The Optimise Your Health trial has been iteratively developed over time following the securement of three successive research grants (see supplementary material: Additional file 1). Each grant relates to separate trial phases and interventions with distinct aims, hypotheses, and outcome measures. Recruitment for the original OPTIMISE protocol commenced in June 2019 with 27 participants recruited and completing the 6 month original iteration. In June 2020, recruitment commenced for the extended 18 month trial, which was inclusive of all phases described above. A SPIRIT checklist for standardized protocol items was followed in writing this manuscript (Additional file 11).
Study design and randomisation
The OPTIMISE Your Health trial is a two-arm individually randomised controlled trial. Randomisation to either the intervention or delayed intervention arm is undertaken in random blocks of sizes 4–8, stratified according to whether participants are taking either 0–1 or 2+ hypoglycaemic medications. Randomisation is automated via REDCap software at the end of the baseline assessment. Due to the nature of the intervention, it is not possible to blind participants or assessors to group assignment.
The trial protocol includes five assessment time-points: baseline, 3- months, 6- months, 12-months, 15- months, and 18- months. Database(s) facilitating intervention management and data collection are undertaken through REDCap — an approved data management platform (version 11.1.25) — with further data stored securely but accessibly to the multidisciplinary multinational research team through The University of Queensland’s Research Data Management System. Ethics approval for both the original and expanded protocols were granted by Alfred Health Human Ethics Committee (Melbourne, Australia), The University of Queensland Institutional Human Research Ethics Committee (Brisbane, Australia), and the University of the Sunshine Coast Human Research Ethics Committee (Sunshine Coast, Australia). The OPTIMISE Your Health trial has been registered with the Australian New Zealand Clinical Trials Registry (ANZCTRN12618001159246; date of registration: 07/03/2018 URL: https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=375487). The trial is undertaken in accordance with CONSORT guidelines (http://www.consort-statement.org/), with Fig. 1 showing the overall study design.
Study setting / population
Participants are adults aged 35–65 years with clinically diagnosed and managed T2D (recent HbA1c test between 6.5 and 10.0%) who are working ≥0.8 full time equivalent (FTE) in predominantly (≥75%) desk-based work in the local Melbourne area. Participants are excluded based on the following criteria: not English speaking, pregnant, using insulin to treat their diabetes, or have high physical activity (maintained for at least 3 months: ≥30 min of moderate-to-vigorous intensity physical activity [MVPA] per day and/or ≥ 30 min of strength training on at least two separate days of the week). Complete details of the eligibility criteria are provided in the electronic supplementary material (Additional file 3). The trial aims to recruit 125 participants per arm to satisfy sample size requirements. In-person clinical assessments are conducted at the Baker Heart & Diabetes Institute in Melbourne, Australia.
Recruitment commenced in June 2019 and is ongoing through to June 2024 using several different recruitment options, including but not limited to:
Disseminating study promotional material into clinics and hospitals within the local community;
Online social media advertising performed both independently and by an external third party recruitment company
Television and newspaper feature articles
Mailing people registered on the National Diabetes Services Scheme (NDSS) database; and,
Contacting patients attending the adjacent diabetes clinic (in person or by telephone).
Screening and consent
Screening for eligibility and gaining consent involves a multistage process. Initially, potential participants are emailed a link to a brief survey containing questions concerning eligibility. Those whose surveys indicate they may be eligible receive a follow up telephone call from research personnel to confirm. Once confirmed, participants are sent further participation information (Additional file 2), and an employer permission form via email to forward to their employer, which is completed online. This employer permission form explains the relevant study requirements and informed consent for the study and permission for installation and use of a sit-stand workstation in the workplace. Participants are requested to send a photo of their current workstation setup to assist in desk installation. If a participant is self-employed or working from home they can sign the form themselves. Thereafter, participants are asked for results of any HbA1c blood test they have had in the last 4 weeks, or are provided with a pathology request slip and asked to arrange a non-fasting blood test (glycosylated haemoglobin; HbA1c%) at their nearest Melbourne Pathology centre (at no cost to the participant). Those within the eligible HbA1c% range (6.5–10.0%) are then requested to give final informed consent (written or electronic; Additional file 2) and have their baseline visit scheduled.
Impact of the COVID-19 pandemic on the trial
Intensive lockdown restrictions occurred in March 2020 leading to the suspension of recruitment. All participants enrolled in the trial during these periods were assessed according to remote COVID-safe practices. In October 2020 the lockdown restrictions ended and resumption of recruitment and the expanded protocol began.
The OPTIMISE intervention is based on the Stand Up Australia workplace intervention [25, 26]. However, it has been extended from a primarily workplace focus to include messaging targeting sitting less and moving more across the whole day. The intervention is grounded in social cognitive theory, with the key constructs targeted being: self-efficacy; socio-structural factors (barriers and facilitators); and outcome expectancies (physical, social and self-evaluative) . Consistent with the socio-ecological model of sedentary behaviour , which highlights the multiple inter-related influences on this behaviour; the intervention targets multiple levels of influence on behaviour (environment, intra-personal, inter-personal). In contrast to the Stand Up Australia intervention , the OPTIMISE intervention does not specifically include workplace organisational strategies. The intervention components are informed by the Behaviour Change Wheel (BCW) and the associated COM-B system, within which capability, opportunity and motivation are postulated as interacting to stimulate behaviour change . Previous feedback collected in the Stand Up Australia intervention  has demonstrated appropriate participant acceptance and feasibility of the key components including adaptations to the physical work environment, and the support of individual behaviour change.
