As a quality improvement project, rather than a randomised controlled trial, this project is inherently at greater risk of confounding. For example, the patient characteristics in each of the baseline audits and PDSA cycles are likely to differ. This may be reflected in the slight changes in ACEI (or ARB) and beta-blocker prescribing seen over time. It is uncertain whether this is explained by the difference in the patient cohort over time, such as patients with more severe LVSD in the pilot phase having a higher sympathetic drive and hence tolerating slightly higher doses of beta-blocker. Alternatively, changes may reflect the challenges of involving more staff members with different levels of experience. There were also significant changes in our health authority over the time period of the intervention, including the closure of three older hospitals and the opening of a new state-of-the-art hospital to provide acute care to a large proportion of the population. Reasons for not achieving target doses were not available for both time periods, so are not described.