Psychosocial interventions for reducing diabetes distress in vulnerable people with type 2 diabetes mellitus: a systematic review and meta-analysis

Abstract

Diabetes distress (DD) disproportionately affects vulnerable people with type 2 diabetes mellitus and interventions targeting this population are therefore relevant. A systematic review and meta-analysis was performed to assess the evidence for an effect of psychosocial interventions for reducing DD, and, secondly HbA1c, depression, and health-related quality of life in vulnerable people with type 2 diabetes mellitus. Vulnerability encompasses poor glycemic control (HbA1c >7.5%) and at least one additional risk factor for poor diabetes outcomes such as low educational level, comorbidity, and risky lifestyle behavior. The interventions should be theoretically founded and include cognition- or emotion-focused elements. We systematically searched four databases for articles published between January 1995 and March 2018. Eighteen studies testing a variety of psychosocial interventions in 4,066 patients were included. We adhered to the Cochrane methodology and PRISMA guidelines. Review Manager 5.3 was used for data extraction and risk of bias assessment, and Grades of Recommendation, Assessment, Development and Evaluation for assessing the quality of the evidence. Data were pooled using the fixed or random effects method as appropriate. We investigated effects of individual vs group, intensive vs brief interventions, and interventions with and without motivational interviewing in subgroup analyses. To assess the robustness of effect estimates, sensitivity analyses excluding studies with high risk of bias and attrition >20% were conducted. We found low to moderate quality evidence for a significant small effect of psychosocial interventions on DD, and very low to moderate quality evidence for no effect on HbA1c, both outcomes assessed at 3, 6, 12, and 24 months follow-up. The effect on depression was small, while there was no effect on health-related quality of life. Exploratory subgroup analyses suggested that interventions using motivational interviewing and individual interventions were associated with incremental effects on DD. Likewise, intensive interventions were associated with significant reductions in both DD and HbA1c.

Keywords: HbA1c; diabetes distress; meta-analysis; psychosocial interventions; type 2 diabetes; vulnerable populations.

Figure 1

PRISMA. Source: From Mohen D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(6):e1000097. doi:10.1371/journal.pubmed1000097. For more information, visit www.prisma-statement.org.

Figure 2

Meta-analysis: intervention vs standard care on diabetes distress at 3, 6, 12, and 24 months follow-up; (1) Means converted from scale with range 0-100 with positive outcomes reflecting a higher number; (2) 24 weeks follow-up; (3) SDs calculated from Cls using Revman 5.3; (4) Measured DDS at 8 months FU (ITT); (5) SDs calculated from SE using Revman 5.3; (6) Significant difference at baseline; PAID 59.9 intervention group versus 42.3 in the control group; (7) Means+SDs from Cochrane review (Chew et al. 2017); (8) Measured with DDS at 9 months follow-up; (9) online care (intervention) versus web training (control); (10) Means+SDs from Cochrane review (chew et al. 2017); (11) Measured with DDS (ITT). Risk of bias legend: (A) Random sequence generation (selection bias); (B) Allocation concealment (selection bias); (C) Blinding of participants and personnel (performance bias); (D) Blinding of outcome assessment (detection bias); (E) Incomplete outcome data (attrition bias); (F) Selective reporting (reporting bias); (G) Other bias.

Figure 3

Meta-analysis: intervention vs standard care on HbA1c at 3, 6, 12, and 24 months follow-up; (1) SDs calculated from CI using Revman 5.3; (2) At 8 months followup (ITT); (3) Mean+SDs calculated from within group differences; (4) SDs Calculated from SE using Revman 5.3; (5) Means+SDs from Cochrance review (Chew et al. 2017); (6) At 9 months follow-up; (7) Means+SDs from Cochrance review (Chew et al. 2017); (8) ITT; Risk of bias legend; (A) Random sequence generation (selection bias); (B) Allocation concealment (selection bias); (C) Blinding of participants and personnel (performance bias); (D) Blinding of outcome assessment (detection bias); (E) Incomplete outcome data (attrition bias); (F) Selective reporting (reporting bias); (G) Other bias