The key intervention messages are to “Sit Less” and “Move More”. The aim of the intervention is to support participants to achieve (by 6 months), and then to either progressively improve or maintain (at 12 and 18 months), their personalised goals for sitting less, actively breaking up sitting time, and moving more. Participants receive coaching to set incremental goals that gradually progress from their pre-existing levels at baseline towards the study targets. Key intervention targets are: at least half of daily waking time in upright postures (50% standing/stepping); at least one ‘active’ break from sitting per hour; and, at least 10,000 steps per day. An ‘active’ break is considered to be at least three minutes of walking (approximately 250 steps) or completion of simple body weight resistance exercise activities that have been adapted from previous experimental trials : calf raises, squats, and single leg kickbacks, done in three sets of three 20-s bouts totaling approximately 3 minutes. Since excessive sitting and excessive standing may both be harmful [31, 32], 50% was chosen as a level that is a simple, heuristic approach that has been safely achieved by participants in our previous trials  and is 2 h/day lower than the average sedentary time reached by adults with type 2 diabetes , assuming a 16 h waking day. ‘Active’ breaks to interrupt sitting were promoted in recognition that not all sitting replacement activities have equal benefit [34, 35] with ambulation and resistance exercise showing particular benefit . The 10,000 steps message is widely used in public health , noting benefits (including glycaemic control) for increased steps can occur below this threshold .
Intervention protocol and core components
The OPTIMISE intervention consists of health coaching (including education), a sit-stand workstation and an activity tracker (both provided by the trial) to encourage participants to sit less, as well as to move more and engage in active breaks. The intervention commences with workstation installation by third party installers approximately 2 weeks following baseline assessment. Immediately after installation, participants meet with their assigned health coach, via zoom, to discuss the intervention, demonstrate correct desk usage, and set up and demonstrate the workings of the Fitbit. The health coaching continues in person and via telephone throughout the first 6 months intensive phase (10 total contacts). They then receive tailored coaching via text message between 6 and 12 months (1–2 messages per week). Participants may seek other concomitant lifestyle interventions to manage their diabetes. After completing the trial, participants retain the sit-stand workstation and activity tracker.
Consistent with the Behaviour Change Wheel intervention functions of enablement and environmental restructuring , all participants are provided with an Ergotrontm WorkFit-T/−TL Sit-Stand Desktop Workstation. The workstation weighs 22.5 kg and is placed on top of the existing work surface. It allows participants to easily and quietly alternate their working posture between sitting and standing whilst still interfacing with their computer. Participants are provided with either a single or dual arm monitor kit that allows their computer monitors to be directly affixed to the monitor platform. Both written and verbal ergonomic instructions are provided by the health coaches to aid in correct workstation usage . Under work-at-home restrictions, the alternative protocol is for participants to take their workstations home when permitted by the employer.
Activity tracker (Fitbit)
In line with the Behaviour Change Wheel intervention functions of education and training, participants are provided with a Fitbit activity tracker and Fitbit smartphone application (app) for Android or Apple. The Fitbit’s main purposes are to encourage participants to move more, and to engage in active breaks from sitting. Participants are given a username and password to be used for the duration of the trial and provide permission for their data to be accessed by the project team. Upon completing the study, the account will be deactivated and participants will be instructed on how to create their own personal account. Each Fitbit is synchronized with Fitabase (Small Steps Labs, LLC; San Diego, CA, USA; http://fitabase.com), a third-party data management platform that supports top down supervision of study participants’ device usage and activity. The model of Fitbit used is the Fitbit Inspire HR, a wireless wrist-worn device that records and displays a range of outputs in real time (including daily step count, approximate resting heart rate, hourly activity / active breaks) — captured through proprietary algorithms applied to tri-axial accelerometery and heart rate (photoplethysmography) inputs — and provide vibration feedback based on the recorded data. The lithium polymer battery powers the device for approximately five days (depending on usage) before requiring recharging and collects data for up to 30 days (at which point it must be synced with the participant’s smartphone).
Using the device and app real-time monitoring, participants self-monitor attainment of their self-selected stepping and active breaks goals throughout the intervention and specifically during health coaching sessions. Hourly active break reminders are sent according to the participant’s preference (up to 14 per day). These reminders are produced by the Fitbit ten minutes before each hour finishes. Participants receive a vibration alert if they have not achieved at least 250 steps in that hour (their active break, if they have chosen walking rather than resistance exercises). Other active break strategies, such as simple resistance activities may be manually recorded using the Fitbit, however these are not automatically monitored by the Fitbit itself. Health coaches remotely monitor participant adherence to steps and active break goals with Fitabase, which supports tailored feedback during coaching sessions.
Face-to-face and telephone health coaching (0–6 months)
The individual health coaching component targets the Behaviour Change Wheel intervention functions of persuasion, education and training. Health coaches, all of whom have at least a bachelor level of training in exercise physiology, nutrition, or nursing, are trained in the delivery of the intervention, including motivational interviewing techniques . A written script is maintained for intervention fidelity and consistency of delivery. Key aims of the health coaching are to build rapport, provide education about the importance of reducing sedentary behaviour and increasing movement throughout the day, encourage sustained self-management, and goal setting. In conjunction with the first health coaching session, participants are provided with an intervention handbook, containing written educational information and instructions (Additional file 4), an email containing basic tips to reduce sitting, a 2-min educational video describing the link between diabetes and sitting and activity levels, a video demonstrating how to perform the simple resistance activity breaks, and an activity feedback report (Additional file 5) generated from their most recent ten days of activity monitor (activPAL) wear. The health coaching consists of ten sessions (two in-person and eight via telephone; approximately 15–30 min in duration) over 6 months. Health coaches follow standardised outlines for each session set out in REDCap forms containing both the coaching script and embedded data-collection fields to further enhance intervention fidelity and collect data for process evaluation. The data collection is also used to guide subsequent health coaching sessions (e.g., goals, goal attainment). All health coaching sessions are delivered by the same health coach wherever possible and conducted at a time selected by the participant.
Coaching session 1 (face-to-face)
The first session, scheduled to take 50–60 min, occurs via video teleconferencing immediately after installation of the workstation. The session includes a brief discussion of the workstation relative to the expectations and needs of the participant, whereby the coach ensures the desk is installed to allow for correct usage habits and ergonomic positioning; demonstration of how to perform the simple resistance activities at the desk; provision, setup and demonstration of the activity tracker (see ‘Activity tracker (Fitbit)’); and, the behavior change motivational interviewing. A script of this interview is available in Additional file 7.
