Lifestyle Strategies for Improving Diabetes Outcomes

This is a single-center, prospective, unblinded, randomized, controlled, clinical trial of 200 subjects randomized to either a Supervised Training Program (STP), Incentivized Weight Loss (IWL) or Standard of Care (SoC).

Subjects randomized to the SoC group will be asked to return for regular 3-monthly clinic visits and for diet and exercise recommendations.The STP group will be offered once-weekly one hour supervised training sessions at no charge for a period of six months in the bcdiabetes.ca fitness center (located in the Vancouver General Hospital Employee Centre gym).

The IWL group will receive a financial reward for achieving monthly weight loss targets. Those who fail to meet their monthly goals will receive no financial reward. Such financial incentives have been considered in previous research. Due to the nature of the intervention, none of the investigative staff, trainers or subjects will be blinded to randomized assignment.

The primary objective of this study is to investigate whether a behavioral modification intervention in the form of either a 6-month supervised training program or a 6-month financial incentive program, results in overall superior clinical outcomes in comparison to standard of care. Specifically, improvement in glucose control, measured by A1c, will be considered. Given the objective nature of the Primary Outcome, there is little possibility of assessment bias of the primary or clinical outcomes.The secondary objective is to investigate the impact of a supervised training program and an incentivized weight loss program on weight-loss, LDL (apoB), blood pressure, waist-hip circumference, and diabetes-specific quality of life measures including behavioral and emotional impact (refer to Secondary Outcomes section).

Patients attending BCDiabetes ( Diabetes Clinic at Diamond Center )will be approached by one of their diabetes caregivers who will describe the study and determine whether they have interest in participating. If the patient expresses an interest, they will be provided additional information and a study consent form and be free to ask questions about the study. They may choose to provide consent at the time or to leave with the consent form and return at a later time to ask further questions. A research nurse, coordinator and/or co/principal investigator will be available to answer any questions the patient may have. When consent is provided, confirmation of the eligibility criteria will be completed and the subject will receive their randomized group assignment.

Eligible subjects will be randomized in a 2:1:1 fashion to the SoC, STP and IWL and groups respectively. Randomization will be stratified by age (≤ 50 vs > 50 years of age) to ensure balance in age groups in all three treatment arms; enrollment will not be stratified and the number of participants in the two age groups will not be restricted. Randomization within strata will be done in a permuted complete block design with random block sizes (blocks no smaller than 6 and no larger than 15). Randomization will be completed in advance and assignments will be kept in individual, separate sealed, sequentially labeled envelopes opened at the time of randomization of each individual.

Withdrawal from Study or Intervention:

Although they should be encouraged to continue participation in their randomized group, subjects can choose to withdraw from the study, or to withdraw from the intervention at any time. Subjects who choose to withdraw from the STP and IWL arms should continue to receive standard of care and will be considered cross-overs to SoC arm; these subjects are not necessarily withdrawing from the study and should be encouraged to continue to return for their scheduled 3 and 6 month study visits so study outcomes can be measured. If a subject chooses to withdraw from the study entirely, all attempts should be made to have them complete an exit visit where all study outcomes can be captured.

Subjects who experience any adverse event related to the STP arm (eg: physical injury while training) or experience the onset or worsening of an signs or symptoms related to a medical condition (eg: shortness of breath, dizziness, vomiting nausea, etc) persisting the day following a supervised training session, or who experience a worsening of a physical condition (eg: degenerative disc disorder, knee or hip pain, plantar fasciitis, etc) will be withdrawn from the STP intervention. As above, these subjects will be asked to return for their scheduled 3 and 6 month clinic visits.

If at any point more than 15% of subjects in the STP arm are withdrawn from the intervention arm due to an adverse event, worsening medical or physical condition, the study will be stopped.

Study data will be collected as part of the subject's electronic medical record at bcdiabetes.ca. Identifiable study data will be housed exclusively on bcdiabetes.ca; if removed from bcdiabetes.ca servers, subjects will be identified only with a study registration number different from their BC CareID number. Remote monitoring by EMMES using confidential authenticated automatically scheduled direct database queries. EMMES Canada will also provide oversight of quality assurance & monitoring.

