A detailed trial protocol has been published previously . All subjects provided written informed consent. The evaluating self-management and educational support in severely obese patients awaiting multidisciplinary bariatric care (EVOLUTION) trial protocol was approved by the University of Alberta Research Ethics Board (PRO00031699) and, prior to patient enrolment, the trial was formally registered at ClinicalTrials.gov (NCT01860131).
We conducted a 9-month pragmatic, prospective, RCT enrolling consecutive, consenting patients with severe obesity newly wait-listed for adult bariatric specialty care. The study was conducted in Edmonton and Red Deer (Northern and Central Alberta, respectively). Patients were randomly assigned (1:1:1) to one of three groups:
In-person, community-based self-management
A one-time provision of printed educational materials
Computer-generated randomization was performed centrally and independently by the EPICORE centre (www.epicore.ualberta.ca) to ensure allocation concealment from all research personnel; randomization was stratified by participating study centre.
The Edmonton Weight Wise program, established in 2005, was the first large-scale, multidisciplinary bariatric program in Alberta. Weight Wise delivers integrated, patient-focused, evidence-based care to the Edmonton Zone of Alberta Health Services (AHS). Treatments are guideline-concordant and based on the Canadian obesity guidelines . This region is one of the largest integrated health delivery systems in Canada, serves a catchment population of 1.6 million residents and has an annual health care budget of 2 billion dollars.
Weight Wise includes a central, region-wide, single-point-of-access referral system. The Adult Specialty Clinic offers intensive multidisciplinary medical/surgical bariatric care to patients with BMI levels of ≥35 kg/m2 (estimated to be at least 125,000 individuals within the region). Patients are referred for both medical and surgical management. About 64% of wait-listed patients are interested in bariatric surgery, and the remainder in medical management alone. All patients receive medical therapy initially and for several months before evaluation for surgery (if interested). The decision to perform surgery depends, in part, on how much effort and commitment the patient has put into lifestyle modification (even if unsuccessful). Therefore, patients interested in surgery cannot forgo medical weight loss recommendation adherence because they wish to have surgery.
At the time that this trial was performed, approximately 1200 new referrals were being processed annually, and 300 bariatric surgeries were being performed. Wait times have fluctuated and varied from as short as 4 months to as long as 30 months. Within the Weight Wise program, nurses are case managers and are responsible for coordinating care with other health care providers. Dieticians provide the bulk of the health behaviour modification counselling. This is reinforced by other care providers (physician, sleep specialist, psychologist, psychiatrist, occupational therapist, physiotherapist and social worker), who see patients for targeted indications within their area of specialization. The program in Red Deer was modelled after the parent Weight Wise program. Red Deer is a smaller centre that serves a more rural population, with approximately 600 referrals and 100 bariatric surgeries per year.
In-person self-management arm
This program varied slightly over the trial period, but, at its core, broadly consisted of 13 sessions (Additional file 1: Table S1) delivered in a group format by a multidisciplinary team at community health centres in Edmonton and Red Deer. Each session was approximately 2.5 hours long and was led by the appropriate content expert(s) from a team of four registered nurses, one Canadian Society for Exercise Physiology (CSEP) specialist, one MSc psychologist and one registered dietician. The program is free of charge and has been designed to educate patients regarding proper diet and exercise; improve their weight management skills by enhancing self-management and self-efficacy; and help them identify/overcome barriers to success. In addition to providing supportive care to wait-listed patients, the program is intended to prime patients for weight management success once they begin bariatric care. The evidence-based curriculum  stresses healthful eating; gives practical tips to increase physical activity; teaches basic behavioural modification techniques such as goal setting and self-monitoring; and includes strategies for dealing with stress and maladaptive eating behaviours such as emotional eating.
Web-based self-management arm
The web-based program was very similar in content to the in-person intervention (Additional file 1: Table S2), with the exception that it was delivered solely in an online format. Each patient had a personalized login that was tracked by the study team. All 13 modules were available to the subject on a single online platform, accessible any time after randomization. Subjects were asked to read all 13 modules (Additional file 1: Table S2) over a 3-month period.
