 The main problem is about getting the target population denominated data because the typical indicator to use when you compare across facilities in the district is about how you are performing related to various health programs like immunization, ANC, etc. And in order to do that you need to have a target population and you need to have of course also enumerated data. We have various problems to actually find out how to do this target population. What to do for example in Rwanda is that since the catchment areas are not following directly the numerator areas for the population census in the district and for the facilities then they have to put the data, I mean the population data, the target data in using their Excel. So here we need to find a way to actually populate DHRs too with estimated catchment population and what has been suggested is to use alternative methods in cases like this also like using ANC first visit as a relatively good sample for the target populations. Another approach which actually it was used in South Africa a long time back and also in Togo is to have head counts for the HMIS reporting from the facilities and distribute the population in the district based on the weight of this head count and how big percentage it's facility have of the total. So that there are different alternative methods and we can just show this example from Mozambique where they use this scorecard of immunization coverage by facility at the district data use meeting and what was the problem in getting that into the DHRs too and why did they use Excel instead of generated from the database. The problem was that the data was not in the DHRs because in Mozambique they are not really agreeing totally on what is the population data in the districts etc. So what they do is then that the district office in each district just make up but not make up estimate population target population for each facility and distribute it. So that is then done locally and also they do it in a special way they are using the target number and not the total number of the target population so that's also a complication that makes it a bit difficult to do it automatically but how that is solved will or being solved that's something that Seferino will tell us. So it was just a brief introduction to the problem of estimating and target population at the local level in the communities and in the facilities in the district health system. So yeah over to the next one thanks I'll stop sharing then Kofi you can start. Okay I'm not seeing my presentation screen just hold on a minute let me use okay having some challenges finding back my screen okay there is it okay yep are you seeing my screen yeah we are it's not in presentation mode but we're seeing your screen okay hi everyone if you click present in the top right hand corner okay okay that's great thanks Kofi. You're welcome. Hi everyone I'm happy to share with you the presentation designed by the East Western Central Africa team and we're going to try to deep into what Jordan just said concerning Togo and Mali that are the outlines we're going to provide a little bit context and then explain how we set up the population in Togo how we updated the target population data annually we're going to share some specificity from Mali and how are the what are the key challenges and way forward and then have a summary to conclude. So we all know how critical is population data regarding decision making as Jordan said when we have some inaccuracy it may result in efficiency inefficiency and program management essentially regarding planning you cannot determine what problems are priority problems you cannot determine the right activities you cannot allocate your resources well and then you have challenges and performance evaluation essentially regarding setting up your targets and your indicators and also when you're conducting the research researches you face some biases in your results and so while counter-level population data is more common place we will face challenges when it comes to sub-level data at these written and then facility areas so in Togo how do we take this challenge so we do each 10 years a general census of the population and houses so we have conducted three of those GCPH and the last one was in 2010 and it was the third one we are currently preparing for them the fourth one and we had not succeeded in doing it last year as it was planned because we were facing some financial challenges and COVID-19 was also there so in that GCPH the data is structured per villages and sometimes per sub-villages which we call enumeration areas so we have the headcounts per villages and for sub-villages and then at the facility level health facility level each health facility has a list of its catchment villages so we can easily reconstitute the facility data adding up the villages of catchment areas that are in that facility list so that is how we constitute at the basic level from the GCPH data the target population for each health facility and for each health program because not only for the facilities but also for them for the programs we do have proportions so we do have the headcount per villages and sub-villages but we also have headcounts for each group ages and we have group ages and then we have that per sex so that we can we can have the proportion for example of adolescent girls we can have the population of children under one or children under five and also all those populations are estimated per group ages and per sex in the GCPH data so from this GCPH data the National Statistics Institute we call it inside here in Togo made projections for the upcoming 10 years based on natural growth rates and then those projections are made annually and you have the big document for the 10 upcoming years and then the data from the GCPH are projected over each year and then they make three scenario low level scenario middle level scenario and then high level scenario and at the MOH we use medium scenario each year because we don't have any indication to take the the high level or the lower level