Clinical reminder alert fatigue in healthcare: a systematic literature review protocol using qualitative evidence

Implementation of desirable clinical behaviours, such as compliance with national guidelines, has taken a variety of forms within healthcare [1, 2]. One popular intervention is the use of an integrated electronic reminder or prompt [3], often taking the form of a ‘pop-up box’ within the clinical health record system [4, 5]. This type of prompt, usually delivered at the point of care, has two broad functions; reminding the user to perform a task, or to alert them about the potential consequences of not performing a task [6].

Approximately 90% of all patient consultations in the UK occur in primary care, equating to some 360 million appointments per year [7, 8]. The treatment of the majority of patients relies on guidelines designed to treat single diseases [9]. However, current estimates indicate that 2.9 million people in the UK will be diagnosed as having two or more chronic conditions by 2018 [10], leading to a rise in the number of reminders for each patient, particularly where treatments are combined [11, 12]. These clinical reminders are used to influence clinician’s behaviour in a number of aspects of care including prescribing [12]. In exploring the benefits of clinical support systems, a number of reviews have reported broadly positive effects. However, frequently these alerts are evaluated individually rather than in a cumulative fashion; furthermore, there has been a dramatic rise in the number of prompts, or reminders. Many are integrated within electronic health records and should be activated at the point of care to promote shared decision-making [13‐15]. However, the numbers of prompts combined with increasing time pressures in the National Health Service (NHS) is increasing the cognitive load within this working environment [16], to such an extent that these reminders are largely being disabled, thus negating their effectiveness [17]. This failure to engage with these digital prompts has been termed ‘alert fatigue’. By manually disabling all prompts, regardless of content, there are implications for patient safety, such as those providing warnings over medication [18], but also the increased interruption to the patient consultation can lead to a potential increase in medicolegal risk [19].

The aim of this research is to identify the barriers and enablers in the use of alerts and reminders within clinical systems as reported by healthcare professionals. Our focus will be within primary care and the electronic alerts found in this setting, including but not limited to alerts for prescribing, practice performance and lifestyle advice. This will be achieved by conducting a systematic review of the qualitative evidence to form our own overarching themes relating to the perceptions and experiences of alerts within computerised clinical decision support systems (CCDSS) [20]. We will form these themes using an inductive approach [21] to facilitate the design of future clinical systems to minimise alert fatigue.

Specifically, we will locate and synthesise all qualitative evidence on the use of research evidence on the experience of electronic alerts in primary care. This will include data on

Current evidence suggests that alerts have a range of effectiveness [34‐38] yet an understanding of the factors that influence discrepancies inefficacy is lacking. There have been several reviews around the use of clinical support tools [34, 35, 39] though there are few high quality reviews which have ultimately attempted to construct a conceptual framework from the qualitative data. Similarly lacking is robust evidence on how the characteristics of alerts and the way they are deployed can impact on the phenomenon of alert fatigue in primary care. Therefore, our review will add considerably to the existing evidence base. The systematic nature of our approach and the comprehensive acquisition of evidence being assessed will provide further insights into how alerts can be altered to improve practitioner and patient outcomes and help explain the heterogeneity in effectiveness and offer suggestions that will increase utility and usability of alerts within CCDSS.