tuncreg failed to converge on observed data

[Ashmita Karki]

Hi team,

I am trying to perform multiple imputation using chained equations and my code is as below. My dependent variables are EQ5D_QOLscore and EQVAS

mi set wide
mi register imputed HbA1c Living_arng Comorbidity Taking_insulin Taking_OHAs Tobacco Alcohol PA EQ5D_QOLscore EQVAS Depression_Yes_No PSS_classification Recode_BMI Recode_hypertension HbA1ctest Foot_care DM_duration Stress_Yes_No
mi impute chained (truncreg, ul(1)) EQ5D_QOLscore (truncreg, ll(0) ul(100)) EQVAS (regress) HbA1c (mlogit) Recode_BMI (logit, augment) Living_arng Comorbidity Taking_insulin Taking_OHAs Tobacco Alcohol PA Depression_Yes_No Recode_hypertension Stress_Yes_No HbA1ctest Foot_care DM_duration = Residential_status Age Gender Marital_status Education Ethnicity Occupation Income, add(20) rseed(1234) savetrace(trace1,replace)

When I run the above code, it throws the following error message:

truncreg failed to converge on observed data
error occurred during imputation of EQ5D_QOLscore EQVAS HbA1c Recode_BMI Living_arng Comorbidity
Taking_insulin Taking_OHAs Tobacco Alcohol PA Depression_Yes_No Recode_hypertension Stress_Yes_No
HbA1ctest Foot_care DM_duration on m = 1

How do I fix this? Any help would be highly appreciated, please.

Thanks!

[Ashmita Karki]

Hi there,

Could anyone please advise on my above query? Should I use truncreg or regress in MI chained equation for my dependent variable "EQ5D quality of life score", the maximum value of which is 1 (best health state). Since truncreg threw the above error, should I use regress instead?