Keeping out and getting in: reframing emergency department gatekeeping as structural competence

A strong tradition in sociology has integrated literature from the sociologies of health and illness, and work and the professions, to examine the role of gatekeeping practices in the social organisation of healthcare (cf. Dodier and Camus 1998; Green and Armstrong 1993; Griffiths 2001, 2003; Hughes 1989; Vassy 2001). These studies have tended to frame medical gatekeeping as an exclusionary social practice, delineating how practitioners and clerical staff police the moral boundaries of medicine by keeping out patients who are categorised as ‘bad’ (Dingwall and Murray 1983), ‘deviant’ (Jeffery 1979), or otherwise problematic (Mannon 1976).1 Yet medical gatekeeping, understood more broadly, can entail not only keeping patients out of particular clinical settings, but also redirecting them to alternative sources of care. This second component of gatekeeping has received markedly less attention from medical sociologists. In this article, I describe medical gatekeeping as a two-step process of, first, categorising certain patient complaints as unsuitable for treatment within a particular clinical setting, which I refer to as the restrictive component of medical gatekeeping, and second, diverting patients to alternative sites for care, which I refer to as the facilitative component of medical gatekeeping. I use these labels with respect to patients’ access to care, recognising that restrictive and facilitative aspects of medical gatekeeping are part of a coordinated organisational strategy for managing resource scarcity.

To analyse facilitative gatekeeping practices in a healthcare setting that serves a large proportion of socioeconomically disadvantaged patients, I build on Metzl and Hansen's (2014) concept of ‘structural competence’. Metzl and Hansen (2014) introduced the concept of structural competence to describe recognition – by clinicians, researchers, and policymakers – of how structural forces shape health inequalities. One motivation was a concern that linking healthcare inequalities to particular cultural identities (e.g. race or ethnicity) may obscure factors that influence the social patterning of disease relatively further upstream, such as food security, urban and rural infrastructure, and immigration laws. Structural competence has attracted substantial scholarly attention, gaining particular traction in discussions of medical education reform (Mansh et al. 2015, Tsevat et al. 2015). The presumption underlying such reform efforts is that healthcare providers may not be attuned to how social structures shape health and illness. However, US safety net providers – institutions that provide medical care to low-income, medically underserved, and racial/ethnic minority communities (Institute of Medicine 2000) – routinely confront the limits of contemporary American medicine to ameliorate sickness due to underlying structural inequities. This article explores how US healthcare providers respond to such structural inequities through facilitative gatekeeping practices that redirect patients to alternative sites of care. Thus, while I view facilitative gatekeeping as a broad category that encompasses a range of efforts to redirect patients to alternative sites of care, my analysis here focuses on a specific subtype of facilitative gatekeeping in which health care providers help vulnerable patients navigate barriers to accessing health care.

In what follows, I draw on data from a study of patient-provider communication in a United States emergency department (ED) to show how ED providers exhibit structural competence in their day-to-day clinical work. Specifically, I illustrate how ED providers reveal intimate knowledge of structural vulnerabilities in diverting socioeconomically disadvantaged patients with chronic back pain to clinical sites that are better equipped to provide care. In doing so, I suggest that we rethink the emphasis on restrictive practices in sociological accounts of medical gatekeeping.

Sociological approaches to medical gatekeeping

Medical gatekeeping entails control over access to healthcare resources or information. Although gatekeeping may occur at various points of health service delivery, including information flow among clinical staff (Riley and Manias 2009, Street 1992), reluctance or refusal to order certain screening tests or treatments (Faulkner 2002, Shuster 2016, Tritter et al. 2014), and pharmacists’ ethical decision-making (Chiarello 2013), most sociological studies have focused on practices that facilitate or constrain patients’ entry to care. In keeping with this understanding of gatekeeping as regulating patients’ access to care, sociologists have emphasised the gatekeeping functions of frontline clinical staff (cf. Jean 2004, Smith 2016, Solimeo et al. 2016). For instance, Vosk and Milofsky (2002) contend that most ED gatekeeping is performed by someone other than the medical provider. However, I will illustrate here that ED physicians and nurse practitioners (NPs), as well as social workers, also play a gatekeeping role in facilitating patients’ access to healthcare resources beyond the ED encounter.

