Anxiety about anxiety: a survey of emergency department provider beliefs and practices regarding anxiety-associated low risk chest pain

Chest pain is one of the most common chief complaints evaluated in the Emergency Department (ED), accounting for approximately 20% of all annual ED visits nationwide [1]. A small minority of these patients have a true cardiopulmonary emergency such as acute coronary syndrome (ACS) or pulmonary embolism (PE). Frequently, these patients have psychosomatic contributors to their symptoms, including panic or anxiety disorders. These disorders are present in up to 58% [2] of patients presenting with chest pain but remain undiagnosed in approximately 80% of these patients [3]. Many adults with chest pain undergo extensive ED evaluations to rule out cardiopulmonary emergencies. In some cases, these evaluations can last 48 h including multiple tests, radiation exposure, and high cost, only to yield no named diagnosis [4]. A significant fraction of these patients go on to develop chronic chest pain and continue to seek medical attention despite negative cardiac evaluations [2]. Prior work has indicated emergency providers are hesitant to diagnose and discuss the role of anxiety [4] as a significant co-factor or etiology of their pain. However, the beliefs of ED providers regarding the prevalence of anxiety among patients with low risk chest pain and how they manage those patients both in the ED and at discharge have not been evaluated together.

This work was a hypothesis generating survey, conducted in accordance with guidelines by Mello et al., including both expert consensus and a modified Delphi technique [5, 6]. This protocol was deemed to be exempt by the IRB (protocol# 1601415405) at Indiana University School of Medicine. Authors PIM, JAL, CLH, and JAK initially determined that the objective of the hypothesis generating survey was to: evaluate the gap between ED provider beliefs regarding the approach and disposition of patients with low risk chest pain thought to be secondary to anxiety compared with their perceived practice patterns. The authors developed an initial set of themes which led to the generation of a 15-item survey which was test-administered to ED providers at Indiana University Department of Emergency Medicine. Responses from this test survey were analyzed along with feedback from participants by an advisory group who determined the final themes, domains, and survey architecture via a modified Delphi technique. This advisory group included individuals with expertise in both the subject areas as well as survey generation and administration in order to ensure face, content, and construct validity. The authors then generated a draft survey consisting of 26 items which were again vetted by the advisory group. Through an iterative process and consensus, edits were made and the items were reduced to a final survey consisting of 22 items. These included a mixture of multiple choice and visual analogue scale (0–100%) questions (Additional file 1: Appendix 1).

The survey was then administered to a convenience sample of Emergency Medicine (EM) providers visiting the exhibitor hall at the 2016 American College of Emergency Physicians Scientific Assembly (ACEP16) in Las Vegas, Nevada from October 16–19, 2016. This forum was chosen because it represents the largest annual gathering of EM providers [7]. Eligible survey participants included advanced practitioners (i.e.: nurse practitioners and physician’s assistants), emergency medicine residents, as well as physicians at the fellow or attending level who were practicing in academic and/or community settings. The investigators obtained a booth in the ACEP16 exhibitor hall and invited all eligible participants passing by the vicinity of the booth to complete the survey. A description and invitation to the booth was printed in the ACEP16 program and the meeting’s website. Additionally, the investigators handed out invitation cards at the meeting. To encourage participation, we provided a material incentive in the form of a raffle for a FitBit® activity monitor. The number of passers-by who were approached but declined participation was not tracked.

Participants first viewed a description and purpose of the survey prior to starting, and then completed the survey using either a laptop computer or electronic tablet. Study data were collected and managed using REDCap electronic data capture tools hosted at Indiana University [8]. Exported data were analyzed using Microsoft Excel for Mac, Version 15.14, and IBM SPSS Statistics for Mac, Version 22.0. The survey data was exported from REDCap into both software packages, and both univariate and multivariate analysis was completed. Where appropriate, 95% confidence intervals are reported. All visual analog-scale data were found to be non-normally distributed by both Kolmogorov-Smirnov and Shapiro-Wilk normality testing. All data generated or analyzed during this study are included in this published article as a supplemental excel dataset (The complete dataset generated from this survey and the basis for this manuscript is available from the authors upon request).

Four-hundred-and-nine surveys were completed, representing approximately 5% of ACEP16 conference attendees. There was diversity in respondent geographic location (Fig. 1), years of experience, and primary practice environment. The majority of respondents practiced at academic centers (52%), had less than 10 years of experience (68%), and were male (72%) as shown in Table 1. Twenty-nine respondents (7.0%) identified their primary practice location as outside of the United States.

