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

Abstract

Background: Effective management of patients with diabetes mellitus (DM) can be time-consuming and costly. One patient-centred quality improvement strategy is to generate reminder letters to prompt patient action(s), but this strategy's effect on DM outcomes is uncertain.

Aim: To determine whether using the electronic medical record to automatically generate reminder letters for patients not meeting recommended DM targets is associated with improvement in practice level quality metrics for DM management.

Methods: Over 15 months, letters were sent monthly to all patients with DM in a large, urban, primary care teaching practice whose records for haemoglobin A1c (HbA1c), low-density lipoprotein (LDL) or blood pressure (BP) indicated non-compliance with recommended levels and testing intervals. Logistic regression was used to analyse cross-sectional, practice-level differences in the proportion of patients meeting DM quality metrics (HbA1c < 7%, LDL < 100 mg/dl and BP < 130/80 mmHg; rates of checking each value within the last 12 months; and a composite of these five measures) across four time points: six months before the intervention, start of the intervention, end of the 15-month intervention period and six months after the intervention.

Results: The number of letters sent per month ranged from 284 to 392, representing 28-38% of all patients with DM. At the end of the intervention, patients' odds of being at goal were higher than before the intervention began for LDL < 100 mg/dl, and for HbA1c and LDL tested once within the last 12 months (or 1.24, P = 0.005; or 1.35, P = 0.03; or 1.48, P < 0.001, respectively). Post intervention, declines were seen in LDL checked within the last 12 months (or 0.76, P = 0.003) and in the composite endpoint (or 0.78, P = 0.005).

Conclusions: The automated patient-reminder letter intervention was associated with modest improvements in several, but not all DM measures. This approach may be an effective tool for improving quality of care for patients with DM.