Sociological understandings of how patients’ access to care is shaped by extra-clinical judgments can be traced to Jeffery's (1979) pioneering study of how clinical staff in a UK ED categorised patients as ‘good’ or ‘rubbish’ using social and moral criteria. While Jeffery's classification scheme has since been criticised for neglecting routine cases that carry a neutral moral valence (Hughes 1989; see also Dingwall and Murray 1983), the enduring legacy of his study for the sociology of health and illness is to highlight how moral judgments can, in certain situations, influence patients’ medical treatment as much as a priori clinical criteria. Moreover, in typifying patients, clinicians not only draw preexisting social norms and values into the clinical realm, but also enact and produce the social structure de novo (Latimer 1997).

The clinical constitution of the socio-moral order through patient typifications is particularly salient in the emergency setting (cf. Dingwall and Murray 1983, Dodier and Camus 1998, Hillman 2014, Hughes 1989, Jeffery 1979, Mannon 1976, Roth 1976, Vassy 2001). As Dingwall and Murray (1983) note, patient typifications are crucial to the social organisation of the emergency department, because ‘bad’ patients – that is, patients who do not comply with normative expectations regarding the sick role – will consume more clinical resources and their treatment may be prolonged until there is more time and space to accommodate them. Moreover, to manage caseloads that are perpetually in flux, ED clinicians must distinguish ‘real emergencies’ from ‘non-urgent’ cases (Vassy 2001). Typifications thus reflect judgments about the ‘doctorability’ of a problem – a term proposed by Heritage and Robinson (2006) to describe the extent to which a problem is treatable within a particular medical context – and embed moral judgments about whether the patient is making an appropriate claim on the provider's expertise and authorised work roles (Roth 1972).

One consequence of adjudicating legitimacy in this way is that patients deemed to be inappropriate for ED treatment are disciplined in some fashion (Hillman 2014). In one study, ED patients low in ‘mobilising worth’, another term for legitimacy, were subjected to longer wait times, limited diagnostic testing, or denied care altogether (Dodier and Camus 1998). For this reason, Vassy (2001) describes categorising patients as a covert form of micro-level rationing in which both the quantity and quality of medical care may be determined by social factors. From this perspective, gatekeeping and treatment constitute intertwining features of the ED encounter, with the former serving primarily to put limits on the latter.

Categorising patients as non-urgent cases does not necessarily result in reduced access to treatment, however. Vassy (2001) demonstrated that ED typecasting could also facilitate patients’ access to care, in that staff tended to accept patients with social problems or limited knowledge of the healthcare infrastructure for ED treatment, regardless of their clinical status, because their needs were viewed as pressing. This finding departs from those of Roth (1972) and Jeffery (1979), who found that ED providers favoured patients with high social value. Vassy hypothesised that socioeconomic differences between France, where her research was conducted, and the US and UK, where Roth and Jeffery worked, respectively, may account for this difference. However, Vassy did not fully explore the potential contribution of historical factors: both Roth's and Jeffery's studies took place in the 1970s, while Vassy's was conducted in 1998.

Along with Vassy (2001), my research indicates that ED patients without ‘real emergencies’ may be treated more favourably than Roth (1972) and Jeffery (1979) suggest. As I will show, providers in my study subjected patients with non-urgent complaints to a more nuanced form of gatekeeping in which patients were first characterised as inappropriate for ED treatment, and then diverted to alternative sources of care. This reveals how gatekeeping may sometimes entail practices that facilitate patients’ access to care.

Research overview and setting

This article is based on a collaborative mixed-methods study of patient-provider communication about back pain and analgesics in an emergency department (ED) based in a US teaching hospital. EDs are important sites for pain treatment in the US because they are always open and provide a critical safety net for access to care among uninsured populations (Institute of Medicine 2011, Todd et al. 2007). Patients with structural vulnerabilities (e.g. uninsurance or underinsurance) often seek treatment for chronic or non-emergent pain in the ED because they do not have a primary care provider (PCP), and because certain EDs are known to provide care to low-income populations at a reduced cost. Pain treatment in the ED thus provides a particularly good site for examining how structural competence mediates access to care among socioeconomically marginalised populations.

