An integrative review of adult patient-reported reasons for non-urgent use of the emergency department

Numerous studies have specifically examined factors influencing use of the ED on a non-urgent basis. Since 2009, five reviews with different methods and foci have captured varying aspects of this literature [3, 5‐8]. Carret et al. [3] and Uscher-Pines et al. [8] both conducted a systematic literature review of quantitative research studies (retrospective and prospective) and examined variables (i.e., sociodemographic and clinical factors) associated with non-urgent use. The reviews by Kraaijvanger et al., [7] Coster et al., [6] and O’Cathain et al. [5] included both qualitative and quantitative studies of both adult and pediatric populations; they all specifically explored patient-reported reasons for ED use. Kraaijvanger et al. [7] performed a systematic review and meta-analysis and Coster et al. [6] performed a rapid (non-systematic) review; O’Cathain et al. [5] performed a realist (non-systematic) review, expanding on these two reviews.

While the academic literature contains various syntheses related to the subject of non-urgent ED use, limitations exist in terms of the 1) heterogeneity of the included population, 2) type of methods employed (systematic versus non-systematic literature searching), 3) assessment of quality and completeness of data culling and abstraction, and 4) subjective reporting of reasons for non-urgent ED use (i.e., patient reported versus inferred). For example, inclusion criteria varied significantly between reviews; the pediatric population was included in three of the five reviews, [5‐7] patients arriving at the ED via ambulance were included in one review, [5] and specific disease categories were included for two reviews [3, 7]. Combining these different populations is problematic as there are clear differences in the decision-making process for medical care of children by parents and caregivers, as well as for taking an ambulance ride (versus walking in the front door). Non-systematic methods were used in two reviews [5, 6] and another review [8] included only American articles. It was not always clear in the reviews how patients were triaged “non-urgent.” Assessments of methodological quality were only reported by Coster et al. [6]. In some reviews, tabled data and reference lists were incomplete. Given the heterogeneity of methods, design, quality appraisal of resources, and synthesis approach, it is difficult to draw meaningful conclusions on patient-reported reasons for ED use. A comprehensive review with defined criteria may better inform practice and policy moving forward.

The objective of this review was to conduct an integrative review of the scientific literature to explore adult patient-reported reasons for using the ED non-urgently.

A study protocol was not previously registered for this review.

An integrative review of the evidence was performed using the methodology described by Whittemore and Knafl [9]. An integrative review is a review method which summarizes empirical or theoretical literature in an effort to comprehensively understand phenomena or a health care problem [10]. They are often used in nursing science where a review of the state of science may directly inform research, practice and policy [9]. The integrative review method allows for the inclusion of a number of different methodologies (i.e., experimental, non-experimental, qualitative, and quantitative) and therefore may be applicable to problems of importance in nursing [9]. Given the nature of the topic, as well as the heterogeneity of design types, this method was deemed most appropriate. The five-stage review approach by Whittemore and Knafl [9] was undertaken: (1) problem identification (i.e., introduction, background, aim), (2) literature search, (3) data evaluation, (4) data analysis, and (5) presentation.

A literature search of multiple databases (i.e., CINAHL, Cochrane, Embase, PsycINFO, MEDLINE, and Scopus) was conducted by applying a systematic approach consistent with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). The search strategy was developed in an iterative form in consultation with a health sciences librarian. It involved three overarching constructs related to the ED, non-urgent care, and decision-making, with MeSH terms, key words, and subject headings used for each database, as appropriate. Filters were applied for the following restrictions: studies involving humans, published between January 1, 1990 and September 1, 2021, and in the English language. Supplementary search techniques consisted of scanning the reference lists of retrieved articles and reviews on the topic for missed citations.

Articles retrieved from each database search were downloaded to EndNote (Version 9.0). After removing duplicates, each article title was assessed for relevance by two screeners (AMc, SJ). Relevant abstracts and subsequent full-text articles were then screened according to the following four [4] a priori inclusion criteria:

The following types of studies were excluded: general, non-systematic reviews, expository/textbook chapters, conference proceedings, program reviews/descriptions (without a study sample), continuous learning/education modules, and clinical practice guidelines. If a study included a sample with only one specific medical condition or disease (e.g., asthma or congestive heart failure or epilepsy), or the study examined reasons for taking an ambulance to the ED, it was excluded. Studies assessing frequent, repeat or high-use ED users only were excluded. There was no minimum sample size required for inclusion.