The primary outcome measure is the percent of subjects who achieve target for A1c. The null and research hypothesis supporting the primary objective of investigating the impact of whether a behavioral modification intervention in the form of either a 6-month supervised training program or a 6-month incentivized weight loss program, results in overall superior clinical outcomes in comparison to standard of care is then:

H0: pSoC ≥ pIWL and pSoC ≥ pSTP vs. H1: pSoC < pIWL or pSoC < pSTP

where pSoC , pSTP and pIWL are the proportion of subjects achieving the target for all three disease measures. Here, rejecting the null hypothesis implies that the percent of patients achieving targets is higher in either of the STP arm or the IWL arm than in the SoC arm. A one-sided, 0.05 alpha-level simultaneous test for multiple contrasts of binomial proportions with an "Add-2" adjustment will be used.

The primary analysis cohort will be an ITT cohort including all randomized subjects and considering the arm to which they were originally randomized regardless of adherence to the regimen or the study protocol. Subjects who cross-over to the other intervention prior to Month 3 (ie: complete less than 3 of the required 6 months) will be considered failures in the arm to which they were originally randomized. Clinical data from the last, or exit visit will be used to construct the primary outcome for subjects who withdraw from the study, or from the intervention, subsequent to Month 3.

To corroborate the primary, unadjusted analysis, a logistic regression model will be fit with baseline clinical and demographic data as well as measures of compliance to intervention to adjust for possible imbalance of important variables. Also, the primary analysis will be repeated on a per-protocol cohort of subjects who completed the study according to the protocol and did not cross-over or withdraw from therapy. Graphical techniques and summary statistics for the two randomized groups at baseline and months 3 and 6 will also be presented.

Summary statistics and graphical presentations will be used to summarize data for all secondary outcomes. All hypothesis tests will be 2-sided and with a 5% Type I error probability. There will be no adjustment for multiple comparisons among the secondary outcomes with the exception that multiple domains within a single psychosocial instrument (eg: emotional burden, physician-related distress, regimen-related distress, and diabetes-related interpersonal distress domains within the DDS) will be subject to a Benjamini-Hochberg multiple comparison adjustment. Analyses will be conducted on the ITT cohort and corroborated by analysis on the per-protocol cohort.

Each clinical outcome measure is captured as a continuous and dichotomized variable (eg: A1c absolute level and those < 7.0%). Dichotomized variables will be analyzed as described in the Primary Analysis.

Continuous variables will be analyzed in three ways. First, summary statistics and graphical presentation of results will be performed for each of baseline and months 3 and 6. Second, change from baseline to month 6 will be compared using a paired t-test, or a Mann-Whitney test if distributional assumptions do not hold. A subject's last available, or exit visit, will be used if their month 6 visit is not available. Third, longitudinal random-effects models will be used to model the trajectory of each outcome over time and to determine the influence of randomized intervention group on that trajectory. Variables may be transformed to satisfy the assumptions of normally distributed random effects, but results will be presented on the original scale.

Sample Size Justification:

The primary null hypothesis compares the probability of a patient being in control in either experimental group to the Standard of Care group:

H0: pSoC ≥ pIWL and pSoC ≥ pSTP vs. H1: pSoC < pIWL or pSoC < pSTP

Based on data observed within the clinic from which cases are to be recruited, the success rate for Standard of Care patients is estimated to about 75%. A sample size of 200 subjects in a 2:1:1 randomization to the SoC, STP and IWL and groups respectively will provide a power of over 95% to reject the null hypothesis if the true difference between SoC and either experimental arm is 25% or greater (eg: 37.5% compared to 62.5%).

In addition, this sample size provides a power of more than 74% to reject the null hypothesis in favor of the alternative hypothesis that the percent of patients achieving control is greater among both the STP and IWL than among the SoC, if the true difference between arms is 25% or greater.