Menu ENN Search
Language: English Français

SQUEAC Coverage survey

This question was posted the Coverage assessment forum area and has 20 replies. You can also reply via email – be sure to leave the subject unchanged.

» Post a reply

Anonymous 2264

Normal user

Normal user

8 Oct 2013, 11:03

We did 2 SQUEAC coverage assessments recently and I have confused with something with in it
1. What is the difference between Alpha and Beta prior ? What is Alpha and Beta stands for?

2. In giving values for /weights / for Boosters and Barriers for example Lack of awareness of malnutrition and good awareness of malnutrition if we give 5 for good awareness and 0 for lack of awareness It is rule to give like this ? In my understanding this affects the value for each b/s we went in some places the community have awareness and in the other places they don't have awareness P/s give me any clarification for these.

Mark Myatt

Consultant Epideomiologist

Frequent user

8 Oct 2013, 16:13

I'll try to deal with (1) and (2) in turn ...

(1) The terms "alpha" and "beta" are called "shape parameters" and these define the shape of the distribution of prior belief (for which we use a Beta distribution - see below). There is only one prior and it has both alpha and beta parameters. For example, alpha = 1 and beta = 1 describes no prior belief (or a belief that any coverage proportion is as likely as any other coverage proportion). You can try this in BayesSQUEAC. Another example, alpha = 18, beta = 12 describes a belief in coverage being a little below 60% (our best guess) with a credible range of between 30% and 85% (although we think those extremes are not likely to be true). You can also try this is BayesSQUEAC.

We use a Beta distributed prior because this is (i) a flexible distribution for modelling prior belief and (ii) it allows for simple combination with survey data (the likelihood) in a process called a Beta-Binomial conjugate analysis see here and here and the SQUEAC / SLEAC Technical Reference for more details. The basic idea is that we make an informed guess about program coverage from (e.g.) routine program monitoring data; interviews with clinic staff, cares, community leaders; and small area surveys and small studies. We then test and update that guess with some survey data. The updated guess is the posterior estimate.

BTW : The posterior of the type of Bayesian inference that we do in SQUEAC is also Beta distributed and has it's own alpha and beta parameters which we can turn into a posterior estimate of the coverage proportion. The required calculation can be done in BayesSQUEAC or by hand using the methods outlined in the SQUEAC / SLEAC Technical Reference.

(2) I think more appropriate weights might be +5 for good awareness and -5 for a more-or-less complete lack of awareness since good awareness will boost coverage and poor awareness will depress coverage. You describe a situation of patchy awareness (i.e. good in some places and poor in other places). In such a situation you need to make an informed guess as to the overall effect. If (e.g.) coverage is good in very few places and poor in vary many places then you might score (weight) awareness with a negative number. If awareness was good in most places and poor in very few cases then you might score it with a positive number. Anything that is patchy is usually scored with a negative number particularly if it is related to distance from CMAM delivery facilities.

I hope this helps.

Please do not hesitate to ask follow-up questions if you are still unsure about this keey ideas in SQUEAC.

Lio

CMAM advisor

Technical expert

8 Oct 2013, 16:50

Just an addition to what Mark already said about your question 2:
in the reality it is rare that something is totally black and something is totally white. In the event that you really have the impression that a "barrier" or a "booster" have the same weight (lets say + 5 and - 5 or + 3 and - 3, etc.) you have two choices:
1. Not adding it to the list because the weight on the total will be nil
2. Write in both column the BB, meaning +5 an - 5, or + 3 and - 3, etc.
I prefer the 2nd option because it is important to point out that this specific BB need attention for 50% of the people interviewed

However, in most cases, you can see a difference, even if small, this is what I call "try to see who is strongly pulling the blanket" and in most cases you will find that the you can chose to fit your BB in the + or in the - column. When this is the case, your weighing necessarily decrease: you cannot give the maximum weight (e.g 5) because you know that the opposite is "almost true", so you may decide to just give 1 or 2 (depending how strong the blanket pull).