and this population is updated each year from that projections so we take the projection of them from the GCPH and then we update the population of these rates and then once the population of this rate is updated facility catchment area population was also updated based based on village we have to to notice that inside the district slight adjustments are made if needed so they use the head counts that we make regularly either during prior prior to LNI and distribution mosquito nest distribution or prior to negative tropical diseases drug distribution so we make some we make head counts that is made by community health workers village by village and then we use that data to make slight adjustment inside the district so that we can make sure that we don't have very very very low or very high population inside the facilities and formally the the scenario or the process is almost the same but their last GCPH was in 1929 the update is due each 10 years but because of insecurity in finance gap they have not been able to complete it so they do the same process and but they don't have the kind of institute of national institute of statistic to perform that that projection and then and they also undergo the annual update based on the population estimates what are the key challenges when you use those these kind of system it is population mobility as I said you may notice that for from year to years you have slight changes in the head counts because of population mobilities one other one other challenge is the result the final result of GCPH because of financial consideration or political consideration we may face some challenges in publishing those results so generally we take over a year to collect the data but it takes one or two years to have the the final result of that because there are a lot a lot of discussion a lot of data analysis before you get the final result of GCPH the third challenge is the creation of new villages new health facilities so it's kind of made to adjust the the catchment area of all health facilities so you you need some time to split the village lease so that you can make sure that this new health facility have catchment areas and as inside the district you have adjustment to be made sometimes you can you can come to outlier results result in certain educators for example you can have DTP that is above 100% or some some district that our facility that are complaining that the denominator is too high or is too low so this is because of those adjustments so that we need to make sure that adjustment that are made inside this which are appropriate enough and it we we may face sometimes disparity in certain problems because they don't use the the official data and they they tend to to to create or to estimate their own population essentially when it comes to certain donors that are very demanding on having very very narrowed population so sometimes you can face challenges when you have to compare data across programs because you don't have all the time the same denominators so what can we conclude from that and the the first thing which should take over is periodic national census is the basic when you have that 10-year GCPH then you have a strong basis for the next projections for the upcoming years but when you don't have that you kind of you kind of you kind of ending in estimate and estimate and that you lose accuracy the second thing is you need a robust national reference institution to issue the projections if you don't have that official projection so this is the door open for everybody to create his target population and then things run in many directions the third thing is you have to to have for each facility its health facility you need a catchment area that is populated by a village list when you have a village list for health facility then your job is easier when it comes to adjust and estimate and finally you have to to have appropriate district agility to make slight adjustment within district based on regular head counts that we have ahead of certain interventions I'm going to stop there and then just show you an image of what kind of what kind of data we have each year do you see my screen so this is the example of the population that has been updated for one of our districts you have the the district this is the all those all the the facilities that we have and then you have the the population of 2020 the updated population of 2021 that is shared among the health facilities and then you have the main group main group for programs so you have the internet health care api and you have birth birth attendance and then you have a family planning and mother to child transmission preventing all those proportions that had been estimated from the gcph that are applied to determine to determine the program target for the year thank you so much for hearing me and i'm heading over to elaine elaine i'm done elaine or hi do hear me we can hear you seferino are you there are you ready to be next to present thank you so much kofi it was very very interesting elaine do you have a contact with seferino looks like elaine might have left briefly but elaine you should be on co-host again now you can unmute yourself oh thank you very much sorry i was i was not co-host i wasn't able to unmute yes so thank you very much kofi and we'll keep the questions until the end but if people could post questions within the chat channel in in the meantime and we can deal with them if we time at the end and so is seferino on hello hello good afternoon hello seferino so i will hand over to you seferino um and you should be happy to share that's great um that's great i hope you can see my slides yes that's fine thank you very much for this opportunity to be able to contribute to the discussion around the the the the the denominator problem um i think most most of the issues that we have been facing here in mozambique have been highlighted by urn and also by kofi so uh they they you know we we know uh that the the the denominator is the the key to measure surface interventions and then almost all the programs