Scholarlink

Aims <p>This study examined psychosocial, clinical, and healthcare system-related factors associated with medication adherence among patients with Type 2 Diabetes Mellitus (T2DM) in Bengaluru, India, using Structural Equation Modelling (SEM).</p> Methods <p>A hospital-based cross-sectional study was conducted at a tertiary care hospital in Bengaluru between March and May 2025, enrolling 600 adult T2DM patients through purposive sampling. Validated instruments were used to measure medication adherence (Morisky Medication Adherence Scale, MMAS-8), self-efficacy (Diabetes Management Self-Efficacy Scale, DMSES; 4 items retained after measurement refinement), perceived social support (Multidimensional Scale of Perceived Social Support, MSPSS; 10 items), health literacy (Health Literacy Questionnaire, HLQ), and health-related quality of life (EQ-5D-5&#xa0;L). Given its very low internal consistency in this sample (Cronbach’s α = 0.235), MMAS-8 was treated as an observed adherence index rather than a latent construct. Adherence was classified using the standard MMAS-8 scoring convention: scores of 8 were classified as high adherence, scores of 6 to &lt; 8 as medium adherence, and scores &lt; 6 as low adherence; for bivariate analyses, scores ≥ 6 were categorised as adherent and scores &lt; 6 as non-adherent. A two-stage SEM approach was applied, combining preliminary PLS-SEM for measurement refinement with covariance-based SEM (CB-SEM) estimated via the robust Maximum Likelihood estimator (MLR) for structural inference.</p> Results <p>Self-efficacy was the strongest and only statistically significant direct predictor of medication adherence (β = 0.512, z = 8.892, <i>p</i> &lt; 0.001). Social support was strongly associated with self-efficacy (β = 0.551, z = 14.312, <i>p</i> &lt; 0.001; R² = 0.304) but did not directly predict adherence (β = −0.098, z = − 1.829, <i>p</i> = 0.067). Health system access (β = −0.063, <i>p</i> = 0.097) and medication complexity (β = −0.025, <i>p</i> = 0.506) were non-significant. The model explained 22.1% of the variance in adherence (R² = 0.221). Model fit was excellent: χ² (116) = 131.024, χ²/df = 1.13, CFI = 0.997, GFI = 0.972, RMSEA = 0.015, PNFI = 0.839, PGFI = 0.737.</p> Conclusions <p>Self-efficacy is the pivotal proximal determinant of medication adherence among urban Indian T2DM patients. Social support operates as a distal facilitator through its enhancement of self-efficacy. Interventions targeting diabetes self-management education, psychological empowerment, and structured family-centred counselling are recommended.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Assessment of factors influencing medication adherence in patients with Type 2 Diabetes Mellitus in Bengaluru using structural equation modelling

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月8日

查看全文
  • L. Revati,
  • Chitra Selvan,
  • Vivek Verma,
  • Philip Morisky,
  • Denny John

摘要

Aims

This study examined psychosocial, clinical, and healthcare system-related factors associated with medication adherence among patients with Type 2 Diabetes Mellitus (T2DM) in Bengaluru, India, using Structural Equation Modelling (SEM).

Methods

A hospital-based cross-sectional study was conducted at a tertiary care hospital in Bengaluru between March and May 2025, enrolling 600 adult T2DM patients through purposive sampling. Validated instruments were used to measure medication adherence (Morisky Medication Adherence Scale, MMAS-8), self-efficacy (Diabetes Management Self-Efficacy Scale, DMSES; 4 items retained after measurement refinement), perceived social support (Multidimensional Scale of Perceived Social Support, MSPSS; 10 items), health literacy (Health Literacy Questionnaire, HLQ), and health-related quality of life (EQ-5D-5 L). Given its very low internal consistency in this sample (Cronbach’s α = 0.235), MMAS-8 was treated as an observed adherence index rather than a latent construct. Adherence was classified using the standard MMAS-8 scoring convention: scores of 8 were classified as high adherence, scores of 6 to < 8 as medium adherence, and scores < 6 as low adherence; for bivariate analyses, scores ≥ 6 were categorised as adherent and scores < 6 as non-adherent. A two-stage SEM approach was applied, combining preliminary PLS-SEM for measurement refinement with covariance-based SEM (CB-SEM) estimated via the robust Maximum Likelihood estimator (MLR) for structural inference.

Results

Self-efficacy was the strongest and only statistically significant direct predictor of medication adherence (β = 0.512, z = 8.892, p < 0.001). Social support was strongly associated with self-efficacy (β = 0.551, z = 14.312, p < 0.001; R² = 0.304) but did not directly predict adherence (β = −0.098, z = − 1.829, p = 0.067). Health system access (β = −0.063, p = 0.097) and medication complexity (β = −0.025, p = 0.506) were non-significant. The model explained 22.1% of the variance in adherence (R² = 0.221). Model fit was excellent: χ² (116) = 131.024, χ²/df = 1.13, CFI = 0.997, GFI = 0.972, RMSEA = 0.015, PNFI = 0.839, PGFI = 0.737.

Conclusions

Self-efficacy is the pivotal proximal determinant of medication adherence among urban Indian T2DM patients. Social support operates as a distal facilitator through its enhancement of self-efficacy. Interventions targeting diabetes self-management education, psychological empowerment, and structured family-centred counselling are recommended.