Collaborative care, associated with the psychological intervention and educational programs for patients with diabetes and comorbid depression, has been an area that has not been thoroughly explored, with limited evidence on the effects it has on clinical outcomes and adherence. This meta-analysis suggests that numerous interventions are required to maintain adherence to long-term treatment in conditions such as diabetes.

Diabetes Control

Diabetic control was measured using HbAc1 levels in all 10 RCTs to indicate if there was any change in the levels between baseline and following the interventions. The primary aim of all these studies was to determine the effectiveness of care interventions on adherence measured in outcomes of depression and diabetes control.

A total of 6 out of 10 studies showed an improvement in HbA1c levels. Bogner et al. [28] found patients randomized to their integrated care intervention showed higher rates of adherence to diabetic medication. Glycaemic control was also improved supporting a link between adherence and improvement in diabetic control. Similarly, Cummings et al. [29] found that patients in their intervention groups experience an average decrease of 1.0% in HbAc1. This finding is consistent with a prior study conducted in the same year by Lauffenburger et al. [27] which also concluded that glycaemic control was improved with high-intensity interventions compared to usual care. Sharif et al. [30] and Weinger et al. [33] found that HbA1c mean variation was significant in intervention groups with Sharif et al. [30] stating that the change in blood glucose changes was significant but had uncertainty regarding intervening factors that could have accounted for this large change such as patients knowing their blood glucose levels before the interventions due to ethical reasons. Comparably, Huang et al. [26] showed that participants displayed statistically significant improvements in HbA1c 90 days after interventions which shows the potential for long-term improvements as well as short-term.

Conversely, 4 out of the 10 chosen studies showed no improvements in diabetic control following interventions. Piette et al. [31] stated that although there was no improvement in haemoglobin A1c values, they were initially reasonably good at the beginning and were therefore unaffected by the intervention. This was identical to findings from Williams et al. [35] as they found that patients had good glycaemic control at baseline so there was limited power to detect small clinical changes in HbA1c levels, hence no improvement was observed in both studies. Newby et al. [34] and Ell et al. [36] also showed no differences in HbA1c levels with levels increasing in the intervention group compared to usual care. However, both of these studies measured HbA1c through patient self-reporting. This could be a factor in why there were no visible improvements.

Depression Control

Eight out of 10 studies showed an improvement in depressive symptoms following the intervention. Bogner et al. [28] identified that compared to patients who are not depressed, depressed patients who have diabetes are more likely to be non-adherent to medication regimens, including their antidepressants and diabetic medications. Cummings et al. [29] postulated a hypothesis that patients with T2DM are twice as likely to have depressive symptoms and that the relationship between the two could be bidirectional. Shariff et al. [30] and Newby et al. [34] and agreed, stating that proactive detection and treatment of depression is essential to reduce the burden of depression in people with T2DM and that depression interferes with metabolic control of diabetes leading to weak control of blood glucose levels. All of these studies showed improvement in depression symptoms with intervention. Two of the studies [26, 28] showed a decrease in depression scores and improvement in QoL© immediately after the intervention, and with Huang et al. [26] improvements were seen 90 days later in follow-up routine appointments. Newby et al. [34] and Williams et al. [35] had similar results whereby patients were seen to have less severe depression and greater progress in overall functioning than in usual care measures with a comparison from baseline to after intervention showing substantial developments. On the contrary, Weinger et al. [33] showed no improvements in depression outcomes. Since this difference has not been found elsewhere it can be assumed that this is possibly due to other factors and not solely based on intervention measures not being effective.

Adherence

Adherence is the main factor identified to have an impact on clinical outcomes in all chronic diseases [37]. All the studies evaluated whether intervention on depression had an impact on adherence to medication and overall improvement in clinical outcomes. Weinger et al. [33] identified that an important reason for poor glycaemic control is the patient’s difficulty in following treatment and self-management recommendations. This is directly linked to adherence, as adherence is the patient’s commitment to their treatment and if they find this difficult then it will lower disease control in those patients. Lauffenburger et al. [27] implied that effectively targeting adherence to the patients most likely to benefit, has the potential to improve efficacy in treatment but it has not been widely assessed.

Newby et al. [34] suggested that depression could be a major factor in lower adherence and Ell et al. [36] looked at enhancing adherence by integrating depression and diabetes care as they projected that providing care for both diseases will aim to improve adherence overall for diabetes with comorbid depression. Bogner et al. [28] agreed as they explored an adherence-based approach as they stated that even though pharmacological treatment is highly efficacious, many patients do not adhere to the treatment and this is a high risk especially in patients with comorbid diabetes and depression. They hypothesized that integrating the care of the two with interventions would enhance adherence.

Cummings et al. [29] suggested that depression symptoms are a risk factor for poor adherence, and their results showed that there were significant reductions in depressive symptoms and consequent improvement in medication adherence. Similarly, Huang et al. [26] found that their intervention group facilitated an increase in adherence rates as significant improvement in the psychological state helped strengthen diabetes by improving glycaemic control. Sharif et al. [30] discussed how CBT interventions were effective for treating depression but they have been poorly utilized in depression associated with physical illness. However, it has proven effective in diabetic patients and has been proven to improve adherence measured in depression outcomes which were evident in results observed in this study.