The purpose of the larger study was to use audio-recordings of patient-provider interactions to characterise patients’ requests for pain medications and examine how communication behaviors influence treatment decisions. Drawing on an ecological model of patient-provider communication adapted from Street et al. (2007), we hypothesised that an interplay of patient, physician, and systems-level factors would shape communication and influence decisions about prescriptions. The study received approval from the Institutional Review Board at the University of North Carolina at Chapel Hill.

Participants were ED providers who made primary treatment decisions about pain medications (attending physicians, medical residents, and NPs) and ED patients who reported back pain as their primary complaint. Providers were recruited via a combination of methods, including a presentation at a staff meeting, an informational email explaining the study's goals, and an in-person review of study protocol. Patients were identified using the electronic medical record and screened prior to enrollment. Only patients of providers who had agreed to be in the study were approached about the study. Providers and patients received a US$5 and US$25 gift card, respectively, per visit recorded. Research design and methods have been described more fully elsewhere (Buchbinder et al. 2015).

I collected data over eight months (September 2012–April 2013) with assistance from five research assistants (RAs). Five-hour research shifts were scheduled several times per week during the study period, balancing daytime/nighttime and weekday/weekend ED shifts. Hand-held digital audio-recorders were used to record all communication between patients and their primary ED provider. We remained present in the exam room during the patient-provider encounter to take field notes, unless asked to leave to maintain privacy during a physical exam. Following each encounter, we abstracted information about visit characteristics (e.g. length of stay, number of prior visits for pain) and treatment decisions from the electronic medical record. Patients completed a pre-visit questionnaire and a telephone survey (assessing satisfaction and shared decision-making) approximately 24 hours following the ED visit. Providers completed a brief post-visit survey immediately following the encounter.

Seventy-four ED encounters were recorded. The 74 patients were nested within 30 providers, who contributed up to 10 recordings with different patients. The large majority of encounters (70.3%) involved 11 NPs. This reflects the distribution of professional labour in this ED, in which NPs provided a large proportion of medical care for non-emergent conditions like back pain. The 74 patients were predominantly from low-income backgrounds, with over one-third reporting earning less than US$10,000 per year, and more than half unemployed. Most participants had no insurance (52.7%) or Medicaid/financial assistance (24.3%). The mean age for patients was 39.6, and most were white (59.5%) or black/African American (35.1%).

This article focuses on audio-recordings of clinical interactions, which lasted an average of 17.5 minutes and typically included history-taking, a physical examination, and the diagnosis and treatment plan. Audio-recordings were transcribed verbatim and de-identified by a trained transcriptionist. I reviewed transcripts iteratively using an inductive analytic approach guided by the tenets of grounded theory (Corbin and Strauss 2007), and developed a coding scheme that focused on three overarching thematic codes, which divided patient-provider communication into the following content areas: (i) discussions of medication; (ii) discussions of imaging/diagnostic testing; and (iii) gatekeeping. I defined gatekeeping as work undertaken to maintain the boundaries of a legitimate ED visit and redirect patients with inappropriate ED complaints to alternative sources of medical care. I then trained two RAs to code all of the transcripts using this coding scheme and Dedoose (2013) data management software. During an initial training period, each RA coded the same transcript independently, and any discrepancies were discussed and resolved. Changes were made to the codebook to accommodate and reduce discrepancies. After the training period, the remaining transcripts were divided and coded by a single coder. For this article, I reviewed ‘gatekeeping’ excerpts to identify patterns and themes within this broad category. I summarised these analytic insights in descriptive memos, which I reviewed iteratively alongside the original transcripts, field notes recorded by the research team, and the sociological literatures on gatekeeping and emergency medical care.

‘ERs are not gonna fix anything’

My initial choice of research site was motivated by an interest in patient-provider negotiations surrounding prescription analgesics. However, I noticed early on in my fieldwork that, in treating patients with back pain, ED providers spent a great deal of their time engaged in a form of boundary work (Allen 1997, Mizrachi et al. 2005; Nancarrow and Borthwick 2005) – that is, policing the borders of appropriate emergency care. For these patients, the question of ‘doctorability’ hinged less on whether they were deserving of care as on whether the ED was the appropriate place to seek it. A major organisational challenge facing providers in this context was that the ER was not structurally equipped to address many of the patients’ needs; consequently, patients were often told that the ED was not an appropriate place for chronic pain management. The case of Randy Stewart, a patient in my study, is particularly instructive here.2

Randy was a 52-year-old male who had come to the ED on a weekday afternoon in the fall of 2012 with a primary complaint of pain in his lower back that radiated into his left hip, leg, and kneecap. After a two-hour wait, Randy was taken to an examination room and assigned to the caseload of Tracy Reinhart, an NP who had worked in the ED for more than ten years. Accompanying Randy to the hospital was an aunt who indicated that Randy was ‘like a son’ to her. Randy relayed that the pain, which had intensified over the course of a year, was constant: ‘It's got where I can't hardly walk. It hurts bad. If I sit down, I'm hurting. If I lay down I'm hurting’.