Two independent reviewers (AMc, SJ) assessed each study for methodological quality using two commonly used quality assessment tools. Qualitative studies were assessed using the Critical Appraisal Skills Programme Qualitative Checklist (CASP) [11]. The CASP is a 10-item questionnaire which allows one to evaluate qualitative studies among three broad areas: 1) Are the results of the study valid (Section A); 2) What are the results (Section B); and 3) Will the results help locally? (Section C). Items were rated either ‘yes’ or ‘not reported’ (i.e., not reported or could not tell). The CASP authors do not suggest scoring the items. Quantitative studies were assessed using the National Institutes Health (NIH) Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies [12]. This tool includes 14 items evaluating a wide range of quality measures; however, some items are more relevant to cohort studies (items 6–10, 12, 13). We defined quantitative, cross-sectional studies as those using methods such as in-person or postal surveys and/or structured interviews where statistical analysis was performed. The CASP and NIH tools have been previously used to evaluate studies on non-urgent use of the ED in a previous rapid review by Coster et al. [6].

Two reviewers (AMc, SJ) abstracted the following data from each study included for review: author(s), year of publication, country of first author’s origin, study design (i.e., quantitative, qualitative, mixed methods, or review), sample size, method of triage (i.e., triage system, definition or list of explicit criteria), patient characteristics (e.g., age, gender), study aim/objective, data collection method, and themes/reasons for ED use. The reasons for ED use were abstracted in the manner and language in which they were reported by the original authors. Data were abstracted and summarized in tabular form.

Whittemore and Knafl [9] describe the importance of identifying themes in the data abstraction and synthesis process. As such, the two reviewers coded and identified themes (data analysis stage) from each of the study’s key results. A structured, six-phase thematic analysis was applied using the approach by Braun and Clark [13] (Table 1). This method can be adapted for different types of data (including reviews) [14]. This approach has recently been used for assessing both qualitative and quantitative studies in an integrative review [15]. Specific data synthesis actions are outlined in Table 1. For qualitative studies, the original themes identified by individual studies (with supporting quotes and examples) were reviewed, coded and iteratively compared until large overarching themes between studies were uncovered. For quantitative studies, authors primarily reported results in tabular or list format whereby patient reported reasons were given (usually as statements), along with the proportion of the sample reporting this reason. Similar to qualitative articles, the list of reasons reported by quantitative studies were reviewed iteratively, coded and compared between studies until overarching themes emerged.

The literature review returned 3,268 studies; 401 abstracts and 171 full-length texts were reviewed, from which a total of 93 studies met the previously stated inclusion criteria (Fig. 1).

Among the 93 articles included, 69 were quantitative studies [16‐84], 21 were qualitative studies [85‐105] and 3 were reviews [5‐7] (Tables 2 and 3). CASP and NIH quality criteria were applied to included articles to consider their rigour. Articles were not excluded based on rigour, rather, areas where there were concerns about rigour within an article were then discussed amongst the research team and themes did not rely solely on such articles. Overall, the included articles were shown to be of very high quality.

In examining five-year intervals, 11 studies were published in 1995–1999, 8 studies in 2000–2004, 18 studies in 2005–2009, 23 studies in 2010–2014, 24 studies in 2015–2019 and 9 in 2020–2021. Studies were published across 16 different countries, the majority of which originated from just five countries (64.5%): United States (N = 20), United Kingdom (N = 18), Canada (N = 8), Australia/New Zealand (N = 7), and the Netherlands (N = 6). Among the remaining 33 studies, 18 originated from Europe, 3 from South America, 5 from Asia, 6 from the Middle East, and 2 from Africa. Some studies evaluated reasons for using the ED among patients with all types of medical severity. However, for this review, only reasons for attending were collected on non-urgent patients. The total number of patients included in the studies used for the review (excluding review papers) was 49,238. Approximately one quarter of studies had samples with either less than 100 patients (28.8%), or more than 500 patients (26.6%); the bulk of studies (44.6%) recruited between 100 and 500 patients. Fifteen studies (16.7%) did not provide information on the sex of patients (N = 7730 total patients). Among the 75 studies that reported sex ratios, there were 21,044 males (50.2%) and 20,864 females (49.8%) in total. A wide variety of formal triage classification systems were used in 39 studies to assign a severity and urgency of patients’ presenting complaints (Tables 2 and 3). A total of 44 studies did not specify which triage system was used and instead reported that “non-urgent” patients were recruited for participation. In seven studies, a 3-, 4-, or 5-level triage system was described but it was not formally named. The most commonly used triage system was the Canadian Triage and Acuity Scale (CTAS; N = 12) and this was used within and outside of Canada.