Controls received a one-time provision of printed educational materials at the time of enrolment, consisting of Canada’s Guide to Healthy Living and Canada’s Physical Activity Guide. These materials are guideline-based, user-friendly and visually appealing. They are produced by the Canadian government and made available free of charge.
Of note, both for purposes of ensuring fidelity and maintaining some degree of ’attention control’, subjects in all three study arms were contacted by telephone 1 week after enrolment for troubleshooting and to encourage them to read the printed materials, register for in-person sessions or access the web modules.
We enrolled patients with BMI levels ≥35 kg/m2 who were newly wait-listed for adult (age >18 years) bariatric specialty care at the Edmonton or Red Deer clinic. Patients with one or more of the following characteristics were excluded: (1) completed more than four Weight Wise Community Modules (web-based or group session) in previous 3 months, (2) pregnant, (3) unable to read/write/comprehend English, (4) unable to access the web, (5) unable or unwilling to attend in-person sessions, (6) uncontrolled severe personality disorder, active psychosis, active substance dependence and/or major cognitive impairment, (7) deemed unsuitable by the study investigators, (8) participated in concurrent trial related to obesity management, (9) resided more than 1 hour driving time away from Weight Wise clinic or (10) declined to participate.
Enrolment procedures and timelines
Consecutive, consenting, eligible patients were enrolled between 2013 and 2015. Study participants were instructed to not begin any other new lifestyle modification or weight loss-related interventions during the first 3 months of the study. Those randomized to the in-person and web-based interventions were instructed to complete the entire program within a 3-month period. Thereafter, all subjects, including controls, immediately entered the bariatric clinic and commenced multidisciplinary bariatric care. This meant that the usual wait-listed period for entry into the Weight Wise clinic was waived for trial participants. Subjects were followed for an additional 6 months after entry into the clinic. Thus, overall, the trial consisted of a 3-month period during which the interventions were applied, followed by a 6-month follow-up period, with ascertainment of outcomes at baseline, 3 months, 6 months and 9 months (see below). To prevent between-arm contamination, subjects received specific information on their assigned treatment group only. Access to the web-based program was controlled using a hidden URL, which was given only to participants randomized to the web-based intervention arm, and patients in the web-based or control groups were asked not to attend the in-person community sessions.
Data collection procedures including detailed case report forms have been published previously . Although clinic staff could not be blinded to allocation status, all outcome assessments were performed by research assistants working independently from regular clinic staff. Data were collected at baseline and at 3, 6 and 9 months post-randomization, with the primary analysis focusing on 9-month outcomes.
Body weight was measured using a validated, calibrated Scale-Tronix bariatric scale and recorded to the nearest 0.1 kg, with the subject wearing light indoor clothing with empty pockets, no shoes and an empty bladder. Height was measured to the nearest 0.1 cm using a wall-mounted stadiometer. Blood pressure was measured with a Microlife Watch BP automated monitor, with three readings taken simultaneously in each arm, the first reading discarded and the latter two averaged. The arm that had the highest mean blood pressure at the baseline visit was used to calculate the 9-month blood pressure change.
The primary outcome was the proportion of patients achieving 5% weight loss, considered a clinically important degree of weight loss by experts and contemporary guidelines . Absolute and relative weight loss and BMI change were also assessed. Additional clinical outcomes included change in blood pressure, fasting lipids and A1c and the change in prevalence of hypertension, diabetes and dyslipidemia as previously described. Hypertension was considered present if self-reported, if blood pressure levels were ≥140/90 mm Hg or if antihypertensive medications were prescribed. Diabetes was defined based upon self-report, a baseline A1c ≥6.5% and/or antidiabetic drug therapy. Dyslipidemia was diagnosed in the presence of one of the following: self-report, treatment with a lipid-lowering agent or an abnormality on the baseline fasting lipid panel (low-density lipoprotein (LDL) cholesterol ≥5.2 mmol/L, high-density lipoprotein (HDL) cholesterol <0.9 mmol/L or triglyceride level ≥2.8 mmol/L).
To ensure that utilization was equivalent once patients entered the clinic, we tracked the mean number of clinic visits to each type of health care provider over the follow-up period.
