Data collection
Data was collected by authors (MA, BA, MB, and MW) prospectively from patient interviews, electronic health records, and through discussion with the health care team. All data collectors were trained on the use of a standardised proforma for data collection. This proforma included patient demographics, BOOST score, and LACE index with individual subsections. The proforma was adapted and agreed by the authors (MA, VS, GL) prior to starting the study. Patient demographics consisted of age, gender, and presenting complaint. The BOOST scoring system consisted of 8 parts (Table 1 ).
As the BOOST 8Ps scoring system does not specifically define the criteria required to score in each category, to prevent bias we implemented an objective threshold for each category using a methodology previously used [11]. In the case of ‘poor health literacy’, we defined this as an inability to answer 2/3 of the teach back assessment questions. For the purposes of further analysis, we collected information on which criteria each patient met to earn a point in each category. For example, data collectors were asked to circle each medication a patient was taking when scoring the problem medication section. The total BOOST score was summed in the end out of a total of 8 points. The total LACE index score was calculated out of 19.
All proforma were kept secured on-site in conformance with trust information governance policy.