The influence of personal communities in understanding avoidable emergency department attendance: qualitative study

Abstract

Background: Use of emergency department (ED) care globally seems to be increasing at a faster rate than population growth (Baker, House of Commons Library. Accident and Emergency Statistics, Demand, Performance, 2017). In the UK there has been a reported 16% rise in emergency admissions over the past 5 years. Estimates that between 11 and 40% of ED attendances are non-urgent, with 11% of patients being discharged from the ED without treatment (NHS Digital 2017), and a further 44% require no follow-up treatment (NHS Digital, Hospital Accident and Emergency Activity 2016-17, 2019) is cited as evidence that these patients did not require this level of care. The solution to not using the most appropriate point in the system has traditionally been seen as a knowledge problem, requiring, improved sign-posting and information to enable people to self-manage or use health care management for minor ailments. However research about help-seeking behaviour suggests that the problem may not be an informational one. A considerable literature points to help seeking as a social process influenced by a range of contingencies and contextual factors including the way in which lay people influence health care utilisation (Giebel et al. BMJ Open 9:1, 2019). Personal communities comprise a variety of active and significant social ties which have potential to influence individual capacity to seek help. Here we extend and unpack further influencing decisions about seeking formal health care with reference to how they are shaped and informed by and within personal social networks.

Methods: We undertook a personal network mapping and qualitative interview-based study to look at, problematize and understand attendance for non-urgent problems. We used network analysis and methods to map and characterise the personal communities of people seeking help from ED for minor ailments and semi-structured interviews with 40 people attending a single ED and associated GP hub providing equivalent care. Interviews were built around an ego network mapping activity and a topic guide structured to explore attender's narratives about why they had visited the ED. This ego network activity uses a diagram consisting of three concentric circles (Fiori et al. J Gerontol B-Psychol 62: 322-30, 2007), representing closest social network members (in the centre) and those at further distance. Participants were initially presented with one of these diagrams and asked to write names of people or resources that had played a role in their attendance and the interviewer probed the interviewee to discuss the actions, input and value of the people and services that supported the visit to the ED.

Results: We analysed number and type of network connections and undertook a thematic analysis to identify how imagined and actual network members and influences were implicated in ED attendance. The network maps created during the interviews were examined and a typology of networks was developed and used to distinguish different types of networks informed by our reading of the data, and a Network Typology Scoring Tool, a measure of frequency of contact and relationship type in networks.

Conclusions: Our study suggests that faced with acute minor illness or injury people's networks narrow: they do not (and perhaps cannot) mobilise their imagined care network because the resources or connections may not be there or are difficult to engage. In addition we identified important system drivers of behaviour, notably that these patients are often directed to the ED by 'professional influencers' including health services staff.

Keywords: Emergency care; Emergency department; Healthcare service; Help-seeking; Inappropriate attendance; Qualitative methods; Social networks.

Conflict of interest statement

The authors were members of the NIHR CLARHC Wessex and there are no financial or non-financial competing interests in relation to this manuscript.

Figures

Fig. 1

Example of participant’s actual network being very small and their imagined network being family supported

Fig. 2

Example of participant’s actual network being very small and their imagined network being diverse

Fig. 3

Example of participant’s actual and imagined networks as very small