For the motivational interviewing, the health coach revises the participant’s personalised feedback report (Additional file 5) covering pre-existing sedentary behaviour and physical activity. Their behaviour at work, over the whole waking day, plotted by date and time is used to help set realistic goals. It is also used to identify and reflect on their “danger zones” (periods throughout the day of prolonged, unbroken sitting time), and choose relevant behavioural strategies targeting those danger zones. Participants are guided to form an action plan consisting of two goals (number of active breaks at work; total step count per day) and at least one strategy each to: sit less at work; sit less outside of work; move more during work; and, achieve the daily step count goal. Participants choose from a recommended list of strategies, or choose their own (Additional file 6). The coach supports the participant to make selections that conform to the SMART framework (Specific, Measurable, Achievable, Relevant, and Time-based), consider barriers and solutions, and for which the participant has adequate readiness and confidence to achieve (i.e., 8+ on a scale from a 1 (not ready / not confident) to 10 (ready / confident)). The health coach reviews the final action plan and any anticipated barriers with the participant. Following the session, a summary of the discussion is provided to the participant in a personalised email from their coach.
Sessions 2–6, 8–10 (telephone)
The seven telephone-delivered sessions occur on a tapered schedule: one call per week for the first three weeks reducing to one call every three weeks for the remaining 23 weeks (approximate timeframe). Each session lasts for a duration of 15 to 30 min. The sessions are designed to check on the participant’s progress (both subjectively and objectively with their Fitbit data as in Fig. 2), address problems with adherence, revise goals as needed, and reinforce goal attainment. Adoption of new strategies, and the progression of goals is encouraged by health coaches when participants report high engagement and success with their current plan.
Session 7 (face-to-face)
Coaches schedule an in-person session to coincide with their attendance at the Baker Institute, Melbourne for their 3-month clinical assessment. In a 30 min session, coaches enquire regarding participants’ overall experiences and their readiness to continue with the intervention, and provide feedback on how their their daily steps and active breaks have tracked over time for each week since the first health coaching session according to their Fitbit. Similar to the telephone sessions, participants’ strategies and goals to sit less and move more are revisited and revised as required. This session is offered via telephone when in-person visits are not possible (e.g., due to the COVID pandemic) or declined.
Extended care (tailored text messaging; 6–12 months; expanded protocol) – OPTIMISE
The 6–12 month intervention phase is designed to support maintained behaviour change (or continued improvement) after the 10-session intensive phase has been completed. This program, which is administered remotely, transitions participants from the face-to-face and telephone delivered health coaching to a tailored text message service using the web-based semi-automated platform, propelo™ (www.propelo.com.au). Coaching involves 24 weeks of text message contact and one telephone coaching session (midway) with their health coach.
The text message content has been informed from formative discussion with research staff, health coaches, and completed OPTIMISE intervention participants who were recruited before the expanded protocol was initiated. This formative research involved qualitative interviews and user-testing of draft text content. Key findings from this formative work highlighted the need for the text message program to: use supportive, non-judgemental language; continue to leverage off the rapport established between the health coach and participant during the intensive phase; and, ensure the texts are tailored to the participant’s Fitbit data. The texts target three key behaviour change strategies in three different types of text messages (see Table 1): prompt actions in real-time (via ‘Sit less prompts’ and ‘Move More promps’); promote ongoing self-monitoring of behaviour (via ‘Check-ins’); and, monitor and reward goal attainment (via ‘Check-ins’).
During the final OPTIMISE telephone coaching sessions (#8), the health coach notifies participants that they will be receiving text messages over the next six months. Following the session, the coach completes a tailored survey translating the participant’s coaching goals into messaging-friendly “danger zones” (periods of the day or activities when prolonged sitting commonly occurs), strategies to sit less and move more, and goals for steps per day and active breaks per day. The appropriate days and times to send each type of message is also captured. Text messages are sent at a minimum frequency of once per week (see Table 1). The text message content is kept under 160 characters and is initially based on information from the tailoring survey, and subsequently, throughout the intervention, based on data from the participant’s Fitbit data accessed weekly through Fitabase. Either one or two check-in texts are sent monthly (on Mondays), based on whether the participant has met (yes/no/almost) their step goal and their active break goal in the previous week (Monday-Sunday) or their goal attainment could not be verified because the participant did not wear the Fitbit in the week prior. The telephone session with the coach midway through the text message program, involves a review of progress, and retailoring of the goals, danger zones, strategies, message days and times as needed.
Control – delayed intervention
To minimise attrition prior to receiving their delayed intervention (at 12 months), the control participants receive contact in the form of a thank-you letter commencing two weeks after the baseline assessment then, monthly emails containing diabetes fact sheets, along with a follow-up phone call (verifying they received the fact sheet and enquiring whether they have any questions). The fact sheets are published by the NDSS, and provide a non-tailored diabetes-management advice. The fact sheets do not cover sedentary behaviour. Participants may seek other lifestyle interventions to manage their diabetes during this period. Following the completion of the 12 month assessment, control participants are provided with a modified version of the intervention (12–18 months). In the original protocol, delayed intervention (6–9 months) participants were provided with a sit-stand workstation and written education materials only. In the expanded protocol this is referred to as OPTIMISED (12–18 months). As with OPTIMISE, OPTIMISED includes environmental support via a sit-stand workstation, provision of a Fitbit, and health behaviour change coaching (one delivered by online video teleconferencing and two delivered by telephone). However, the behaviour change health coaching and educational materials are modified to be more suitable for scale-up and external delivery, based on consultation and collaboration with end users and stakeholders. The intervention adaptation process, and the resultant OPTIMISED intervention protocol, will be reported elsewhere.
The primary outcomes are overall daily sitting time, and glycosylated haemoglobin (HbA1c) at 6 months, which were originally chosen in order to evaluate the intensive phase of the intervention (OPTIMISE). With the addition of the expanded protocol, cognitive function measures: visual learning and new memory (Paired Associates Learning Total Errors [adjusted]) were added as a primary outcome at 12 months.