We used previously validated instruments to assess health-related quality of life (Short Form Survey (SF-12)  and EuroQol five dimensions (EQ-5D) ), preference-based utilities , satisfaction with medical care (Patient Satisfaction Questionnaire (PSQ)  scored on a 5-point Likert scale), self-efficacy (Weight Efficacy Life-Style Questionnaire (WEL) , depression (Patient Health Questionnaire (PHQ-8)  and readiness to change (assessed using a visual analogue scale (VAS) ranging from 0 to 10).
Resource use and costs for each study arm
The overall annual and per-patient costs of the two interventions relative to the control group were calculated using a methodology conforming to the three-step micro-costing technique of identification, measurement and valuation of resources [24, 25]. Estimates of health care professional time for creation and delivery of the interventions were obtained from AHS Chronic Disease program managers. In 2012–2013, 1707 patients were referred to the Edmonton and Red Deer programs and had an initial consult; for costing it was assumed that annually 1707 patients would receive the in-person strategy or the web-based strategy or be mailed printed educational materials.
The health care professional cost to develop and update the materials used in the in-person and web-based study arms as well as delivery of the in-person strategy were determined by obtaining estimates of time required and use of AHS wage rates for each category of staff (nurses, dietician, exercise therapist, psychologist). The resources to develop both the web-based and in-person interventions consisted primarily of the health care professional time required for literature review, summarizing the content and preparation of the content for delivery. The one-time development cost was amortized over a 5-year period. For the in-person strategy, a registered dietician and registered nurse performed the bulk of the work, with smaller contributions from an exercise specialist and a psychologist. No overhead costs were assigned to the in-person strategy, as sessions were conducted in health care facilities at off-peak hours. A registered dietician developed the web-based modules.
Further, both the in-person sessions and web-based modules were updated every 2 years; therefore, the health care professional time required for this work was estimated and included in the overall costs. Finally, the costs of delivering content, including generating patient lists and mailing out instructions in the web-based group as well as hosting and delivering in-person sessions, were estimated. Current wage rates in Alberta, Canada for each type of health care professional delivering content were used. Hourly wage rate assumptions were $42.62 for a registered dietician, $42.62 for a registered nurse, $44.44 for an exercise specialist and $46.88 for a psychologist.
Resources required for the control arm included patient list generation, administrative work required to prepare and address the envelopes and mailing costs . Costs associated with development of the materials were not included; the mailed literature (Canada’s Food Guide) is provided free of cost by Health Canada.
All costs are reported in 2013 Canadian dollars.
Descriptive analyses were performed including calculation of means, medians and standard deviations (SDs). Baseline variables were compared between the three study groups using one-way analysis of variance (ANOVA) for continuous outcomes and chi-squared tests for dichotomous ones.
Between-group change scores were compared using multivariable logistic regression for dichotomous variables (including the primary) and linear regression for continuous outcomes, adjusting for age, sex, site and baseline BMI. An intention-to-treat analysis was performed using a baseline-observation-carried-forward approach, designated a priori , for the primary analysis. A completers analysis, limited to those participants with full baseline and 9-month data, was also reported for weight and BMI-based outcomes. Multiple imputation was not performed because the data were not missing completely at random . The primary outcome comparisons of interest were between the in-person strategy and controls and the online strategy and controls. Subsequently, the in-person and online interventions were then compared.
One-way ANOVA was used to compare health care provider visit frequency and costs across study arms, with t tests used to compare between-study arms if significant differences were found. We considered p values less than 0.05 statistically significant. No adjustment for multiple testing was performed . All analyses were performed using SAS® (Version 9.3, Cary, NC, USA).
Sample size estimate
The study was powered to detect a 15% difference between the two interventions and controls in the proportion of 5% weight responders with an alpha level of 0.05 and a power of 0.90. We assumed that the control arm would result in a 5% weight loss in 20% of subjects (i.e. the control event rate = 20% at one year) . The initial sample size estimate was for ~180 patients per arm or 540 patients total. This was adjusted upwards to account for potential attrition and permit secondary and subgroup analyses to arrive at the final sample size.