Abstract
AIM: To investigate the impact of medication beliefs, illness perceptions and quality of life on medication adherence in people with decompensated cirrhosis. METHODS: One hundred adults with decompensated cirrhosis completed a structured questionnaire when they attended for routine outpatient hepatology review. Measures of self-reported medication adherence (Morisky Medication Adherence Scale), beliefs surrounding medications (Beliefs about Medicines Questionnaire), perceptions of illness and medicines (Brief Illness Perception Questionnaire), and quality of life (Chronic Liver Disease Questionnaire) were examined. Clinical data were obtained via patient history and review of medical records. Least absolute shrinkage and selection operator and stepwise backwards regression techniques were used to construct the multivariable logistic regression model. Statistical significance was set at alpha = 0.05. RESULTS: Medication adherence was "High" in 42% of participants, "Medium" in 37%, and "Low" in 21%. Compared to patients with "High" adherence, those with "Medium" or "Low" adherence were more likely to report difficulty affording their medications (P < 0.001), lower perception of treatment helpfulness (P = 0.003) and stronger medication concerns relative to medication necessity beliefs (P = 0.003). People with "Low" adherence also experienced greater symptom burden and poorer quality of life, including more frequent abdominal pain (P = 0.023), shortness of breath (P = 0.030), and emotional disturbances (P = 0.050). Multivariable analysis identified having stronger medication concerns relative to necessity beliefs (Necessity-Concerns Differential ≤ 5, OR = 3.66, 95%CI: 1.18-11.40) and more frequent shortness of breath (shortness of breath score ≤ 3, OR = 3.87, 95%CI: 1.22-12.25) as independent predictors of "Low"adherence. CONCLUSION: The association between "Low" adherence and patients having strong concerns or doubting the necessity or helpfulness of their medications should be explored further given the clinical relevance.