A multicentre randomised controlled trial of an augmented exercise referral scheme using web-based behavioural support in individuals with metabolic, musculoskeletal and mental health conditions: findings from the e-coachER study.

Talk Code: 
Adrian Taylor
Wendy Ingram, Rod Taylor, Nana Anokbe, Lucy Yardley, Kate Jolly, Nanette Mutrie, Jeff Lambert, John Campbell, Sarah Dean, Colin Greaves, Mary Steele, Paul Little
Author institutions: 
University of Plymouth, University of Exeter, University of Birmingham, University of Edinburgh, Brunel University, University of Southampton


There is no rigorous evidence that exercise referral schemes (ERS) increase physical activity (PA) of individuals with chronic health conditions. E-coachER is a novel web-based behavioural support package (https://www.ncbi.nlm.nih.gov/pubmed/30244214 ) to augment ERS. This is the first RCT to report whether adding web-based support to ERS improves objectively-assessed PA after one year compared to usual ERS.


450 low active adults with obesity, diabetes, hypertension, osteoarthritis or depression were recruited by primary care or exercise practitioners in Glasgow, Birmingham and Devon, and randomised to receive a pedometer, PA recording strips and a guide to access the 7-step web-based support package. The primary outcome was minutes of moderate and vigorous PA (MVPA) in ≥10 min bouts measured by accelerometer over one week at 12 months, with a wear-time threshold of ≥16 hours per day for ≥4 days including ≥1 weekend day. Other accelerometer-derived PA measures, self-reported PA, ERS attendance, EQ-5D-5L and HADS were collected at 4 and 12 months.


The sample had a BMI mean (SD) of 32.6 (4.4), and were primarily referred for weight loss, though 54% reported having low mood as a referral reason. The provisional ITT complete case adjusted comparison of groups at 12 months showed a weak effect in favour of the Intervention group (N=236; 95% CI -1.4 to 25.9; P = 0.08), adjusting for site, participant's reported main reason for referral, IT literacy level, age and gender. Among the intervention participants, 64% logged on to the on-line support at least once, and 36% progressed to review a PA goal (Step 5). Reaching the goal review stage (or not) did not influence the findings in a CACE analysis. The intervention had no significant effect on ERS attendance, EQ-5D-5L or HADS scores, but in complete case repeated measures analyses (including both 4 and 12-months follow up) the intervention participants reported lower depression (P < 0.05) and anxiety (P = 0.05) scores compared with the control group. Challenges were faced in the primary outcome analysis: Few people accumulated ≥10 min bouts of accelerometer accumulated MVPA, resulting in a poor fit for the predefined model. Further sensitivity analyses are warranted.


Adding e-coachER to usual ERS had only a small effect on long-term objectively assessed MVPA. Engagement in the intervention was acceptable but whether or not they completed a goal review did not influence the findings. Web-based support overcame the need to train ERS professionals to provide consistent, evidence-based, behaviour change support, but further analysis is needed to understand differences in those who did and didn’t use the support. The trial involved a number of steps to identify and recruit participants which would not be needed if patients received similar support in a routine primary care referral.

Submitted by: 
Adrian Taylor
Funding acknowledgement: 
Funding was received from the Department of Health as part of the National Institute for Health Research (NIHR) Health Techology Assessment (HTA) programme (project number 13/25/20)