uh when it comes to evaluating the intervention they rely on denominators uh and most of them they are they are they are calculated or they are they are uh yeah so all all these uh let's say the the the indicators are calculated based on that information which is uh they they are receiving uh those estimates from usual national uh boards like national statistics we are here for example in mozambique the municipal health receives that information from the national statistics and uh sometimes in mozambique they receive in the beginning of the year uh through the the hMIS unit but we do have countries for them they don't they don't get that uh on on periodically and and then they don't even update they they they they they system on time and then uh most of the outcomes that when we when we come to the looking at the out the data is the use of the data the the the effort that we put on our uh DHIS to uh dashboards is sometimes not it doesn't have the the the the desired outcome because uh when you when you go to the dashboard you won't see is do you miss some indicators there uh visualization that are not presented uh because uh the that the information uh that is not as the the users are expecting that is the reason for example you are mentioned during the the work that we did here in mozambique when we went to the facilities we found that uh some of them they download data from DHIS too and then they do their calculations outside because when okay when they get to the system they don't find what they they they they they what they they would like to use at the denominator to measure the services based on uh they uh definition for example of what they consider denominator for that specific area so that is is a is a is a is a challenge that we we we face when we when it comes to implementation of these systems and special that has impact on the use of the data so we found in the systems we we we usually uh we often see on dataset that are created and then storing on population estimates by by by district and uh what we are doing as part of addressing some of these uh challenges is to uh design uh the datasets that can store uh that will be used to store target populations uh this by sub-national depending on the what we were were who they provided for example during the the covid vaccine we did have experience from the countries we had to design um the the datasets that were going they are storing information related to the target population for that for this camp for the campaign for example so that the at the national level at the province level this is they are able to measure how much they have they have accomplished uh with regard to the what we were expecting or what was written in the in the in the in the uh covax plan uh um immunization plan uh it's also mentioned about this the definition of these target groups and the data which is entered depending on the analysis period the way as i mentioned now uh for the for the api we can consider we did we did define some target group like for example health professionals people that that are living for them that are the general public and there's people that are working for in the point of entry all of these are target groups and then they are they are they we usually add them in the system based on the analysis period if they the the campaign for three three weeks three months or six months we have to discuss discuss with them and then make sure that the data is going to be entered during that period then so that they can use it um uh to to evaluate that specific service of course this brings a lot of other discussions on how we can continuously maintain yeah this how we can continue and maintain maintain the system so that uh so that we will be able to to to to do at least the users will be able to always get the data that they want so uh this is what we we we we we uh uh special we are working on trying to see uh how we can uh the the the keep design the system so that the the the lower level that are able to uh make some uh adjustment in the system we uh in special uh is not specifically related to the denominator problem but which is also uh we can consider that the denominator if for example i want to know how many health facilities have reported so if the the health facilities are not complete in the system so we won't be able to evaluate so what we have done here we did develop some uh uh application that allows uh what we call decentralization we mean that the the lower level they are able to suggest changes in the system for example adding users adding adding services to specific facilities adding new facilities for example what kofi mentioned about splitting the the the the structure like the one district that where was the the split didn't only created two districts so it's it's possible uh to to to for the lower users or the lower demonstrator that the district to suggest those changes and then the national level what they do is just to uh confirm and approve that previously it was taking long to have those changes uh made in the system because they were sending letters to of course even now they send letter but while the the letter is sent is going through all these administrative levels the the request has already arrived from in in the dji system so that the the the unit at the national level can take a look and then approve or even follow up those requests and then see make sure that the the the the the system is is responding what the they're they're they're expecting so the other challenge is about this maintaining we did find also uh this situation where in some country like for example in guinebisau you find uh each of organization or unit program maintain is on population estimates and there is also no they don't have a specific agreement on how to to to um the coefficient that can be used for example to estimate a population for the next year so it's not there's no agreement that is something also the challenge that that that that is always bringing uh coming it's also posing the the system not to to respond to to be used uh to to