After comparing and contrasting major reasons for non-urgent ED use among studies, a total of seven major themes were identified:

Each theme and sub-theme will be described. Some patients reported that they had no specific reason for attending the ED [32, 35, 36, 63, 79, 81]. Several studies stated that there were “other, unspecified reasons” reported by patients; however, there were no further details provided [19, 28, 32‐36, 38, 42, 47, 51, 60, 61, 64, 65, 69, 76, 77, 79, 81, 91, 103].

One of the primary reasons reported in the literature for presentation at an ED was the tendency of patients to be risk averse in terms of their health issue. There was a self-perceived sense of severity or urgency to their medical matters, despite that their presenting complaints were deemed non-urgent [16‐54, 85‐90]. Many patients described having feelings of anxiety, uncertainty or significant concern about their health problem [22, 45, 65, 70, 88, 92‐95]. Often patients had experienced pain or other discomforts which impacted their function and they desired immediate relief [22, 42, 45, 70‐73, 88, 93, 94, 96, 97]. In some cases, they had attempted self-treatment at home, without good effect [73, 86, 96, 98], or had sought out primary care without resolve [21, 42, 53, 66, 74, 75, 88, 92]. Even when patients knew their condition was non-urgent, they still wanted reassurance, advice, or a second opinion [26, 33, 37, 43, 44, 48, 53, 55, 56, 64, 69, 72, 88, 94, 97]. One study found that patients had a self-perceived inability to cope [88]. The ability to leave the ED with a confirmed diagnosis or answer to their health problem (attestation) was particularly helpful in mitigating their fears of a real emergency [54, 55, 100, 102, 103].

Studies reported that some ED patients had limited knowledge and awareness of alternative sources of medical care. They were unaware or unsure of the differences between services [16, 27, 33, 35, 43, 51, 52, 55, 62, 66, 76, 78, 87, 89, 95] or simply had not considered going to their PCP [64, 90]. Patients reported that they believed the ED was the only and most appropriate option [22, 24, 31, 34, 40, 45, 54, 61, 63‐69, 86, 90, 91]. Some people did not want to bother their PCP and did not feel it was necessary to seek primary care first [35, 43, 44, 53, 55, 77, 79].

One of the most prevalent themes was related to patients’ dissatisfaction with primary care services. Within this theme there were three sub-themes: availability, competency, and preference.

Satisfaction with the ED was a highly cited reason for attending non-urgently. This theme included two sub-themes related to benefits of the ED, namely quality of care and access to ED-specific services.

Compared to other primary care options, the ease, accessibility and convenience offered in the ED provided patients with a low burden of access to medical care [17, 18, 21, 22, 26, 30, 31, 34‐38, 43, 47, 51, 59, 61, 63, 65, 68, 70, 72, 76, 79, 85, 87, 90, 92, 105]. Patients reported that it saved them time and overall the wait was short in order to receive help [18, 19, 26, 27, 30‐33, 39, 42‐44, 55, 60, 61, 64, 66, 81, 82, 94, 95, 99, 100, 103, 105]. Some patients faced transportation barriers getting to their PCP so it was easier to access the ED [54, 55, 89, 96]. Similarly, a great number of patients reported geographical proximity to the ED as a motivating factor for attending non-urgently [16, 19, 22, 27, 30, 32, 35, 36, 41‐43, 45‐48, 53‐57, 60‐62, 64, 66, 72, 74, 75, 77, 81‐83, 92, 94, 95]. Other convenience factors, such as not requiring an appointment [18, 22, 33, 37, 43, 45, 48, 53, 56, 58, 76, 95] and unrestricted availability (open day and night) [25, 40, 72] were cited as important indicators for ED use. For patients seeking care where medical insurance coverage may be problematic, EDs were often sought out for relief of any financial burden [19, 20, 22, 24, 30, 32, 34, 35, 39, 41, 45, 46, 48, 51, 53, 54, 57, 65, 68, 72, 75, 78, 82, 84, 87, 89]. In a small number of studies, the reason was circumstantial. For example, patients reported being on vacation or were from out of town [56, 58] whereas others just happened to be at the hospital for an unrelated reason [53, 81].

