Seite - 168 - in Applied Interdisciplinary Theory in Health Informatics - Knowledge Base for Practitioners

correspond to actual practice; how they are influenced by feedback; and, ultimately, how feedback changes intentions to improve practice. To that end we designed an online two- step experiment, driven by CT, to elicit these beliefs and intentions before and after receiving first-time feedback. The experiment took place upon first login into the dashboard; 83 intensive care professionals from 21 units participated. In step 1, professionals were presented with the indicator descriptions whilst withholding all performance feedback (that is, no performance scores or benchmark comparisons were displayed). Professionals were asked to estimate for each indicator their own unit’s performance score, the national average score, the minimum score they would consider “good performance” (target), and whether or not they would perform actions to improve. The study found that half of the time professionals overestimated their own performance and rarely underestimated it. Targets were set very high. In step 2 professionals received feedback on their performance. Feedback included the unit’s own performance, median and top 10% peer performance, and improvement recommendations based on peer comparisons (good performance; room for improvement; or improvement recommended). Professionals were asked again, but now given the performance information at hand, what their performance target was and whether they intended to improve practice. If improvement intentions were at odds with CT in step 1 (score < target and no intention; or score ≥ target and still intention) or in step 2 (e.g. room for improvement but no intention to improve) we asked professionals to explain their choice. Also, if there were discrepancies between intentions in the first and second step, professionals were asked what feedback elements drove them to change (e.g. measured score or benchmark was higher/lower than expected). Even before receiving any feedback some 68% of professionals’ intentions corresponded with the feedback recommendations. In other words, while professionals were not very good at estimating absolute performance, they had good intuitions about whether it was on target or not– without seeing any numerical information. After receiving the feedback, this number increased to 79%. In more than half of the cases in which units were already top performers, professionals still wanted to improve. In 8% of cases professionals lacked improvement intentions because they did not consider the indicators important; did not trust the data; or deemed benchmarks unrealistic. This research concluded that audit and feedback does indeed help healthcare professionals to work on those aspects for which improvement is recommended because it increases the accuracy of their clinical performance perceptions. However, given the abundance of professionals’ prior good improvement intentions, efforts to optimise A&F interventions should focus on translating those intentions into (effective) actual change in clinical practice. 4. Discussion Control theory (CT) provides a conceptual framework for self-regulation and human behaviour and has already demonstrated its usefulness for the field of HI and in particular A&F interventions. In the A&F literature CT has been used to synthesise evidence of interventions, enhance their design, explain why interventions were or were not successful, and generate hypotheses about how feedback mechanisms work in practice. Nevertheless, the majority of studies have not explicitly reported the use of CT (or other relevant theories) for such purposes. The simplicity of CT’s negative feedback loop makes for an elegant framework that is widely applicable, but it also has limitations. Individuals may compare feedback to W.T.GudeandN.Peek /ControlTheory toDesignandEvaluateAuditandFeedback Interventions168