I think you are better positioned to comment on this on what approach we should use. In the file attached, you can see that there are 49 different conversion factors to come up with approximations of different data elements like expected pregnancies and infant population.
If these factors were the same for all regions, we could have added them as constants and used them to define the denominators for the indicators. However, as you can see each region has different conversion factors and there’s a big difference among them which makes defining indicators very difficult.
I see two options here:
Capturing the population and calculating each data elements for each facility by multiplying it with its respective region’s factor (hence coming up with at least 49 data elements)
Defining different indicators for different regions using different constants as conversion factors (which I think may complicate things)
In my view the first option is the best option for us because once we capture the total population, it is easier to generate data for the other data elements using these conversion factors. Of course, as I stated in my mail yesterday, there are two sets of population data used in the structure (MoH uses the official data from Central Statistical Authority dis-aggregated by Wereda while the regions use data collected from the ground. The data the regions use may be the most accurate but we have to accomodate both options I think because the two parties use their own population data to come up with the population figures, effectively having two different population data (administrative population and facility catchment population)
If we are using two sets of population data, we may have to define the same for all the 49 other data elements with the factors as well.
Would you please deliberate over it and suggest? Once we have a concrete plan, we can discuss with M&E people here on how to proceed.
This is the summary of our discussions regarding the conversion factors and other issues
Regarding the conversion factor, we take the first option you suggested.
Capturing the population and calculating each data elements for each facility by multiplying it with its respective region’s factor (hence coming up with at least 49 data elements)
For the two level indicators of vaccine wastage rate, we use the expression ((100*Dose opened)-Dose given)/Dose opened
The indicators should be revised. most of the denominators are defined because of lack population data and estimates
Seid please contact those who are working with Phem (IDSR) regarding how often they collect data. If weekly, is that including Pagume. We should also ask if they are using the International or Ethiopian calendar when the weekly data collection. Because as IDSR is an international program, there is a possibility that they are using the International Calendar to compare data across countries
Seid please contact Solomon from gate foundation and respond to the emails of Dykki
Write your technical problems directly to the mailing list
I think you are better positioned to comment on this on what approach we should use. In the file attached, you can see that there are 49 different conversion factors to come up with approximations of different data elements like expected pregnancies and infant population.
If these factors were the same for all regions, we could have added them as constants and used them to define the denominators for the indicators. However, as you can see each region has different conversion factors and there’s a big difference among them which makes defining indicators very difficult.
I see two options here:
Capturing the population and calculating each data elements for each facility by multiplying it with its respective region’s factor (hence coming up with at least 49 data elements)
Defining different indicators for different regions using different constants as conversion factors (which I think may complicate things)
In my view the first option is the best option for us because once we capture the total population, it is easier to generate data for the other data elements using these conversion factors. Of course, as I stated in my mail yesterday, there are two sets of population data used in the structure (MoH uses the official data from Central Statistical Authority dis-aggregated by Wereda while the regions use data collected from the ground. The data the regions use may be the most accurate but we have to accomodate both options I think because the two parties use their own population data to come up with the population figures, effectively having two different population data (administrative population and facility catchment population)
If we are using two sets of population data, we may have to define the same for all the 49 other data elements with the factors as well.
Would you please deliberate over it and suggest? Once we have a concrete plan, we can discuss with M&E people here on how to proceed.
I tried to think over it, but I don’t think the current Indicator implementation does not accommodate this requirement.
Wastage rate is of course a strange indicator (they could have created an indicator called usage rate). But somewhere somebody may come up with another strange indicator like Infant survival rate which may be calculated as:
1000 - (Number of <1 dead / Number of live births)
Please check again because my calculations could not show me a viable path. If your formula works, it will work for all the ‘strange’ indicators as well.
This is the summary of our discussions regarding the conversion factors and other issues
Regarding the conversion factor, we take the first option you suggested.
Capturing the population and calculating each data elements for each facility by multiplying it with its respective region’s factor (hence coming up with at least 49 data elements)
For the two level indicators of vaccine wastage rate, we use the expression ((100*Dose opened)-Dose given)/Dose opened
The indicators should be revised. most of the denominators are defined because of lack population data and estimates
Seid please contact those who are working with Phem (IDSR) regarding how often they collect data. If weekly, is that including Pagume. We should also ask if they are using the International or Ethiopian calendar when the weekly data collection. Because as IDSR is an international program, there is a possibility that they are using the International Calendar to compare data across countries
Seid please contact Solomon from gate foundation and respond to the emails of Dykki
Write your technical problems directly to the mailing list
I think you are better positioned to comment on this on what approach we should use. In the file attached, you can see that there are 49 different conversion factors to come up with approximations of different data elements like expected pregnancies and infant population.
If these factors were the same for all regions, we could have added them as constants and used them to define the denominators for the indicators. However, as you can see each region has different conversion factors and there’s a big difference among them which makes defining indicators very difficult.
I see two options here:
Capturing the population and calculating each data elements for each facility by multiplying it with its respective region’s factor (hence coming up with at least 49 data elements)
Defining different indicators for different regions using different constants as conversion factors (which I think may complicate things)
In my view the first option is the best option for us because once we capture the total population, it is easier to generate data for the other data elements using these conversion factors. Of course, as I stated in my mail yesterday, there are two sets of population data used in the structure (MoH uses the official data from Central Statistical Authority dis-aggregated by Wereda while the regions use data collected from the ground. The data the regions use may be the most accurate but we have to accomodate both options I think because the two parties use their own population data to come up with the population figures, effectively having two different population data (administrative population and facility catchment population)
If we are using two sets of population data, we may have to define the same for all the 49 other data elements with the factors as well.
Would you please deliberate over it and suggest? Once we have a concrete plan, we can discuss with M&E people here on how to proceed.