Mark Myatt

Consultant Epideomiologist

Frequent user

8 Oct 2013, 17:01

I really like the idea of putting in both columns. If (e.g.) "awareness" looks to be a barrier for many and a booster for some then we could have a -5 in the barriers' column and a +1 in the boosters' column. This would have the same effect on the prior mode as just having -4 in the barriers' column but would act to narrow the range of the prior in, I think, an appropriate way. This is what we want it to do ... it would lift the lower limit a little (a booster for some), drop the upper limit quite a bit (a barrier for many) and shift the mode down. I think I will be doing this in future.

Anonymous 2264

Normal user

Normal user

9 Oct 2013, 08:36

Thanks for your reply to , I got a lot of things from your explanation, in the 2 question to be clear It is possible to give value for the same idea in both Booster and Barrier for example lack of awareness of malnutrition and good awareness malnutrition ?

Anonymous 2371

Normal user

9 Oct 2013, 10:12

Dear Mark,
I am much interested to learn much on SQUEAC and SLEAC. Are trainings on this and when?

Jose Luis Alvarez Moran

ACF Senior Technical Advisor

Normal user

9 Oct 2013, 10:34

Hello,

In the Coverage Monitoring Network we organize Training and Coverage Assessment's support. Many development partners interested in Coverage and specially SLEAC and SQUEAC work with the CMN.

We are planning no training per se right now in 2013 but we are planning a few workshops, conferences, publications and SQUEACs. Most people will tell you that the best way of doing a SQUEAC is by doing one so we have helped colleagues to join one of the SQUEACs currently planned or ongoing. You can have a look at them here.

http://www.coverage-monitoring.org/events/

The CMN website is also a good place to get updated about other events regarding coverage

http://www.coverage-monitoring.org/

as are Brixton Health and CMAM forum

Regards

Jose Luis

Mark Myatt

Consultant Epideomiologist

Frequent user

9 Oct 2013, 13:30

It is legitimate to take either approach. That is:

(1) You can decide the "balance" and use a positive weight (boosters' column) or negative weight (barriers' column), or ...

(2) Put a weight in each column depending on the "balance".

I have always used the first approach but I can see advantages to the second approach. I think the second approach should be favoured as it keeps more data (i.e. you know that a given factor can be a barrier and a booster depending on who or where you are).

Mark Myatt

Consultant Epideomiologist

Frequent user

9 Oct 2013, 13:38

CMN (see Jose Luis' post above) are now doing a lot of work on training and supporting SQUEAC. I hope that this will continue. I have done a couple of SQUEAC trainings for CMN (in Kenya and Pakistan) and Lio has done some elsewhere (Francophone Africa I think). VALID International also offer SQUEAC services and SQUEAC support and were part of the team that developed SQUEAC.

Jose Luis is correct ... the best way to learn SQUEAC is to participate in one. There are many SQUEACers on this forum and known to CMN, VALID, &c. If you let us know where you are based then we can, perhaps, put you in touch with someone doing a SQUEAC in your area.

Anonymous 2371

Normal user

9 Oct 2013, 14:28

Hi Mark,

I am based in Yemen Hodeidah Governorate. I will really appreciate if I get this training based the context of the areas that we are working. Some of our staffs also are interested in this training.

Regards,
ENUNU, Henry

Mark Myatt

Consultant Epideomiologist

Frequent user

9 Oct 2013, 14:53

I do not know of anyone currently doing SQUEAC in Yemen.

Issakha (VALID International) has done CTC and survey work in Yemen. Melaku (CMN) has, I think, done CTC coverage and survey work in Yemen. I think that a special training may have to be organised.

CMN may be able to help with this. Another option is that FMoH in Khartoum and UNICEF in Sudan have some experienced SQUEACers - It may be possible to arrange for a brief secondment via UNICEF.

