Background
Maintaining blood sugar levels in a healthy range is essential for people with type 2 diabetes in order to avoid long-term complications such as CVD, retinopathy, neuropathy and nephropathy. Performing self-management activities including eating a healthy diet, being physically active, measuring blood glucose levels and taking medication, is required to keep recommended HbA1c levels. Such efforts derive from motivation and skills. Self Determination Theory (SDT) is a theory of human motivation that defines intrinsic and extrinsic sources of motivation. SDT suggests that humans are more volitional motivated when experiencing autonomy, competence and relatedness. In relation to glycemic control a SDT process model has been developed suggesting that autonomy support is an important determinant of motivation and perceived competence in diabetes.  

Aims
The aim of this study was to investigate how the relations between autonomy support, well-being, perceived competence, diabetes distress, motivation and glycemic control in a population of people with type 2 diabetes correspond with the hypothesized SDT process model: 
Autonomy support → Autonomous motivation →      
Perceived competence in diabetes → Self-care activities → HbA1c level	
              ↓		          
Diabetes distress → Well-being 

Method
This study uses baseline questionnaire data from an RCT designed to investigate the impact of using dialogue tools in consultations. Data include HbA1c measurements and behavioural and psychosocial measures in type 2 diabetes.
Participants were recruited from a specialist diabetes clinic in the Capital region of Denmark from October 2014 to June 2016. Inclusion criteria were: > 18 years of age, duration of type 2 diabetes ≥ 1 year; and HbA1c ≥ 64 mmol/mol (≥ 8%) in the last three measurements at routine consultations. Moreover, participants should be in medical treatment (oral/injection) and be able to speak, read and understand Danish. Exclusion criteria included: participating in other research/clinical projects, attending psychological/psychiatric treatment and severely impaired vision or blindness. A total of 154 participants completed the questionnaire and were afterwards randomised to either dialogue tool intervention or routine consultations.

The questionnaire measured demographics, disease duration and treatment, well-being (WHO-5), perceived competence in diabetes (PCD), diabetes distress (PAID), autonomous and controlled motivation (TSRQ), autonomy support from health care professional (HCCQ) and self-care activities (SDSCA). 
The latest HbA1c measurement, before or on the day of questionnaire completion, was used. 

SPSS version 22 was used for the data analysis. Mean values, standard deviations and reliability were calculated on the key variables. Pearson correlations were used to measure bivariate correlations between all variables. Exploratory factor analyses, using the correlation matrix with principal components and direct oblimin rotation, were applied for TSRQ scores for validation check.  Mediation analysis was conducted using multivariate linear regression analysis to reveal the direct and indirect associations of relevance for elucidation of the SDT process model.
Level of significance was p < 0.05 for all correlations. 
	
Results
Table 1: Pearson correlation coefficients for the variables in the model
Variables	WHO-5	PAID	SDSCA	HCCQ	PCD	HbA1c
WHO-5	1	-,528**	,194*	,065	,333**	-,057
						
						
PAID	
	1	,012	-,116	-,462**	-,045
						
						
SDSCA			1	,109	,272**	-,255**
						
						
HCCQ				1	,438**	-,102
						
						
PCD					1	-,159*
						
						
HbA1c						1
						
						
          *Correlation is significant at the 0.05 level (2-tailed). 
         **Correlation is significant at the 0.01 level (2-tailed). 
    
Factor analysis of TSRQ data gave invalid/inconsistent results and we excluded these results from further analyses.
Regression analyses revealed that the correlation between perceived competence and HbA1c was mediated by self-care activities (-.23**). Likewise, the correlation between perceived competence and well-being was mediated by diabetes distress (-.48**). The results support the following process model: 

Autonomy support → Perceived competence in diabetes → Diabetes distress → Well-being
             		 	 ↓
Self-care activities → HbA1c level	                         	              				
Discussion
This study essentially supports the hypothesized SDT process model, in that autonomy support and perceived competence are determinants of well-being and HbA1c levels. However, we could not establish the role of autonomous motivation in our study. 
This study provides additional support for the SDT variables as important determinants of diabetes self-management for patients with type 2 diabetes.

Determinants of glycemic control among patients with type 2 diabetes: testing a self-determination theory process model