 align:middle line:90%
reflects the truth and--

00:01:35.610 --> 00:01:38.910 align:middle line:84%
of the particular data
that is being used.

00:01:38.910 --> 00:01:41.520 align:middle line:84%
So I'd like us to
think about what

00:01:41.520 --> 00:01:46.350 align:middle line:84%
happens if any of these elements
of data quality are not met.

00:01:46.350 --> 00:01:51.480 align:middle line:84%
And I'd like you to relate
this concept of implementation

00:01:51.480 --> 00:01:56.040 align:middle line:84%
fidelity-- that is the degree
to which an intervention is

00:01:56.040 --> 00:02:00.030 align:middle line:84%
delivered the way it's supposed
to be delivered like the polio

00:02:00.030 --> 00:02:04.400 align:middle line:84%
program with the plan-- the
microplans and the activities

00:02:04.400 --> 00:02:07.950 align:middle line:84%
of the health workers are
supposed to actually carry out.

00:02:07.950 --> 00:02:11.430 align:middle line:84%
So each of those
activities is linked

00:02:11.430 --> 00:02:16.020 align:middle line:84%
to data and the fidelity--
the implementation fidelity

00:02:16.020 --> 00:02:18.750 align:middle line:84%
is strongly linked to
the initial microplan.

00:02:18.750 --> 00:02:21.270 align:middle line:84%
So I'd like you to think
about this and relate it.

00:02:21.270 --> 00:02:24.580 align:middle line:90%


00:02:24.580 --> 00:02:30.340 align:middle line:84%
So I'd like us to talk about
what causes poor quality data.

00:02:30.340 --> 00:02:33.250 align:middle line:84%
Before you look at
the various options

00:02:33.250 --> 00:02:37.030 align:middle line:84%
that we've given that you need
to discuss amongst yourselves

00:02:37.030 --> 00:02:39.280 align:middle line:84%
and come up with
some of these things

00:02:39.280 --> 00:02:43.750 align:middle line:84%
that, in your experience,
result in poor quality data.

00:02:43.750 --> 00:02:48.340 align:middle line:84%
So from my end, we're looking
at no standards for data

00:02:48.340 --> 00:02:50.830 align:middle line:84%
collection or the
standards are not

00:02:50.830 --> 00:02:54.010 align:middle line:84%
enforced as they're
supposed to be enforced.

00:02:54.010 --> 00:02:57.460 align:middle line:84%
This usually occurs
when there's no training

00:02:57.460 --> 00:02:59.980 align:middle line:90%
or the training is inadequate.

00:02:59.980 --> 00:03:03.010 align:middle line:84%
So it is essential
that the health workers

00:03:03.010 --> 00:03:07.030 align:middle line:84%
or the people involved in data
collection are well-trained.

00:03:07.030 --> 00:03:09.460 align:middle line:84%
Then there's also the
aspect of integration

00:03:09.460 --> 00:03:13.930 align:middle line:84%
of data from systems with
different data standards.

00:03:13.930 --> 00:03:17.920 align:middle line:84%
So we need to harmonize the
data standards to ensure

00:03:17.920 --> 00:03:22.870 align:middle line:84%
that the quality of the data
that we collect is acceptable.

00:03:22.870 --> 00:03:26.500 align:middle line:84%
Then, data quality issues are
perceived as time consuming

00:03:26.500 --> 00:03:28.550 align:middle line:90%
and expensive to fix.

00:03:28.550 --> 00:03:31.750 align:middle line:84%
And that is something that
training can take care of

00:03:31.750 --> 00:03:34.300 align:middle line:84%
and then reiterating
to the people that

00:03:34.300 --> 00:03:38.470 align:middle line:84%
are involved in data collection
that it is important,

00:03:38.470 --> 00:03:43.570 align:middle line:84%
that we need to get that data
right right from the beginning,

00:03:43.570 --> 00:03:47.740 align:middle line:84%
also to explain to them that
it's more expensive to come

00:03:47.740 --> 00:03:50.260 align:middle line:90%
back and try and fix it.

00:03:50.260 --> 00:03:53.380 align:middle line:84%
So today's work
focus people may not

00:03:53.380 --> 00:03:56.950 align:middle line:84%
be incentivized for
high quality data

00:03:56.950 --> 00:04:00.290 align:middle line:84%
or they may just have
perverse incentives.

00:04:00.290 --> 00:04:03.820 align:middle line:84%
So if people are rewarded
for showing high coverage,

00:04:03.820 --> 00:04:07.060 align:middle line:84%
they may adjust the data
to show high coverage

00:04:07.060 --> 00:04:10.180 align:middle line:84%
and I'm sure most of you
have experienced that.

