Introduction
Methods
Data Sources and Searches
Study Selection
Data Extraction and Quality Assessment
Data Synthesis and Analysis
Results
Study Selection
Study Characteristics and Risk of Bias
Study | Methods | Participants | Interventions | Outcomes | Risk of bias |
---|---|---|---|---|---|
Aspinall et al. (2011) [11] | –Controlled before-after trial –Conducted at 21 veterans affairs’ networks in the USA | –Community-dwelling veterans (mean age 77 years, 99.5% male) on glyburide, age ≥65, serum creatinine ≥2 mg/dL –Intervention group n = 4368 (baseline HbA1C 7.2%) –Control group n = 1886 (baseline HbA1C 7.2%) | –Intervention group (“targeted” cohort): information regarding risk of hypoglycemia in older persons on glyburide and instructions for switching to alternative agent provided to pharmacists, who could then contact patients’ physicians to deprescribe –Control group (“non-targeted” cohort) received usual care | –Discontinuation rate for glyburide up to 135 days after index prescription –HbA1C at 3–9 months –Hypoglycemia over 9 months | –Serious due to contamination of intervention in control group, and insufficient matching of intervention and control group |
Sjöblom et al. (2008) [10] | –Controlled before-after trial –Participants were patients at 17 different nursing homes in Sweden | –Nursing home patients (mean age 84 years, 42% male) with a HbA1C level of ≤6.0% on any diabetes medication(s) –Intervention group n = 32 (baseline HbA1C 5.2% (33) –Control group n = 66 (baseline HbA1C 7.1%) | –Intervention group: all oral antihyperglycemics and insulin ≤20 units per day were discontinued, insulin >20 units per day was reduced by 50% –Control group received usual care | –HbA1C at 3 and 6 months –Discontinuation rates –Mortality at 6 months | –Serious due to selection bias and confounding which was not adjusted for |