A stratified, parallel group, randomized controlled study was conducted in family physician clinics and community pharmacies across Canada from July 2006 to April 2010. The University of Western Ontario Centre for Studies in Family Medicine served as the Coordinating Centre. Ethics approval was obtained from academic ethics committees across Canada including The University of Western Ontario.
Family physician enrollment occurred from July 2006 to September 2008. Physicians invited to participate were randomly selected from a published national directory [34
]. Physicians provided written, informed consent, and were screened for eligibility and included: full time status [>25 hours/week]; ability to generate a list of patients with T2DM [ICD-9, 250 billing code]; minimum of 35 patients with T2DM, 4 insulin-eligible patients in the practice; and attendance at a mandatory workshop. Physician study exclusion criteria included: academic practice; participation in a diabetes behaviour-change intervention trial; planned retirement or moving practice to another city; or planned extended locum coverage beyond 4 weeks during the subsequent 12 month intervention period. Enrollment ended when all physicians assessed for eligibility consented, refused consent, or were deemed unable to follow-up (i.e. moved).
Community pharmacist recruitment was conducted utilizing a national registry at the Department of Family Medicine at McMaster University. Potential pharmacists in geographic proximity to the postal codes of consenting physicians were identified and surveyed to determine their diabetes education training and pharmacy services and resources.
Physicians generated a list of all patients with T2DM in their practice and mailed study information to all. Study auditors reviewed the charts of patients who provided written informed consent and collected year of birth, year of T2DM diagnosis, glycosylated haemoglobin (HbA1c) values and glucose-lowering medications. The remaining charts of patients who either, did not consent or did not respond, were reviewed by the physician to determine insulin-eligibility. Patients were considered insulin-eligible if they had an HbA1c ≥7.5% (most recent laboratory value) and their oral anti-diabetes drug (OAD) score was ≥1.5 (OAD score is the sum of all OADs prescribed; OAD ½ to maximum dose
1 OAD) [35
]. Lantus® or NPH insulin was available free of charge for 6 months to all insulin-eligible patients.
All eligible physicians were stratified by the study geographic site and their level of comfort prescribing insulin (determined by questionnaire) and randomly allocated (1:1) in a blocked manner to an insulin initiation strategy (intervention) or usual care (control) by the Coordinating Center. Sanofi-aventis generated the mechanism used to implement the random allocation sequence. A follow-up chart audit commenced 15 months post workshop (defined for each physician; January 2007 – March 2010).
Study sites hosted insulin initiation workshops for all physicians enrolled in the study to ensure comparable knowledge on the appropriate use of insulin therapy in T2DM. All physicians received a complete registry of insulin-eligible patients in their practice. Pharmacists attended the same program but were educated separately to avoid contamination. The workshop focussed on education regarding glycemic control, the rationale for insulin prescription, and common patient barriers. Additional activities included viewing of an insulin initiation video and hands-on experience with an insulin pen. For physicians, summary chart audit data were presented and individual practice-specific ‘report cards’ distributed. Physicians were notified of their randomization status at the workshop by the Coordinating Center. Intervention physicians had the opportunity to meet pharmacist(s) with whom they were matched.
Intervention – insulin initiation strategy
Physicians randomized to the intervention group were provided with a 12-month insulin initiation strategy consisting of (1) diabetes specialist/educator consultation support (active diabetes specialist/educator consultation support for 2 months [the educator initiated contact every 2 weeks] and passive consultation support for 10 months [family physician initiated as needed]); and (2) community retail pharmacist support (option to refer patients to the pharmacist(s) for a 1-hour insulin-initiation session). The session checklist included education on insulin action, injection sites, the pen device, and hypoglycemia awareness and treatment.
The primary outcome was the physician’s insulin prescribing rate (IPR)—the number of insulin starts per the 12-month intervention of insulin-eligible patients.
Secondary clinical outcome measures, limited to the data from the audit of consenting patients’ charts, included: HbA1c; fasting plasma glucose (FPG); OAD prescription and score; insulin prescription and dosage; proportion of patients at HbA1c target (≤7.0%); and proportion of patients with intensification of diabetes management (i.e. increased dose of OAD or insulin, increased OAD score, or the addition of insulin). For those patients prescribed insulin during the intervention, additional outcomes included: number of days from study start to insulin initiation; the change of HbA1c, FPG, and OAD score; and type(s) of insulin prescribed from initiation of insulin to 3 and 6 months post-initiation.
The physician knowledge, attitude and self-efficacy questionnaire was used to create change scores (pre – post intervention) to measure physician knowledge, attitude and self-efficacy [36
] for both glycemia control and insulin initiation and titration. This questionnaire was developed by the Coordinating Centre for this project and tested for content validity (example questions are provided in Table ).
Example questions from the knowledge attitude and self-efficacy questionnaire
Sample size calculation
With a standard deviation of 7.1 estimated from Poisson regression, two-sided alpha of 0.05 and power of 0.90, the sample size required was 89 physicians per group (n
178). Protocol modifications due to recruitment challenges included inclusion of group practice physicians (excluded in original protocol), and reduction of the minimum required number of patients with T2DM (original protocol
50) and insulin-eligible patients (original protocol
The unit of analysis was the physician. Intention-To-Treat analysis was performed on the primary outcome with the IPR imputed as zero if data were not available. The analyses of secondary outcomes were based on all available data. A p
0.05 was deemed statistically significant. Analyses were generated using SAS Version 9.1. All data were examined and the appropriate analyses were employed.
The IPR was analyzed using Poisson regression with intervention group as a class effect and mean HbA1c at baseline as a covariate from which the mean number of people started on insulin per 12 months, standard error (SE) and 95% confidence intervals (CIs) were calculated.
Continuous variable changes from workshop to 15 months post-workshop were examined using analysis of covariance (ANCOVA) with the intervention group, baseline mean HbA1c, insulin comfort stratum and pooled site as class effects, and corresponding baseline value of the variable of interest as a covariate.
ANOVA procedures were used to compare intervention and control baseline physician and practice demographics for continuous variables and chi-square tests for categorical variables. Knowledge, attitude and self-efficacy change scores were analyzed using one-way ANOVA.