00:04:10.180 --> 00:04:13.270 align:middle line:84%
If people are told, oh,
you'll get extra money

00:04:13.270 --> 00:04:16.829 align:middle line:84%
based on the number of
children that you immunize--

00:04:16.829 --> 00:04:20.005 align:middle line:84%
so they might just
go out there and--

00:04:20.005 --> 00:04:23.530 align:middle line:84%
in the case of polio-- when
they we're using the oral polio

00:04:23.530 --> 00:04:28.390 align:middle line:84%
vaccine, just keep on using the
vaccine drops of the vaccine

00:04:28.390 --> 00:04:30.580 align:middle line:84%
and write it up that
they immunized children

00:04:30.580 --> 00:04:33.400 align:middle line:84%
when they actually did
not, just because they were

00:04:33.400 --> 00:04:36.660 align:middle line:90%
being incentivized to do so.

00:04:36.660 --> 00:04:40.340 align:middle line:84%
So in terms of implementation
fidelity, which I had already

00:04:40.340 --> 00:04:44.060 align:middle line:84%
talked about before, the
degree to which an intervention

00:04:44.060 --> 00:04:47.210 align:middle line:84%
is delivered as it's
intended, polio data

00:04:47.210 --> 00:04:50.540 align:middle line:84%
illustrates the root
cause of poor fidelity,

00:04:50.540 --> 00:04:52.400 align:middle line:90%
including pervasive incentives.

00:04:52.400 --> 00:04:55.070 align:middle line:84%
We've talked about that
with the oral polio virus

00:04:55.070 --> 00:04:58.800 align:middle line:84%
that people could just go out
and give two drops to what--

00:04:58.800 --> 00:05:01.520 align:middle line:84%
even put them on the
ground just because they

00:05:01.520 --> 00:05:03.800 align:middle line:84%
were being incentivized
to immunize

00:05:03.800 --> 00:05:05.570 align:middle line:90%
as many children as possible.

00:05:05.570 --> 00:05:07.280 align:middle line:84%
So we need to look
at the strategies

00:05:07.280 --> 00:05:11.840 align:middle line:84%
for tackling this, including
technology-based strategies.

00:05:11.840 --> 00:05:16.400 align:middle line:84%
And then this is also important
in other immunization programs.

00:05:16.400 --> 00:05:20.810 align:middle line:84%
It's-- there's no use saying
that people have immunized

00:05:20.810 --> 00:05:22.160 align:middle line:90%
children when they haven't.

00:05:22.160 --> 00:05:23.720 align:middle line:90%
For instance, in measles--

00:05:23.720 --> 00:05:26.120 align:middle line:84%
people go out on
measles campaigns

00:05:26.120 --> 00:05:28.140 align:middle line:84%
and they say, oh,
they have immunized

00:05:28.140 --> 00:05:30.755 align:middle line:84%
this number of children
and, at the end of the day,

00:05:30.755 --> 00:05:32.630 align:middle line:84%
the truth of the matter
is that they have not

00:05:32.630 --> 00:05:34.172 align:middle line:84%
immunized these
children, but there's

00:05:34.172 --> 00:05:36.660 align:middle line:90%
a false sense of security.

00:05:36.660 --> 00:05:39.350 align:middle line:84%
So we need to address
this and make sure

00:05:39.350 --> 00:05:42.110 align:middle line:84%
that there's fidelity
in the implementation

00:05:42.110 --> 00:05:44.640 align:middle line:90%
of our various programs.

00:05:44.640 --> 00:05:47.850 align:middle line:84%
So we're looking at
the use of technology

00:05:47.850 --> 00:05:51.780 align:middle line:84%
versus a manual way of
estimating the number

00:05:51.780 --> 00:05:54.120 align:middle line:90%
of children in Afghanistan.

00:05:54.120 --> 00:05:58.500 align:middle line:84%
So if you're looking at this
slide-- in the 2017 GI--

00:05:58.500 --> 00:06:02.040 align:middle line:84%
Geographical Information
Systems, GIS,

00:06:02.040 --> 00:06:04.740 align:middle line:84%
under five estimates of
the population of children

00:06:04.740 --> 00:06:09.270 align:middle line:84%
in Afghanistan was about 5.9
million-- almost 6 million,

00:06:09.270 --> 00:06:13.590 align:middle line:84%
whereas the under five
polio targets in Afghanistan

00:06:13.590 --> 00:06:17.440 align:middle line:90%
was 10.2 million.

00:06:17.440 --> 00:06:20.340 align:middle line:84%
So what do you think,
in your opinion,

00:06:20.340 --> 00:06:24.090 align:middle line:84%
might have resulted in the
divergence in the numbers

00:06:24.090 --> 00:06:29.280 align:middle line:84%
and why might the geographical
information system numbers be

00:06:29.280 --> 00:06:30.930 align:middle line:90%
lower?