When Tracy inquired about Randy's clinical history, Randy explained that he was unemployed, uninsured, and did not have a regular doctor, so the pain had not been treated systematically. It was not clear from the interaction whether Randy's unemployment was consequent to his long-term chronic pain, although his aunt insinuated that it was. Randy's insurance status, however, was almost certainly a result of his unemployment, because this study took place prior to many of the reforms introduced by the US Patient Protection and Affordable Care Act, which made it easier for patients to obtain affordable health insurance outside of employer-based markets.

Randy reported that when he did not have any prescription medications on hand, he had occasionally resorted to using cocaine to manage the pain. At one point, his back had been x-rayed, but his aunt felt strongly that what he really needed was an MRI. Two of Randy's sisters had suffered from hip dysplasia that caused tremendous pain; the family was worried that Randy might have the same problem. They had travelled to an ED more than 70 miles from their hometown with the hope that Randy might be able to get an MRI.

After a thorough physical examination, Tracy explained that while she thought a herniated disc or degenerative disc disease might be causing Randy's pain, the only time they performed MRIs in the ED setting was if there was a potential surgical emergency. Otherwise, the MRI results would not change the ED treatment plan. Because Tracy did not see any evidence of surgical emergency, Randy could not obtain an MRI today. Instead, Tracy offered to repeat Randy's x-rays, which she viewed as a reasonable option, and she assured Randy that she would treat his pain with medication. But the biggest thing to accomplish, she noted, was to get him a PCP, who would likely begin by ordering physical therapy. ‘I'll have somebody talk to you about options for follow-up, and that's really going to be the key for everything, to kind of move forward. ‘Cause ERs are not gonna fix anything’, Tracy said.

Randy's aunt was visibly distraught by the news that an MRI would not be forthcoming. ‘We didn't figure that you were going to be able to fix anything’, she said. ‘We were just hopin’ and prayin’ that he could get the MRI. That he has needed now for well over a year’. Once again, Tracy explained that the MRI results would not change the treatment plan. The aunt countered that it would give him a diagnosis, revealing a tension between the value of diagnostic knowledge for its own sake, and the expectation that diagnostic knowledge will serve an instrumental purpose in the ED setting. She also suspected that this diagnostic test was being withheld because Randy was uninsured. In response, Tracy explained that the decision about the MRI had nothing to do with Randy's insurance status; hospital protocol simply precluded her from ordering an MRI on a non-emergent basis: ‘It just doesn't get done’.

Ultimately, Randy was given Percocet, an opioid analgesic, and Valium, a muscle relaxant, and sent home with prescriptions for each. His x-ray showed degenerative disc disease and gallstones, but nothing definitive regarding hip dysplasia. He also spoke with an ED social worker about options for follow-up medical care given his insurance status. His length of stay in the ED was three hours and 40 minutes. Tracy wrote on the end-of-visit survey, when asked what, if anything, was challenging about this patient encounter: ‘Patient and family had unrealistic expectations for the visit today. Couldn't quite understand why an MRI could not be done and why it was not the solution to the problem. Patient however felt better upon discharge’.

Randy's treatment experience was similar to other study participants. No patient received a surgical consultation, and of the 23 patients with whom the possibility of ordering an MRI was discussed, only one patient received an MRI in the ED – because the treating provider was concerned about an abnormal neurological examination. Tracy's response to Randy and his aunt demonstrates gatekeeping in the ED as a two-step process of, first, laying out a clear justification for why the patient could not obtain the desired diagnostic test or treatment, and, second, offering alternatives to the desired form of care.