Secondary and exploratory outcomes
Secondary outcomes include: sitting time accumulation patterns (sitting time in prolonged bouts [≥30 min bouts], prolonged bouts at work, usual bout duration, alpha [unitless measure characterising the frequency distribution of sedentary bout durations]); sitting time at work; physical activity (active breaks from sitting, time spent standing, stepping, light-intensity physical activity, moderate-to-vigorous intensity physical activity) overall and during work; sleep (total sleep time, sleep efficiency, sleep onset latency, wakefulness after sleep onset); anthropometry (weight, waist, and hip circumference); body composition (fat mass, visceral fat mass, lean mass); fasting and postprandial glucose metabolism assessed as incremental area under the curve (iAUC) for glucose and insulin during a 2 h 75 g oral glucose tolerance test (OGTT); fasting lipid levels (LDL, HDL, total cholesterol, cholesterol ratio, triglycerides); blood pressure (systolic and diastolic pressure); vascular function (flow-mediated dilation and shear rate); inflammatory markers (high-sensitivity c-reactive protein [hs-CRP], interleukin [IL]-6, IL-1β, tumour necrosis factor – alpha [TNF-α], adiponectin, and leptin); neurotrophic factors (brain-derived neurotrophic factor [BDNF], insulin-like growth factor 1 [IGF-1]); and, additional cognitive function domains including reaction time, visual matching ability and visual recognition memory, reaction time, and working memory. Some of the survey measures are considered as outcomes, including the measures of anxiety and depression, musculoskeletal health, fatigue, workplace performance and satisfaction, and motivation. Outcomes regarding longer term trends in physical activity and sleeping are extracted from the Fitbit worn by intervention participants only.
Data collection occurs at baseline, 3, 6, 12, 15 and 18 months via: in-person clinical assessments; cognitive testing; device-based monitoring; and online questionnaires administered through REDCap; and, process data collected in REDCap and by the health coach in REDCap health-coaching forms. Table 2 shows the timepoints at which each component of the data-collection occurs under the original, expanded, and COVID-adapted protocols. Table 3 and Additional file 9 provide extensive details of the objective and questionnaire measures, as well as their intended purpose in the trial (e.g., primary outcome, secondary outcome, process outcome). The data collection procedures used for each of these components are indicated below.
Device-based measurement of physical activity, sedentary behavior and sleep
Device-based measures of physical activity, sedentary behavior, sedentary behavior accumulation and sleep are collected using the activPAL4 activity monitor (PAL Technologies Limited, Glasgow, UK; default settings) worn on the thigh and the Actigraph GT3X+ (Actigraph, Pensacola, FL, USA) worn on the wrist. Both devices are intended to be worn 24 h per day for 10 days. Eleven days before their clinic visit, the activPAL4 — previously initialised and waterproofed with a nitrile sleeve — and the GT3X+ monitor and wrist band are sent to participants along with instructions and materials for affixing and wearing both devices (transparent polyurethane acrylate adhesive patches and wrist band, respectively). It has been previously shown that participants attaching their activPAL devices is feasible, acceptable, and results in adequate placement on the anterior midline of the thigh . A sleep diary is provided to participants for 10 consecutive days to define sleeping and waking periods. Incomplete records for sleep and wake times are inferred from using one or more published methods (visual estimation; automated estimation; average values, for example with, Edwardson et al. ; Winkler et al. ; LaCroix et al.  respectively) and checked against movement visually.
The primary outcome of total daily sitting time (sitting or lying while awake and wearing the device, h/16 h day) will be derived from the activPAL4 data, as will secondary measures of physical activity, standing time, and sedentary behaviour accumulation. The activPAL4 uses triaxial accelerometry (30 Hz) and activPAL proprietary algorithms (here, VANE) to measure sitting/lying, standing, and stepping, stepping cadence, as well as transitions between sitting/lying and upright posture with high accuracy, reliability and excellent responsiveness to change for sedentary reduction interventions [45,46,47,48]. Further measures, with less published data on validity, are available using the CREA algorithm . Relative energy expenditure will be estimated as metabolic equivalents (MET) from cadence  or acceleration  with a similar level of accuracy to most accelerometers. Consistent with standard procedures for the field , sleep time and device non-wear time (based on the daily log and device non-movement) will be excluded from measures of physical activity and sedentary behavior, as will days with insufficient wear. Sufficient-wear criteria for waking days will be 10h hours wear while awake, with evidence of movement (≥500 steps/day and not ≥95% of the day in any one activity). The activPAL data (activPAL VANE algorithm primarily) are processed in SAS (9.4 or higher) using a bespoke program adapted from previous studies . Measures are extracted overall and for specific timeframes of interest identified from the log — work and non-work days, work and non-work times, and time at the workplace — and as detailed time-series data (daily; hourly).
The wrist-worn Actigraph GT3X+ activity monitor is used to estimate sleep duration (i.e., total sleep time) and sleep quality metrics (i..e, sleep efficiency, wakefulness after sleep onset, sleep onset latency) and also enables harmonisation of data to other studies that have used wrist-worn devices [53, 54]. Actigraph GT3X+ data are sampled at 30 hz. Behaviour classification from the raw wrist-worn accelerometer data is performed in GGIR, an open source R package . This package implements methods providing valid and reliable estimates sleep duration and quality  as well as physical activity [57, 58] from wrist-worn devices. The 24-h wrist monitor data will be separated into sleep periods (for measuring sleep quality and duration) versus waking periods (for other measures) based on sleep and wake times collected in the online-administered daily logs. Sleep quality and duration measures exclude sleep periods during which the device is not worn, which will be identified via the Choi algorithm .