to what we are expecting also coffee mentioned about the same thing when we were talking about using their outputs uh to public to to to share this information as official information in order to have that you need to have a kind of agreement that we are going to use the same set of of of for the denominator should be the same so that the everyone that goes to the system is going to generate that so this having this uh the duality or differences on the the way the information is is stored the more the way the denominator is stored can bring uh also challenges uh that we still uh looking and then see how we can address those challenges and uh those also mentioned about the catchment areas that also bring this challenge we are talking about we have been involved in the community information system both in muslimic and angola we did we do find some challenges with the related what the definition of catchment area for each uh for example uh uh community of worker where they are working and how the information for them should be uh send and then at the end of the when which uh is are we are they going to use for them as a basis for defining the the denominator and also uh final that the issue which related to updating that the denominator um uh periodically and per target group and this is also one of the the the the challenges and which we hope that with this uh uh that what called the centralizations along the the the the lower levels to suggest even if they cannot make changes but they at least suggest that the the uh changes it will help to have this somehow uh addressed so this is what uh i have for the disc to contribute to the discussion related to the denominator problem that we all face uh on the date they uh work with the with the with the countries uh or with the that with the hs users thank you very much uh back to you uh uh Elaine thanks very much um Zafrino um and um i just want that Arthur is on on online but he's having difficulty with his internet connectivity so he was suggesting that he might try um to just say something first so we'll try if if Arthur can actually just talk um um about the issues in in Zambia and if not then we've got plenty of questions coming in and plenty of items for discussion so we'll try Arthur i'm not sure if you can try to connect now and see if we'd be able to hear some of your um um experience within Zambia can't we can hear um some keyboards of our oh you're going to share the screen as well great if that doesn't work Arthur we can we can probably listen to you hello are you hearing me we are yes yes Elaine are you hearing me yes we are okay fine sorry folks but i'm just okay um i want to talk very briefly about an experience that we've had in Zambia um just to say we're we're sharing your sorry Arthur to interrupt there you're sharing your screen on a a a slide on population data do you want to show the presentation from the first slide in presentation mode seeing that now yeah not if you just click on presentation are you seeing my population data yeah so if you if you click on presentation mode if you can or from your slideshow or down your right hand screen yeah so if you just we can see your screen if you can just and we can see you going through it but we're not seeing it in presentation mode yeah so if you just go over to the right there further right oh all right okay our slideshow okay if not we can we can not just go ahead we can see it yeah come on yeah just okay so this is about Zambia and a year ago just over a year ago when we started we had real problems with population data and Arthur we've lost you again maybe at every level had no data because the district boundaries had been reconstituted so we had no data for the district and it took like four months I mean sorry we finally got which was roughly a thousand people was projecting um up with an organizing in most countries in Africa this seems somehow to have bought the rights to all the GIS of all the dwellings you probably people sorry Arthur we're actually having great difficulty to hear you maybe very comprehensive maps you can see the top right hand corner of my map this is Chinseali district put sorry Arthur we're having difficulty hearing you there's a there's a delay maybe if you stop sharing your screen yeah so maybe if you just talk it might be better try to talk now Arthur we can't hear you have we lost Arthur completely yeah so um let's say we there's a couple of questions in the chat but I just want to thank very much Dr Coffey Sillidan for the hispan western central africa and Dr Sovina Stavdun from South Digitus who looks after Mozambique as well as a number of other countries and I think there's a couple of questions there that have been posted in the chat um I think if we some of those have been responded to already by Coffey and Jorn um I'm not sure Coffey if you wanted to add to any of the other questions you responses you've given um I think you've looked at should we have villages and DHIS too um you're saying something around your community health worker catchment areas do you want to expand on that Coffey yes just a few comments um as I said um for community health workers we have to map for each community health worker which are the villages where he or she works so when we were designing our um community health information system based in DHS too Jerry is correct connected maybe he may add some things for Dr Chanile we have mapped the all the villages of um health facilities inside the DHS too so that we can attribute a number of a certain number of villages to a community health worker sometimes we are um obliging to uh to assign one really too many community health workers because uh the the head count of population is uh too much for one community health worker so we have already mapped that in our DHIS too integral for the community health information system and uh we're in currently in discussion with a Côte d'Ivoire for something like this not completely like what we did in Togo but something that is close to that so yes and essentially when it comes to community