This two-step process likewise draws on structural competence, which I understand as consisting of, first, contextual sensitivity toward the structural limitations facing socioeconomically marginalised ED patients in obtaining medical care, and second, an intimate knowledge of resources for fostering possible workarounds. In what follows, I describe three components of ED medical gatekeeping that demonstrate gatekeeping as a facilitative process of diverting patients with structural limitations to alternative healthcare resources: (i) referral to a PCP or usual source of care; (ii) assistance with financial assistance enrollment; and (iii) additional support with navigating the medical bureaucracy.

Referral to primary care

The default practice in the ED I studied was for providers to encourage patients to follow up with their PCP if their pain continued. Therefore, facilitative gatekeeping practices were most apparent among patients who did not already have a doctor. Providers often began the process of soliciting a patient's clinical history by inquiring about whether the patient had a PCP, and if so, whether the patient had called this person before coming to the ED. As a blunt screening mechanism, this line of questioning enabled providers to evaluate patients’ healthcare resources and their capacity to extend pain management beyond the ED encounter.

About half of the patients in my study reported that they did not have a PCP and had exhausted other avenues of treatment.3 Some of those patients who did have a PCP had lost their insurance and came to the ED because they did not have money to pay for an appointment with their regular doctor. As one patient explained, ‘They wanted a hundred and seventy-five dollars today and I just don't have that kind of money’. Another patient reported that he had come to the ED at the advice of his doctor: ‘She's the one that suggested that I come up here because she wasn't going to give me anything for pain because people abuse the pain medicine system. And, uh, she wouldn't authorise an MRI, she wouldn't request one, she wouldn't do anything like that. I don't have any insurance so she was very, you know – how would I put that? Short’.

Dennis Riverton, a 36-year-old uninsured patient who worked in manual labour, developed chronic back pain at the age of 18, when he injured himself hauling wood. The night before his ED visit, his legs had gone numb and gave out on him while using the bathroom. When Jackie Sierra, the NP assigned to Dennis’ case, inquired about the medications he tended to take for his pain, he told her that he had been given 20 Percocet pills the last time he was seen in the ED, but ‘I really don't like taking pain pills unless I absolutely have to ‘cause I've seen a lot of people get addicted to them’. Later in the visit, Dennis admitted that some of his reluctance to take opioids stemmed from his history of drug addiction. Despite his self-restraint, his last ED visit had been three weeks ago and his supply had run out, while ibuprofen did nothing to relieve his pain. Without a PCP or a regular source of medical care, Dennis had told his girlfriend that he needed her to drive him to the ED.

After examining Dennis, Jackie told him that, because of the numbness in his leg, he probably needed an MRI, but that this could not be done in the ED. ‘You know as well as I do that you really need a primary care doctor to manage this for you’. She gave him the phone number of a community health organisation in his county, and explained: ‘If you tell them you were seen in the emergency room, they usually can give you an appointment with a primary care doctor, and it's a sliding scale, so it's meant for people without insurance. You just kind of take a proof of income and they kind of bill you based on what you can afford’.

Jackie's treatment involved both restrictive and facilitative dimensions of medical gatekeeping. Jackie denied Dennis an MRI despite recognising his need, noting that his symptoms did not qualify him to receive the MRI on an emergency basis. Yet rather than turn Dennis away to navigate the healthcare system on his own, Jackie offered Dennis detailed information about a sliding scale clinic in his community and specific instructions about how to pursue medical care there and obtain a PCP. Such supportive resources constitute a crucial, yet under-acknowledged, component of medical gatekeeping through which ED providers like Jackie demonstrated awareness of the structural constraints and access barriers facing socioeconomically marginalised patients. In fact, getting the patient ‘plugged into the system’ – that is, redirecting them to alternative sites for routine care – often became the primary therapeutic objective.