Patients were often referred by others to attend the ED for their problem; there were two types of referrals discussed, those made by health care professionals and those made by non-health care professionals.

There are often dynamic interactions or relationships between patients and health care providers. In certain groups and geographical regions, use of the ED was an automatic, habitual behavioural or cultural practice shared by many patients [26, 33, 42, 45, 47, 48, 51, 52, 56, 66, 76, 83, 87]. For hospital staff or members of their family, the ED was a logical place to attend given their proximity to place of employment; the relationships these patients had with the ED (and the health care system at large) facilitated its use [19, 32, 42, 46, 47]. Attending the ED, even non-urgently, also made ‘sense’ for those who were currently (or previously) receiving treatment from that hospital already [19, 22, 27, 32, 34, 36, 45, 51, 65, 72, 78, 79, 95]. Conversely, for others, the ED acted as a place of anonymity because no relationship existed. The possibility of obtaining medical care from a doctor they did not know [37, 48, 56] or from someone of the same or opposite sex [48] was appealing.

The aim of this study was to conduct an integrative review of the scientific literature to explore patient-reported reasons for using the ED non-urgently. The studies included for review reported that attending the ED was an intentional decision based on several influential factors. Seven main themes were identified: 1) Need to be risk averse with respect to the health issue; 2) Knowledge and awareness of alternative sources of care; 3) Dissatisfaction with PCP (Subthemes: availability, competence, preference); 4) Satisfaction with ED (Subthemes: quality care, access to ED-specific services); 5) ED accessibility and convenience resulting in low access burden; 6) Referred to the ED by others (Subthemes: health care professionals, non-health care professionals); and 7) Relationships between patients and health care providers. For many patients, there was a very clear problem which needed to be addressed, whether it was physical, psychological, or social. After weighing several options, from their perspective their need was real and the ED as an option for care was rational and justified, not just their last resort.

The results reported here are well-aligned with other reviews [5‐7], but also extend the current knowledge of the subject by providing a comprehensive synthesis of all extant literature of reasons for non-urgent ED use. Recently, O’Cathain et al. [5] examined non-urgent ED use using a ‘realist review.’ Building on earlier reviews [6, 7, 106], they performed an updated literature search to the end of January 2017. They compiled and compared results from 29 quantitative studies, existing health behaviour theories, and 32 qualitative studies. Our integrative review was able to validate and supplement the ten program theories and six mechanisms of decision-making as described by O’Cathain et al. [5] with a larger compilation of studies. With respect to program theories, we did not uncover the theme of ‘fear of consequences when responsible for others’ found by these authors. This theme potentially relates to individuals’ responsibility to care for children, and we did not include studies on the pediatric population. Five mechanisms of decision-making were described by O’Cathain et al. [5] and shared with our integrative review (i.e., the need to be risk averse with respect to a health issue, ED accessibility and convenience resulting in low access burden, satisfaction with ED, dissatisfaction with PCP, and referral to ED by others). However, they reported that there was either limited or no support at all from the quantitative literature with respect to experiences of past traumatic events, anxiety, stress, coping, and need for immediate pain relief. In contrast, we found significant support for these reasons within the quantitative literature included here for review. Further, two themes not emphasized by O’Cathain et al. [5] were found to be highly influential in this review (i.e., Knowledge and awareness of alternative sources of care and Relationships between patients and health care providers). The additional studies incorporated in this integrative review (N = 60), not previously captured in other reviews, serve to both validate and enhance our previous understanding of the context surrounding decision-making for non-urgent ED.

While there is the wealth of knowledge on this topic, the majority of studies were published from highly resourced nations (i.e., USA, UK, Canada, Australia); as such, the results should be considered in light of this context. For example, Canada has a publicly funded healthcare system which contrasts with the private health care model utilized in the United States, and various two-tiered systems adopted in Europe and Australia. American studies have reported financial barriers to primary care as a common reason for attending the ED [107]. Financial barriers are not particularly relevant to individuals from nations with public or semi-public health systems since they, in part, have a reduced (direct) financial responsibility for medical care. Our understanding of reasons for non-urgent ED use in less resourced nations is currently limited.

