Diabetes Distress and Illness Perceptions in Tunisian Type 2 Diabetes Patients [Letter]

Dear editor

The article with the title “Diabetes Distress and Illness Perceptions in Tunisian Type 2 Diabetes Patients” attracted our attention because it produces a nomogram prediction model for diabetes distress (DD) and illness perceptions in type 2 diabetes mellitus patients which can be used to predict weaknesses in self-management and psychological distress.¹ This nomogram involves four factors (economic income, age at diagnosis, comorbid illness, and HbA1C status) which were obtained from the result of analysis. In the initial stage of analysis there were six significant risk factors for DD and illness perception, but in the final stage there were only four significant factors. Meanwhile, factors that were not significant were marital status and insulin. Is it true that these last two factors have no effect on DD or illness perception? Could these factors have an indirect influence?

Researchers used a logistic regression test to analyze the significance of DD and illness perception factors, so that indirect effects could not be detected, because this test assumes that all factors have a direct influence.² If we suspect that insulin conditions have an indirect influence, then we have to go through intermediate variables, for example comorbidities. So first insulin affects comorbidities and then comorbidities affect DD and illness perception.

Therefore, we recommend that further analysis be carried out to prove the existence of indirect effects from the six factors. In this case, the method that can be used is path-analysis. Because researchers use nominal scale data, one program that can be used is Smart-PLS.^3–5 We hope that the results of this advanced analysis provide more accurate information for future improvements to the predictive model.