One ED encounter I recorded that was particularly noteworthy in this regard involved Stuart Jackson, an attending physician, and a 33-year-old unemployed, uninsured male patient. The patient had chronic back pain resulting from injuries sustained during two separate motor vehicle collisions, and was returning to the ED for the second time in three weeks. At his first visit, the patient had been instructed to call the hospital's internal medicine clinic every day at 8:00 am to secure one of the clinic's limited primary care appointment slots for uninsured, low-income patients. Despite following this advice, he had been unable to obtain an appointment, and was starting to feel like his efforts were ‘a waste of time’. In response, Dr. Jackson told the patient about a sliding scale clinic in his county that would likely be easier to get into. He instructed the patient not to call the general number listed online, but rather to call the number listed on the pamphlet, which belonged to a care coordinator assigned to accept referrals from the academic hospital. Dr. Jackson then had the patient call the clinic while he was still waiting in the ED. Notably, while the patient left the ED with only a limited supply of Percocet and a warning that this would be ‘the last time you're going to get a prescription here’, the patient left the ED feeling ‘completely satisfied’ with his visit.4

Of the 37 patients who indicated that they did not have a PCP, only three did not receive a referral for such a provider during their ED stay. Two of these patients were treated by the same provider, an NP with three years of ED experience, while the third was treated by a medical resident (i.e. physician trainee). In the latter case, the provider's experience appeared to play a role in his failure to offer supportive resources. When the patient indicated that he had been unable to find a local physician after relocating, the resident told him, ‘I don't know what to offer you except to say keep at it. Sometimes that's hard. I mean, there's a lot of doctors around here, you should be able to find someone’.

There were also cases in which a referral to PCP was more perfunctory, like the physician who said, ‘So we can give you the phone number of (name) health services. And you know, it's kind of up to you’. Moreover, in two cases in which patients reported that they had PCPs but were unable to pay for an office visit because they lacked insurance, the providers (an NP and physician, respectively), missed an opportunity to offer the patients additional resources for accessing affordable care. In general, providers offered more detailed information and advice to patients with more complicated medical histories or those who were more forthright in requesting imaging. Nevertheless, in light of the recognisable limits of EDs to offer effective treatment for chronic pain, referral to a PCP constituted an achievable therapeutic outcome, albeit a modest one.

Enrolling in financial assistance programmes

One of the earliest ED encounters that I observed involved Seth Harper, a 31-year-old uninsured man with chronic pain in his upper back and neck that resulted from a fall off a roof while doing construction work. At his initial ED visit, he was referred to the hospital's neurology clinic, which referred him on to the pain clinic. However, the pain clinic wanted US$180 up front to pay for the visit, which proved cost-prohibitive for Seth. When I met Seth and his girlfriend, Janine, on a quiet weekend morning in the ED, he was waiting to hear back from the hospital's financial assistance programme, which he had applied to several months earlier.5 For the past few months, his pain management strategy had been to wait as long as he could and go to the ED when he could no longer bear the pain. Nancy Bales, the ED attending physician assigned to his care, told me that Seth exemplified everything wrong with the US healthcare system. She would try to get a social worker to come in and discuss the financial assistance programme enrolment with him; that would likely be the best thing he would get out of the visit, she said, aside from a few days’ supply of pain pills.

While Seth and Janine waited for the social worker, I chatted with them about their experiences in the ED. They had very positive things to say about Seth's treatment thus far. The major frustration, for them, was that it only provided stopgap relief. Janine told me that she had wondered if she should call and inquire about Seth's status with the financial assistance programme, but had ultimately decided that there were probably other people ‘further up the list’ and that she should try to be patient. Concerned that the paperwork might have been lost or overlooked, I encouraged her to call again, noting that the ‘squeaky wheel’ approach has its benefits. Later, the social worker – who usually worked in a different department but was filling in for the ED social worker – told Dr. Bales that Seth had ‘let the ball drop’ on the financial assistance programme, and should have called to follow up.

Although the social worker whom Seth saw (who, notably, was not a regular ED clinician) was quick to blame Seth for his delays in accessing care, Janine's inclination to wait reflects the practical barriers and knowledge gaps that many patients face in navigating medical bureaucracy. It also suggests a deficit of what Janet Shim (2010) has called cultural health capital: the repertoire of explicit skills and tacit cultural knowledge that may predispose certain patients (and providers) to more optimal healthcare encounters. From this perspective, the major therapeutic outcome of Seth's ED visit was not a biomedical treatment but a structural intervention: encouragement to follow up with the financial assistance programme on Monday.

Had Seth's ED visit transpired on a weekday, he might have seen the regular ED social worker, who spent much of her time helping low-income patients enroll in the financial assistance programme.6 An explicit part of this work involved educating people like Seth and Janine how to be ‘activated’ consumers in the healthcare market (Shim et al. 2016). That is, rather than passively waiting to make it to the top of ‘the list’, as Janine suggested, successful enrollment into financial assistance programmes required persistency and frequent check-ins.

