Update model of BMI associated risks

[HSalat]

• Change BMI array from 12 parameters to 16, see fitData.csv attached.
• Update every calculation of oddBMI from

oddBMI = params->bmi_multipliers[originNew * 3] + params->bmi_multipliers[originNew * 3 + 1] * [!!! BMI VARIABLE !!!] + params->bmi_multipliers[originNew * 3 + 2] * pown([!!! BMI VARIABLE !!!],2);

to

oddBMI = params->bmi_multipliers[originNew * 4] + params->bmi_multipliers[originNew * 4 + 1] * [!!! BMI VARIABLE !!!] + params->bmi_multipliers[originNew * 4 + 2] * pown([!!! BMI VARIABLE !!!],2) + params->bmi_multipliers[originNew * 4 + 3] * pown([!!! BMI VARIABLE !!!],3);

replacing [!!! BMI VARIABLE !!!] accordingly

• update scenarios to BMI = 26.7 / if (new_bmi > 26.7) / if (new_bmi >= 27.7)
• add else if (new_bmi > 50){ //oddBMI is calculated using 50 for [!!! BMI VARIABLE !!!]}

[mfbenitezp]

lets define !!! BMI VARIABLE !!! appropriately., that would be the dynamic new_bmi, am I right?

[mfbenitezp]

So no need for scenario where new_bmi = 40?, no need for a scenario where else if (new_bmi > 29.0)

[HSalat]

lets define !!! BMI VARIABLE !!! appropriately., that would be the dynamic new_bmi, am I right?

what is currently inside the brackets, which is different every time

[HSalat]

So no need for scenario where new_bmi = 40?, no need for a scenario where else if (new_bmi > 29.0)

No, you replace the ones we modified the other day because the BMI model was skewed to new values without deleting anything

[mfbenitezp]

@HSalat would you mind take a quick look at this branch, at the moment I got 6 scenarios, including the one with BMI of 50., but also validate if the new equations are correct, so I get the green flag to run all of them.

uatk-aspics/aspics/kernels/ramp_ua.cl

Line 198 in 82186a8

//new_bmi = 26.7; //Scenario 2

[HSalat]

I believe it is, except the BMI > 50 condition is not a scenario. It needs to be there all the time (so at the beginning), this is because the equations can curve down above that value (you might even end up with a negative odd for extreme value), and that part of the model is not supported by the data anyway.

PS: I'll check the stuff printed when you run the scenarios, it's the best way to check if it works as intended

[mfbenitezp]

Gotcha, on my way to the run the scenarios, fingers crossed

[mfbenitezp]

The outcomes still don't make sense. Ill keep checking tonight and Mon morning. Here is the latest branch in case you have spare time to double check

uatk-aspics/aspics/kernels/ramp_ua.cl

Line 198 in 939a676

//new_bmi = 26.7; //Scenario 2

[HSalat]

You need to move the BMI > 50 condition before the scenarios, otherwise it gets ignored