Fixing Epidemiological development of COPD Module

[mnjowe]

This is an outline of questions surrounding modeling of COPD module. @andrew-phillips-1 , @tbhallett and @tdm32 feel free to add more questions if any.

1. Can people transition lung function backwards without health system interaction i.e. IF they naturally get better, they stop smoking or they stop being exposed to wood burn smoke?
2. Is COPD linked with Lung Cancer?
3. Should we say lung functionor lung obstruction?

copd_presentation - ap.pptx

[mnjowe]

Hi @andrew-phillips-1 . kindly find attached copd plots for your review. For the parameters refer to the resourcefile here and the latest commit here for the code. Thanks copd_plots.docx

[andrew-phillips-1]

Thanks @mnjowe I will get back to you as soon as I can

[andrew-phillips-1]

Hi @mnjowe I've now studied the outputs.

We have to decide on the effect of wood burning stove. I don't think there is much evidence that the effect is as substantial as the effect smoking, so I would not raise the effect of wood burning stove to 7 (the same as smoking) just so that it gets us closer to the GBD model estimates for women. Are you aware of evidence on the effect of wood-burning stove ? If not, I suggest reducing to 2 for now. It also feels to me that the prob_progress_to_next_cat should be lower than what we have, so that lung function is generally good in non smokers.

I think we can break down the death rate by age a bit more also.

So this below is what I would suggest we change, but let me know if you disagree.

prob_progress_to_next_cat,0.01 prob_will_die_sev_exacerbation_ge80,0.95 prob_will_die_sev_exacerbation_7079,0.7 prob_will_die_sev_exacerbation_6099,0.5 prob_will_die_sev_exacerbation_5059,0.3 prob_will_die_sev_exacerbation_4049,0.2 prob_will_die_sev_exacerbation_4049,0.1 prob_will_die_sev_exacerbation_3039,0.05 prob_will_die_sev_exacerbation_lt30,0.0 rel_risk_tob,7.0 rel_risk_wood_burn_stove,2.0

Once we are happy with the lung function by age and the deaths I think we can return to specify the initial values in 2010.

Thanks

[mnjowe]

Thanks for these suggestions @andrew-phillips-1. Let me first apply these then we can keep the discussion going. Thanks

[mnjowe]

@andrew-phillips-1 . One thing on the above suggestions, I see there is a repetition on probability of dying with severe exacerbation for ages 40-49 prob_will_die_sev_exacerbation_4049,0.2 prob_will_die_sev_exacerbation_4049,0.1

what should we go with?

[andrew-phillips-1]

sorry @mnjowe lets go with 0.15 for that parameter !

[mnjowe]

great, thanks

[mnjowe]

I guess we will do the same with the effect of oxygen as in we will be multiplying the effect with these new age category risk, right?

[andrew-phillips-1]

yes right thanks

[andrew-phillips-1]

copd - write-up.docx

Hi @mnjowe Here is a draft of the write-up. See what you think. Please insert the figures when final. Also, it would be good at some point to be able to fill in the model outputs for Table 3. We can discuss how to specify those outputs if it is not clear,

[mnjowe]

Thanks @andrew-phillips-1.

[tbhallett]

copd - write-up.docx

Hi @mnjowe Here is a draft of the write-up. See what you think. Please insert the figures when final. Also, it would be good at some point to be able to fill in the model outputs for Table 3. We can discuss how to specify those outputs if it is not clear,

I've pushed a commit that brings that document onto the PR #998 and references it on the website.

[mnjowe]

copd - write-up.docx

Hi @mnjowe Here is a draft of the write-up. See what you think. Please insert the figures when final. Also, it would be good at some point to be able to fill in the model outputs for Table 3. We can discuss how to specify those outputs if it is not clear,

Hi @andrew-phillips-1. Thanks for the write-up. It represent well what we're working on. kindly find attached some few comments(mainly on outputs).

One general question , Should we still include aminophylline in HSI? I understand we haven't found evidence that it has any effect on COPD mortality and again I cant see it mentioned anywhere in the write-up. Thanks

copd.-.write-up.-.em.docx

[andrew-phillips-1]

Thanks @mnjowe

copd.-.write-up.-.em - ap.docx

I seem to recall that aminophylline was used so we probably should keep in aminophylline, even though we have not included any beneficial effect

[mnjowe]

Thanks @andrew-phillips-1 for the clarification(both here and in the document attached). I will wait for the COPD analyses file I've submitted on azure to finish and then, we can take a look again at the outputs(I'm running a simulation of 250,000 individuals from 2010-2030)

[mnjowe]

Hi @andrew-phillips-1. kindly find attached COPD plots for your review. I've run a simulation of 250,000 individuals from 2010 to 2030

__copd_plots.docx

[andrew-phillips-1]

Thanks @mnjowe

I suggest changing as follows: prob_will_die_sev_exacerbation_5059 =0.10 prob_will_die_sev_exacerbation_4049=0.05

The plot of lung_function by gender still does not look right - there should be poorer lungfunction in men due to smoking. (Or have you made the wood-burning stove effect only apply in women ?)

I think you can also change the values for prob_not_tob_lung_func_15_39 etc to match the proportions in 2019 in the top graphs.

Does this seem OK ?

Thanks

[mnjowe]

Thanks for the suggestions @andrew-phillips-1, they all seem okay. You're right on the second plot, I indeed made wood burn stove apply for women only. let me correct that and re-run the simulation

[andrew-phillips-1]

HI @mnjowe Well you may in fact be able to justify that gender effect - it does make sense to me also - so perhaps leave it in if you think it reflects a reality of greater exposure in women ?

