Temporal dynamics of daily diabetes-specific burdens using intensive longitudinal data: the central role of feelings of guilt within a multilevel network analysis.

AIMS/HYPOTHESIS: The aims were to: (1) analyse the day-to-day mechanisms through which daily diabetes-specific negative emotional experiences (burdens) accumulate and potentially affect emotional well-being; and (2) identify the specific triggers of these dynamics in people with type 1 and type 2 diabetes. METHODS: Data were derived from three prospective observational studies employing ecological momentary assessment for up to 17 days. Multilevel vector autoregression models were used to estimate within-person temporal networks of nine daily diabetes-specific burdens. Out-strength and in-strength indices quantified each variable's role as a trigger (i.e. influencing others) or receiver (i.e. being influenced by others). Individual network density was computed to characterise how strongly daily burdens predicted one another. Logistic regression models tested whether these dynamic features predicted questionnaire-assessed incidence and remission of elevated diabetes distress and depressive symptoms at follow-up. RESULTS: Among 741 participants (50.1% with type 2 diabetes, 50.7% male, aged [mean ± SD] 50.4±14.2 years, HbA1c [mean ± SD] 65±18.6 mmol/mol [8.1±1.7%]), 'feelings of guilt' regarding suboptimal self-management emerged as the most significant trigger of other negative emotional experiences/burdens. In type 1 diabetes, greater 'feelings of guilt' on one day were temporally associated with increased levels of 'feeling alone' (β=0.09, p<0.001), 'feeling overwhelmed' (β=0.05, p=0.016), 'feeling restricted' (β=0.05, p=0.015) and 'feeling burdened by hyperglycaemia' (β=0.05, p=0.030) on the following day. In type 2 diabetes, 'feelings of guilt' functioned primarily as a receiver. Greater 'feelings of guilt' on one day were predicted by 'feeling alone' (β=-0.07, p=0.016) and 'feeling burdened by hypoglycaemia' (β=0.04, p=0.017) on the previous day. Higher individual network density predicted greater odds of incident elevated diabetes distress (OR 1.50, p=0.025) and depressive symptoms (OR 2.08, p<0.001), with 'feelings of guilt' and 'feeling restricted' as the most relevant predictors. CONCLUSIONS/INTERPRETATION: The results provide evidence of how daily diabetes-specific burdens can accumulate and potentially affect the course of emotional well-being. Feeling guilty regarding diabetes self-management may play a central role and represent an important clinical target as well as outcome. TRIAL REGISTRATION: ClinicalTrials.gov NCT03811132, NCT04438018, NCT05548699.