UNLABELLED: Abstract Aims: To develop and evaluate a screening tool to identify people with diabetes at increased risk of medication problems relating to hypoglycaemia and medication non-adherence. METHODS: A retrospective audit of attendances at a diabetes outpatient clinic at a public, teaching hospital over a 16-month period was conducted. Logistic regression was undertaken to examine risk factors associated with medication problems relating to hypoglycaemia and medication non-adherence and the most predictive set of factors comprise the Diabetes Medication Risk Screening Tool. Evaluating the tool involved assessing sensitivity and specificity, positive and negative predictive values, cut-off scores, inter-rater reliability, and content validity. RESULTS: The Diabetes Medication Risk Screening Tool comprises seven predictive factors: Age, living alone, English language, mental and behavioural problems, comorbidity index score, number of medications prescribed, and number of high-risk medications prescribed. The tool has 76.5% sensitivity, 59.5% specificity, and has a 65.1% positive predictive value, and a 71.8% negative predictive value. A score of 27 or more out of 62 was associated with high-risk of a medication problem. The inter-rater reliability of the tool was high (κ = 0.79, 95% CI 0.75-0.84) and the content validity index was 99.4%. CONCLUSION: The Diabetes Medication Risk Screening Tool has good psychometric properties and can proactively identify people with diabetes at greatest risk of medication problems relating to hypoglycaemia and medication non-adherence.
UNLABELLED: Abstract Aims: To develop and evaluate a screening tool to identify people with diabetes at increased risk of medication problems relating to hypoglycaemia and medication non-adherence. METHODS: A retrospective audit of attendances at a diabetesoutpatient clinic at a public, teaching hospital over a 16-month period was conducted. Logistic regression was undertaken to examine risk factors associated with medication problems relating to hypoglycaemia and medication non-adherence and the most predictive set of factors comprise the Diabetes Medication Risk Screening Tool. Evaluating the tool involved assessing sensitivity and specificity, positive and negative predictive values, cut-off scores, inter-rater reliability, and content validity. RESULTS: The Diabetes Medication Risk Screening Tool comprises seven predictive factors: Age, living alone, English language, mental and behavioural problems, comorbidity index score, number of medications prescribed, and number of high-risk medications prescribed. The tool has 76.5% sensitivity, 59.5% specificity, and has a 65.1% positive predictive value, and a 71.8% negative predictive value. A score of 27 or more out of 62 was associated with high-risk of a medication problem. The inter-rater reliability of the tool was high (κ = 0.79, 95% CI 0.75-0.84) and the content validity index was 99.4%. CONCLUSION: The Diabetes Medication Risk Screening Tool has good psychometric properties and can proactively identify people with diabetes at greatest risk of medication problems relating to hypoglycaemia and medication non-adherence.