health workers you need to have your villagers uh mapped into the DHIS too and her facility catchment area and the second uh point I want to make is for Federica and she was asking if for the population estimates uh for example pregnancy in a given year we use the district or provincial crude birth rate estimates uh no we don't um use specifically those um crude birth estimates but rather the standard option that you have um provided you have the Statistic National Institute that uh have those two the three scenarios the high-sale scenario the medium scenario and the low scenario by provincial and district areas so by the end by the beginning of each year they say okay this is what we're going to use for the estimates so we don't make new estimates from our side but we have the instructions direction from this institute and as I said during my presentation we have a big book over 10 years and then and that provide us with those estimate yearly uh province and uh district so they don't make a new estimate it uh I have never seen that since I've been here maybe uh Ghaniou can say something on that but or maybe Samaqe for Mali but they don't need to make a new estimate we already have those estimates over 10 years from the result of the CGPH so they does indicate us which estimate we choose to move forward so those are the two comments that I can make okay thanks thanks Kofi and I think you're and you're yourself and then Randy have responded around this um village listing and this idea of facility profiles you want to expand on that journey yeah generally that is something uh which we we need to find out how to address and exactly how to deal with it indeed Shai Steve I mean in Laos they have John Lewis and company they have used option sets for organizing the villages and linked villages then in doing case-based case-based immunization registry and other other case-based case-based applications and they claim to have a relatively standardized village list in in in in Laos they they have of course problems with this spelling and one thing is that one one ministry will have a standard list that is not necessarily shared with those who are telling the illustrator her address in this case then so so there are there are challenges and also Rwanda claimed to have a standardized village list so they want to uh link then their immunization register I mean each immunize the child and to to the village for thereby to be able to find out I mean where where where the coverage is low yeah so yes so this is a very very important and and and uh much raised requirement yeah over and so Frino do you want to add anything to the discussion going on there I think one of them at the issue of you know you mentioned kind of new districts and being able to split those and I know that was a question that Kofi responded in in the chat there in terms of a question for from was it Philip McKay on adding new facilities and our new areas and I know you mentioned something about that Seferino in your presentation yes yes I did mention about that what inside technical what we did in DHS too we did develop business of course is not strictly linked to the denominator but of course what we did is the to be able to allow the the administrators what who knows usual DHS the the native the administration was done at the central level at the province at national level so together with the minister we decided that they should create the team expand the team that are going to do the administration of certain parts of the system so we did develop that app that is we call decentralization model which has two modules one that is used by the subnational administrators and then the other one which are national administrators so what they can do are the subnational is just suggesting some changes that are required in the at their levels so that if there is a district that has new facility so they can suggest that that help facility to be created and then they do it in the system we're using the module that they have so the national level when that's that's that this it is suggested all that facility suggested they receive that request and then through that the system they cannot approve that request once they approved this is automatically created in DHS too so in that way it's possible for example to make sure that what we have is always always updated because sometimes what we realize is that they do if you go to the the province they say no we have request we have new facilities that were created but they are not in the system so this is also impacting on our when we when when you calculate the coverage is influenced by this because you are at the end of the day we are counting that we have 10 11 facilities but we have 11 or 12 so that is what I was I was mentioning with regard to that how the flexibility that we have added to make sure that lower level they can also do administration of the system yeah thanks very much Stefano I'm running coffee I know you have your hand up I'll go to you now and then there's a questioning from Randy that we'll address so Kofi thanks Helen I wanted to come back on the master village list which is very nice and interesting and for example in Togo we do have one the challenge with that master village list is very very dynamic that's something that need to be keeping in mind that even though you have a master village list then you can be sure that by the end of the year you have new villages that are created or some really that are expanded or getting integrated so this is something that is very very dynamic so we need to to keep that in mind when we're dealing with it in the DHS too maybe we can let Randy ask his question before because I got some element for him okay so maybe if we can allow Randy to ask the question if that's possible Randy I'm not sure if you could unmute you should be able to now yeah I think so yeah I think Kofi is opening it up so that he that he can he can ask you a question in return so can you hear me yes yes Randy please go ahead very interesting discussion now I'm just looking at the