To convert it to percentage probably you have to divide it by a denominator and then multiply it by 100. In this case the way the indicator is conceptualized, the fact that from a start something is subtracted from 100, the final result (99.42) is already percentage.
If this value doesn’t make sens, then we have to devise another expression. Otherwise, the formula is Selam gave you is correct (at least mathematically).
And yes, in DHIS2 we can put any mathematical expression and handle complex things. We don’t have to limit ourselves with the simple numerator/denominator expression.
This is the summary of our discussions regarding the conversion factors and other issues
Regarding the conversion factor, we take the first option you suggested.
Capturing the population and calculating each data elements for each facility by multiplying it with its respective region’s factor (hence coming up with at least 49 data elements)
For the two level indicators of vaccine wastage rate, we use the expression ((100*Dose opened)-Dose given)/Dose opened
The indicators should be revised. most of the denominators are defined because of lack population data and estimates
Seid please contact those who are working with Phem (IDSR) regarding how often they collect data. If weekly, is that including Pagume. We should also ask if they are using the International or Ethiopian calendar when the weekly data collection. Because as IDSR is an international program, there is a possibility that they are using the International Calendar to compare data across countries
Seid please contact Solomon from gate foundation and respond to the emails of Dykki
Write your technical problems directly to the mailing list
I think you are better positioned to comment on this on what approach we should use. In the file attached, you can see that there are 49 different conversion factors to come up with approximations of different data elements like expected pregnancies and infant population.
If these factors were the same for all regions, we could have added them as constants and used them to define the denominators for the indicators. However, as you can see each region has different conversion factors and there’s a big difference among them which makes defining indicators very difficult.
I see two options here:
Capturing the population and calculating each data elements for each facility by multiplying it with its respective region’s factor (hence coming up with at least 49 data elements)
Defining different indicators for different regions using different constants as conversion factors (which I think may complicate things)
In my view the first option is the best option for us because once we capture the total population, it is easier to generate data for the other data elements using these conversion factors. Of course, as I stated in my mail yesterday, there are two sets of population data used in the structure (MoH uses the official data from Central Statistical Authority dis-aggregated by Wereda while the regions use data collected from the ground. The data the regions use may be the most accurate but we have to accomodate both options I think because the two parties use their own population data to come up with the population figures, effectively having two different population data (administrative population and facility catchment population)
If we are using two sets of population data, we may have to define the same for all the 49 other data elements with the factors as well.
Would you please deliberate over it and suggest? Once we have a concrete plan, we can discuss with M&E people here on how to proceed.
Anyway, the way to do this in DHIS2 is to define an “Indicator type” with a factor of 100 (like this from the demo site)
and then the indicator would be
(Doses used - Doses opened) / Doses opened
(60-35) / 60 = 60 = 0.41667 which is equivalent to 41.667%
Just be sure you indicator (if it is a percent) has an indicator type of “Percent” (which you will need to create if it does not exist).
Regards,
Jason
···
On Thu, Aug 27, 2015 at 9:50 PM, Abyot Gizaw abyota@gmail.com wrote:
Hi Seid,
Selam’s formula is correct - it works.
5965 / 60 is 99.42 not 9942 %
To convert it to percentage probably you have to divide it by a denominator and then multiply it by 100. In this case the way the indicator is conceptualized, the fact that from a start something is subtracted from 100, the final result (99.42) is already percentage.
If this value doesn’t make sens, then we have to devise another expression. Otherwise, the formula is Selam gave you is correct (at least mathematically).
And yes, in DHIS2 we can put any mathematical expression and handle complex things. We don’t have to limit ourselves with the simple numerator/denominator expression.
This is the summary of our discussions regarding the conversion factors and other issues
Regarding the conversion factor, we take the first option you suggested.
Capturing the population and calculating each data elements for each facility by multiplying it with its respective region’s factor (hence coming up with at least 49 data elements)
For the two level indicators of vaccine wastage rate, we use the expression ((100*Dose opened)-Dose given)/Dose opened
The indicators should be revised. most of the denominators are defined because of lack population data and estimates
Seid please contact those who are working with Phem (IDSR) regarding how often they collect data. If weekly, is that including Pagume. We should also ask if they are using the International or Ethiopian calendar when the weekly data collection. Because as IDSR is an international program, there is a possibility that they are using the International Calendar to compare data across countries
Seid please contact Solomon from gate foundation and respond to the emails of Dykki
Write your technical problems directly to the mailing list
I think you are better positioned to comment on this on what approach we should use. In the file attached, you can see that there are 49 different conversion factors to come up with approximations of different data elements like expected pregnancies and infant population.
If these factors were the same for all regions, we could have added them as constants and used them to define the denominators for the indicators. However, as you can see each region has different conversion factors and there’s a big difference among them which makes defining indicators very difficult.
I see two options here:
Capturing the population and calculating each data elements for each facility by multiplying it with its respective region’s factor (hence coming up with at least 49 data elements)
Defining different indicators for different regions using different constants as conversion factors (which I think may complicate things)
In my view the first option is the best option for us because once we capture the total population, it is easier to generate data for the other data elements using these conversion factors. Of course, as I stated in my mail yesterday, there are two sets of population data used in the structure (MoH uses the official data from Central Statistical Authority dis-aggregated by Wereda while the regions use data collected from the ground. The data the regions use may be the most accurate but we have to accomodate both options I think because the two parties use their own population data to come up with the population figures, effectively having two different population data (administrative population and facility catchment population)
If we are using two sets of population data, we may have to define the same for all the 49 other data elements with the factors as well.
Would you please deliberate over it and suggest? Once we have a concrete plan, we can discuss with M&E people here on how to proceed.