key: cord-0738368-z9xo4lv9 authors: Buntine, Paul; Aldridge, Emogene; Craig, Simon; Crellin, Dianne; Stella, Julian; Wright, Breanna; Mitchell, Rob D; Arendts, Glenn; Rawson, Helen; Rojek, Amanda M title: A mixed methods investigation of behavioural drivers influencing Emergency Department attendance in Victoria during the 2020 COVID‐19 pandemic date: 2022-03-23 journal: Emerg Med Australas DOI: 10.1111/1742-6723.13973 sha: 1995e33a64b1092c1281180a8f38adec7770b122 doc_id: 738368 cord_uid: z9xo4lv9 OBJECTIVE: To identify behavioural drivers and barriers that may have contributed to changes in ED attendance during the first 10 months of the COVID‐19 pandemic in Victoria. METHODS: We conducted a mixed methods analysis of patients who attended one of eight participating EDs between 1 November 2019 and 31 December 2020. A random sample of patients were chosen after their visit and invited to participate in an online survey assessing behavioural drivers and barriers to attendance. The study timespan was divided into 4 periods based on local and world events to assess changes in attitudes and behaviours over this period. RESULTS: 5600 patients were invited to complete the survey and 606 (11%) submitted sufficient information for analysis. There were significant differences in participants' attitudes towards healthcare and emergency departments, levels of concern about contracting and spreading COVID‐19 and the influence of mask wearing. Patients expressed more concern about the safety of an ED during the largest outbreak of COVID‐19 infections than they did pre‐COVID, but this difference was not sustained once community infection numbers dropped. General concerns about hospital attendance were higher after COVID than they were pre‐COVID. 27% of patients specifically stated that they had delayed their ED attendance. CONCLUSION: Patients expressed increased concerns around attending ED during the first 10 months of the 2020 COVID‐19 pandemic and frequently cited COVID‐19 as a reason for delaying their presentation. These factors would be amenable to mitigation via focussed public health messaging. During the first year of the coronavirus disease 2019 (COVID-19) pandemic, early Australian data suggested a reduction of between 25% and 37% in patients attending Australian emergency departments (ED) for non-COVID-19-related conditions. [1] [2] [3] This mirrored international findings related to both the COVID-19 pandemic [4] [5] [6] [7] and severe acute respiratory syndrome (SARS, 2003) epidemic. [8] [9] [10] The reasons for this reduction are unclear; however, in addition to potential changes in illness and injury patterns, it is possible that behavioural changes related to government-mandated restrictions on socialisation, concerns about unnecessary exposure to potentially infectious patients, fears of overburdening hospitals and changes in patterns of health care consumption may have also contributed. Patient behaviours relating to ED attendance can be influenced by factors unrelated to illness. Previous experience with ED care 11 and pre-existing sociodemographic, socioeconomic and psychosocial factors 12-15 appear to be important. External factors such as the time of day, season and weather conditions also play a role. [16] [17] [18] Additionally, different population groups vary in the degree to which they seek medical care from a primary care physician (PCP) versus an ED. 11 A decrease in PCP capacity, such as that which was observed during the initial phases of the COVID-19 pandemic 19 , might logically be expected to increase ED utilisation even when offset by conversion to telemedicine, 19 -21 yet the sharp drop in ED attendances suggests that this did not occur. The extent to which ED attendance was impacted by changes in illness and injury epidemiology (such as fewer workplace or sporting injuries, fewer non-COVID communicable diseases and a reduction in road traffic-related accidents) and altered patient behaviours due to the pandemic is unknown. This study aimed to identify behavioural drivers and barriers that may have contributed to changes in ED attendance during the first 10 months of the COVID-19 pandemic in Australia, when community transmission remained comparatively low. 22 Patients who attended one of eight participating EDs (Table 1) for emergency care between 1 November 2019 and 31 December 2020 were invited to complete a survey following their ED visit (Appendix 1). The survey evaluated behaviours, attitudes and decision-making processes relating to their ED attendance during this period. The study timespan was divided into four periods pragmatically, based on local and world events: a pre-COVID period characterised by infrequent cases amongst overseas visitors and few cases of local transmission (1 November 2019 to 10 March 2020); an initial wave of infections that commenced with the World Health Organization (WHO) declaring COVID-19 a global pandemic (11 March to 12 May 2020); a second wave of infections that coincided with the largest Victorian outbreak in 2020 (13 May to 31 August 2020); and "COVID normal", a period in which there were no local COVID cases (1 September to 31 December 2020). Ethical approval was obtained from the Alfred Hospital Human Research Ethics Committee (project number: 474/20) and governance approved at all participating sites. Survey questions were adapted from an established behavioural questionnaire and created using behavioural survey design methodology by a behavioural science expert (BW), reviewed and modified by content experts within the research group and reviewed by a community representative for readability. 