another piece of the denominator issue not just the crude numbers of population but the the way that you will select you know the percentage of the population that fits into a specific target group in Rwanda we we typically would store those percentages as constants so you know 24 percent of the population women of reproductive age or something like that and that that was then applied across all districts etc you know a couple issues with that particularly because that we found that those numbers as we went from one DHS to another those numbers changed and if we change that constant then all the historical data was recalculated at the new rate so just wondering whether people had any innovative ways of storing those types of semi-permanent target population percentages so that they could be applied uniformly across facilities you're new at some reply to that yeah I know that some are storing it as as a percentage and others are using the percentages for for for for estimation for calculating the target populations so then they will use the typically four percent for expected pregnancies across the country whether it's as a percentage in the DHS to or as a way to calculate the the target population that's my experience I'm I see in this four percent or whatever percentages used across Africa for that matter I know obviously not not the optimal way of doing it and I see nor responding nor is responding under there I don't know nor if you want to come into the conversation if you're able to nor a stooped and then I'll get back to coffee and Suleiman yes yeah um yes it's one of my hobby horses population denominator and how every rephrase it how so often it is danced around um and nobody wants to take responsibility for sorting it out and actually getting it done and uh you know and because population figures are so controversial nobody you know nobody says okay I'm putting my hand up I'm putting it in and let's start and let's go let's get going um the use of these percentages I think is very problematic because a four percent for expected antenatal deliveries or births and under one just doesn't to me make logical sense and we also know with the senses is that your population normally is broken down to under five and then we do some or some interesting mathematical calculations are done to get the smaller groups and they sometimes really are problematic those smaller group um you know you're under ones and you're one year old's because we need those so crucial the other thing I know is that if your org hierarchy is not set up correctly I know a country that puts its org it's has put its big hospitals into their own sort of level in the org hierarchy so when you try and calculate delivery coverage or bcg coverage it's just a complete disaster so there's many issues around population denominators and I think we need to start looking and making suggestions about best practice I'm sorry I'm wondering on okay okay thanks Nora sorry just to have coffee if you come back come back and then the last um hand is up to Suleiman but I think this is certainly a topic that needs to go on and continue within the community of practice and probably organize another session on so coffee we only have three minutes left so if I just hand over quickly to you and then if you can hand over to Suleiman I think you have said it Ellen we need to continue that passion and then that's all that's all I can say because as I said as this has been said when we we come to that kind of discussion this is we take days to discuss that issue and then for example in in in Togo we we have decided to give the authority to the National Statistics Institute so they they used to come up with this kind of uh you can see my screen this kind of table uh and then they use the estimate from the synthetic fertility rate uh from the mortality and then from migration and from urbanization and then using that they can come up with this kind of scenario over the years and then for the total population they can estimate for the upcoming years the the lower the the middle the middle intermediate and the high-level scenario for age groups and for for sex and then they can split that into different groups for example when age groups when you take this kind of so you can see for each part each portion of five years then you can see from zero to four from five to to nine and those kinds of estimates helps a lot when it comes to discussions on the target population so I will really support um organizing the session and that and discuss continued discussion thank you okay okay the last word then to Suleyman and we're running out of time you've only got a minute there Suleyman no okay but I think certainly I don't know Jørn if you want to say some concluding words on this but really thanks for the presentation I think there's obviously with the amount of questions going on and the amount of movement within the chat I think we can continue this on the community of practice but it would be really nice to kind of document these approaches and share what people have done and what currently is happening so Jørn do you want to just say a couple of words before we close yeah just just that this initiative we have we have called it facility profile where we look at how to include the population denominators how to include villages etc in the DHIs we call it facility profile not to come into all this political mess with the more I mean all the competing MFL we have around so that is an ongoing project and we are actually experimenting with the search facility profile and village listing in Indonesia and see how that can be linked to the DHIs in what we call health facility profile yeah thank you we continue in other fora yeah okay great thanks very much everyone and it was great particular thanks to Kofi and Zafrino and Arthur for really trying I'm sorry the internet didn't allow you but we certainly need to continue this conversation and look at different ways in which we can actually overcome the denominator challenge so thank you very much and see you in the following sessions