Psychological self-care in chronic conditions : development of digital interventions in type 1 diabetes and atopic dermatitis

Abstract

Background: Psychological interventions for somatic conditions have a long history with many successful examples, often based on cognitive behavioural therapy (CBT) However, psychologists are not commonly found in somatic healthcare. Therefore, it is of interest to develop self-guided interventions based on CBT for somatic disorders, for easy dissemination in regular healthcare. In this thesis, intervention in different stages of development were investigated for people with type 1 diabetes mellitus (T1DM) and for people with atopic dermatitis (AD) in both cases based on CBT.

Aims: The overall aim in this thesis was to document the development of self-guided psychological interventions for somatic conditions, by a novel intervention for T1DM and a shortened and adapted intervention for AD.

Method: The interventions developed for T1DM were inspired by earlier CBT research on T1DM and other somatic conditions. The self-guided intervention for AD was adapted from an evaluated clinician-guided intervention. Study I was a case report describing the treatment of two patients with T1DM who finished an early version of a 10-week CBT programme. Study II was a feasibility study in the setting of a diabetes clinic. Eight participants used two digital tools, focused on problem solving and exposure for four weeks. Study III was a feasibility study about a self-guided CBT-based intervention for AD with 21 participants. Study IV was a report on the secondary outcomes of study III. Study V was a randomised clinical non-inferiority trial with 168 participants randomised to a self-guided intervention for AD or to a clinician-guided intervention.

Results: In study I, the 2 participants reported clinically significant improvements. Case 1 reported reduced blood glucose levels and improvements in his health behaviours. Case 2 reported clinically significant improvements in fear of hypoglycaemia. Both cases reported high satisfaction with their treatments and reported no adverse events. In study II, the results did not unreservedly support feasibility. All participants used the problem-solving tool, and three out of eight participants had used the exposure tool. No serious adverse events were reported. Results on feasibility were inconsistent, with credibility generally rated low, and usability below cut-off. However, some participants reported small improvements in outcome measures. As an additional tentative finding, a preliminary analysis of participant’s comments and reported problems suggested two types of users: One that may have more use for a problem-solving tool and one with fear of hypoglycaemia, that may be more helped by exposure. In study III, participants had a significant decrease on AD symptoms at post measurement immediately after intervention with a moderate within-group effect size (d=0.61 95% CI 0.11-1.12) which had improved to a moderate to large within-group effect size at 3-month follow up (d=0.84 95% CI 0.38-1.37) No serious adverse events were reported. The revised intervention consisted of 16 726 words, shortened from 111 142 words. In study IV, secondary outcomes on study III were reported. The intervention was feasible in all measures, except that system usability was rated 67 on average on the System Usability scale (range 0-100), below the scale’s authors’ cutoff of 70. At 3-month follow-up, within-group effect sizes on secondary outcomes were moderate to large. In study V, no significant differences were found between the groups. The estimated mean difference in change of 0.36 points (1-sided 97.5 % CI, −∞ to 1.75) which was lesser than the 3-point noninferiority margin. A clinician administering the clinician-guided group spent on average 50 minutes compared to 16 minutes for the self-guided group.

Conclusions: Digital self-guided tools for people with T1DM may work for some people but would need improvements before a larger randomised trial. It’s possible that a general intervention based on problem-solving, and a specific intervention based on exposure for fear of hypoglycaemia would be a way to move forward in the development of interventions for T1DM. A self-guided CBT based intervention for AD was found feasible and effective and could be easily implemented in primary or specialist care. The results also support the feasibility of adapting an existing clinician-guided intervention into a self-guided variant. Self-guided interventions may have the advantage of less time spent by clinicians and therefore less use of resources. An additional possible advantage may be the less time spent by patients on reading written material.