An electronic prompt multiplied diabetes education referrals

Design

Cluster-randomised controlled trial, 18 months

Population

Adults with type 2 diabetes attending six primary care practices (four intervention, two control)

Primary outcome

Referral to diabetes self-management education and support

Effect

12.9% vs 1.2% at 12 months (OR 10.55, 95% CI 4.65 to 23.94); 13.7% vs 1.6% at 18 months (OR 9.44, 95% CI 4.91 to 18.17)

In the PROMPT cluster-randomised trial, four primary care practices in central North Carolina received a best practice advisory in the electronic record that appeared during visits by eligible patients with type 2 diabetes; two practices served as controls. After 12 months, standing orders were added so the prompt also reached nurses or medical assistants before the clinician visit.

At 12 months, referral to diabetes self-management education and support was 12.9% in intervention practices against 1.2% in controls. At 18 months it was 13.7% against 1.6%. Adjusted odds ratios were 10.55 (95% CI 4.65 to 23.94) at 12 months and 9.44 (95% CI 4.91 to 18.17) at 18 months.

The outcome was referral, not attendance, glycaemic change or complications, and only six practices took part. Even with the prompt, about 87% of eligible patients were not referred. Still, the cost of a prompt is low compared with the usual underuse of self-management education. Where there is no electronic record, a nurse-led checklist or standing referral order can do the same job.

• Build a referral trigger into the visit: an electronic prompt, a template field or a nurse checklist.
• Let nurses or assistants start the referral before the clinician sees the patient.
• Track referral and attendance rates by month so gaps are visible.
• Do not count referral as success: follow up whether the patient actually attended.
• Make sure a local education service exists to receive the referrals.

Why it matters

It shows the barrier to self-management education is partly a workflow problem that a small system change can shift.

Don't overread it

Referral rate is a process measure; the trial does not show that more referrals improved glucose control or complications.

The statistics, in plain English

An odds ratio of about 10 describes referral odds, and because referral was rare in controls (about 1 to 2%), the absolute change was about 12 percentage points. The confidence intervals are wide because only six clinics were randomised, and outcomes in the same clinic are correlated.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.