23,24 Questions were grouped into eight discrete subgroups that related to behaviours and attitudes that the participant had at the time of the specific ED attendance in question: demographic information (7 questions), reason for decision to attend ED (3 questions), behaviours relating to the individual's health care utilisation (5 questions), attitudes to EDs (5 questions), social expectations around ED attendance (1 question), emotions experienced at the time of ED attendance (1 question), general factors that might influence ED attendance (9 questions) and the physical environment of the ED (6 questions). Items consisted of a combination of free-text responses, binary options, selection from lists and 5-point Likert scales, with responses for the latter ranging from "strongly agree" to "strongly disagree" (Appendix 1). The survey was collated in a secure web-based research data collection and management tool called REDCap (Research Electronic Data Capture). 25, 26 Accepted Article A convenience sampling method was employed to recruit participants who had attended a wide range of hospitals (Table 1) , including a large regional hospital and a dedicated paediatric hospital. Six health services (eight EDs) across Melbourne and regional Victoria contributed: Eastern Health (Box Hill Hospital, Angliss Hospital, Maroondah Hospital), Monash Health (Monash Medical Centre, Dandenong Hospital, Casey Hospital), Barwon Health (University Hospital Geelong) and The Royal Children's Hospital. Planned data collection at two additional sites (Alfred Hospital and The Royal Melbourne Hospital) was aborted due to delays in obtaining necessary approvals and an information technology issue that prohibited participant interaction with the REDCap survey tool. Local administrative data were used to identify all ED presentations retrospectively at each site during the study period. The list of presentations was then manually filtered to exclude deceased patients, current inpatients, patients requiring translators, patients transferred to other hospitals, patients brought to the ED by police and patients from residential aged care facilities. A final random sample of fifty patients from each month was then created from each of the eight participating sites using an Excel-based random number generator. These patients were contacted via short messaging service (SMS; seven sites) or registered post (one site) and invited to complete the survey by opening a web link to the REDCap survey instrument. For the seven sites that used SMS messaging, a single reminder message was sent seven days later. Designated guardians were contacted for paediatric presentations. Due to the need to capture comparative behaviours and attitudes that pre-dated the COVID-19 pandemic, there was a large variation ranging from less than 1 month to 12 months between a patient's ED attendance and invitation to participate in the study. This varied from site to site and between time periods, but it was generally longer for patients who had attended during the first 6 months of the study period. As the study was exploratory and no precedent existed prior to its planning, it was not possible to power this study to detect a specific outcome. A 10% response rate was anticipated. Data were extracted from REDCap. Quantitative analysis was performed in SPSS (version 27, IBM, Armonk, NY, USA). Responses were removed if they contained insufficient data for Accepted Article analysis or were unable to be interpreted. Categorical data (gender, age, education level, discharge status, and symptoms prior to ED presentation) were examined using summary statistics. Chi 2 (χ 2 ) analysis was used to determine differences between the four time periods, ANOVA was used to compare each Likert scale variable for the four time periods and Tukey's post hoc analysis was conducted on variables with significant p-values to determine differences between time periods. The use of parametric statistical measures in conjunction with Likert scales is commonly accepted presuming a sample size of sufficient size. 27 A content and thematic analysis was conducted for qualitative variables. Two authors (PB and EA) independently coded the data and developed themes, which they discussed until a consensus was obtained. A total of 5600 patients were contacted to complete the survey, 700 via mail and 4900 via text message. 1205 (22%) opened the link, of whom 606 (11%) submitted sufficient usable information for inclusion in the analysis. A detailed breakdown of participant demographics for all time periods is shown in Table 2 . Most respondents were adults aged 18 and over (98%), female (70%) and discharged home from the ED (50%). The respondents across time periods did not differ significantly according to gender, education, discharge status or between children and adults. Fifty-three percent (n=323) of the respondents experienced symptoms for less than six hours prior to presenting, with the majority (65%) stating that they did not delay their presentation to the ED. Across the four time periods, there were significant differences in participants' attitudes towards healthcare and emergency departments, levels of concern about contracting and spreading COVID-19, and the influence of mask wearing (Table 3 ). Participants reported that they considered emergency departments to be safer before COVID-19 (mean Likert = 4.1) than in time period 3 (May -August 2020, mean Likert = 3.7). Participants were also less concerned Accepted Article about attending health services generally before COVID (mean Likert = 3.6) compared with all post-COVID time periods (mean Likert scores = 3.6, 3.1, 3.2, 3.2 respectively). Participants were more concerned about