The results from this review suggest that ED patients are heterogenous and that many factors influence their decision-making. Considering the complexity of patients that EDs care for, treating them as a single entity may be problematic. Thus, a multi-pronged approach may be required to limit excessive non-urgent visits. For example, simply redirecting non-urgent patients to other settings has been shown not to be wholly effective [108]. Instead, ensuring health care providers (at both PCP clinics and ED) understand how and why patients make decisions may help to provide insight and direct patient education. Health education should be explicitly and intentionally embedded in all ED health care provider roles [109]. This involves communicating, managing knowledge, mitigating errors, and supporting decision-making [109]. Research suggests that basic educational expertise, fundamental knowledge and reasoning, as well as emotional self-regulation are all critical components of health [110]. Thus, education is a social determinant of health which can potentially impede or enhance patients’ health [110]. Routinely educating patients on the role of the ED, as well as alternatives in the community, is a critical aspect of improving the public’s health.

This review found that many patients were anxious, uncertain, or fearful of their health problem. They had decreased ability to manage their discomforts and some reported the inability to cope. Guidance and support should be provided to patients with respect to managing recurring symptoms which may be directly or indirectly (e.g., anxiety, stress) related the presenting condition. Discharge teaching could include problem solving techniques for decision-making (e.g., accessing information) as well as self-management strategies (e.g., pain relief). While these “common sense” strategies may be commonplace among health care providers, it should not be assumed they are shared with lay persons. A recent systematic review highlighted and confirmed the disparity in patients’ and clinicians’ mutual understanding [111]. The authors examined the effectiveness of different methods of providing discharge instructions in the ED and found that communicating discharge instructions verbally may be insufficient; greater success could be achieved with the addition of video or written information [111], or via social media.

Finally, this integrative review demonstrated that there are notable deficiencies in various design and functioning of health care systems, where the literature was drawn. Many patients reported significant issues with accessing primary care, and were dissatisfied as a result. Simultaneously, patients were satisfied with the ED due to increased accessibility and quality of care, thus driving their attendance. This result has been supported by Van den Borg et al. [112] who examined the relationship between attending the ED and accessibility and continuity of primary care among 34 countries (60,991 patients). They found that ED visits had a significant and negative relationship with better primary care accessibility [112]. Systematically improving deficiencies in primary care may reduce non-urgent ED visits. Policy makers and practitioners should reflect and consider the complexities of their given health care environments to adequately design systems which are responsive to patients needs.

There has been a significant amount of inquiry generated on patient-reported reasons for non-urgent ED use. Regardless, there are a few areas that should be targeted for deeper inspection which would assist in filling gaps in the knowledge and addressing certain methodological considerations. Future research should aim to explore, in greater depth, specific themes identified in this review. For example, the role of health knowledge, emotions, beliefs, attitudes, and behaviour response patterns have been indicated as influencing the decision-making process, specifically with respect to perceived severity and urgency of presenting condition. New studies should explicitly evaluate ED users’ health literacy, health-related personal beliefs, stress and coping ability using validated outcome measures. This approach has received little to no attention in the literature. Psychosocial factors (e.g., stress, coping) have been explored in only a dearth of studies, largely as an afterthought to the primary objective [2]. Linking this subjective data to large, objective administrative health data could provide greater context than simple patient-reported reasons. Researchers should endeavour to use standardized criteria to evaluate triage acuity, when possible, and to fully describe their patient population and geographic region for accurate interpretation of results and comparisons with others.

This integrative review is not without its own limitations. The strict inclusion and exclusion criteria may have limited some articles from being included (e.g., all non-English studies). Further, specific populations (e.g., ambulance riders, pediatrics, specific presenting complaints, frequent ED users) have been cited as using the ED non-urgently but studies focused specifically on such subgroups were excluded from this review. This was intentional in an attempt to create a more homogenous sample for review. These ED subgroups may contribute unique results which could be informative to this topic. Nevertheless, results were drawn from a very large pool of general ED population studies. Finally, integrative reviews have the potential to suffer from lack of rigor given the process of combining diverse, complex methodologies [9]. The methods described herein were conducted using an iterative coding process by two individuals following well-cited, formulated guidance [9, 13].