Interestingly, Piette et al. [31] observed depression symptoms improving greatly, however, this was not reflected in the patient’s medication adherence as there were no important differences found. Medication adherence was measured using the Morisky© medication adherence scale. This raises the question of whether improvement in depression symptoms has an impact on medication adherence. However, this could be linked to the no change in glycaemic control also. Correspondingly, Williams et al. [35] expected effective treatment for depression could benefit adherence to self-care regimens which therefore would improve diabetic control in patients. They found that patients reported almost perfect adherence to medication however, lower adherence rates were reported for glucose testing and foot inspections. This could be explained by looking at the measure for these self-care behaviors which showed ceiling effects, meaning that the participants already had high scores for the medication adherence so there was little room for improvement.

Intervention

Cognitive behavior therapy is a form of psychological treatment that has been proven effective for depression. It helps change the way of individual thinking and behavior [38]. Out of the 10 studies, 6 used CBT as a form of intervention to assess its effectiveness in improving adherence and consequently clinical outcomes. Cummings et al. [29] stated that an integrated care model involving delivering CBT plus lifestyle counseling to patients with T2D is feasible in primary care practice and has the promise to be highly effective. Newby et al. [34] performed a study on online CBT (iCBT). They stated that it is as efficacious as face-to-face CBT and can be delivered at a fraction of the cost and clinical time. CBT has been shown to help patients feel better when participating in enjoyable activities when they can discuss their mental well-being as it can help the patient be their therapist and use self-management skills to improve adherence and control depressive symptoms [29]. The rest of the studies looked at educational interventions as opposed to psychological measures. They looked at using counseling and collaborative care with physicians offering education and tailored guidance to participants. Bogner et al. [28] looked at addressing patient-subjective factors that could affect adherence and ensured they addressed them during the intervention.

Piette et al. [31] found that telephone-delivered CBT was more likely to increase depression remission at 58% compared to 39% in usual care.

The risk of bias summary table (Figure 4), shows 4 studies having a relatively high risk of bias [26, 31, 35, 36]. This is mirrored in discrepancies in the results, for example, Piette et al. [31] and Williams et al. [35] were the only 2 studies to have no improvement or worsening of glycaemic control, and both had SMD of <0.01. Interestingly, the study with the highest risks of bias, Weinger et al. [33] was the only study not to show improvements in depression outcomes.

Huang et al. [26] projected that motivation enhancement therapy combined with CBT would improve HbA1c and depressive symptoms in patients from baseline to follow-up. The interventions included enhancing patients’ motivation to improve their self-care and introduced stress-coping strategies to help patients cope with hyperglycemia and other symptoms of their conditions. This study found that these interventions significantly helped to improve diabetes and depression control.

Lauffenburger et al. [27] looked at delivering intensive insulin adherence interventions in patients with type 2 diabetes which included regular phone consultations and a weekly text message program that focused on reminding patients on taking their medication and essentially improve medication adherence. They found that high-intensity intervention improved glycaemic control compared to low intensity.

Limitations

This analysis had limitations. Firstly, using combined results from different RCTs, showed differences in measured outcomes, specifically for the depression outcome as different psychometric measures were used to measure depression in participants. This could have been the main cause of the high heterogeneity.

Another limitation is the use of different interventions in each of the studies. Some studies looked at collaborative care whilst others looked at psychotherapy and although this was an aim of this study to assess; the methods to each of these interventions were different e.g., some studies had nurses conduct counseling and educational sessions whilst others had doctors and pharmacists. In addition to that, the number of sessions differed between all the studies. This could have created a bias in the results.

This study included only 10 RCTs with a total of 5759 patients with all except for Lauffenburger et al. [27] having small sample sizes and mainly being conducted in the USA which means that results are not representative of the population as a whole, limiting generalisability. Many studies found for this analysis were not accessible without payment, for which no funding was available.

Four of the studies included results for both Type 1 and type 2 diabetes which could have taken the focus out away from the type 2 diabetes sub-group. Also, some of the studies used self-reporting as a form of measuring depression symptoms and HbA1c % which is demonstrated in the table of bias as a category (Figures 2 and 3) which could have introduced detection bias.

Finally, sub-group analysis was not conducted for short-term effects compared to long-term; only long-term effects were analyzed. This may have identified differences in how interventions work by time and their estimated time of response.

CONCLUSION

In this study, the aim was to investigate whether intervention strategies in the form of psychological (CBT) and educational programs had an impact on the clinical outcomes of depression and diabetes compared to usual care concerning adherence determined by improvements in both outcomes. The consensus for diabetes outcome showed 6 out of 10 studies that had an improvement in HbA1c values in the intervention group compared to usual care. Four of the studies showed either no improvement with values being indistinguishable or an increase in HbA1c value in the intervention group. Depression outcomes showed 8 out of the 9 studies showing a significant improvement in depression outcomes in the intervention group compared to the usual group with only 1 study showing no improvement.

These findings indicate that intervention with psychological and educational measures has a positive impact on depression outcome and is therefore reflected in the improved diabetic HbA1c control which shows a link between the two and adherence measured in diabetic control highlights an improvement.

Recommendations from this study include the investigation of cost-effectiveness to examine the costs and clinical outcomes of depression and diabetes by comparing the intervention to usual care and estimating the cost.  Also, consistent CBT or educational measures need to be investigated over a longer period to investigate the long-term effects compared to the short-term effects this intervention will have. In addition to this, it is recommended that the implementation of depression screening in all diabetic patients be carried out regularly, to manage early depressive symptoms if any, and try preventative measures before resulting in treatment which can be difficult and result in adherence issues from patients.

In conclusion, adherence follows a complex process and is not a singular occurrence, therefore adherence support should be integrated into all health consultations in diabetic patients with or without comorbid depression.

ACKNOWLEDGMENTS: None

CONFLICT OF INTEREST: None

FINANCIAL SUPPORT: None

ETHICS STATEMENT: None