However, assistance with the enrollment process depended on the time of the ED visit and whether it coincided with the ED social worker's part-time schedule. Helena Townsend, a 40-year-old uninsured hairstylist who had recently lost her job, came to the ED for the explicit purpose of filing the paperwork for the financial assistance programme and seeking a regular source of care for her chronic back pain. This strategy had been recommended by a physician who had performed a tubal ligation procedure on Helena several months earlier, when Helena, as result of her recent pregnancy, still qualified for Medicaid.7 The NP, Jackie Sierra, referred Helena to the hospital's spine clinic, but Helena did not complete the paperwork for the financial assistance programme, presumably because the social worker was not in the ED that day.

Support with enrolling in the hospital's financial assistance programme was thus an important target of facilitative gatekeeping, albeit one limited by ED staffing constraints. Despite such limitations, ED socialisation about medical bureaucracy also included additional practical advice, which NPs and physicians were well equipped to provide, about navigating the broader healthcare infrastructure and accessing care outside the ED. I describe such assistance next.

Navigating medical bureaucracy

Louisa Reynolds, a 35-year-old woman who had suffered from back pain for four years, had obtained an MRI several years back that showed a disc problem, but had been unable to continue care with the doctor who ordered it after losing her health insurance. Since then, she had bounced around to different EDs, the last of which had initiated the process of enrolling her in a charity care clinic. Unfortunately, she was still compiling the necessary paperwork and had not yet established care. A co-worker had advised her to come to this ED because it was known to offer financial assistance.

In the encounter that I recorded, Louisa told Tracy, the NP assigned to her care, that she didn't like to take pills, and that she wanted to see a doctor who could ‘fix it’. When Louisa's sister, who had accompanied Louisa to the hospital, said that Louisa had already been told that she needed to see a surgeon, Tracy was clear that Louisa would not see a surgeon in the ED: ‘You're going to find the same treatment in almost every emergency room. Because there, what we're looking for is, do you have any signs of a cervical [spine] emergency? Acute cord compression, they'd maybe take you to the operating room today. ‘Cause if not, there's not a whole lot we can do’. She added, ‘And actually, all we can do is give you medication’.

Tracy compensated for the restrictive side of ED gatekeeping by offering advice for navigating the healthcare system. First, Tracy gave Louisa information about a clinic that would accept payment for medical care on a sliding scale. Tracy advised Louisa to call the number listed every day at seven or eight in the morning because they only schedule appointments a few days in advance. Tracy then offered several practical strategies for Louisa to prepare herself for her next clinical visit. Because Louisa had been having a hard time recalling details from her medical history, Tracy suggested that she write down significant events with their dates before her next appointment: ‘You can just say, ‘Hey, this is what my history is on my back, and it gives you a little synopsis’. So it gets the person that is taking care of you a synopsis of kind of where things have gone, how everything has progressed’.

Tracy also advised Louisa to obtain her MRI images from her doctor's office so that her new provider could see how her condition was progressing. When Louisa asked how she would do that, Tracy explained: ‘Call first, wherever you had the images done. Ask them how to get a copy of the images. Who do you need to contact? Don't go drive anywhere until you actually get confirmation that somebody is actually holding a disc in their hand. Otherwise you might be wasting a trip’. Finally, Tracy gave Louisa general advice about seeking financial assistance for medical care: ‘Every single place you go to … there's gonna be a ton of paperwork, including tax statements, and all kinds of things. So my recommendation is whatever you send in, to any place, you better have a copy of it. Just in case it gets lost, doesn't get to the right place. I recommend that you make copies of every piece of paper’.

This example stands out for the level of detail in Tracy's instructions: this was the only time that I heard a provider advise a patient to write down their medical history or make copies of paperwork. More frequently, providers advised patients to be persistent in calling a clinic that reserved slots for underserved patients, noting that it might take several days or weeks to secure an appointment. Tracy's thoroughness with Louisa may have been in response to Louisa's bald request to see a surgeon, which likewise was somewhat rare in my data. To prevent Louisa from returning to the ED yet again to demand a surgical consultation, Tracy systematically walked Louisa through the steps she would need to complete to have a surgeon evaluate her back pain.