[mnjowe]

Hi @andrew-phillips-1. Kindly find attached copd lung function plot by gender( I don't think it is any different from the previous ones?). I've also attached a document with values showing lung function categories grouped by gender for both smokers and non smokers combined and lung function categories grouped by gender for smokers only. lung_function_categories.docx

[andrew-phillips-1]

Hi @mnjowe Thanks.

As discussed, I suggest plotting "proportion of people with lungfunction >= 5 by gender and year.

Also to change the value of the parameters as we discussed.

Then we need to think about how you can reproduce the values in the Table in the write-up to compare with real data. Let me know when you have tried to do that. I can see if I think your code seems to be giving what we want.

Thanks !

[mnjowe]

Hi @andrew-phillips-1. Following our discussion yesterday, kindly find attached copd plots. Find here the updated resource file. You can also find the latest commits here.

Thanks copd_plots-em.docx

[andrew-phillips-1]

Thanks @mnjowe I will take a look and get back to you in the next few days. Have a good weekend.

[andrew-phillips-1]

Thanks @mnjowe

Hmmm. I wonder if we need to consider increasing the probability of exacerbation rather than the rate of progression to the next category. Shall we try to change these parameters below like this ?:

prob_mod_exacerb,"[0,0,0,0.01,0.05,0.20,0.30]" prob_sev_exacerb,"[0,0,0,0.0,0.0,0.10,0.20]" prob_progress_to_next_cat,0.010 prob_will_die_sev_exacerbation_5059,0.05 prob_will_die_sev_exacerbation_4049,0.02 prob_will_die_sev_exacerbation_3039,0.01 prob_will_die_sev_exacerbation_lt30,0.0

Sorry I wasn't clear. For the figures on the first page I meant to plot the proportion of all people who were in lungfunction category 5 or 6.

[mnjowe]

Hi @andrew-phillips-1. Thanks for the suggestions. I will modify the parameters as per your suggestion and re-run the model however I still can't get your comment(s) on the first and second page graphs. Do you mean amongst proportions of all categories, I should plot only proportions for categories 5 or 6? for which they would not add-up to 1? sorry for taking long to get your point on this

[andrew-phillips-1]

Thanks @mnjowe Sorry I think it is me that is not being clear. I suggest that you continue to produce the figures on page 1 as you had them (distribution of the 7 categories by year and smoking, distribution of the 7 categories by sex) but also separate versions in which the y axis is "percent of people with moderate or severe lung obstruction (lungfunction 5 or 6)" - so the y axes might go up to a maximum of 20%. That way we should be able to more clearly see the effects of smoking and sex on moderate/severe lungfunction. Do you see what I mean ?

[mnjowe]

Ah I see. Thanks for the clarification @andrew-phillips-1.

[mnjowe]

Hi @andrew-phillips-1 . Kindly find attached copd plots following our latest conversations. Seems we have increased the deaths to a higher extent? copd_new-em.docx

[andrew-phillips-1]

Hi @mnjowe Thanks. The graphs all look good. I think we are making progress. How about we reduce

prob_sev_exacerb,"[0,0,0,0.0,0.0,0.10,0.20]"

to

prob_sev_exacerb,"[0,0,0,0.0,0.0,0.05,0.10]"

what do you think ?

[mnjowe]

@andrew-phillips-1. I feel we will definitely reduce the overall deaths to some extent. Not sure though how deaths by age groups plot will look like. Lets run the simulation and see.

[andrew-phillips-1]

@mnjowe yes I agree it it unlikely to be that it looks anywhere near a perfect alignment with the GBD data but hopefully close enough. We might need to reduce the exacerbation risk a bit more, but let's see.

[mnjowe]

Hi @andrew-phillips-1. This is how the last two plots look like. I was just thinking, should we consider plotting deaths by age groups plot based on the age groups we have defined i.e. <30, 30-39, 40-49, 50-59, 60-69, 70-79, >80?

[andrew-phillips-1]

@mnjowe yes good idea. I think this is close enough. Do we now need to think about the other estimates to compare with data in the write-up ?

[mnjowe]

@andrew-phillips-1 yes, I agree.

[mnjowe]

@andrew-phillips-1 . I'm attaching the agreed version of copd plots. I guess we will discuss the observed data in the write-up in our epi call today?

copd_new-em.docx

[andrew-phillips-1]

Yes thanks @mnjowe

[andrew-phillips-1]

copd - write-up.docx @mnjowe Here is the latest version of the write-up

[mnjowe]

@andrew-phillips-1 thanks.

[mnjowe]

Hi @andrew-phillips-1. Kindly find attached COPD write-up with a comment and some few suggested changes on parameter values for your approval. Thanks

copd.-.write-up-em.docx

[mnjowe]

I'm suggesting the parameter value changes based on our conversation here and here

[andrew-phillips-1]

Thanks @mnjowe for updating the parameter values in the write-up. I will go through your comments and upload a new clean version.

[andrew-phillips-1]

copd.-.write-up.-.em - ap.docx Here is the updated version @mnjowe Thanks for the edits.

[mnjowe]

Thanks @andrew-phillips-1 for the document but it seems to me like that first version(without plots and also the comments and suggestions are not addressed).

copd.-.write-up.-.em - ap.docx Here is the updated version @mnjowe Thanks for the edits.

[andrew-phillips-1]

Sorry @mnjowe copd.-.write-up - 18 aug 2023.docx

[mnjowe]

Thanks @andrew-phillips-1

Sorry @mnjowe copd.-.write-up - 18 aug 2023.docx

[mnjowe]

Hi @tbhallett and @andrew-phillips-1 . Kindly find attached COPD write-up with updated plots. Thanks copd.-.write-up-new.docx

[mnjowe]

This is the latest version. Thanks copd.-.write-up-new.docx

[andrew-phillips-1]

Thanks @mnjowe Note that the Alupo et al mortality rates are for all cause death and not only COPD death