00:06:30.930 --> 00:06:35.280 align:middle line:84%
What is GIS modeling missing
that the administrative data

00:06:35.280 --> 00:06:36.840 align:middle line:90%
has?

00:06:36.840 --> 00:06:39.570 align:middle line:84%
What's the importance
of the knowledge

00:06:39.570 --> 00:06:41.975 align:middle line:84%
that the administrative
system has?

00:06:41.975 --> 00:06:44.640 align:middle line:84%
And why do you think that
the administrative numbers

00:06:44.640 --> 00:06:45.630 align:middle line:90%
might be higher?

00:06:45.630 --> 00:06:47.910 align:middle line:84%
Could it be as a result
of the incentives that

00:06:47.910 --> 00:06:49.900 align:middle line:90%
are being given?

00:06:49.900 --> 00:06:51.430 align:middle line:84%
This could
potentially have to do

00:06:51.430 --> 00:06:54.400 align:middle line:84%
with people inflating
numbers so that they

00:06:54.400 --> 00:06:55.540 align:middle line:90%
can get more resources.

00:06:55.540 --> 00:06:56.920 align:middle line:90%
We've talked about this.

00:06:56.920 --> 00:06:59.200 align:middle line:84%
Could you give examples
in your old environment

00:06:59.200 --> 00:07:01.290 align:middle line:90%
how this might work?

00:07:01.290 --> 00:07:06.260 align:middle line:84%
So what would you do with
this if you were a planner?

00:07:06.260 --> 00:07:08.840 align:middle line:84%
For further discussion,
here's a comment

00:07:08.840 --> 00:07:10.990 align:middle line:90%
from the Indian context.

00:07:10.990 --> 00:07:13.160 align:middle line:84%
There are other reasons
for polio planners

00:07:13.160 --> 00:07:18.610 align:middle line:84%
to use higher figures than any
demographic or GIS estimates.

00:07:18.610 --> 00:07:22.970 align:middle line:84%
In India's program, we always
took the reported coverage

00:07:22.970 --> 00:07:26.630 align:middle line:84%
as the base for planning
the next campaign.

00:07:26.630 --> 00:07:29.960 align:middle line:84%
Though we normally expect
under five population

00:07:29.960 --> 00:07:33.860 align:middle line:84%
to be 14% of the
total population,

00:07:33.860 --> 00:07:38.930 align:middle line:84%
polio program population
used to be around 20%.

00:07:38.930 --> 00:07:42.230 align:middle line:84%
This was a very practical
approach-- many reasons

00:07:42.230 --> 00:07:43.610 align:middle line:90%
for that.

00:07:43.610 --> 00:07:48.680 align:middle line:84%
Vaccinators would like to err
on the overage rather than

00:07:48.680 --> 00:07:52.280 align:middle line:84%
the underage, really
giving overestimates

00:07:52.280 --> 00:07:57.680 align:middle line:84%
rather than underestimates
for children five years below.

00:07:57.680 --> 00:08:00.890 align:middle line:84%
So they would rather have
more population on ground

00:08:00.890 --> 00:08:03.320 align:middle line:90%
than predicted by the estimates.

00:08:03.320 --> 00:08:06.800 align:middle line:84%
Some accounts
unaccounted wastage--

00:08:06.800 --> 00:08:11.210 align:middle line:84%
planning vaccine with
higher figures for the fear

00:08:11.210 --> 00:08:16.080 align:middle line:84%
of running out of
vaccine and [INAUDIBLE]..

00:08:16.080 --> 00:08:18.210 align:middle line:84%
How can public health
programs better

00:08:18.210 --> 00:08:20.970 align:middle line:90%
incentivize quality data?

00:08:20.970 --> 00:08:24.360 align:middle line:84%
I'd like you to reflect
on this question.

00:08:24.360 --> 00:08:28.110 align:middle line:84%
The polio program has not fully
cracked this particular nut.

00:08:28.110 --> 00:08:31.200 align:middle line:84%
And I think it would be an
excellent final project for one

00:08:31.200 --> 00:08:34.309 align:middle line:84%
of you who might want
to look in that area

00:08:34.309 --> 00:08:36.549 align:middle line:90%
and who's interested in data.

00:08:36.549 --> 00:08:39.780 align:middle line:84%
One thing that has been
tried is using technology

00:08:39.780 --> 00:08:41.440 align:middle line:90%
to achieve this.

00:08:41.440 --> 00:08:44.880 align:middle line:90%
Let's look at that next.

00:08:44.880 --> 00:08:47.630 align:middle line:90%
[MUSIC PLAYING]

00:08:47.630 --> 00:08:53.000 align:middle line:90%