Clinic visits occur at every data collection timepoint, wth the exception being at 15 months. Prior to each visit, participants are reminded to abstain from engaging in any moderate- to vigorous-intensity physical activity, from consuming caffeine and alcohol in the 24 h prior to the assessment, to fast for at least 8 h, and to take their medications for the day of the visit at lunchtime during the visit rather than in the morning. For each clinic visit, participants report to the Baker Institute Clinic at 8 am. Over a 5-h period, participants have their body weight, body composition, blood pressure and vascular function assessed, a fasting blood collection, and undergo a 2 h oral glucose tolerance test (OGTT) during which they complete their questionnaires. Upon completion of the OGTT, participants receive lunch (e.g., sandwich/coffee) followed by cognitive assessments. Those participants whose assessment occurs during COVID-19 related social isolation mandates are emailed the questionnaires and referred to off-site pathology for glycosylated haemoglobin (HbA1c) testing, all other measures are omitted.
Clinic measures and their procedures are described below.
A peripheral intravenous catheter is inserted near the antecubital fossa and a fasting blood sample is taken to measure glucose, serum insulin, HbA1c, high-sensitivity C-reactive protein (hs-CRP), full blood examination (FBE), cholesterol, and triglycerides. Thereafter, the participant is guided through a standard 75-g 2 h OGTT , with plasma glucose and serum insulin collected at half-hourly intervals. An additional volume of blood is collected for verification of results if required. Further methodological information is available in Table 4.
Inflammatory and neurotrophic factors
Markers of inflammation (e.g. hs-CRP, TNF-α, IL-6, IL-1β, adiponectin, and leptin) and neurotrophic factors (BDNF, IGF-1) will be ascertained from serum/plasma samples collected at all clinical assessments except the 3 month assessment. Samples will be analysed using multiplex microsphere-based immunoassays (MAGPIX, Merck Millipore) and traditional ELISA methods (R&D Quantikine) at the Sunshine Coast Health Institute. Apolipoprotein E 4 (ApoE-4), a commonly assessed genetic risk marker for cognitive decline is assessed at baseline only from DNA isolated from blood collected into PAXgene® tubes using QIAGEN DNA extraction kits. Custom-designed primers will be used for PCR amplification of the ApoE gene (Taqman qPCR,Thermofisher Scientific). Refer to Table 4 for further detail on inflammatory and neurotrophic factor methodologies.
Global cognitive function is first screened with the Addenbrooke Cognitive Examination-Revised (ACE-R) and then assessed using the Cambridge Neuropsychological Test Automated Battery (CANTAB) . The ACE-R is administered by the research staff at baseline only, as it provides a global measure of cognition. ACE-R evaluates six cognitive domains (orientation, attention, memory, verbal fluency, language, and visuospatial ability) and includes an inbuilt Mini Mental State Exam .
Following the ACE-R test, CANTAB is completed by the participant at baseline, and then tested additionally at the 6, 12 and 18 month assessments. Specific CANTAB tests (and the domain they measure) are: Paired Associates Learning (visual memory and new learning); Reaction Time (motor and mental response speeds, as well as measures of movement time, reaction time, response accuracy and impulsivity); Delayed Matching to Sample (visual matching ability and short-term visual recognition memory); and, Spatial Working Memory (strategy and working memory). The Motor Screening Task is included as an introduction to the CANTAB assessments. The key primary outcome is the total number of errors on the Paired Associates Learning task adjusted (PAL TEA) shown in Fig. 3. The outcome is automatically calculated by CANTAB. The PAL TEA score represents the number of times a participant chose the incorrect option during the task. Participants that terminate the task early have their scores adjusted; accounting for the estimated number of errors they would have made on all the problems. This allows for comparison with those who completed the final stage of the task. The CANTAB is an iPad administered cognitive battery and results are downloaded from the official Cambridge Cognition platform.
Anthropometry: height, weight, waist and hip circumferences and body composition
Height is obtained using a stadiometer (Seca, Germany) at the baseline clinic visit only, measured in duplicate, with a third measurement taken if the difference between the first two measurements is ≥0.5 cm. Body mass is measured using a platform scale (JAC-929; Nuweigh, Australia) taking measures in duplicate, with the average of two reading recorded at each assessment. BMI is calculated using height and weight measurements (weight / height2). Waist circumference is obtained using a non-elastic tape measure around the midpoint between the iliac crest and the lowest rib. Hip circumference is measured at the maximum circumference in the horizontal plane, over the buttocks. Both measurements are recorded according to the nearest 0.1 cm, and completed twice unless the first two differ by ≥1 cm, in which case a third measure is obtained. A whole-body dual energy x-ray absorptiometry (DXA) scan (Lunar iDXA; GE Healthcare, Australia) is used to assess body composition variables, including total and regional fat mass, and lean mass, and total body fat percentage. Visceral adipose tissue (volume and mass) is estimated by the DXA software using the android region of the body. Where participants’ body size exceeds the DXA scan area, a half body scan is performed, which has previously been shown to provide valid estimate of body composition . Where possible, the same research staff member will perform each participant’s follow-up scan.
Blood pressure is measured via a digital blood pressure monitor (OMRON HEM-907; Omron Healthcare, Japan) placed on the non-dominant arm. Participants are requested to lie down in a dimly lit, temperature controlled (approximately 22–24 C) room. After 10 min of rest, the monitor takes three measurements with an interval of 2 minutes between each measurement. The monitor measures systolic pressure, diastolic pressure, and heart rate and determines the average of the three measurements separately. The three measurements, as well as the average values, are recorded.
Arterial function is assessed by flow-mediated dilation (FMD). Participants lie in a supine position in a dimly lit temperature controlled room for at least 20 min before the baseline ‘steady state’ recording of FMD is obtained. Recording is made on the brachial artery of the dominant arm (opposite to blood pressure) using a high-resolution ultrasound machine (Terason t3200, Teratech, Burlington, MA). A rapid inflatable cuff (SC-12-D, D.E. Hokanson Inc., Bellevue, WA) is placed distally to the antecubital fossa and the ultrasound probe applied to the brachial artery. After an optimal image of the brachial artery has been established, a 1 min recording of continuous resting vessel diameter and blood velocity (shear rate) is collected. The cuff is then inflated to > 200 mmHg for 5 min. Thereafter, the cuff is released to induce reactive hyperemia. An additional 3 min of ultrasound recording is completed to determine the post-deflation peak vessel diameter, which is compared to resting vessel diameter (FMD%).
