Table 1

Variables within the predictive framework

Framework data and, for outcomes only, MIDMeasure, cut-offs as appropriate and completion mechanismsJustification for inclusion
Demographic dataIMD: participant's home postcode was used to identify IMD ward. Deprived areas are those ranked lower than 6562 (the 20% most deprived wards in the UK)34
Ethnicity: self-report
Gender, age, time since diagnosis and occupation are not known to be related to DSME effects so were not included in the framework
Deprivation is linked to less successful management of type 2 diabetes35 36
Ethnicity is linked to type 2 diabetes prevalence23 and poorer self-management/diabetes outcomes37
Knowledge of diabetesRDKS37 is a 19-item multiple choice scale assessing diabetes-related knowledge. Correct responses are coded as ‘1’ and incorrect responses are coded as ‘0’, and these are summed to give a possible score between 0 and 19Knowledge is modifiable by intervention and people with low knowledge at baseline might expect to increase their knowledge, and subsequently improve their behavioural outcomes, following DSME38
Self-efficacyDMSES tool is a 15-item scale that has been validated with UK populations.37 39 Items are scored on a 0–10 Likert-type scale to indicate how confident they are at the task described. Responses are summed, giving possible self-efficacy scores between 0 (no self-efficacy) and 150 (very high self-efficacy)Self-efficacy is modifiable by DSME.8 People with low self-efficacy might be expected to raise this following DSME
Diabetes distress
The MID=half a SD.40 The SD was 20.84, therefore the MID=10 scale points
PAID41 is a 20-item scale measuring emotional functioning relating to diabetes, with each item scored on a 5-point Likert-type scale (0–4). Responses are summed and multiplied by 1.25 to give an overall score between 0 and 100. In order to categorise baseline levels of distress, PAID scores were categorised as either high distress (over 40 scale points), medium distress (20–40), or no distress (under 20)Diabetes distress is modifiable by DSME.8 25 People with high distress might be expected to have lower diabetes distress following DSME
MID=1.5 scale points.42
MID=1.5 scale points42
HADS43 is a 4-point Likert-type scale to indicate the extent of 14 anxious and depressive feelings over the past week. Responses are coded from 0 to 3 and the total is computed for each subscale (giving total scores between 0 and 21). Scores over 8 indicate clinical levels of anxiety/depression.44 The HADS has been used and validated with diabetic populations45Depression is known to compromise self-management efforts24 and so it is likely that no changes in depression or key self-management outcomes will be seen in someone with depression
The clinical cut-off for uncontrolled diabetes, and for participation in the study, is 7.4% or 57 mmol/mol46DSME has been shown to have an effect on HbA1c, and this is a key clinical marker of disease control5 6
Waist circumference
For white and black men waist measurement should be below 94 cm, for Asian men it should be below 90 cm, and for all women it should be below 80 cm48Waist circumference is a better predictor of health, and particularly type 2 diabetes, than is overall weight or BMI49 50
Physical activity
MID=an increase of 2500 steps per day.51
Yamax Powerwalker accelerometer was used for 3 days (including one weekday and one weekend day) to record (1) the number of steps (2) the number of kilometres walked and (3) number of calories burned. Data from the accelerometer was averaged for the 3 days. The recommended average steps per day is 10 00051DSME focus on physical activity was high so the potential for 2500 step increase was change easy to observe
Change talk: changes madeProcess measure identified during QA audio-recordings and mid-point interviews for 17/27 participants. A brief description of changes already made since starting the DSME was included. For example, patient 1 said: “I go to the gym three times a week now […] which I haven't done for about 20 years”Patient-led change talk indicates readiness to initiate/sustain behaviour changes52
Change talk: changes plannedProcess measure identified during QA audio-recordings and mid-point interviews for 17/27 participants. A brief description of changes planned during the QA consultation was included. For example, patient 24 said that she planned to increase her exercise so that she made herself out of breath: “Just when I do walk to step it up, yes to make sure to stop walking on the flat and taking it nice and easy, just to pick a few hills and go for it [laughs]. […] just make myself out of breath [for] more than five or ten minutes”Patient-led change talk indicates readiness to initiate/sustain behaviour changes52
Treatment satisfaction with DSMEProcess measure identified during QA audio-recordings and mid-point interviews for 17/27 participants. A brief description of comments made about the DSME was included. For example, patient 4 was very positive about the DSME: “I'm feeling actually much better and after going through my manual I felt it was quite informative…I enjoyed reading it”Assessment of treatment satisfaction of an indicator of usefulness to participant
  • BMI, body mass index; DMSES, Diabetes Management Self-Efficacy Scale; HADS, Hospital Anxiety and Depression Scale; HbA1c, glycated haemoglobin; IMD, Index of Multiple Deprivation; MID, minimal important difference; PAID, Problem Areas In Diabetes Scale; RDKS, Revised Diabetes Knowledge Scale; QA, Quality Assurance.