attending the ED in time periods 2 and 3 (11 March to 12 May and 13 May to 31 August 2020, mean Likert scores = 3.0, 3.1) compared with time period 1 (mean Likert score = 3.5), but not time period 4 (mean Likert = 3.2). Participants were more concerned of coming into contact with an infectious person post COVID-19 (mean Likert scores = 3.2, 3.2, 2.9 time periods 2-4 respectively), than before (mean Likert score = 2.7) However, participants did not report any significant difference between any time periods with respect to concerns about becoming ill from contact with other patients and spreading COVID-19 unknowingly. Patients who presented to an ED between 13 May and 31 August 2020 (time period 3) expected face masks to be mandatory for all patients and visitors within the ED, differing significantly from pre-COVID periods. One hundred and sixty-two participants (27%) reported delaying their ED visit. Of these, 159 respondents (26%) provided free text information detailing one or more reasons for the delay. From these responses, seven main factors were identified which increased the likelihood of a delayed ED presentation. These included a belief that their condition was not serious or would improve, concerns around being exposed to COVID-19, expectation of a negative hospital experience based on previous experience, logistics, seeking alternative prior medical advice, hospital avoidance due to anxiety or apathy, and not wanting to burden the health system (Table 4 ). This survey of adult and paediatric patients from eight Victorian EDs compared patient behavioural responses and attitudes in relation to ED attendance, sampled from a period that extended from three months before to ten months after the WHO declared COVID-19 a pandemic. Participant attitudes and behavioural outcomes were similar across the study population but varied by time period. In particular, despite also flagging that mandatory mask wearing provided a degree of reassurance, patients expressed more concern about the safety of an ED during the largest outbreak of COVID-19 infections than they did pre-COVID, but this Accepted Article difference was not sustained once community infection numbers dropped. General concerns about attending a hospital or an ED were higher during all periods after COVID than they were pre-COVID, and no differences were observed in attitudes around general hygienic measures (handwashing, physical distancing). Approximately one quarter of patients provided specific reasons for delaying their ED attendance. In line with our findings, Nab et al. 28 specifically reported COVID-19-related delays due to behavioural changes in Dutch patients seeking ED care during the Netherlands' initial COVID-19 wave. Nab et al. reported a smaller proportion of their participants indicated that they delayed their presentation (20% vs 27%), than our sample but a greater proportion mentioned specific COVID-19 related concerns when doing so (45.4% vs 17%). Likewise, a small qualitative study of paediatric caregivers in London by Watson et al. described delays in decision to seek care during the COVID-19 pandemic being driven by fear of exposure to COVID-19, driven by community perception, shared lived experiences and media portrayal. 29 Other reasons for delay captured in our study may also relate indirectly to concerns around COVID-19. A thematic exploration at a community hospital in America by Wong et al. 5 describes five main themes associated with decreased ED attendance during the COVID-19 pandemic: hospitals seen as infectious reservoirs, patients not being informed about current risk mitigation efforts by hospitals, need for confirmation from a trusted source that it was "ok to come in", national focus on extreme cases skewing perceptions of risk and delays amongst vulnerable groups due to social disconnection. It is likely that many of these concerns were captured under the additional themes identified in our study. To our knowledge, ours is the largest study focusing specifically on behavioural drivers relating to COVID-19 ED attendance. However, it is unclear whether patients from areas with different rates of COVID-19 infection, or from rural or disadvantaged areas, may have behaved differently. In their survey of community residents during the first four weeks of the South Korean COVID-19 outbreak, Lee et al. found that females, those aged in their fifties and patients from urban areas were all more likely to avoid healthcare due to COVID-19-related concerns. 30 These differences in age and sex were not apparent in the specific behavioural concerns expressed amongst the ED attenders who responded to our study, although it is possible that this Accepted Article was due to inherent limitations in our study design. We did not include a rural subset in our data collection. Our study has significant limitations which are inherent in the subject matter, study design and population. In aiming to study behavioural drivers, it would have been best to study both ED attenders and non-attenders, with a validated survey tool. The survey used in this study was purpose built and while drew from validated tools, was not tested prior to use. A parallel study designed to explore behavioural drivers in ED non-attenders was unable to be completed due to complexities in obtaining relevant approvals in the short time frame afforded by the evolving pandemic. Without this data, we can only describe the experience of patients who actually presented to the ED at some stage and cannot comment on the behaviours of patients who failed to attend. Furthermore, we do not know whether the differences that we observed have real clinical significance. There