Surveys are administered either during the clinic visit, or emailed to the participants for self-completion. All surveys, unless specified otherwise, are combined together into one survey, referred to hereon as the OPTIMISE survey. Contingent on COVID-19 and associated social distancing mandates, all surveys are administered electronically. Information pertaining to questionnaire source, rationale for inclusion, reliability and validity is described below and summarised in Additional file 9.
Age and gender are ascertained in the eligibility survey. Sociodemographic characteristics (ethnicity, household type, education, occupation) are obtained at the baseline visit, with question formats based upon previous studies such as the Australian Diabetes, Obesity and Lifestyle (AusDiab) study . Changes in occupation are recorded in follow-up assessments.
Health history checklist
A health history checklist recording pre-existing health conditions is recorded by research staff with participants at baseline, and updated throughout the trial in the instance it changes. The checklist includes questions pertaining to any history of angina, heart attack, heart bypass operation, stroke, angioplasty for peripheral vascular disease, kidney damage from diabetes, eye or retinal damage from diabetes, nerve damage from diabetes, numbness burning or tingling in feet, foot ulcer, and lower limb amputation.
Self-reported physical activity, sitting time, and sleep
The Active Australia questionnaire is used to measure the number of minutes per week engaged in walking, vigorous gardening, moderate activity, vigorous activity (time multiplied by two), and strength training, with amounts totalled to determine MVPA . The IPAQ questionnaire (two-items) is used to measure overall sitting time obtained in hours and minutes across week days and weekends. Sitting time is contextualised as part of work, travel, recreation, on a screen device, online chores (e.g. emails), or ‘other’ sitting that occurs during waking hours . Sitting time and activity are also assessed using the Sedentary, Transport and Activity Questionnaire (STAQ) , which assesses time spent in transport, whether driving, on public transport, walking, and cycling. STAQ has acceptable reliability and validity . Sitting and activity are further quantified using the Occupational Sitting and Physical Activity Questionnaire (OSPAQ) . Here, participants are asked to divide their work day into percentages spent sitting, standing, stepping, and performing heavy labour tasks. OSPAQ has demonstrated acceptable validity and reliability [70, 71]. Additional questions asked include the percentage of the work day that is occupied by prolonged sitting (sitting bouts equal to or greater than 30 min) is also obtained , as well as the proportion of the workday spent sitting in common occupational tasks . Participants are asked to identify their desired levels of sitting, standing, and stepping during work and home hours . A 15-item checklist is used to assess participants’ current use of sit less and move more strategies (at home and at work) and a 6-item questionnaire assesses barriers to sitting less and moving more . Knowledge (5-items) and perceived organisational norms (4-items) about sitting, activity, and health are assessed using a questionnaire adapted from a previous trial . Participant’s self-regulation strategies for sitting less and moving more are both assessed with 9-item questionnaires, which are adapted from previous trials . Participants are asked to evaluate how supportive their peers are of reducing sedentary behaviour, as well as engaging in physical activity [77, 78]. Finally, participants are assessed for their perceived changes across 16 domains of sitting in the last 6 months (assessed at 6, 12 and 18 months only). This measure has been adapted from previous research for this trial .
Each morning over a 10-day monitoring period (coinciding with the wear of the activity monitor devices), participants are emailed links to a REDCap administered online daily log. Participants can also request to use a handwritten copy of the diary. Using a modified version of the Consensus Sleep Diary , the log enquires about their wake (and out of bed) times that morning, their sleep times and sleep quality the night before (into bed, lights out, time to fall asleep, number and duration of awakenings), and over the previous day, their work hours and work location (workplace / home / other), whether each device was worn that day (yes/no) and the start and end times of any removal greater than 10 min. The consensus sleep diary is a validated measure of sleep .
Anxiety and depression
The Hospital Anxiety and Depression Scale (HADS)  is used to measure anxiety and depression. It is a valid and widely accepted measure for determining level of anxiety and depression and can reliably differentiate between the two . The questionnaire consists of 14 items: seven for anxiety (e.g. “I feel tense or wound up”), and seven for depression (e.g. “I feel cheerful”).
The Problem Areas in Diabetes (PAID)  scale assesses perceptions of the intervention and its influence on subjective diabetes management. The measure has demonstrated high reliability  and sensitivity to change over time , and has demonstrated use for screening depression and emotional problems in people with diabetes . The 5-point scale asks participants to rank aspects of their diabetes management from “Not a problem” to “Serious problem”. Questions pertain to feelings about diabetes treatment and treatment goals, diabetes and social situations, diabetes and adverse events, burden and acceptance of diabetes, and coping with diabetes complications.
Fatigue is evaluated with a Fatigue Symptom Inventory (FSI) . The inventory questions the participant on their current and previous level of fatigue in the last week. It examines how their fatigue interferes with activities of daily living, cognition, relationships, and mood. Participants are asked to rate the level of fatigue on a 0 (no fatigue) to 10 (extreme fatigue) scale. The measure has demonstrated validity with other fatigue symptom measures [87, 88].
Musculoskeletal pain and health
Musculoskeletal pain and health is measured using the 27-item modified Nordic Musculoskeletal Questionnaire, which surveys both the last 3 months and the last 7 days . Participants indicate the body locations they have experienced ‘trouble’ in muscles or joints. Pain is ranked with an adapted 0–10 scale , where 0 indicates ‘no pain’, and 10 indicates the ‘worst pain imaginable’. This measure has been demonstrated to be repeatable and sensitive to change .
