Use of automated reminder letters to improve diabetes management in primary care: Outcomes of a quality improvement initiative

Background Effective management of patients before the intervention, start of the intervention, end of with diabetes mellitus (DM) can be time-consuming the 15-month ntervention period and six months after and costly. One patient-centred quality improve- the intervention.ment strategy is to generate reminder letters to Results The number of letters sent per month prompt patient action(s), but this strategy's effect ranged from 284 to 392, representing 28-38% of on DM outcomes is uncertain. all patients with DM. At the end of the intervention, Aim To determine whether using the electronic patients' odds of being at goal were higher than medical record to automatically generate reminder before the intervention began for LDL < 100 mg/dl, letters for patients not meeting recommended DM and for HbA_lc and LDL tested once within the last targets is associated with improvement in practice- 12 months (or 1.24, P = 0.005; or 1.35, P = 0.03; or level quality metrics for DM management. 1.48, P< 0.001, respectively). Post intervention, Methods Over 15 months, letters were sent monthly declines were seen in LDL checked within the last to all patients with DM in a large, urban, primary 12 months (or 0.76, P= 0.003) and in the composite care teaching practice whose records for haemo- endpoint (or 0.78, P - 0.005). globin A_1c (HbA_1c), low-density lipoprotein LDL Conclusions The automated patient-reminder letter or blood pressure (BP) indicated non-compliance intervention was associated with modest improve with recommended levels and testing intervals. ments in several, but not all DM measures. This Logistic regression was used to analyse cross-sectional, approach may be an effective tool for improving practice-level differences in the proportion of patients quality of care for patients with DM.