As a display of structural competence, Tracy's attunement to the implications of a ‘wasted trip’ to the doctor's office for a patient with chronic pain revealed sensitivity to the financial, temporal, and physical burdens of navigating healthcare access. Although Louisa's request to see a surgeon was refused due to standard ED protocol, she left with a concrete plan and practical advice for accessing medical care elsewhere and managing bureaucratic challenges.

Conclusion

In this article, I have argued for expanding the scope of sociological work on medical gatekeeping to include attention to how it may facilitate, as well as restrict, patients’ access to treatment. While the sociological literature has historically emphasised the restrictive dimensions of medical gatekeeping, this analysis of audio-recorded ED encounters demonstrates that gatekeeping is more than simply an organisational mechanism for excluding patients from unjustified treatment. Instead, ED gatekeeping also encompasses strategies that divert patients to alternative sites of care. Attention to these additional components opens up new analytic possibilities that depict medical gatekeeping as a set of multidimensional practices that may align with patient interests while serving institutional needs for rationing and resource management.

Using Metzl and Hansen's (2014) concept of structural competence, I have illustrated how ED gatekeeping constitutes an important means by which providers display and enact contextually sensitive understandings of local systems of healthcare delivery and barriers to accessing care. In the US, safety net providers are often acutely aware of the structural limitations and access gaps facing patients from socioeconomically marginalised groups. To help patients avoid repeat visits for problems that are largely ‘undoctorable’ in the ED, they may learn to navigate the patchwork US healthcare system and guide patients to alternative healthcare resources. Stepwise instructions for accessing care (e.g. how to make an appointment, how to obtain financial assistance) reflect providers’ understandings of gaps in patients’ cultural health capital and points at which treatment-seeking is likely to stall out or go astray.

Critics might argue that providers are more interested in moving patients out of the ED than in ensuring their access to care. Here, it is critical to acknowledge that the facilitative dimensions of ED gatekeeping serve important institutional interests even as they ease patients’ access to care. The diversion of patients to healthcare resources outside of the ED will also reduce future ED visits for non-emergent problems, which not only lessens patients’ inconvenience, but also alleviates strains on limited ED resources. This is particularly important because so much ED medical care for low-income US populations is not substantially remunerated. Moreover, to the extent that chronic pain patients are often viewed by medical providers as difficult or emotionally draining (Crawley-Matoka and True 2012, Slade et al. 2009), there are additional incentives for implementing such gatekeeping practices and diverting patients toward alternative sites of care, particularly when the patient's primary complaint is recurrent. Yet, rather than undermining my argument, this position shows how restrictive and facilitative gatekeeping are two sides of the same coin, which holds a similar payoff for patients and staff. Redirecting patients to places where they are more likely to receive the care they seek is organisationally useful for patients and practitioners alike.

Finally, the performance of ED gatekeeping may build on or reinforce existing inequalities in the healthcare labour force. In the ED I studied, much of the detail-oriented facilitative gatekeeping was performed by female NPs, if only because NPs were more likely than physicians to treat patients with chronic back pain, and most (though not all) NPs in this setting were women. More research is needed to determine whether facilitative gatekeeping practices detract from other professional responsibilities and impose gendered burdens on ED staff. At the same time, while acting compassionately towards patients may be a form of gendered emotional labour that disproportionately burdens female clinicians (Wharton 2009), it also serves professional interests to move patients along and out of the ED with minimal overt resistance.

To be sure, facilitative gatekeeping practices may look somewhat different in healthcare settings outside the US, where patients may face different structural limitations and access barriers to care. Nevertheless, illustrating how structural competence is enacted in real-time clinical encounters offers a fresh perspective on ED medical gatekeeping that may be useful in other contexts. I have shown how ED providers demonstrate awareness of the systematic ways in which underserved patients are thwarted in their efforts to seek medical care. This is not to suggest that facilitative gatekeeping practices are adequate for correcting the sweeping inequities in the US healthcare system. Yet analysing the micro-level practices through which individual actors respond to structural inequalities is a helpful reminder that health outcomes are not entirely predetermined by large scale structural forces. Examining the facilitative side of medical gatekeeping can help to elucidate other ways in which healthcare providers and clerical staff innovate creative solutions to deeply entrenched structural barriers to care.