Minimize number of modules used to facilitate debugging of own module

[pascalbirnstill]

Hi everyone,

I want to debug a Synthea module I created for a specific disease. For debugging, I would like to find out whether there are interactions with other modules that influence the data that I get. For example, I am using delay states with exponential distribution, and the generated data does not look as I would expect. The high values that are generated are too low for the given mean values.

[jawalonoski]

You can delete the modules you're not using, or you can use the M-Feature: https://github.com/synthetichealth/synthea/wiki/The--M-Feature. We do not provide assistance when people choose to do this.

What is your delay exponential distribution, and what results are you seeing?

[pascalbirnstill]

Thanks for your replay! In my module, I have delay states like this, for example: "Diagnose_C71.2_to_Sterbedatum_Sterbedatum": {       "type": "Delay",       "direct_transition": "Sterbedatum_Sterbedatum",       "name": "Diagnose_C71.2_to_Sterbedatum_Sterbedatum",       "unit": "days",       "distribution": {             "kind": "EXPONENTIAL",             "parameters": {                   "mean": 331             }       } }, When analyzing the synthetic dataset, I get the histogram shown in survival-time_C71-2_synthetic-dataset.png. I would expect something as plotted in survival-time_C71-2_rng-sample.png from an exponential distribution with mean 331. I already tried the -m feature, but it doesn’t seem to load my module at all (using Synth 3.3.0 built from source). Von: Jason Walonoski ***@***.***> Datum: Montag, 6. Oktober 2025 um 14:04 An: synthetichealth/synthea ***@***.***> Cc: Birnstill, Pascal ***@***.***>, Author ***@***.***> Betreff: Re: [synthetichealth/synthea] Minimize number of modules used to facilitate debugging of own module (Discussion #1621) You can delete the modules you're not using, or you can use the M-Feature: https://github.com/synthetichealth/synthea/wiki/The--M-Feature. We do not provide assistance when people choose to do this. What is your delay exponential distribution, and what results are you seeing? — Reply to this email directly, view it on GitHub<#1621 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ARXXT2MIQBLQA6H2TVJHXOT3WJLDPAVCNFSM6AAAAACIL7XUU2VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTINRQGQZTAMA>. You are receiving this because you authored the thread.

[pascalbirnstill]

Here are the Plots mentioned in the post above.

[jawalonoski]

Well, I can't do the math from the graph, but the shape of the graphs generally looks like I would expect. The number of patients is low, so perhaps the numbers will smooth out with more data points.

[pascalbirnstill]

I managed to run Synthea with only my module (deleting the other modules from the modules folder while keeping the folder lookup_tables, copying my module into the modules folder; this does not require the -m feature). This way the synthetic data looks much better. So it seems that if I run Synthea with my module + all default modules this influences some statistical properties of the generated data - which is fine. However, for my purposes, i.e. evaluating whether my module generates representative data for a specific disease, I assume I can best validate my synthetic data by running Synthea with only my module.