Participants are asked to estimate their hours and days working, and how many leave days they have had in the preceding 3 months due to illness. Perceived workload and caring responsibility are assessed with a valid  two-item questionnaire based on the Borg workload scale , and NASA Task Load Index . Self-reported work satisfaction  and work performance  are also assessed via 7-point scales. Participants are surveyed for how they feel during work with the ultra-short measure for work engagement (UWES-3)  featuring 3-items with a 7 point scale, including energy levels, enthusiasm, and being immersed in their work. The Work Limitations Questionnaire (WLQ)  is used to examine the frequency of difficulty that the participant has with performing specific work-related tasks. The WLQ has demonstated validity and reliability and is correlated with arthritis pain, as well as functional limitations and work productivity .
An audit of the participant’s work environment is conducted at baseline only and made according to the Checklist of Health Promotion and Environments at Worksites (CHEW)  which was successfully modified for sedentary behaviour interventions in BeUpstanding . Information is captured on layout, the nearby physical environment (e.g. stairs, centrally located printers and amenities), and workplace policy (such as flexible work hours). The 26-item questionnaire also includes questions pertaining to whether the workplace provides sit-stand desks, and whether a wearable tracker has been previously provided by employers.
Medication and allied health appointments
A record of the participant’s medication, including dosage, frequency is created at the baseline clinical assessment. Any changes to their medication is noted at each subsequent assessment. Similarly, any appointments with an allied health professional in the preceding months between clinical visits are recorded and updated at each assessment. Both records inform potential confounding effects of medication and allied health treatment on the primary and secondary analyses.
Personal wearables questionnaire
Participants use of wearable activity trackers and/or apps in the preceding months is recorded at each visit. Questions about type of wearable and app have been adapted from previous research , with the inclusion of additional questions pertaining to how long they have been using the device for, what feature they use (e.g., step counts, exercise intensity), how often they follow associated prompts, and whether these are followed at work, home, or both. Similar questions have been used in previous workplace sedentary behaviour interventions  and are important for evaluating prior exposure to health-behaviour trackers.
Dietary intake is measured using the University of Newcastle’s Australian Eating Survey (AES), a Food Frequency Questionnaire which examines eating habits over the previous 3–6 months . It is a semi-quantitative questionnaire of 120 items. The questionnaire has comprehensive validaty and reliability , requires less time to complete compared to a 24-h dietary recall, and is representative of longer term changes in dietary intake. Completion of the questionnaire results in a report detailing total daily energy intake, the contribution of healthy nutrient-rich and unhealthy nutrient-poor food choice to diet, the Australian Recommended Food Score (ARFS) a diet quality score indicative of the participant’s alignment to Australian dietary recommendations, macronutrient intake, micronutrient intake and fibre intake. The results are considered as confounding variables on primary and secondary analyses.
Quality of life
Quality of life is measured using the Australian Quality of Life Survey (AQoL-8D) consisting of eight dimensions (Independent Living, Happiness, Mental Health, Coping, Relationships, Self Worth, Pain, Senses) totaling 35 items  . The questionnaire is validated and has high test-retest reliability .
Motivation for physical activity and motivation to break up sedentary behaviour
Motivation for physical activity and motivation to break up sedentary behaviour are assessed using two separate modified versions of the validated Behavioural Regulation Exercise Questionnaire (BREQ-3)  based on self-determination theory . The original prompt of “Why do you engage in exercise?” has been modified for the two questionnaires separately as “Why do you engage or not engage in physical activity?” and “Why do you break up sitting with standing up and/or moving more versus continued sitting?” respectively. These two questionnaires contain 24 items each with a five-point Likert scale. There are six subscales that represent the average scores for 4 items: amotivation (lack of intention), four subscales reflecting degrees of extrinsic motivation – external regulation, introjected regulation, identified regulation, integrated regulation – and intrinsic regulation. This is the first known study to measure motivation to break up sitting time using a modified version of the BREQ-3 questionnaire.
Menopause status questionnaire
The menopause status questionnaire asks questions to assist in classifying menopause status according to the Stages of Reproductive Aging Workshop (STRAW+ 10) criteria . It is administered only to those participants identifying as female. Participants answer questions pertaining to criteria that fall outside of the STRAW+ 10 including polycystic ovarian syndrome, premature ovarian failure, hypothalamic amenorrhea, oestrogenic malignancy, or endrometriosis. To determine menopause status, participants are asked when (if at all) they had their first menstrual period, and whether they have either temporarily stopped, or finished having their period. Menstrual cycle history, use of hormonal contraceptive or hormonal therapy, and associated symptoms (according to Greene Clinimetric scale)  are recorded.
Experience in the OPTIMISE your health intervention
At the 6, and 12 month assessments for the intervention group, and at the 18 month assessment for the delayed intervention group, questions are included pertaining to the participants’ experience in the intervention. These questions are primarily adapted from previous research  and related to overall experience, as well as experience with the specific components of the intervention. Participants are asked about how the intervention has led to changes in their activities of daily living, and whether the trial changed these activities on a scale of a lot, a little, or no change. The components of the trial are ranked from most important to least important according to their perceived utility to assist with making changes in the domains of sitting less and moving more. Finally, participants are asked about their goal setting, and whether it was realistic and achievable and supported by the research team, as well as the applicability of the intervention to people without diabetes in order to inform the generalisability of the intervention messaging.
COVID-19 impact questionnaires
A questionnaire was originally added to assess the immediate impact of the pandemic and restrictions (Additional file 8). The questionnaire records any changes that the participant incurred with respect to workload, work environment, caring responsibilities, sitting and standing at their workstation, joint and muscle discomfort, motivations to sit less and move more, and physical activity participation changes due to the pandemic and restrictions. Following the temporary ending of restrictions and return of the trial in October 2020, this questionnaire was replaced with a five-item questionnaire. The new questionnaire asks on a ten point scale (not at all – very much) how the pandemic impacts self-management of diabetes, sitting at desired levels, moving at desired levels, and participation in the trial. Finally, participants are asked how the pandemic has changed their work commute.