Anonymous 81

Public Health Nutritionist

Normal user

9 Oct 2013, 14:57

Hi Enunu,
Save the Children did SQUEAC investigation in Hodeidah last Month, September 2013. Partners were invited through Nutrition cluster but only two people from the local NGO and four from MoH were participated. we are planning to have similar exercises in November. So, you are welcome to participate. Make sure you are ready to participate the whole exercise (classroom + field) for about two weeks.

Thanks

Mark Myatt

Consultant Epideomiologist

Frequent user

9 Oct 2013, 15:03

Thanks Kiross!

This just goes to show what a great resource EN-NET can be when we treat it as a network.

Jose Luis Alvarez Moran

ACF Senior Technical Advisor

Normal user

9 Oct 2013, 15:09

Thank you Mark,

Indeed Valid also has all the technical expertise behind SQUEAC and it is worth to visit their site regularly.

The CMN has sent some coverage experts to Yemen. Recently we have worked together with the Nutrition Cluster to do a SQUEAC in Hodeidah (AL Qanawis to be precised) so it is really a pity that you have not contacted us before but the assessment ended last Saturday. During this assessment 4 people from MoH were trained and there are plans where they will continue and implement other coverage assessments before the year ends with different levels of support from the CMN. In order to participate on those the best thing will be to contact the Nutrition cluster. The CMN will be very happy to facilitate the contact.

Otherwise, as Mark said Sudan has good experience and we are planning additional training and different workshops there before the year ends. Again, everything is updated often in the website

Jose Luis Alvarez Moran

ACF Senior Technical Advisor

Normal user

9 Oct 2013, 15:11

Sorry, my reply arrived too late. Kiross seems the best person to contact for further SQUEACs in Hodeidah

Anonymous 2264

Normal user

Normal user

10 Oct 2013, 06:09

Thanks ,Mark you give a great and useful info. about SQUEAC

Anonymous 2371

Normal user

10 Oct 2013, 07:01

Hi Kiross,
Thanks just do update me the dates before Nov for planning from my side

Anonymous 81

Public Health Nutritionist

Normal user

10 Oct 2013, 07:42

Dear Nunu,
PU-AMI is not far from our field office, Hodeidah. There is also a possibility of meeting with our team during the monthly nutrition cluster meeting. Any ways, for any update regarding the incoming SQUEAC investigation, you can contact the Nutrition Manager, Sebsibie.

Thanks

Anonymous 2371

Normal user

10 Oct 2013, 08:23

Hi Kiross,
I know him and I have discussed with him on this issue and he will update on the dates.
Regards

Abu Ahammad abdullah

Normal user

15 Oct 2013, 03:14

Hi

Is there any plan for organizing a SQUEC survey or classroom+field training in Bangladesh or South Asia during 2014? I know there was a SQUEC training in Nepal in earlier 2013. Unfortunately, I was unable to attend the training.

Actually, I am not updated on this topic at all. But several times I have followed the discussion and tried to extract the essence. As a Public Health and Nutrition worker (with highest interest in epidemiology and biostatistics), I would like to know it from up to bottom and apply in my programme and/or my country.

I would like to start from zero e.g. from context to objective and it's application. If there is any online interactive training for beginners, I will be more happy.

Thanks in advance.

Best regards,
Abdulllah

Jose Luis Alvarez Moran

ACF Senior Technical Advisor

Normal user

15 Oct 2013, 10:34

Hello,

From the CMN side, we are planning a SQUEAC evaluation in Bangladesh at the end of November. You can check it here

http://www.coverage-monitoring.org/events/

The approach is slightly different from Nepal since it will be mainly a SQUEAC evaluation but it could work as a field training if you are interested (no classroom or big teams this time), I am sure after following one SQUEAC hand to hand with one of the Coverage experts and additional remote support from the CMN you could be able to apply SQUEAC to your program.

As per 2014, we are still funding pending, but I am sure new activities will be announced soon

Back to top

» Post a reply