Table 1 CONSORT Checklist of items of cluster randomized trials.

Title and abstract

Design

→ Interdisciplinary Diabetes Care Teams operating on the interface between primary and specialty care are associated with improved outcomes of care: Findings from the Leuven Diabetes Project.

Introduction

Background

→ Scientific background and explanation of rationale: see Background section

→ Clustering: randomization at GP-level, primary/secondary outcomes at patient level.

→ Randomized per practice; stratified

Methods

Participants

→ All 379 primary care physicians (PCP's) that actively execute their profession in the project region were invited to participate.

→ The only inclusion criterion for the PCP's is the agreement to bring in all their known patients with type 2 diabetes mellitus. In this way selection bias is prevented. Patients had to provide informed consent before their data could be transmitted for collection and analysis. Only patients with type 2 diabetes mellitus were be included in the study, regardless of their age. Patients who were not capable to provide informed consent were excluded from the study.

→ Data were collected on paper files and from medical records

Interventions

→ See methods section

Objectives

→ See methods section

Outcomes

→ The primary endpoints of the study were the proportion of patients reaching three clinical ADA-targets: (1) HbA1c < 7%; (2) SBD ≤ 130 mm Hg; (3) LDL-C < 100 mg/dl. Secondary endpoints were the mean improvements in individual parameters of 12 validated parameters, i.e. HbA1c, LDL-C, HDL-C, Total Cholesterol, SBP, Diastolic Blood Pressure (DBP), weight, physical exercise, healthy diet, smoking status, statin and anti-platelet therapy.

Sample size

→ The financer to the project imposes a sample size of minimal one third of the potential PCP's. Using the calculator of the university of Aberdeen, sample size for cluster trials was computed. With a significance level of 0.05 and assumed Intra Cluster Coefficient of 0.1, we calculated that 114 clusters with a cluster size of 20 gave 80% power to detect between AQIP and UQIP a 10% in the absolute difference in the proportion of patients achieving a 10% improvement in the primary biochemical endpoints. Based on the fitted mixed models the observed ICC values are: HBA1C: 0.0445, SBD: 0.0466, LDL Cholesterol: 0.0399.

Randomization

Sequence generation

→ After the recruitment period, using computer-generated numbers, a researcher not involved the study and blind to the identity of the practices will perform a randomization stratified by practice size (solo/duo/group practice) and the presence/absence of an electronic medical recording system.

Allocation concealment

→ Program Manager - invitation, stratified.

→ To minimize the possibility of selection bias all patients within a cluster were included

Implementation

→ Allocation: Van Den Broeke Carine, researcher to the scientific team,

→ Enrollment: Borgermans Liesbeth, researcher to the scientific team,

→ Assignment: program manager

Blinding (masking)

→ No blinding was possible at physician level, (both groups presented as 'intervention'), but patients didn't know to which intervention arm their physician belonged.

Statistical methods

→ See methods section, sub-heading statistical analysis.