Any adverse events encountered during the trial are recorded by the research staff when they arise. Adverse health events may lead to a pause in participation or withdrawal from the study. Information collected includes the type of adverse event, the date of onset and resolution (if applicable), the maximum intensity of the adverse event (mild, moderate, severe), action taken, the adjusted medication due to event, the likelihood of relationship to the trial (scale of 1–4; 1 indicating unrelated; 4 indicating definitely related), and whether the event has resolved. Adverse events are reported to Baker Governance and Alfred Ethics Committee (Additional file 10). The participant that reports the adverse event is followed up with a phone call in order to monitor their health and participation in the trial.
A process evaluation is undertaken for the intervention and its components by assessing direct implementation indicators and surveying participants at the end of the intensive (6 months) and extended phase (12 months) timepoints for the intervention group and at 18 months for the delayed intervention group. Elements of context (e.g. workplace support), implementation (e.g., number and duration of health coaching sessions), and mechanisms of impact (including potential mediators) are explored in accordance with the Medical Research Council process evaluation framework . Process evaluation for the intervention components is summarised in Table 5.
Incremental cost-effectiveness analysis is to be undertaken to determine whether the intervention represents “value for money” compared to the “usual care” control. The economic evaluation is conducted from a limited societal perspective, using detailed pathway analysis to specify all relevant intervention activities and costs. A detailed accounting of resource use required to deliver the two arms in the trial is undertaken to allow accurate costing of the interventions. The limited societal perspective incorporates cost impacts on government as the provider of healthcare services, healthcare costs to individuals and costs to workplace organisations. The intervention is costed assuming it is in ‘steady state’ (i.e. excluding research-related costs).
Additionally, a within-trial economic evaluation is to be undertaken as a cost-consequence analysis, comparing the costs of the intervention with the primary outcome measures (e.g. cost per change in HbA1c) at 6 months and 12 months timepoints. A modelled economic evaluation is to be undertaken, extrapolating intervention costs and effects and extending the target population, time horizon and decision context. The model will be extrapolated to lifetime to incorporate benefits accruing to both T2D and dementia prevention. As 5 years is insufficient to observe outcomes of dementia in a prevention study, the intermediate outcomes of the trial (cognitive function, neurotrophic factors and inflammatory markers) will be incorporated.
into final outcomes for economic evaluation, e.g., dementia diagnosis, time to dementia.
Incremental health benefits, reported as quality-adjusted life years (QALYs) saved, and incremental healthcare cost offsets attributable to T2D prevention and dementia are reported. The commonly accepted cost-effectiveness threshold of AUD$50,000 per QALY saved will be used to determine cost-effectiveness over the lifetime. Utility is calculated from the Assessment of Quality of Life AQOL-8D using the Australian algorithm  and multiplied using the area under the curve method by the time in the trial to derive the QALYs.
Finally, an incremental cost-effectiveness ratio will be determined by calculating the difference in cost between the intervention and usual care, divided by the difference in QALYs between intervention and usual care.
With an estimated 20% attrition based on our previous intervention trials, 2-tailed significance of 2.5% (correcting for two primary outcomes for the original OPTIMISE protocol), 125 participants per group (250 total) are required for 80% power to detect minimum differences of interest in HbA1c and sitting of 0.5% and 0.5 h/16 h-day, assuming standard deviations (SD) of 1.6 and 1.3 and a pre-post correlation (r) of 0.7 and 0.6 respectively. A 0.5% HbA1c decrease is clinically meaningful corresponding to to an approximate 10% reduction in diabetes-related mortality . Assumptions concerning attrition, SD and r for primary and most secondary outcomes were based on Living Well with Diabetes , AusDiab  and Stand Up Victoria  studies. Recruitment projections indicate that a sample of 250 participants is feasible within the allocated timeframe. Recruitment will stop when the trial reaches either the required sample size or the maximum number of participants who can be recruited in the trial’s allotted recruitment timeframe, while complying with the unforeseeable restrictions and conditions related to the COVID-19 pandemic.
The sample size for the cognitive function outcomes are dictated by the original sample size calculations described above. From a sample size of 250, we anticipate 160 participants (80 in each arm) will complete the 12-month assessment. This sample size of 80 per group provides > 80% power (5% 2-tailed significance) to detect our minimum difference of interest in the primary outcome of visual memory (PAL TEA) at 12 months (1/3 SD, d = 0.33) assuming r = 0.7, based on a previous trial . Changes in cognitive function have been detected in a similar sample with 145 participants participating in a 24-week RCT assessing pharmacotherapy to improve metabolic control. Minimum detectable differences will be recalculated if the actual sample size is lower than projected due to the potential impact of the COVID-19 pandemic.
Outcomes are all continuous and expected to be either normal or log-normal, in which case log-transformation will be used. Analyses will be performed in STATA version 15 or higher. Significance is set at p < 0.05 two-tailed (except for p < 0.025 when testing co-primary outcomes based on α = 0.025). To test the differences between groups in primary and secondary outcomes during (3 months), end of the intervention (OPTIMISE; 6 months) and end of the extended maintenance phase (12 months – OPTIMISE-extended) mixed models statistical analysis will be employed, accounting for repeated measures, adjusting for baseline values, randomisation strata, and potential confounders. Within-group changes for intervention participants occurring during the no health coaching contact phase from 12 to 18 months will be tested and compared to intervention phases. Within-group changes will also be tested in the delayed intervention participants receiving the OPTIMISED intervention after 12 months, at the 18 month timepoint, using paired t-tests (or non-parametric paired tests). Possible confounders will be identified a priori from the literature and narrowed down to a number that can be modelled without overfitting, based on an objective criterion not open to manipulation (backwards elimination: retaining age, sex, and p < 0.2 association with the outcome). Analyses will follow intention-to-treat principles. Sensitivity to handling missing data will be evaluated by comparing results from the main analyses (evaluable case analysis for mixed models; complete-case analysis for t-tests) with alternative methods that are appropriate for different missing data scenarios (e.g., multiple imputation; selection-covariate adjustment). Descriptive statistics will be used to describe the implementation indicators of the different phases of the intervention.