was a significant delay between ED presentation and sending out the questionnaires, which may have created recall bias within responses. Due to logistical challenges, the delay was between 1 month and 12 months from presentation and invitation to participate. It is possible that individuals' responses were further influenced by the media and events that occurred during the period of time between ED discharge and invitation, and therefore the responses given may not truly reflect the participants true feelings and beliefs at the time of presentation. It is also possible that as the pandemic continued, individuals became less concerned about COVID-19 and the values recorded during later time periods are underestimations of participants' true concerns. There is clear response bias with 70% of the participants being female, however, it has been demonstrated that females are more likely to participate in online surveys than males. [31] [32] [33] No rural EDs participated in this study, so it is not possible to know if differences exist in this setting. Similarly, due to study resource limitations, we excluded patients who required translators which may have created selection bias due to differences in the degree by which people from diverse culture and language backgrounds utilise EDs and healthcare in general. Our study had a low response rate of 11%, which may have improved through participants receiving the invitation to participate closer to their ED discharge date. Another important limitation is the fast pace at which the pandemic unfolded, and with it the variables which we studied. We concluded data collection on 31 December 2020, when the majority of the population were still unvaccinated, and before the prominent COVID-19 variants (Delta and Omicron) arrived in Australia. Today's post COVID-19 world is different and this snapshot of behavioural drivers during the early COVID-19 pandemic may not be applicable to current ED attendance behaviours. Nevertheless, it is likely that many of the behaviours observed in this study are not COVID-19 specific and have relevance to any population events that receive significant media attention. In particular, future public health messaging around EDs being safe places to seek medical treatment could assist in mitigating concerns during these events. Also, as we were informed by the SARs and MERs outbreaks, future pandemic planning and response will draw on COVID-19 literature, and the results from our study can assist decision makers and public health messaging. Patients attending a Victorian ED during the first 10 months of the 2020 COVID-19 pandemic expressed increased concerns around attending an ED or health service. Concerns relating to the safety of EDs were greatest during the period of highest community infections, and while mask wearing appeared to provide some level of reassurance, it did not fully mitigate these concerns. COVID-19-related concerns were frequently cited as reasons for delayed presentation to ED, which has the potential for associated adverse health consequences. All these factors would be amenable to mitigation via focussed public health messaging. This article is protected by copyright. All rights reserved. Accepted Article Accepted Article Impact of COVID-19 State of Emergency restrictions on presentations to two Victorian emergency departments Fewer presentations to metropolitan emergency departments during the COVID-19 pandemic Emergency department mental health presentations before and during the COVID-19 outbreak in Western Australia Do not stay at home: We are ready for you. NEJM Catalyst Innovations in Care Delivery Where are all the patients? Addressing COVID-19 fear to encourage sick patients to seek emergency care. NEJM Catalyst Innovations in Care Delivery Changes in emergency department activity and the first COVID-19 lockdown: A cross-sectional study The impact of COVID-19 pandemic on psychiatric emergency department visits -A descriptive study Impact of SARS on an emergency department in Hong Kong Impact of severe acute respiratory syndrome (SARS) outbreaks on the use of emergency department medical resources Declining emergency department visits and costs during the severe acute respiratory syndrome (SARS) outbreak Nonurgent emergency department patient characteristics and barriers to primary care When I visited the Emergency Department I didn't want to take a bed from someone who needed it ( This article is protected by copyright. All rights reserved. "thought the bleeding was temporary" (ID 440) "thought I was overreacting" (ID 447) "thought it may have been muscular" (ID 230) "thought it might get better" (ID 19) "tried to get over it" (ID 501) "waiting to see if sleep improved symptoms" (ID 10) Concerns about being exposed to COVID-19 27 (15%) "fear and uncertainty around the COVID situation at the hospital, potential delays and exposure to virus" (ID 73)More common in time periods 3 (n = 11; 41%) and 4 (n = 11; 41%), than time period 2 (n = 5; 18%) Hospital avoidance due to anxiety or apathy 6 (3%) "I was worried that they would want to admit me, and I couldn't do that as I would be leaving my husband at home with 6 children" (ID 244) "GP appointment was in the evening, children were already tired and I didn't really think this is an emergency, so waited for the next day" (ID 419) Not wanting to burden the health system 10 (5%) (Of these, 30% involved uncertainty as to whether the condition warranted an ED visit) "concerned it was nothing and did not wish to take up valuable resources" (ID 296) 90% of these responses corresponded to time periods 3 and 4. This article is protected by copyright. All rights reserved.