This study incorporated both a diverse geographical representation, and a wide range of practice settings and experience. The main message of this work is that physicians believe that approximately 30% of patients seeking emergency care for low risk chest pain have anxiety or panic as a primary problem, yet fewer than half provide any treatment or information to help the patient manage anxiety. The authors recognize that the primary imperative of emergency care is to protect and intervene against threats to a patient’s life. However, in patients without a serious cardiopulmonary disease, untreated anxiety can degrade quality of life and worsen perceptions of wellness [9, 10], contributing to systemic inflammation [11], which ironically creates the pathophysiology of coronary artery disease [12, 13]. Additionally, this is associated with an increased health resource burden [14‐18]. We would argue that the most important of these to the emergency clinician is the detrimental effect on patient quality of life and unnecessary ED resource utilization and recidivism. Prior work in a large cohort of ED patients with chest pain discovered that only 0.2% were given an ICD9 diagnosis of anxiety; however, none of the 8% of patients who self-identified anxiety as likely the cause for their chest pain symptoms at follow-up were given this diagnosis even secondarily [4]. This low rate of anxiety diagnoses in patients with non-life threatening chest pain in the ED was further demonstrated in a national sample using the National Hospital Ambulatory Medical Care Survey, where only 2.3% received this diagnosis [19]. Whereas, a systematic review by Webster et al. incorporated nine studies and found the likely prevalence of diagnosable anxiety in patients with non-cardiac chest pain to be between 21 and 58% [2]. Our current work is consistent with that of Al-Ani et al., who found that 30% of chest pain patients who had been risk-stratified to be low-risk for ACS were identified as suffering from anxiety and 80% of those patients were untreated [20]. This highlights the fact that at least in this evaluation, EM provider gestalt regarding anxiety in the presence of low-risk anxiety is in line with objective measures shown previously. Taken together, these data indicate the need for a more patient-centered approach to communicating with and managing anxiety in patients with low risk chest pain. Helpful methods could include using formal anxiety risk stratification tools such as the Hospital Anxiety Depression Scale (HADS) or Generalized Anxiety Disorder scale (GAD-7). Both of these validated tools have been used to evaluate for the presence of anxiety in this patient population previously [21‐23].

Furthermore, even after workup and provider reassurance, patients who present with possible cardiac symptoms are often left with “residual anxiety” despite normal test results [24]. Reassurance of negative test results alone may not help reduce patient concerns about the potentially life-threatening causes of their chest pain [24‐26]. Ackerman et al. defined the ideal content for ED discharge communication for patients with chest pain including directed follow-up suggestions and advice on self-care, [24] which are all important parts of the patient-provider interaction, particularly in the emergency department. Our data suggest these steps to be absent in current emergency care and may represent a missed opportunity to affect the trajectory of care [10]. This low rate of discharge communication is likely heavily influenced by the fact that the majority of respondents (61%) found their hospital resources for appropriate follow-up to be inadequate. Not surprisingly, 82% of providers wanted a specific clinic to direct their patients for further evaluation for anxiety.

Implications of these data are that patients could benefit from targeted interventions which minimize psychological distress, improve quality of life [9] and prevent futile emergency department visitation. There are a number of psychological interventions which are available for this patient population which include cognitive behavioral therapy (CBT), relaxation therapy, hyperventilation, hypnotherapy etc. [15]. Foremost and most well established is cognitive behavioral therapy (CBT), which has been shown to be both acceptable to patients and effective in reducing chest pain and improving quality of life [15, 27]. Even brief interventions have been successful as van Beek et al. randomized subjects with non-cardiac chest pain and anxiety after psychological testing to usual treatment vs an abbreviated CBT course (6 sessions) showing a significant decrease in both depression and anxiety symptoms [28]. Unfortunately, CBT interventions are not a readily available resource for ED referral and can be expensive [29]. However, it appears the intervention need not be extensive or expensive, as simply providing self-help information explaining the connection between anxiety and chest pain can make a difference in both psychological and/or physical symptoms [18, 30, 31]. Other anxiety management strategies include training in breathing techniques and mindfulness-based exercises may show effectiveness in this patient population [32‐35]. However, any intervention requires a frank conversation about anxiety and appropriate discharge instructions.

ED providers believe 30% of patients seeking emergency care for chest pain stratified to be low risk for ACS have anxiety as a primary problem, yet fewer than half discuss this concern or provide information to help the patient manage anxiety. These data support the need for a more patient-centered approach to communication and management of anxiety in patients with low risk chest pain. Further investigation to elicit reasons why providers would be hesitant to discuss suspected anxiety in the setting of low risk chest pain and contributing biases is needed.